0
$\begingroup$

Can someone give a precise definition of a direct summand of an $R$-module $M$. (You can assume $R$ is commutative with unity).

Here is what I thought till date:

"We say an $R$-module $N$ is a direct summand of $M$ if there exists an $R$-module $N'$ such that $M$ is isomorphic to $N \oplus N'$".

But while going through some articles on commutative algebra/ homological algebra, I don't think this is taken as a defintion.

For example, in my definition, $2\mathbb{Z}$ is a direct summand of $\mathbb{Z}$. But is it really so according to standard literature? Can someone point me to a definition of direct summand in some popular textbook?

$\endgroup$
2
  • $\begingroup$ A submodule $N$ is a direct summand of $M$ if there exists a submodule $N'$ of $M$ such that $M$ is the (internal) direct sum $N\oplus N'$, i.e., for every $m\in M$ there exists unique $(n,n')\in N\times N'$ such that $m=n+n'$. $\endgroup$ Apr 11 at 14:53
  • $\begingroup$ Your definition is good. To be extremely precise, one should probably distinguish between internal direct summands and external direct summands but since this is usually/always clear from the context, it isn't done. In your case, yes, $2 \mathbb{Z}$ is an (external) direct summand but I wouldn't phrase it that way as the $2$ can be confusing and the important thing here is rather that $2 \mathbb{Z}$ is embedded via $2 \mathbb{Z} \cong \mathbb{Z}$ in $\mathbb{Z}$. $\endgroup$
    – Qi Zhu
    Apr 12 at 6:10
1
$\begingroup$

A submodule $A$ of $B$ is a direct summand if there exists a $C$ such that $B\cong A\oplus C$ where the canonical map of the first summand is the inclusion map.

$\endgroup$
3
  • $\begingroup$ So does direct summand always refer to submodules? Can you give a reference to this definition? $\endgroup$
    – Anupam
    Apr 11 at 14:57
  • $\begingroup$ @Anupam That depends on the context, for most purposes it suffices to say that it is any module $A$ for which there exists a module $C$ such that $B \cong A \oplus C$. (And certainly, one could then identify $A$ with a submodule of $B$.) Ok, I just saw that this was exactly your textbook definition as well. $\endgroup$
    – Qi Zhu
    Apr 12 at 5:39
  • $\begingroup$ @Qi In Rotman's "Advanced Modern Algebra", the definition (or rather one of the equivalent criteria) for a submodule to be a direct summand is if there exists a retract onto the submodule (i.e., if the inclusion map splits). So in that setup, I believe we should better talk in terms of internal direct sums. $\endgroup$
    – Anupam
    Apr 12 at 8:32

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.