Showing the socle of a module exists and is unique

382 Views Asked by At

I have been set the following exercise:

For every $R$-module $M$, show that there exists a unique semi-simple submodule $sM$ $\subset$ $M$ which contains every semi-simple submodule of $M$.

After looking around online, it seems that this is called the socle of $M$? And this question is essentially asking me to show that it exists and is unique?

Am I right in thinking that this is how to the view the problem? And if so, how would I go about showing this?

Thanks in advance!

1

There are 1 best solutions below

0
On BEST ANSWER

The key is the fact that any semisimple module is the sum of its simple submodules and, conversely, the sum of simple submodules is semisimple.

Modules will be left $R$-modules.

Theorem. For a module $M$ the following conditions are equivalent:

  1. $M$ is a direct sum of simple submodules,
  2. $M$ is a sum of simple submodules,
  3. $M$ is the sum of its simple submodules,
  4. every submodule of $M$ is a direct summand.

The implications $(1)\implies(2)\implies(3)$ are obvious. The proof of $(3)\implies(4)$ relies on the following lemma.

Lemma. If $M=\sum_{i\in I}S_i$ with $(S_i)_{i\in I}$ a family of simple submodules of $M$ and $K$ is a submodule of $M$, then there exists $J\subseteq I$ such that

a. $(S_i)_{i\in J}$ is independent (that is, the sum $\sum_{i\in J}S_i$ is direct), and
b. $M=K\oplus\sum_{i\in J}S_i$.

The proof is an application of Zorn's lemma. Take a subset $J\subseteq I$ maximal with respect to $(S_i)_{i\in J}$ being independent and $K\cap\sum_{i\in J}S_i=\{0\}$ (here Zorn's lemma is applied, work it out). Set $N=K+\sum_{i\in J}S_i$: we want to show that $N=M$. Let $k\in I$; since $S_k$ is simple, either $S_i\cap N=\{0\}$ or $S_i\subseteq N$. In the first case $J'=J\cup\{k\}$ would contradict the maximality of $J$, so we have that $S_i\subseteq N$ for all $i\in I$ and so $N=M$.

The lemma also shows that $(2)\implies(1)$, by taking $K=0$, so we only need to prove $(4)\implies(2)$. So, assume every submodule of $M$ is a direct summand. Let $x\in M$, $x\ne0$. Then $Rx$ an has a maximal submodule $H$ (just take a maximal left ideal of $R$ containing $\{r\in R:rx=0\}$). So $M=H\oplus H'$. Then $$ Rx=Rx\cap M=Rx\cap (H\oplus H')=H\oplus(Rx\cap H') $$ (verify it), so $Rx\cap H'\cong Rx/H$ is a simple submodule of $Rx$.

Let now $N$ be the sum of the simple submodules of $M$; then $M=N\oplus N'$; if $N'\ne\{0\}$, then $N'$ would have a simple submodule: contradiction.


As a consequence of this, if $M$ is any module, the sum of its simple submodules is a semisimple submodule of $M$ containing all simple submodules and so all semisimple submodules.