Automorphisms of the category of free Lie algebras (Q1770511): Difference between revisions

An automorphism \(\varphi\) of a category \(\mathcal C\) is called \textit{inner} if there exists an isomorphism of functors \(s:1_{\mathcal C}\rightarrow \varphi\). This means that for every object \(A\) of the category \(\mathcal C\) there exists an isomorphism \(s_A:A\rightarrow \varphi(A)\) such that \(\varphi(\nu)=s_B\nu s_A^{-1}:\varphi(A)\rightarrow \varphi(B)\) for any morphism \(\nu:A\rightarrow B\) in \(\mathcal C\). Let \(\Theta\) be the variety of all Lie algebras over a field \(P\) and \(\Theta^0\) be the category of free Lie algebras over the field \(P\). A semimorphism in \(\Theta\) is a pair \((\sigma ,\nu ):A\rightarrow B\), where \(A\) and \(B\) are algebras in \(\Theta\), \(\nu :A\rightarrow B\) is a homomorphism of Lie rings, \(\sigma\) is an automorphism of the field \(P\), subject to the condition \(\nu (\lambda a)= \sigma (\lambda) \nu (a)\), where \(\lambda \in P\), \(a\in A\). If \(\sigma=1\) then \(\nu\) is a homomorphism of Lie algebras. Semimorphisms are multiplied componentwise. Thus, if \(\mu :A\rightarrow B\) is a homomorphism, and \((\sigma,\nu_1 ):A\rightarrow A_1\), \((\sigma,\nu_2 ):B\rightarrow B_1\) are semi-isomorphisms, then \((\sigma,\nu_2)(1,\mu)(\sigma,\nu_1)^{-1}= (\sigma,\nu_2)(1,\mu)(\sigma^{-1},\nu_1^{-1})=(1,\nu_2\mu\nu_1).\) An automorphism \(\varphi\) of a category \(\Theta^0\) is called semi-inner if for some \(\sigma\in \Aut(P)\) there is a semi-isomorphism of functors \((\sigma,s):1_{\Theta^0}\rightarrow \varphi\). This definition means that for every object \(F\in \Theta^0\) there exists an semi-isomorphism \((\sigma,s_F);F\rightarrow \varphi(F)\) such that \(\varphi(\nu)=s_{F_2}\nu s_{F_1}^{-1}:\varphi(F_1)\rightarrow \varphi(F_2)\) for any morphism \(\nu:F_1\rightarrow F_2\) in \(\Theta^0\). The main result of the paper is the following: Theorem. Every automorphism of the category \(\Theta^0\) of free Lie algebras over an infinite field \(P\) is a semi-inner automorphism. This solves Problem 3.9 from another work of the authors [Electron. Res. Announc. Am. Math. Soc. 8, 1--10 (2002; Zbl 1016.08007)] for Lie algebras.