Two proofs of a Jantzen conjecture for Whittaker modules (Q6958778)

The authors state and prove a Jantzen conjecture for Whittaker modules, generalizing results of Beilinson-Bernstein. They provide two proofs, one algebraic (relating Whittaker modules to Verma modules in category \(\mathcal O\)), and one geometric (using mixed twistor \(\mathcal D\)-modules). The Jantzen conjectures establish that a certain algorithmically defined filtration of a Verma module satisfies remarkable functoriality and semisimplicity properties. The conjectures were proven by Beilinson-Bernstein using mixed structures on categories of holonomic \(\mathcal D\)-modules. The authors define an analogous filtration on a standard Whittaker module. The main result is that their filtration satisfies the same functoriality and semisimplicity properties as the Jantzen filtration of a Verma module. Their theorem contains the Jantzen conjectures for Verma modules as a special case. One notable aspect of their geometric proof is that it relies on the machinery of mixed twistor \(\mathcal D\)-modules. Mixed twistor \(\mathcal D\)-modules can be seen as an instance of Simpson's meta theorem applied to Saito's mixed Hodge modules. The authors work provides one of the first applications of this technology to geometric representation theory. The main result is: (The Jantzen conjectures for Whittaker modules) (i) (Hereditary property) Embeddings of standard Whittaker modules are strict for Jantzen filtrations. (ii) (Semisimplicity property) The filtration layers of the Jantzen filtration of standar Whittaker modules are semisimple.