2
$\begingroup$

I'm watching the video tutorial on spans here: https://www.khanacademy.org/math/linear-algebra/vectors_and_spaces/linear_combinations/v/linear-combinations-and-span

At 8:13, he says that the vectors a = [1,2] and b = [0,3] span R2. Visually, I can see it. But I tried to work it out, like so:

sp(a, b) = x[1,2] + y[0,3] such that x,y exist in R
         = [x, 2x] + [0, 3y] st x,y e R
         = [x, 2x + 3y] st x,y e R

With that said, how do we know that [x, 2x + 3y]spans R2? I tried picking a random point ([19, 6]) and let x=19 and solved for y (2*19 + 3y = 6) and found that when x=19 and y=-10, then I can get the point [19, 6].

But I'm confused as to how would I be able to find out that [x, 2x + 3y] can make any and all points in R2? What's my next step in determining if it spans R2?

$\endgroup$
3
$\begingroup$

Well then, let us pick any point in $\mathbb R^2$, e.g. the point $(a,b)$.

Let us define $x = a$ and $y=\frac{b-2a}{3}$ then you can easily see that:

$$ \begin{align*} (a,[b-2a]/3) = (x,y) \\ \iff (a,b-2a) = (x,3y) \\ \iff (a,b) = (x,2a+3y) \\ \iff (a,b) = (x,2x+3y) \end{align*} $$

So $(a,b)$ really is in the span of $\{(1,2),(0,3)\}$.

As we can choose any $(a,b) \in \mathbb R^2$, we know that those vectors span the whole $\mathbb R^2$.

$\endgroup$
  • $\begingroup$ Thanks for the response. Quick question, how exactly did we get y = b-2a / 3? $\endgroup$ – user2719875 Jan 16 '16 at 23:46
  • 2
    $\begingroup$ I started from the last line=) (Many times in math the proof goes in the other direction compared to the order of the discovery.) Assume there is $x,y$ such that $(a,b) = (x,2x+3y)$.. Can we solve for $(x,y)$ or do we get a contradiction? $\endgroup$ – flawr Jan 17 '16 at 9:56
2
$\begingroup$

For any $(a,b)\in \mathbb{R}^2$, $$(a,b) = a\cdot(1,2)+(-\frac{2a}{3}+\frac{b}{3})\cdot(0,3)$$

$\endgroup$
  • $\begingroup$ Thanks for the response. Quick question, for any (a, b) ∈ R2, doesn't (a, b) = a(1, 2) + b(0, 3)? Where did we get (−2a/3 + b/3) from? $\endgroup$ – user2719875 Jan 16 '16 at 23:49
  • $\begingroup$ If we simplify the RHSwe get $$(a,2a) +(0, -2a+b) = (a,b)$$. Does this make more sense? All we did was multiple the scalars into the vectors and then add component-wise. If we used your suggestion we end up with $$(a,b) = (a,2a+3b)$$ Which is only true if $$b=0$$. $\endgroup$ – fosho Jan 16 '16 at 23:52
  • 1
    $\begingroup$ Right, but how exactly did you arrive at (a,2a)+(0,−2a+b)=(a,b) (or the equation in your answer) in the first place? $\endgroup$ – user2719875 Jan 17 '16 at 0:07

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.