scientific article; zbMATH DE number 16286

From MaRDI portal
Publication:3972939





[For the entire collection see Zbl 0719.00018.] Author's introduction: Let \({\mathfrak g}\) be a complex Lie algebra with enveloping algebra \(U({\mathfrak g})\) and symmetric algebra \(S({\mathfrak g})\). Let \(G\) be a simply connected group with Lie \(G= {\mathfrak g}\). The Nullstellensatz provides a tight link between ideal theory in \(S({\mathfrak g})\) and geometry in \({\mathfrak g}^*\); more precisely, it yields a bijection between prime quotients of \(S({\mathfrak g})\) and irreducible affine varieties in \({\mathfrak g}^*\). A basic goal in any reasonable theory of noncommutative algebraic geometry is to extend this bijection to a (slightly) noncommutative setting; a natural one to consider is that of \(U({\mathfrak g})\). For solvable \({\mathfrak g}\), Dixmier extended the bijection to \(U({\mathfrak g})\) as follows: he showed that primitive ideals of \(U({\mathfrak g})\) correspond \(1-1\) to \(G\)-orbits in \({\mathfrak g}^*\). We call this the orbit correspondence. It reduces the classification of the primitive spectrum of \(U({\mathfrak g})\) to a purely geometric question which in many cases admits a complete solution. In view of the power and generality of the orbit correspondence in the solvable case, it was natural to believe that it should extend to the semisimple case without too much difficulty. Twenty years of effort have unfortunately shown that there are fundamental obstacles to its implementation for semisimple \({\mathfrak g}\). In this paper, the author describes these obstacles in some detail and shows how the attempt to overcome them has led to the notion of a Dixmier algebra, roughly a finite algebra extension of a primitive quotient of \(U({\mathfrak g})\). He surveys the work done so far on Dixmier algebras, briefly reviewing the methods currently available for constructing them and concludes the paper by giving a number of interesting examples of Dixmier algebras. Most of these algebras are completely prime and have nice geometric properties, but a few of them are not. The latter algebras are, however, closely related to certain algebras which we call unipotent, so we call them mock unipotent. Mock unipotent Dixmier algebras help to illustrate a basic principle of any orbit correspondence: only completely prime Dixmier algebras can hope to have a nice geometric theory.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3972939)