# Linear Algebra - Proofs involving Inverses

1. Jan 21, 2009

### descendency

Two fairly simple proof problems. . . why aren't they simpler? :(
1. The problem statement, all variables and given/known data
Let A be an nxn matrix...
If A is not invertible then there exists an nxn matrix B such that AB = 0, B != 0. (not equal to)

2. Relevant equations
None really.

3. The attempt at a solution
Obviously, when A is the zero matrix, AB = 0.

If we call A the coefficient matrix in the system of equations Ax = 0, then x = x1B1 + x2B2 + ... + xnBn, where B = [B1|B2|...|Bn].

I can't seem to explain why that works. Is it obvious enough just to say that or is there a step of explanation I have left out?

--------------------------------------------------------------------------------

1. The problem statement, all variables and given/known data
If A is an m x n matrix, B is an n x m matrix and n < m, then AB is not invertible.

2. Relevant equations

3. The attempt at a solution
Obviously, A and B are not square and are therefore not invertible. Does that fact really matter? The product of invertible matrices is invertible, but is the product of non invertible matrices also non invertible?

2. Jan 21, 2009

### sutupidmath

Well, what i would say is that: We know that an invertible matrix is nonsingular,moreover, a matrix is invertible if and only if it is non-singular. So, since here A is supposed to be non-invertible, it means that A is singular. What this means is that: There exists some non-zero vector, call it b such that Ab=0.
Extrapolating from this, we can argue that, there will be some non-zero vectors, call them

$$\arrow B_1, B_2,...B_n$$ such that $$A*B_i=0$$, for all i=1,2,....n.

So, if we built our matrix $$B=[B_1,B_2,...,B_n]$$

We have actuall proven that AB=0. Where as we can clearly see B is not the zero matrix.

3. Jan 21, 2009

### descendency

How can I guarrantee that B is not zero?

4. Jan 21, 2009

### Dick

You know there is a vector Ab=0 with b not equal to zero, as sutupidmath said. You also seem to know the column space of a matrix represents it's range. So build a matrix whose column space is only multiples of b.

5. Jan 21, 2009

### descendency

Thanks everyone. Think I have it now.