Prove or disprove the following statement:

"If $(v_1,v_2,v_3,v_4)$ is a basis for the vector space ${\Bbb R}^4$ and $\textbf{W}$ is a subspace of ${\Bbb R}^4$ then some subset of $(v_1,v_2,v_3,v_4)$ forms a basis for $\textbf{W}$."

I have an intuition for this and I think it is false because it's saying that I can write $(v_1,v_2,v_3,v_4)$ as a linear combination of each other since the vectors form a basis. However, just because $\textbf{W}$ is in he subspace of ${\Bbb R}^4$ does not mean I can write is with linear combinations of $(v_1,v_2,v_3,v_4)$. There is no guarantee that I can completely span $\textbf{W}$ with $(v_1,v_2,v_3,v_4)$ right?

Is that even the right approach or is my thinking completely off? How would I actually do this mathematically?

  • $\begingroup$ Your intution has not come out in the words you have written. Rewrite the second sentence starting with "However, just because...". $\endgroup$ – P Vanchinathan Feb 3 '18 at 7:40

Your idea is right. A concrete counter example would be good.

Let $v_i = e_i$, the $i$-th standard unit vector.

Let $W = \operatorname{span}\{ v_1 + v_2\}$, then none of the $\{e_i\}$ is a basis for $W$.

  • $\begingroup$ I'm just going to add step to be very concrete about this. So Essentially, you're saying that let $v_i=e_i= \begin{bmatrix} \vdots \\ \vdots \\ 0 \\ 1 \\ 0 \\ \vdots \\ \vdots \\ \end{bmatrix}$ and , $v_1=e_1= \begin{bmatrix} 1 \\ 0 \\ 0 \\ \vdots \\ \vdots \\ \end{bmatrix}$ $v_2=e_2= \begin{bmatrix} 0 \\ 1 \\ 0 \\ \vdots \\ \vdots \\ \end{bmatrix}$. However, if I add $e_1$ and $e_2$ it's impossible for me to get $e_i$ since I can't write $e_i$ as a linear combination of $e_1$ and $e_2$. Is this correct? $\endgroup$ – Future Math person Feb 3 '18 at 7:46
  • $\begingroup$ Clearly the dimension of $\operatorname{span} \{ v_1+v_2\}$ is one. hence if the subset of $\{ v_1, \ldots, v_4\}$ is a basis, it has to be of the form of $\{ e_i \}$, however, we can easily check that none of them can be a basis. since after all, none of them is inside $W$. $\endgroup$ – Siong Thye Goh Feb 3 '18 at 7:54
  • $\begingroup$ I understand this now. Thank you so much! $\endgroup$ – Future Math person Feb 3 '18 at 8:20

Actually, since the set $\{v_1,v_2,v_3,v_4\}$ has $2^4=16$ subsets, there are only $16$ subspaces of $\mathbb{R}^4$ which have a basis which is a subset of $\{v_1,v_2,v_3,v_4\}$. Since $\mathbb{R}^4$ has infinitely many subspaces…


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.