I was watching a lecture on tensors and the professor said that a defining feature of a vector $v$ is that it transforms under a coordinate transformation $x^{\mu} \rightarrow x^{\mu'}$ as

$$v^{\mu'}(x^{\mu'}) \equiv \frac{\partial x^{\mu'} }{\partial x^\mu} v^{\mu}(x^{\mu}(x^{\mu'})) $$

Basically the last last term on the right hand side is the $x^{\mu}$ coordinates expressed in terms of $x^{\mu'}$ coordinates.

I am trying to understand this in $\mathbb R^2$ for the Cartesian and Polar coordinates. If the $x^{\mu'} \equiv (x,y)$ and $x^{\mu} \equiv (r,\theta)$, then I get

$$\bigg(\begin{matrix}x\\y \end{matrix}\bigg) \equiv \bigg[ \begin{matrix} \cos\theta&-r\sin\theta\\ \sin\theta&r\cos\theta \end{matrix}\bigg] \bigg(\begin{matrix}r\\\theta \end{matrix}\bigg)$$

which doesn't work out to be true. Can you please tell me what I'm missing here?

  • 1
    $\begingroup$ Link to lecture? Minute? $\endgroup$
    – Qmechanic
    Commented Nov 14, 2018 at 23:16

1 Answer 1


This transformation equation applies to a vector $v$ that "lives at" at the point $x$ (i.e., is in the tangent space at $x$). The point $x$ is described by the coordinates $(x, y)$ or $(r, \theta)$, but these coordinate tuples are not vectors at $x$. Examples of vectors at $x$ are $(dx, dy)$ and $(dr, d\theta)$, and your matrix equation applies to them:

$$\bigg(\begin{matrix}dx\\dy \end{matrix}\bigg) = \bigg[ \begin{matrix} \cos\theta&-r\sin\theta\\ \sin\theta&r\cos\theta \end{matrix}\bigg] \bigg(\begin{matrix}dr\\d\theta \end{matrix}\bigg)$$

  • $\begingroup$ This is right. So many times I see people thinking that coordinates are vectors, and it leads to so much confusion when we have a vector space defined over those coordinates. $\endgroup$ Commented Nov 15, 2018 at 1:18
  • 2
    $\begingroup$ For a general manifold, as opposed to flat space, why do you have to think of vectors and tensors as living at a point? Well, think about the curved surface of the Earth. A vector pointing north in New York is different from a vector pointing north in Tokyo. A lot of differential geometry is about how to compare vectors at different points. $\endgroup$
    – G. Smith
    Commented Nov 15, 2018 at 1:22
  • $\begingroup$ Right. I see a similar confusion in introductory EM in expressing vectors using the unit vectors $\hat r$, $\hat\theta$, and $\hat\phi$ depends on where the vector is located. $\endgroup$ Commented Nov 15, 2018 at 1:28
  • $\begingroup$ If I have to be pedantic, the vectors living at the point $p=(x,y)$ are $\partial_x$ and $\partial_y$. $\endgroup$
    – Gonenc
    Commented Nov 15, 2018 at 1:38
  • 1
    $\begingroup$ I think you are using mathematicians’ definition of contravariance and covariance, and I am using physicists’. $\endgroup$
    – G. Smith
    Commented Nov 15, 2018 at 1:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.