# Commuting the Lorentz transformation with derivative

elemental09
In the process of proving that the d'Alembert operator

https://www.physicsforums.com/attachments/31306

is invariant under Lorentz transformations, it was required to commute two terms in the following expression for the transformed operator, which was obtained by switching the index on the first del to the contravariant position, applying the Lorentz transformation to each del, then switching the index on the second del to the covariant position:

https://www.physicsforums.com/attachments/31307

Now, so long as the last lambda and the second last del terms can commute, then all the Lorentz and metric terms collapse into the Kronecker delta, and the proof is complete (after switching the index positions to obtain the original form). But what justifies the ability to commute them, i.e. why is the following true (or is it at all?):

https://www.physicsforums.com/attachments/31308

If the Lorentz transformation is just a boost in some arbitrary direction, then I see why it holds - the components of lambda then do not depend explicitly on the time or space coordinates, thus the derivative can commute with that term. However if the transformation involves a rotation, the components of lambda will depend explicitly on the coordinates, so it seems that commutativity would fail.

What am I missing?

EDIT:
It seems these attachments failed to upload...here are the intended equations, in order:
http://www.use.com/ccd724f4259a55704e5f?tc=no#photo=1

Last edited:

## Answers and Replies

Science Advisor
I don't understand your argument why a rotation from SO(3) would depend on the coordinates, but a boost from SO(3,1) not. You should pay attention to that, because it will solve your problem.

The Lorentz group SO(3,1) or Poincare group ISO(3,1) is a global symmetry group. That means that the elements don't involve the coordinates. So, performing a Lorentz rotation (whether a boost or spatial rotation) and taking a derivative commute.

elemental09
A silly mistake of mine - that's true, the components Lambda are just constants that depend on the speed of light, rotation angle, and relative velocity of the two frames. Thanks. Problem solved.