0

Let $G$ be a Lie group and $H$ a closed subgroup. Let $\pi_1(G/H)$ denote the fundamental group of the homogeneous space $G/H$. If $\lambda$ is an irrep of $H$, then under what conditions does there exist an irrep $\rho$ of $\pi_1(G/H)$ such that $$ H/\mathrm{ker}(\lambda) = \pi_1(G/H)/\mathrm{ker}(\rho). $$ In other words, under what conditions does there exist an irrep $\rho$ such that the image of $\lambda$ is isomorphic to the image of $\rho$? I am hoping for something geometric in nature (if possible).


If $G$ is simply connected and $H$ is discrete then it is known that $\pi_1(G/H) = H$ and so we can choose $\rho$ to be $\lambda$. However, if we drop the requirement that $H$ be discrete then more generally we have that $\pi_1(G/H) = \pi_0(H)$ (see this answer). It is less clear in this case when the above equation holds.


Context: This equation has come up regarding the "canonical" or "$H$-connection" one can place on the vector bundle $\mathrm{Ind}_H^G \lambda$ (that is, the "twisted vector bundle" underlying the induced representation). In particular $H/\mathrm{ker}(\lambda)$ is the "monodromy group" of the connection and $\pi_1(G/H)/\mathrm{ker}(\rho)$ seems to be related to classification of the flatness of the connection.

Eric Kubischta
  • 688
  • 3
  • 11
  • what is "irrep"? First time I see the term – Masacroso Jan 20 '24 at 12:15
  • @Masacroso It's short for "irreducible representation" – Eric Kubischta Jan 20 '24 at 14:37
  • @EricKubischta - how do you pass between representations of $\pi_1(G/H)$ and $H$? – hm2020 Jan 23 '24 at 14:26
  • Note: If $G$ is a linear algebraic group over the complex numbers and if $\rho: \pi_1(G/H) \rightarrow GL_k(V)$ is a representation it corresponds to a flat connection $(E, D)$, where $E$ is a vector bundle. If $\eta: H \rightarrow GL_k(W)$ is a representation we get a $G$-linearized vector bundle $E(\eta)$. – hm2020 Jan 23 '24 at 16:00
  • @hm2020 This sounds interesting but I am not sure I totally follow. Perhaps you could expand your comment more as an answer? – Eric Kubischta Jan 23 '24 at 23:38
  • @EricKubischta - It seems to me in your post you indicate there are natural functors between $\pi_1(G/H)$-modules and $H$-modules. Do I understand this correctly? If there are, you should include a definition of these functors in your post. – hm2020 Jan 24 '24 at 09:07

0 Answers0