If A, B are n order square matrices, and AB=0, then BA=0?

  • Thread starter Thread starter Portuga
  • Start date Start date
  • Tags Tags
    Matrices Square
AI Thread Summary
The discussion revolves around the question of whether, for square matrices A and B, the condition AB = 0 implies that BA = 0. Participants highlight that one cannot assume A has an inverse, as not all nonzero matrices are invertible. A counterexample is suggested using specific matrices A and B, which are non-zero but yield AB = 0. It is emphasized that reliance on software like SageMath can lead to misleading conclusions when learning matrix theory. The conversation concludes with a call for further exploration of counterexamples to clarify the relationship between AB and BA.
Portuga
Messages
56
Reaction score
6

Homework Statement


If A and B are square matrices of same order, prove of find a counter example that if AB = 0 then BA = 0.

Homework Equations


A^{-1} A = I_n, ABC = (AB)C

The Attempt at a Solution


AB = 0 \Rightarrow A^{-1} A B = A^{-1} 0 \Rightarrow (A^{-1} A) B = A^{-1} 0 \Rightarrow I_n B = A^{-1} 0 \Rightarrow B = A^{-1} 0 \Rightarrow B = 0 \Rightarrow BA = 0A = 0.
I am not pretty sure if this procedure really solve the problem, so I would like some advices...
Thanks in advance.
 
Physics news on Phys.org
Portuga said:

Homework Statement


If A and B are square matrices of same order, prove of find a counter example that if AB = 0 then BA = 0.

Homework Equations


A^{-1} A = I_n, ABC = (AB)C

The Attempt at a Solution


AB = 0 \Rightarrow A^{-1} A B = A^{-1} 0 \Rightarrow (A^{-1} A) B = A^{-1} 0 \Rightarrow I_n B = A^{-1} 0 \Rightarrow B = A^{-1} 0 \Rightarrow B = 0 \Rightarrow BA = 0A = 0.
I am not pretty sure if this procedure really solve the problem, so I would like some advices...
Thanks in advance.

You cannot assume that ##A^{-1}## exists. Some nonzero ##n \times n## matrices do not have inverses.
 
Ok. I was using sagemath to make some reasonings. So I put there a generic 2x2 A matrix,
\begin{bmatrix} x & y\\ z & t \end{bmatrix}
and solved AB = 0 for B. The software answered that the only solution is a null matrix. That's why I am trying to prove it. Did I miss something?
 
Portuga said:
Ok. I was using sagemath to make some reasonings. So I put there a generic 2x2 A matrix,
\begin{bmatrix} x & y\\ z & t \end{bmatrix}
and solved AB = 0 for B. The software answered that the only solution is a null matrix. That's why I am trying to prove it. Did I miss something?

What about
$$A = \pmatrix{1&1\\1&1}, \: B = \pmatrix{1 & -2 \\-1 & 2} ? $$
Do we have ##AB = 0##? Is ##A \neq 0##? Is ##B \neq 0##?
 
Ray Vickson said:
What about
$$A = \pmatrix{1&1\\1&1}, \: B = \pmatrix{1 & -2 \\-1 & 2} ? $$
Do we have ##AB = 0##? Is ##A \neq 0##? Is ##B \neq 0##?

You should AVOID using software like sagemath or whatever when you are learning the subject. In this case the answers it gives you may be highly misleading, and may cause you to "learn" things that are actually false.
 
Portuga said:
Ok. I was using sagemath to make some reasonings. So I put there a generic 2x2 A matrix,
\begin{bmatrix} x & y\\ z & t \end{bmatrix}
and solved AB = 0 for B. The software answered that the only solution is a null matrix. That's why I am trying to prove it. Did I miss something?

I think what you found there was that the 0 matrix is the only matrix, ##X##, for which:

For all matrices ##A## we have ##AX = 0##
 
Ok, I got the point. So, I should make a counter example. Thanks!
 
Portuga said:
Ok, I got the point. So, I should make a counter example. Thanks!
Let us know what you come up with.
 
Back
Top