Let $\Gamma$ be a finitely generated group and let $A=\mathbb{C}[\Gamma]$ be the corresponding group algebra over $\mathbb{C}$. Let $X$ be the set of all maximal left ideals of $A$ and let $X_0=\{I \in X : \,\, A/I \textrm{ is finite dimensional } \mathbb{C} \textrm{-vector space}\}.$

Now consider $J= \bigcap_{I \in X} I$ and $J_0= \bigcap_{I \in X_0} I$. It is easy to see that $J$ and $J_0$ are two-sided ideals. The ideal $J$ is the Jacobson radical of $A$ and it has been proven that it is the zero ideal.

My question is the following: For which groups $\Gamma$ is $J_0$ also the zero ideal?

Is there some literature on this question? I am interested in necessary and sufficient criteria as well as in examples, but even more in non-examples.

What I know so far: $J_0$ is the zero ideal:

-if $\Gamma$ is finite (that's trivial),

-if $\Gamma$ is commutative,

-if $\Gamma$ is the free group in finitely many generators.

  • $\begingroup$ It will be 0 iff the group is residually finite. I think Taft or Passman proved this. $\endgroup$ – Benjamin Steinberg Jun 27 '14 at 20:50

This question is the same as asking when does $\mathbb C\Gamma$ have enough finite dimensional irreducible representations to separate points.

A necessary condition is that $\Gamma$ has enough finite dimensional irreps to separate points. Since a fg linear group is residually finite by a theorem of Malcev, having enough finite dim irreps to separate points implies residually finite.

Conversely, if $\Gamma$ is residually finite and $0\neq a\in \mathbb C\Gamma$ then there is a finite image $G$ of $\Gamma$ such that the image of $a$ in $\mathbb CG$ is nonzero (since the support of $a$ involves only finitely many elements of $\Gamma$). Since for finite $G$ we have the regular rep of $\mathbb CG$ is a direct sum of irreps we can find a finite dim irrep in which $a\neq 0$ (by projecting to an appropriate irreducible summand of the regular rep).

Thus the intersection of the cofinite dimensional maximal left ideals is $0$ iff $\Gamma$ is residually finite.


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.