$V$ is a finite dimensional vector space. How to prove that any linear map on subspace of $V$ can be extended to linear map on $V$.

I attempted it by taking the basis of the subspace $W$ with $\text{Dim}(m)$ and extending it to basis of $V$ with $\text{Dim}(n)$. Since mapping of all vectors of basis of $W$ exists and taking the remaining $m-n$ vectors as zero - we will get a linear map to an element that is in $V$. But is it sufficient to prove that all vectors in $V$ will also be having a linear map?


If $\{w_1,\ldots,w_m\}$ is a basis of $W$, all you need to be able to do is show that there exist vectors $v_{m+1},\ldots,v_n \in V$ with the property that $\{w_1,\ldots,w_m,v_{m+1},\ldots,v_n\}$ is a basis of $V$. This is a pretty standard result in linear algebra.

The extension you propose $$\tilde L (v) = \tilde L (\alpha_1 w_1 + \cdots + \alpha_m w_m + \alpha_{m+1}v_{m+1} + \cdots \alpha_n v_n) = L(\alpha_1 w_1 + \cdots + \alpha_m w_m)$$ is linear and uniquely defined for all $v \in V$.

  • $\begingroup$ @ Umberto Please correct me if I'm wrong but a linear map on V should be defined for all elements of V. But does the extension in the answer not limit to vectors in V with $ v_{m+1}=.....=v_n= 0 $ thereby excluding vectors in V where these are not zero $\endgroup$
    – Hashtag
    Sep 12 '13 at 18:31
  • $\begingroup$ Since $\{w_1,\ldots,w_m,v_{m+1},\ldots,v_n\}$ is a basis of $V$, every vector $v \in V$ has a unique representation $v = \alpha_1 w_1 + \cdots + \alpha_m w_m + \alpha_{m+1}v_{m+1} + \cdots \alpha_n v_n$. $\endgroup$
    – Umberto P.
    Sep 13 '13 at 13:06
  • 1
    $\begingroup$ Is the extension unique? Or there are other possible extensions? $\endgroup$ Aug 9 '16 at 20:29
  • $\begingroup$ Say you have a map on the vector space of the x-axis that multiplies everything by 2. You can extend this to the whole plane to a map that multiplies everything by 2, or a map that multiplies only the x-axis by 2, etc. $\endgroup$
    – Dapianoman
    Nov 3 '19 at 5:05

Suppose $L:W\to W$ is the linear map that you want to extend to whole of $V$. First consider a basis $\{w_1, \ldots, w_m\}$ of $W$ and extend this to a basis $\{w_1, \ldots, w_m, v_{m+1}, \ldots, v_n\}$ of $V$. Now define $\widetilde{L}:V\to V $ as follows: we define the action of $\widetilde{L}$ on basis vectors by,

$$\widetilde{L}(w_1)= L(w_1),$$ $$\widetilde{L}(w_2)= L(w_2),$$ $$\vdots$$ $$\widetilde{L}(v_{m+1})=0,$$ $$\vdots$$ $$\widetilde{L}(v_n)= 0,$$ and on any arbitrary vectors $v\in V$, we define the action of $\widetilde{L}$ on $v=\alpha_1w_1+ \ldots + \alpha_mw_m + \alpha_{m+1}v_{m+1} + \ldots + \alpha_n v_n$, by $$\widetilde{L}(v):= \alpha_1\widetilde{L}(w_1) + \ldots + \alpha_m\widetilde{L}(w_m) + \alpha_{m+1}\widetilde{L}(v_{m+1})+\ldots+ \alpha_n\widetilde{L}(v_n).$$

Its clear that $\widetilde{L}|_W = L$ and it is easy to check that $\widetilde{L}$ is a linear map (defined on whole of $V$).


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.