Applications of a noncommutative Jacobian matrix

From MaRDI portal





Let \(F=k\langle \langle X \rangle \rangle\) be the free power series ring in a set \(X\) over a field \(k\), and for any \(x \in X\) and \(f \in F\) denote by \(x^{-1}f\) the right cofactor of \(x\) in \(f\). If \(X,Y\) are any non- empty sets of cardinals \(\alpha,\beta\) respectively and \(\varphi:k\langle \langle X\rangle \rangle \to k \langle \langle Y \rangle \rangle\) is a continuous homomorphism, its Jacobian matrix over \(k\langle \langle Y \rangle \rangle\) is defined as the \(\beta \times \alpha\) matrix \(J(Q)=(y^{-1}\varphi(x))\), where \(x \in X\), \(y \in Y\). For a continuous automorphism \(\varphi\) of \(F\), \(J(\varphi)\) is invertible, and the same holds for automorphisms of the free algebra \(k\langle X \rangle\). Easy examples show that the converse does not hold for \(k\langle X \rangle\), but as the author shows, it does hold for the free Lie algebra \({\mathcal L}(X)\) on \(X\). For the proof he shows that any endomorphism of \({\mathcal L}(X)\) with invertible Jacobian matrix is a product of elementary automorphisms (i.e. tame); this also provides another proof of the reviewer's theorem that any automorphism of \({\mathcal L}(X)\) is tame [\textit{P. M. Cohn}, Proc. Lond. Math. Soc., III. Ser. 14, 618--632 (1964; Zbl 0142.27704)]. As a corollary he finds that a free Lie algebra of finite rank is Hopfian. In the case of a free algebra of rank two the author obtains a normal form for continuous automorphisms of \(k\langle x,y \rangle\) analogous to the form for \(mathrm{GL}_2(R)\) described in [\textit{P. M. Cohn}, Publ. Math., Inst. Hautes Étud. Sci. 30, 365--413 (1966; Zbl 0144.26301)], using the map \(x\to y+f\), \(y\to x\), where \(f=f(x)\) is a polynomial in \(x\) without constant term. Finally he describes an inversion formula for continuous automorphisms of \(F\). Let \(\varphi\) be such an automorphism, with Jacobian matrix \(J\) and for each \(x\in X\) define an \(\varepsilon\)-derivation \(D_ x\) by \(D_ x(S)=\sum_ yJ_{x,y}^{- 1}(y^{-1}S)\) (where \(y\) runs over \(X\) and \(\varepsilon\) is the augmentation map). Then for each word \(w=x_ 1...x_ n\) and each \(S\in F\), one has \((\varphi^{-1}(S),w)=(D_{x_ n} \circ ... \circ D_{x_ 1}(S),1)\) [cf. \textit{W. Gröbner}, Monatsh. Math. 66, 129--139 (1962; Zbl 0128.30401)].




Cited in
(28)








This page was built for publication: Applications of a noncommutative Jacobian matrix

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1184034)