Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

Question: Let $V$ be a $K$-Vectorspace and $f: V \rightarrow V$ be linear. It holds that $f \circ f = f$. Show that $V = \mbox{ker}(f) \oplus \mbox{im}(f)$.

My attempt: So i guess that the $\oplus$ denotes a direct sum which means i have to show that (i) $V = \mbox{ker}(f) + \mbox{im}(f)$ and (ii) $\mbox{ker}(f) \cap \mbox{im}(f) = \{0\}$.

I tried to do (ii) first: Let $v \in \mbox{im}(f) \cap \mbox{ker}(f)$

$\Rightarrow \exists u: f(u)=v \wedge f(v) = 0$

(can i put a "Rightarrow" here?) $(f \circ f)(u)=f(f(u)) = f(v) = 0$

As for (i) i am having difficulty with an approach to showing that $V = \mbox{ker}(f) + \mbox{im}(f)$. Should I even be trying to do this in the first place? if so, any advice as to how?

share|cite|improve this question
up vote 6 down vote accepted

Let's see (ii) first: (Yes, your use of the implication symbol $\Rightarrow$ is appropriate.) You want to use that $f\circ f=f$, which you do not mention explicitly. The point is: In (ii) you need to show that the $v$ you are considering is $0$. You have: $f\circ f=f$, so $v=f(u)=(f\circ f)(u)$ etc, as you did. This gives $v=0$.

For (i), let $v$ be any element of $V$. What can we say about $v-f(v)$? Let's call it $w$. We have $f(w)=f(v-f(v))=f(v)-f(f(v))=f(v)-f(v)=0$. Here, I used that $f$ is linear and (again) that $f\circ f=f$. This shows that $w\in{\rm ker}(f)$. So, $v=(v-f(v))+f(v)=w+f(v)$ is the sum of an element of the kernel (namely, $w$) and one of the image (namely, $f(v)$), and that is precisely what you needed.

(A meta-question here is how would one think of considering $v-f(v)$. Ok, let's try to work backwards. Suppose we already know that $v=a+b$ with $a$ in the kernel and $b$ in the range. Then $f(v)=f(a)+f(b)=f(b)$, since $a$ is in the kernel. Also, $f(v)=f(f(v))$, so it is reasonable, as a first approach, to let $b=f(v)$, and see if we can directly prove that $a=v-b$ is in the kernel. And this is precisely what we did.)

share|cite|improve this answer
FYI: use \ker f ($\ker f$) instead of {\rm ker}(f) $({\rm ker}(f)$) – kahen Nov 27 '10 at 21:32
@kahen : Thanks for the tip. – Andrés E. Caicedo Nov 27 '10 at 21:34

$v = (v - f(v)) + f(v)$

share|cite|improve this answer

For 1) let v be an arbitrary element in the intersection of ker f and image f. This second statement, as you have pointed out, means that v=f(u) for some u, and the first statement, means that f(v) = 0. So 0=f(v)=f(f(u)) which is f(u) (or v) by assumption. This shows v=0.

For 2) I believe you can do this with nothing more than set theory. Just prove V is a subset of ker(f) + im(f) and then prove the reverse inclusion (this second inclusion is trivial since ker(f) and im(f) are clearly both subsets of V)

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.