The Stacks project

11.6 Skolem-Noether

Theorem 11.6.1. Let $A$ be a finite central simple $k$-algebra. Let $B$ be a simple $k$-algebra. Let $f, g : B \to A$ be two $k$-algebra homomorphisms. Then there exists an invertible element $x \in A$ such that $f(b) = xg(b)x^{-1}$ for all $b \in B$.

Proof. Choose a simple $A$-module $M$. Set $L = \text{End}_ A(M)$. Then $L$ is a skew field with center $k$ which acts on the left on $M$, see Lemmas 11.3.2 and 11.4.6. Then $M$ has two $B \otimes _ k L^{op}$-module structures defined by $m \cdot _1 (b \otimes l) = lmf(b)$ and $m \cdot _2 (b \otimes l) = lmg(b)$. The $k$-algebra $B \otimes _ k L^{op}$ is simple by Lemma 11.4.7. Since $B$ is simple, the existence of a $k$-algebra homomorphism $B \to A$ implies that $B$ is finite. Thus $B \otimes _ k L^{op}$ is finite simple and we conclude the two $B \otimes _ k L^{op}$-module structures on $M$ are isomorphic by Lemma 11.4.6. Hence we find $\varphi : M \to M$ intertwining these operations. In particular $\varphi $ is in the commutant of $L$ which implies that $\varphi $ is multiplication by some $x \in A$, see Lemma 11.4.6. Working out the definitions we see that $x$ is a solution to our problem. $\square$

Lemma 11.6.2. Let $A$ be a finite simple $k$-algebra. Any automorphism of $A$ is inner. In particular, any automorphism of $\text{Mat}(n \times n, k)$ is inner.

Proof. Note that $A$ is a finite central simple algebra over the center of $A$ which is a finite field extension of $k$, see Lemma 11.4.2. Hence the Skolem-Noether theorem (Theorem 11.6.1) applies. $\square$


Comments (3)

Comment #4168 by Patrick Chu on

lemma 11.6.2 (074R) seems incorrect

Field extensions with non-trivial galois groups have automorphisms which aren't inner, but are still simple -algebras.

Any automorphisms which are linear over the center field are certainly inner by the theorem (so matrices of fields are still fine), but an automorphism could be linear over the original field and not linear over the center.

Comment #4370 by on

Yep, I agree the word "central" was missing. Thanks and fixed here.

Comment #9548 by Kevin Buzzard on

Near the beginning of the proof of 074Q you give M two right B tensor_k L^op-module structures but I have a student trying to type this into Lean and they say that this doesn't seem to work. They suggest that instead you give M two left B tensor_k L-module structures (or two right B^{op} tensor_k L^{op}-module structures).


Post a comment

Your email address will not be published. Required fields are marked.

In your comment you can use Markdown and LaTeX style mathematics (enclose it like $\pi$). A preview option is available if you wish to see how it works out (just click on the eye in the toolbar).

Unfortunately JavaScript is disabled in your browser, so the comment preview function will not work.

All contributions are licensed under the GNU Free Documentation License.




In order to prevent bots from posting comments, we would like you to prove that you are human. You can do this by filling in the name of the current tag in the following input field. As a reminder, this is tag 074P. Beware of the difference between the letter 'O' and the digit '0'.