4
$\begingroup$

I know, that a tensor is a mathematically entity that is represented using a basis and tensor products, in the form of a matrix, and changing a representation doesn't change a tensor, is kind of obvious.

So does the invariance of a tensor under coordinate transformation mean what I stated above or does it mean that under a set of particular transformation the representation of a particular tensor also doesn't change.

Quoted from Wikipedia:

A vector is invariant under any change of basis, so if coordinates transform according to a transformation matrix $L$, the bases transform according to the matrix inverse $L^{−1}$, and conversely if the coordinates transform according to inverse $L^{−1}$, the bases transform according to the matrix $ L$.

Can someone please shed some light on this?

$\endgroup$
  • 2
    $\begingroup$ My answer from a few days ago might be what you're looking for. If not, I am sure we already have discussed what you are asking somewhere else on the site. $\endgroup$ – ACuriousMind Jun 30 '14 at 20:55
2
$\begingroup$

Regarding what you quoted: a vector is represented by the sum of a set of basis vectors times the vector components. If the components transform according to $L$, then the bases will transform according to $L^{-1}$, which means that when you multiply the bases with the components (to make the vector), you will get the same result every time (since $L\cdot L^{-1}=I$). This is what is meant by invariance.

Invariance of a tensor means basically what you stated above- the tensor itself doesn't change under a change of coordinates (like I explained). However, the tensor's components can very well change.

$\endgroup$
1
$\begingroup$

I guess there is two different notions of invariance of tensors. First notion is that if you look at a tensor as a mapping then the first notion of invariance is what you mentioned above. The other notion of invarance is that you do transformation but the " component" of metric does not change. For instance, if we do Lorentz transformation then the Minkowski metric is invariant, meaning that the component will be +1 , -1 , -1 ,-1 .

I might be wrong!!

$\endgroup$
  • 1
    $\begingroup$ Malinowski should be Minkowski? $\endgroup$ – Kyle Kanos Nov 18 '15 at 21:48

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.