Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

The problem, and my solution is outlined below. I think my solution is correct, but I feel as if I went about my solution in an awkward way, and that there may be a better/cleaner way to solve the given problem. Would appreciate some feedback on what I did.

enter image description here

share|cite|improve this question
Looks OK to me. The nullspace and the row space are orthogonal complements to each other, so the nullspace of a matrix whose row space is $U$ will be the row space of a matrix whose nullspace is $U$. – Gerry Myerson Sep 29 '13 at 1:04
Agreed, but is there any way to avoid the part where I generated a bunch of conditions on aij, and then manually used those conditions to construct E, where Ex=0? I thought that was a bit cumbersome, and am wondering if there is a better way to construct E? – dlaser Sep 29 '13 at 2:01
I've posted an answer which you may find to be a slicker way to construct $E$. – Gerry Myerson Sep 29 '13 at 5:27
up vote 1 down vote accepted

I'll take up the story at the point where we have the matrix $$B=\pmatrix{1&0&0&-4&\phantom{-}3\cr0&1&0&-2&\phantom{-}4\cr0&0&1&\phantom{-}0&-1\cr}$$ and we want to find a basis for the nullspace. The non-pivot columns are the 4th and 5th. The nullspace will have a vector with 4th component 1, 5th component zero; by inspection, that vector is $(4,2,0,1,0)$. It will also have a vector with 4th component zero, 5th component 1; by inspection, that's $(-3,-4, 1,0,1)$. So we can take $E$ to be $$\pmatrix{\phantom{-}4&\phantom{-}2&0&1&0\cr-3&-4&1&0&1\cr}$$

Note that the rows of my $E$ add up to the 1st row of OP's $E$, while 3 times row 1 plus 4 times row 2 of my $E$ gives 4 times row two of OP's $E$, so both matrices have the same row space.

More generally, there is a basis for the nullspace containing, for each non-pivot column, one vector where that component is 1 and the other non-pivot components are zero, and the exact value of such vectors is easily read off from the reduced row-echelon form.

share|cite|improve this answer
So the null space of the null space of a matrix, is that matrix's row space? Makes sense; that's what I was missing, this understanding. Thanks! – dlaser Sep 29 '13 at 21:15
That's what I meant when I wrote, "the nullspace and the row space are orthogonal complements to each other" (which I think is a better way to say it, as I am comfortable talking about the nullspace of a matrix, but not the nullspace of a vector space). – Gerry Myerson Sep 29 '13 at 23:03

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.