Suppose A is a 3 × 5 matrix, and that $dim\ Nul\ A = 2$. Prove the following;

(a) There exists a basis for $R^3$ made up of some of the columns of $A$.

(b) The rows of $A$ form a linearly independent set.

Part A;

$5 = Dim\ Nul(A) + Rk(A), Rk(A) = 3$ (rank nullity theorem)

So with rank of 3, $rref(A)$ contains 3 pivot columns, and the corresponding columns of $A$ (also the column space of A) form a basis of $R^3$.

Part B;

If the rows are linearly independent, then $Ax = 0$ IFF $x = 0$.

However since $Dim\ Nul(A) = 2$, there must be two $x \ne 0$ such that $Ax=0$. (contradiction?)

Am I doing this properly? Both answers seem intuitive, and I can't tell if I've proven them sufficiently (if at all). Is there anything I can or should add to improve these proofs?


Part A is alright, although there is no need to mention the reduced row echelon form of $A$. The rank of $A$ is the dimension of the column space of $A$ (the span of the columns). You have shown that $A$ has rank $3$ so the span of the five columns of $A$ is a three dimensional subspace of $\mathbb{R}^3$ which means this is the whole of $\mathbb{R}^3$. You can always throw away linearly depend elements from a spanning set and get a basis so you can find a basis of $\mathbb{R}^3$ that consists of three columns of $A$. In practice, in order to find it you can indeed row reduce and look at the pivots.

Part B lacks some explanation. What is (more or less) obvious is that if the columns of $A$ are linearly independent then $Ax = 0$ iff $x = 0$. The reason is that $Ax$ is a linear combination $C_1 x_1 + \dots + C_5 x_5$ of the columns of $A$ and not of the rows so if the columns are linearly independent, this combination is zero iff the coefficients $x_1 = \dots = x_5$ are zero. However, the column rank of a matrix $A$ (the dimension of the column space) is the same as the row rank of $A$ (the dimension of the row space) so in your case, the span of the three rows is three dimensional and hence they form a linearly independent set.


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.