Spin module versions of Weyl's reciprocity theorem for classical Kac- Moody Lie algebras. An application to branching rule duality (Q912964): Difference between revisions

In this big paper the author studies tensor products of the spin modules (Fermion Fock space representations) for classical (simple or affine) Kac-Moody Lie algebras. He finds that there are mutually commutant pairs of classical Kac-Moody Lie algebras acting on the spin modules, and describes the irreducible decompositions in terms of Young diagrams. For a finite dimensional reductive complex Lie algebra \({\mathfrak g}\), let \({\mathfrak g}^\wedge ={\mathfrak g}\otimes {\mathbb{C}}[t,t^{-1}]\oplus {\mathbb{C}}c\) be the corresponding affine Lie algebra. Let W be an N-dimensional vector space equipped with a nondegenerate symmetric bilinear form \(\{.,.\}\), and put \(W^{{\mathbb{Z}}+h}=W\otimes t^ h{\mathbb{C}}[t,t^{-1}]\) for \(h=0\) or 1/2. The bilinear form \(\{.,.\}\) in \(W^{{\mathbb{Z}}+h}\) is defined by \(\{w\otimes t^{\nu},w'\otimes t^{\nu'}\}=\{w,w'\}\delta_{\nu +\nu ',0}.\) Consider the exterior algebra \(\wedge (W^{{\mathbb{Z}}+h,-})\) over the maximal isotropic subspace \(W^{{\mathbb{Z}}+h,-}\) with respect to the above inner product. This exterior algebra has the spin module structure over \({\mathfrak o}(W)^\wedge\), and is called the Fermion Fock space of the Ramond (if \(h=0)\) or the Neveu-Schwarz (if \(h=1/2)\) sector. Next the Lie algebra pairs (\({\mathfrak g}_ L,{\mathfrak g}_ R)\) and their embeddings \({\mathfrak g}_ L\oplus {\mathfrak G}_ R\hookrightarrow {\mathfrak o}(W)\) are given as follows. \[ (1)\quad {\mathfrak g}_ L\oplus {\mathfrak g}_ R={\mathfrak gl}(V_ 1)\oplus {\mathfrak sl}(V_ 2),\quad W=(V_ 1\otimes V_ 2)\oplus (V_ 1\otimes V_ 2)^*; \] \[ (2)\quad {\mathfrak g}_ L\oplus {\mathfrak g}_ R={\mathfrak sp}(V_ 1)\oplus {\mathfrak sp}(V_ 2),\quad W=V_ 1\otimes V_ 2; \] \[ (3)\quad {\mathfrak g}_ L\oplus {\mathfrak g}_ R={\mathfrak o}(V_ 1)\oplus {\mathfrak o}(V_ 2),\quad W=V_ 1\otimes V_ 2. \] The problem is to decompose the \({\mathfrak o}(W)^\wedge\)-spin module \(\wedge (W^{{\mathbb{Z}}+h,-})\) \((h=0,1/2)\) with respect to the subalgebra (\(\mathfrak g_ LL+\mathfrak g_ R)^\wedge \hookrightarrow \mathfrak o(W)^\wedge\). The author solves this problem completely as well as its finite dimensional version. If we use the notion of the ``dual pairs'' due to R. Howe, the result can be summarized as follows. In each case, the pair \[ ({\mathbb{C}}[\Sigma {\hat{\;}}_ L]\ltimes U({\mathfrak g}{\hat{\;}}_ L),\quad {\mathbb{C}}[\Sigma {\hat{\;}}_ R]\ltimes U({\mathfrak g}{\hat{\;}}_ R)) \] forms a dual pair on \(\wedge (W^{{\mathbb{Z}}+h,-})\), where \(\Sigma^\wedge_*\) is the group consisting of Dynkin diagram automorphisms corresponding to \({\mathfrak g}^\wedge_*\). As applications the author describes the ``branching rule duality'' and the ``S-transformation rule duality''.