2

I was trying to come up with a proof of why: $AA^{-1} = I$.

If we know that: $A^{-1}A = I$, then $A(A^{-1}A) = A \implies (AA^{-1})A = A$.

However I don't like setting $AA^{-1} = I$ for fear that it might be something else at this point, even though we know that $IA=A$. For example, could $A$ times its inverse equal something other than the identity leading back to the original matrix $A$.

Does anyone have a another proof for why $A$ times its inverse would give you the identity or could explain something I'm missing?

Tom Collinge
  • 7,981
  • 25
  • 59
Will
  • 429
  • 5
    How do you define $A^{-1}$ if not that it is $AA^{-1}=I$? – Hirshy Jun 16 '16 at 13:54
  • 1
    You have to use that these are effective matrices: associativity and existence of the identity are not enough to prove that right invertibility is equivalent to left invertibility. – Crostul Jun 16 '16 at 13:55
  • How do you know that these are effective matrices then? – Will Jun 16 '16 at 13:56

1 Answers1

3

We say a matrix $B$ is an inverse for $A$ if $AB = BA = I$, and the notation for $B$ is $A^{-1}$.
So it's by definition $AA^{-1}=I$, you cannot really prove it.

Alexa
  • 289
  • 2
    The question is asking why $AB=I$ implies $BA=I$. The use of the word "inverse" or the symbol $A^{-1}$ is a red herring. – almagest Jun 16 '16 at 14:15