Let's say I have a matrix A of arbitrary size, and I perform a finite number of both elementary row/column operations on it, obtaining matrix B.

Are there any unique properties of matrix B that would be the same as matrix A? Such that I will be able to recognize that matrix B is a result of these operations performed on matrix A.

Here are some of the stuff that I've managed to find so far:

From Wikipedia, "Elementary row operations do not change the kernel of a matrix" "Elementary column operations do not change the image, but they do change the kernel."

And I've been told on StackOverflow that "if you don't consider scaling a row/column as an elementary operation a lot more structure is maintained".

Thank you.

  • $\begingroup$ Cross-posted here, right? $\endgroup$
    – draks ...
    Aug 29, 2012 at 11:54
  • $\begingroup$ Yup, I was advised to ask here instead. $\endgroup$ Aug 29, 2012 at 22:10

2 Answers 2


You will not be able to recognize that matrix $B$ originated from matrix $A$. There are other matrices that can be reduced to matrix $B$ via performing elementary operations.

  • $\begingroup$ If I rephrased my question slightly, instead of recognizing that B originated from A, but for Matrix A and B to still share some common un-changed properties as a result of those elementary operations, will there still be an answer? $\endgroup$ Aug 29, 2012 at 22:16
  • $\begingroup$ I think all other relevant properties preserved by elementary operations are listed in @Gerry Myerson's answer. $\endgroup$
    – Godot
    Aug 30, 2012 at 10:49

I think the only things you can count on are that you won't change the number of rows, the number of columns, the rank, or the nullity. If it's a square matrix, you won't change its invertibility.


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.