4
$\begingroup$

Let $R$ be a ring.

Qeustion: Is it true that the first Hochschild cohomology of $R$ is zero if and only if the outer automorphism group of $R$ is finite?

(It is not true, by the two answers. Is it at least true that the Hochschild cohomology is zero if the outer automorphism group is finite?)

This is true for finite dimensional algebras but possibly true in a much more general sitation (maybe even for more general categories than module categories of rings). Is there a suitable reference for this (at least for finite dimensional algebras)?

$\endgroup$
4
  • $\begingroup$ How is the outer automorphism group of a ring (= associative unital ring?) defined? $\endgroup$
    – YCor
    Commented Oct 12, 2021 at 11:49
  • $\begingroup$ @YCor The automorphism group of the ring modulo the normal subgroup of inner automorphisms. $\endgroup$
    – Mare
    Commented Oct 12, 2021 at 11:55
  • $\begingroup$ What is an inner automorphism of a ring $R$? Is it the set of $x\mapsto uxu^{-1}$ for $u$ ranging over invertible elements of $R$? $\endgroup$
    – YCor
    Commented Oct 12, 2021 at 14:06
  • $\begingroup$ @YCor Yes, that is correct. $\endgroup$
    – Mare
    Commented Oct 12, 2021 at 15:43

2 Answers 2

5
$\begingroup$

Another easy counter-example: take $X = \mathbb{N}$ as discrete topological space and $R = C(X, \mathbb{R})$ as continuous functions on it. These are just all functions. Equivalently, you can view them as smooth functions on the smooth manifold $X$. Now the first Hochschild cohomology is trivial since $X$ is zero-dimensional. Any automorphism is outer since $R$ is commutative. The pull-back with any bijection of $X$ gives an outer automorphism, quite many.

$\endgroup$
2
  • $\begingroup$ You can just take $X$ to be any topological space with I infinitely many homeomorphisms. planetmath.org/derivationsonaringofcontinuousfunctions $\endgroup$ Commented Oct 12, 2021 at 21:30
  • $\begingroup$ @BenjaminSteinberg Yes, of course. This will give a huge class of examples. I just wanted to indicate one particularly simple one. $\endgroup$ Commented Oct 13, 2021 at 6:43
4
$\begingroup$

I don't believe this is true. Consider $D_{\mathbb{A}_{\mathbb{C}}^{1}}=:R$. Then we have $HH^{1}(R)=0$ by a standard computation. In fact $HH^{*}(D_{X})\cong H^{*}_{dR}(X)$ holds more generally. Now $Out(R)=Aut(R)$ as $R^{*}=\mathbb{C}^{*}$ is central. $Aut(R)$ includes all the transformations $$x\mapsto x+f(\partial), \partial\mapsto \partial,$$ where $f$ is any polynomial in one variable.

$\endgroup$
2
  • $\begingroup$ You mean $=\colon$. $\endgroup$ Commented Oct 12, 2021 at 13:43
  • $\begingroup$ Sure. Edited now $\endgroup$
    – user108998
    Commented Oct 12, 2021 at 13:45

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .