5
$\begingroup$

Let $ V $ be a finite dimensional representation $ \pi $ of a group $ G $. For any $ k $, there is a natural representation of $ G $ on $$ V^{\otimes k} $$ There is also a natural representation of $ G $ on $$ Sym^k(V) $$ If $ V $ is $ n $ dimensional then, after a choice of basis, $ Sym^k(V) $ can be thought of as the space of homogeneous degree $ k $ polynomials in $ n $ variables. Some polynomials are left invariant by this natural action of $ G $. Such a polynomial is called an invariant of $ G $ (really an invariant of the representation $ \pi $). The span of such a polynomial is a trivial irreducible representation of $ G $. Now what if we have a polynomial whose span is invariant under the $ G $ action but the polynomial itself is not invariant with respect to the $ G $ action. In other words, the span of the polynomial is a nontrivial irreducible one. dimensional representation of $ G $.

Is there a name for a polynomial which is a "twisted" (transforms as a nontrivial character) invariant in this sense?

The (regular) invariants form a (finitely generated for many cases of interest) graded subring of $$ Sym(V)=\oplus_k Sym^k(V) $$

These "twisted" invariants would be closed under sums but almost never closed under products. It seems like a major drawback that there is no ring structure for such a theory of "twisted" invariants.

But if we take all "twisted" invariants for all 1d irreps of $ G $ together then this should be a ring since the product of two characters is another character.

Question: Is there a theory of these "twisted" invariants, say for $ G $ a finite group and $ V\cong \mathbb{C}^n $? If so where would I go to learn about such a thing? Also does anyone study the ring of "twisted" invariants I describe above?

$\endgroup$
1
  • $\begingroup$ Oh of course! Feel free to add that as an answer. $\endgroup$ Aug 5 at 18:13

1 Answer 1

4
$\begingroup$

This is exactly the ring of invariants of the commutator subgroup $[G, G]$, with an additional grading given by the character group $\text{Hom}(G, \mathbb{C}^{\times})$ of $G$, so all the standard invariant theory stuff continues to apply. A well-known example is $G = S_n$ acting on $\mathbb{C}^n$ by permutation matrices, whose commutator subgroup is $A_n$, and where the extra "twisted" invariant polynomial is the Vandermonde determinant.

$\endgroup$
4
  • $\begingroup$ I'm loving this Vandermonde determinant example. Feel free to leave the answer as is, but of course I would love to hear any other concrete examples that come to mind for you! $\endgroup$ Aug 5 at 18:23
  • 1
    $\begingroup$ This is essentially the only example I understand so that would be difficult :P The Wikipedia article says that the $S_n$ example is a special case of a more general phenomenon related to Weyl groups and the Weyl character formula but I can't claim to understand this material in enough detail to explain it. $\endgroup$ Aug 5 at 18:28
  • 1
    $\begingroup$ That's fine! Also just wanted to say I'm a huge fan of yours and have been reading your posts on MSE since I was a teenager, your stuff is so good! I definitely appreciate so many of your great posts and I know a lot of other people do too! In fact my officemate math.stackexchange.com/users/444923/geoffrey-sangston walked in right as I was posting here and we both had a complete fanboy moment when you answered my question. Hope you keep doing what you're doing, you are awesome! $\endgroup$ Aug 5 at 18:45
  • $\begingroup$ @Ian: aw thanks, I appreciate that! $\endgroup$ Aug 5 at 21:31

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.