This question came in my exam, and is by no mean a homework.

Let $\{v_1,v_2\}$ be a linearly independent subset of a vector space $V$ and let $w$ belong to $V$ but not to $\operatorname{Span}(\{v_1,v_2\})$. We need to show that the set $T= \{v_1 + w, v_2 + w\}$ is linearly independent.

My reasoning so far:

$\{v_1,v_2\}$ are linearly independent, thus there exist $c_1$ and $c_2$ such that $c_1v_1 + c_2v_2 = 0$, $c_1=c_2=0$.

$w$ is not in $\operatorname{Span}(\{v_1,v_2\})$, thus the set $\{v_1,v_2,w\}$ is linearly independent.

Thus there exist another $c_1,c_2,c_3$ such that $c_1v_1+c_2v_2+c_3w = 0$.

$c_1v_1+c_2v_2+c_3w = 0 \implies$

$c_3w = -c_1v_1 - c_2v_2 \implies w = (-c_1/c_3)v_1 + (-c_2/c_3)v_3$

$T= \{v_1 + w, v_2 + w\}$. Let there be $d_1$ and $d_2$ scalar such that

$d_1(v_1 + w) + d_2(v_2 + w)= 0$

$\implies d_1(v_1 + (-c_1/c_3)v_1) + d_2(v_2 + (-c_2/c_3)v_3) = 0$

$\implies v_1(d_1 + (-c_1/c_3)d_1) + v_2 (d_2 + (-c_2/c_3)d_2) = 0$,

but $v_1$ and $v_2$ are linearly independent.

Thus, $d_1 + (-c_1/c_3)d_1 = d_2 + (-c_2/c_3)d_2 = 0$

We conclude that the set $T$ is linearly independent.

Is this correct?


$c_1v_1+c_2v_2+c_3w = 0$

T linearly independent, there exit $d_1(v_1 + w) + d_2(v_2 + w)= 0$

$\implies d_1.v_1 + d_1.w + d_2.v_2 + d_2.w = 0$

$\implies (d_1.v_1 + d_2.v_2) + (d_1.w + d_2.w )= 0$

But $ d_1.v_1 + d_2.v_2 = 0 $ since $v_1$ and $v_2$ are linearly independent

thus $d_1 = d_2 = 0$

$\implies d_1(v_1 + w) + d_2(v_2 + w)= 0$ with $d_1 = d_2 = 0$

T is linearly independent


Note that $v_1,v_2,\ldots,v_k$ are linear independant does not mean (as you write) that there exist $c_1,c_2,\ldots,c_k$ such that $c_1v_1+\ldots +c_kv_k=0$ and $c_1=\ldots =c_k=0$. Rather it menas that for all $c_1,c_2,\ldots,c_k$ with $c_1v_1+\ldots +c_kv_k=0$ we have that $c_1=\ldots =c_k=0$. However, in your reasoning you seem to make use of the correct form.

Nevertheless, you divide by $c_3$ without knowing whether $c_3\ne 0$. Can you modify your calculation to avoid this division?

Apart from this a remark about notation: I am always unhappy when people talk about a set of vectors being linearly independant. For example (not related to the problem statement at hand), assume $\{v_1,v_2\}\subset V$ is linearly indpendant and let $w=v_1+2v_2$. Can it happen that the set $\{v_1,v_2,w\}$ is linearly independent? Maybe surprisingly the answer is: Yes, it can! (Can you find an example for this?)

  • $\begingroup$ I edited my reasoning, is this correct? $\endgroup$ – abedzantout May 10 '15 at 12:22
  • $\begingroup$ If we take the vectors (1, 1) and (1, 2), then w = (1,1) + 2(1,3) = (3,2). The three vectors are linearly independent. And yes, these kind of question are annoying, and created some problems to my colleagues during their exam. $\endgroup$ – abedzantout May 10 '15 at 12:24

you have shown that $\{v_1, v_2, w\}$ is linearly independent. you want to show $\{v_1+w, v_2 + w\}$ is linearly independent.

suppose $$c_1(v_1+w) + c_2(v_2 + w) = 0.\tag 1$$ we need to show that $$c_1 = 0, c_2 = 0 $$

rewrite $(1)$ as $$c_1v_1 + c_2v_2 + (c_1+c_2)w = 0.$$ use the fact that $\{v_1, v_2, w\}$ is linearly independent to conclude $c_1 = 0, c_2 = 0.$ that concludes the proof.


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.