5
$\begingroup$

For $A$ a finite-dimensional algebra over a field $K$
Does there exist a finite group $G$, such that $A$ is a sub-algebra of $K[G]$ ?
Where $K[G]$ denotes the group-algebra of $G$ over $K$.

In case that the answer is no, would there be a way to "detect" when it is the case ?

I would not mind an answer under some "nice" conditions such as commutativity, associativity, etc.

I tried to do few low-dimensionnal examples through the use of the structure coefficients but it became quickly untractable by hand.

One the one hand, I have a feeling that possibility of taking $G$ as large as one wants would give some trivial construction but did not find any.
On the other hand I feel like it would imply quite a lot and that some algebraic structure such as hypergroups would loose a bit of interest.

$\endgroup$
1
  • 4
    $\begingroup$ If you're fine with a non-unital subalgebra, can't you take a group with an irrep of dimension at least that of $A$, and identify $A$ with a matrix representation? $\endgroup$
    – Kevin
    Aug 2, 2022 at 15:02

1 Answer 1

11
$\begingroup$

Assuming that by "sub-algebra" you mean "unital sub-algebra":

Every group algebra has a one-dimensional module (the trivial module), so any subalgebra has a one-dimensional module.

But many finite-dimensional algebras (e.g., $M_n(K)$ for $k>1$) have no one-dimensional modules.

But there may be more interesting restrictions.

$\endgroup$
4
  • 1
    $\begingroup$ I may not understand what you mean by "having a module" for that I am not that versed in algebra but isn't $K Id_n$ a one-dimensonal module of $M_n(K)$ ? $\endgroup$
    – Hugo MTV
    Aug 2, 2022 at 14:25
  • $\begingroup$ @HugoMTV No, multiplying an element of $K\text{Id}_n$ by an element of $M_n(K)$ doesn't usually give an element of $K\text{Id}_n$. $\endgroup$ Aug 2, 2022 at 14:33
  • $\begingroup$ Nevermind, I thought that "having a module" meant having a sub-module when it meant "having a module over oneself. Thanks $\endgroup$
    – Hugo MTV
    Aug 2, 2022 at 14:39
  • 1
    $\begingroup$ Nice answer! The same argument works for any augmented algebra as well. $\endgroup$ Aug 2, 2022 at 18:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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