The centralizer of a \(C^1\)-generic diffeomorphism is trivial (Q928654): Difference between revisions

This interesting and well written paper gives the answers on some questions of \textit{S. Smale}, see [Math. Intell. 20, No.~2, 7--15 (1998; Zbl 0947.01011)]. More detailed, let us consider a compact, connected manifold \(M\) and consider the space \(\text{Diff}^r(M)\) of \(C^r\) diffeomorphisms of \(M\). The centralizer of \(f\in \text{Diff}^r(M)\) is defined as \(Z^r(f)=\{g\in \text{Diff}^r(M): fg=gf \}\). The question is: Consider the set of \(C^r\) diffeomorphisms from \(\text{Diff}^r(M)\) with trivial centralizer. {\parindent=6mm \begin{itemize}\item[(1)] Is this set dense in \(\text{Diff}^r(M)\) ? \item[(2)] Is it residual in \(\text{Diff}^r(M)\) ? \item[(3)] Does it contain an open and dense subset of \(\text{Diff}^r(M)\)? \end{itemize}} For the case \(r=1\) the authors give the following answer: For any compact, connected manifold \(M\), there is a residual subset of \(\text{Diff}^1(M)\) consisting of diffeomorphisms with trivial centralizer. This set does not contain any open and dense subset. Let us note that knowing the centralizer of a diffeomorphism gives answer to more concrete questions as well, such as the embeddability of a diffeomorphism in a flow and the existence of roots of a diffeomorphism.