1
$\begingroup$

Let $V$ be an $n$-dimensional $\mathbb{K}$-vector space with basis $\{e_i\}_{i=1}^n$, and consider the map $\phi:V\to \mathbb{K}$ given by $e_i\mapsto 1$. Let $K=\ker\phi$ and set $N$ to be the subspace of vectors whose coefficients are all equal. I'd like to show that $V=K\oplus N$ (provided $char(\mathbb{K})\nmid n$).

So far I've shown that $N\cap K=0$, but how can I write a general $v\in V$ as a sum of elements in $K$ and $N$?

$\endgroup$
  • 2
    $\begingroup$ Is $\phi$ also assumed to be linear? $\endgroup$ – GEdgar Sep 19 '12 at 20:21
  • $\begingroup$ @GEdger: Yes! I forgot to include that $\endgroup$ – Bey Sep 20 '12 at 3:14
2
$\begingroup$

Let $v=a_1e_1+\cdots+a_ne_n$, so $\phi(v)=a_1+\cdots+a_n$. Let $x=\frac{a_1+\cdots+a_n}{n}$, $u=(a_1-x)e_1+\cdots+(a_n-x)e_n$ and $w=xe_1+\cdots+xe_n$. Then $v=u+w$ and $u\in K$ while $w\in N$.

| cite | improve this answer | |
$\endgroup$
  • $\begingroup$ Thanks! There's a small typo in your $w$; the first term should be $xe_1$, not $x_e1$ :) $\endgroup$ – Bey Sep 19 '12 at 19:57
  • $\begingroup$ @Bey Thanks for pointing that out. $\endgroup$ – Alex Becker Sep 19 '12 at 20:15
  • 1
    $\begingroup$ Alternatively, one may show $N\cap K = \{0\}$ (NOT $=\emptyset$ as is written in the original question) and determine the dimensions of $K$ and $N$. $\endgroup$ – Oliver Braun Sep 19 '12 at 20:30
  • $\begingroup$ Oh whoops. Thanks for that $\endgroup$ – Bey Sep 20 '12 at 3:10

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.