1
$\begingroup$

True or False: Matrices with linearly independent row and column vectors are square.

Here is the answer of my textbook:

True; if the row vectors are linearly independent then $\text{nullity}(A)=0$ and $\text{rank}(A)=n=\text{the number of rows}$.
But since $\text{rank}(A)+\text{nullity}(A)=\text{the number of columns}$, $A$ must be square.

Why must a matrice with linearly independent vectors have $\text{nullity}(A)=0$?
That is where I lose track of the question.
Are zero rows considered to be linearly dependent?

$\endgroup$
  • $\begingroup$ Yes, any set containing the zero vector is linearly dependent. $\endgroup$ – Andrés E. Caicedo Aug 30 '14 at 2:12
  • $\begingroup$ Hint: if you have $m$ linearly independant vectors in ${\mathbb R}^n$, what can you say about $m$ and $n$? $\endgroup$ – Taladris Aug 30 '14 at 2:13
  • $\begingroup$ m</=n Where does this lead me? $\endgroup$ – yolo123 Aug 30 '14 at 2:13
  • $\begingroup$ Precisely $m\leqslant n$. I misread the first sentence and didn't see the link before commenting and thought you were asking for a solution to the title rather than an explanation on a particular proof. Sorry. $\endgroup$ – Taladris Aug 30 '14 at 2:15
  • 1
    $\begingroup$ My idea for a proof was: assume the matrix $A$ has $m$ rows and $n$ columns. Rows (that represents vectors in ${\mathbb R}^n$) are linearly independant so $m\leqslant n$. Similar reasoning for columns gives $n\leqslant m$, hence a square matrix. $\endgroup$ – Taladris Aug 30 '14 at 2:17
1
$\begingroup$

Let's say the matrix has $m$ rows and $n$ columns. Either $m < n$, $m > n$, or $m = n$.

If $m < n$, then we have $n$ columns which lie in $\mathbb{R}^m$. Since $\mathbb{R}^m$ has dimension $m$, we can't have more than $m$ linearly independent vectors in $\mathbb{R}^m$. So the $n$ columns must be linearly dependent, a contradiction. Thus, we cannot have $m < n$.

If $m > n$, then we have $m$ rows which lie in $\mathbb{R}^n$. Since $\mathbb{R}^n$ has dimension $n$, we can't have more than $n$ linearly independent vectors in $\mathbb{R}^n$. So the $m$ rows must be linearly dependent, a contradiction. Thus, we cannot have $m > n$.

The only remaining possibility is $m = n$, which means that the matrix must be square.

$\endgroup$
  • $\begingroup$ If I have zero rows in a matrix, does that mean that the rows vectors are linearly DEPENDENT? $\endgroup$ – yolo123 Aug 30 '14 at 2:17
  • 2
    $\begingroup$ If you have zero rows in a matrix, then you don't really have a matrix. $\endgroup$ – JimmyK4542 Aug 30 '14 at 2:18
  • $\begingroup$ Sorry, I meant a row WITH 0's in it. $\endgroup$ – yolo123 Aug 30 '14 at 2:19
  • 1
    $\begingroup$ A set of vectors $\{v_1,\ldots,v_m\}$ is said to be linearly dependent if you can find scalars $\alpha_1, \ldots, \alpha_m$ which are not all zero such that $\alpha_1v_1+\cdots+\alpha_nv_n = 0$. Suppose we have a row of all zeros, i.e. $v_1 = [0,0,\ldots,0]$. Then, pick $\alpha_1 = 1$ and $\alpha_i = 0$ for $i = 2,\ldots,m$. It is easy to see that $\alpha_1v_1+\cdots+\alpha_nv_n = 0$, so that set of vectors is linearly dependent. $\endgroup$ – JimmyK4542 Aug 30 '14 at 2:24
  • $\begingroup$ Does that mean every non-square matrix has row vectors which are linearly dependent? $\endgroup$ – Krishna Deshmukh Jan 20 at 10:16

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.