I have a question which says the following:

Let $V$ be the span of $v_{1}=(0,1,2)$, $v_{2}=(-1,0,1)$ and $v_{3}=(-1,1,3)$.

Construct an orthonormal basis $B'$ for $V$ (usual dot product).

I know how to do the question. That's easy. I need to apply the gram schmidt procedure thrice and I have my answer. My problem is the 3rd iteration.


Orthonormalizing, we obtain $V_{1}'=(0,\frac{1}{\sqrt{5}},\frac{2}{\sqrt{5}})$

Then $U_2=V_2-\frac{V_2 \cdot U_1}{||U_1||^2}U_1$

Doing so gives: $U_2=(-1,\frac{-2}{5},\frac{1}{5})$

Orthonormalizing, we get $V_{2}'=(0,\frac{-2}{\sqrt{30}},\frac{1}{\sqrt{30}})$

Finally, $U_3=V_3-\frac{V_3 \cdot U_1}{||U_1||^2}U_1 - \frac{V_3 \cdot U_2}{||U_2||^2}U_2$

However this time, if I plug in everything, I get $U_3=(0,0,0)$. While that is indeed orthogonal to both ${v_1}$ and ${v_2}$ there is no way I can orthonormalize $U_3$ since I get division by $0$.

Is the question just badly asked? Because there is simply no way to obtain an orthonormal basis.


Hint: what is the dimension of $V$? This should tell you how big a basis should be.

  • $\begingroup$ Forgive my elementary question but is it not just $3$? There are $3$ vectors after all which span V is there not? $\endgroup$ – Future Math person Dec 12 '18 at 6:12
  • $\begingroup$ This doesn’t mean that the dimension is exactly $3$, only that it is at most $3$. For example, $span\{(1,0),(2,0)\}$ does not have dimension $2$. $\endgroup$ – platty Dec 12 '18 at 6:14
  • $\begingroup$ Notice that $v_1+v_2=v_3$, so they are not linearly independent. Dimension is $2$ or less. $\endgroup$ – Larsson Dec 12 '18 at 6:15
  • $\begingroup$ Yes I was just about to comment that haha. I only have 2 basis vectors so the dimension would be 2. Thanks! $\endgroup$ – Future Math person Dec 12 '18 at 6:16

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.