Sur des représentations indécomposables de dimension finie de \(SL(2).{\mathbb{R}}^ 2\). (On indecomposable finite dimensional representations of \(SL(2).{\mathbb{R}}^ 2)\) (Q1089084)

Let \({\mathfrak g}.a\) be a semi-direct product of a semi-simple Lie algebra \({\mathfrak g}\) and an abelian algebra a. Consider its embedding into a semisimple algebra \({\mathfrak f}\). A dilatation relative to this embedding is an element \(d\in {\mathfrak f}\) such that the restriction of ad(d) on a is identical and [d,\({\mathfrak g}]=0\). Such dilatations are of interest to physicists in the case of \({\mathfrak g}.a\) being the Poincaré group and \({\mathfrak f}\) being a conformal group. The author studies the ''simplest'' case \({\mathfrak g}=sl(2,{\mathbb{R}})\), \(a={\mathbb{R}}^ 2\) as a model example. He proves, in particular, that the embedding sl(2).a\(\hookrightarrow sl(35)\) possesses no relative dilatations. Other interesting results along these lines are also obtained in the paper under review. For example, an embedding j: sl(2).a\(\hookrightarrow sl(2n+1)\) exists such that j(sl(2).a) is not contained in any proper semisimple subalgebra of \(sl(2n+1)\). The way to such examples is long and sufficiently difficult. An indecomposable representation \(\phi\) : sl(2).a\(\to End(V)\) is said to be cyclic if there exists an irreducible sl(2)-invariant complement to \(\phi\) (a).V in V. All cyclic representations are completely classified, they are parametrized by the arbitrary rows of natural numbers. The proof of this fact is done by direct but elaborate calculations. The author also obtains some partial results on the classification of indecomposable representations in the far more complicated non-cyclic case. As described above, this ''hard work'' yields various interesting examples.