4
$\begingroup$

Let $G_\mathbb{C}$ be a connected reductive group over $\mathbb{C}$ with Lie algebra $\mathfrak{g}_{\mathbb{C}}$. For any algebraically closed field $k$, let $G_k$ denote the connected reductive group with the same root datum as $G_{\mathbb{C}}$. Let $\mathfrak{g}_k$ denote the corresponding Lie algebra.

Question: Suppose the characteristic of $k$ is large enough. Is it true that nilpotent orbits in $\mathfrak{g}_{\mathbb{C}}$ are in (canonical?) bijection with nilpotent orbits in $\mathfrak{g}_k$?

My impression is that various classification theorems (such as Bala-Carter or Dynkin) which are classical for $\mathbb{C}$ work equally well if $k$ has large characteristic. This indicates that the answer to above question is YES. Am I right? If so, can one give a proof that avoids the classification theorems?

$\endgroup$
6
  • 2
    $\begingroup$ I think it should be true in good characteristic, and you can take a look at results in "A. Premet; Nilpotent orbits in good characteristic and the Kempf-Rousseau theory; J. Algebra 260 (2003), no. 1, 338-366." for proofs avoiding case-by-case considerations. $\endgroup$ Commented Feb 20 at 4:14
  • $\begingroup$ @testaccount I did look at that paper, but it wasn't clear how to conclude what I wanted from their discussions. If you have digested that, it would be great if you share your thoughts. $\endgroup$
    – Dr. Evil
    Commented Feb 20 at 4:50
  • 1
    $\begingroup$ The statement "there are exactly $n$ nilpotent orbits" can be characterized by one first-order sentence (in the language of fields). Hence, being true for $\mathbf{C}$, it is true in large enough characteristic ($p>p_0(G)$ depending on $G$, with no explicit estimate on $p_0(G)$). $\endgroup$
    – YCor
    Commented Feb 20 at 7:39
  • $\begingroup$ @YCor It would be great if you could expand on this in an answer and put references for beginners. $\endgroup$
    – Dr. Evil
    Commented Feb 20 at 22:48
  • 1
    $\begingroup$ Just to support the comment by @testaccount. In Premet's paper, the class of nilpotent orbits "coming from" characteristic zero (i.e. Dynkin's classification via weighted Dynkin diagrams) are called "standard". He proves (Thm. 2.7) that all nilpotent orbits are standard in good positive characteristic ($>5$ will do in all types). He also establishes Pommerening's theorem (i.e. Bala-Carter, Thm. A) and Sommers' generalisation of Bala-Carter (Thm. B). Premet's paper wasn't the first to prove these results, but his uniform arguments make it a good reference (and see his paper for more history). $\endgroup$
    – Paul Levy
    Commented Feb 23 at 10:27

0

You must log in to answer this question.

Browse other questions tagged .