I am confused with the definition of Christoffel symbols for the dual space.

Let $M$ be some manifold, $x_i$ local coordinates

The Christoffel symbols are defines as

$\nabla_{\partial_i} \partial_j = \Gamma^k_{ij} \partial_k$

where $\nabla$ is the Levi Civita connection on $M$.

Now I read that

$\nabla_{\partial_i} dx_j = - \Gamma^j_{ik} dx_k$

but what is $\nabla$ here? It can't be the Levi-Civita connection right?

  • $\begingroup$ It is a connection defined on $T^*M$ $\endgroup$ Commented Jul 8, 2020 at 19:47
  • $\begingroup$ But a specific one? $\endgroup$
    – User1
    Commented Jul 8, 2020 at 19:58
  • $\begingroup$ Yes, the one defined by your equation. It is also called the dual connection $\endgroup$ Commented Jul 8, 2020 at 19:59
  • 1
    $\begingroup$ Oh, so the connection is defined by the second equation in my question? $\endgroup$
    – User1
    Commented Jul 8, 2020 at 20:09

1 Answer 1


Whenever one has a connection $\nabla$ on $TM$ there is an induced connection on the dual bundle $T^*M$ (which we will also call $\nabla$ because the abuse of notation is consistent), and it defined in such a way that the connection satisfies a sort of product rule when taking the covariant derivative of $(\alpha,v)=\alpha(v)$.

$$\nabla_X (\alpha,v) = (\nabla_X\alpha,v) +(\alpha,\nabla_X v)$$

But then recall that the action of the covariant derivative on a function $\nabla_X(f)=X(f)$ so we have

$$X \left((\alpha ,v)\right)=(\nabla_X\alpha,v) +(\alpha,\nabla_X v)$$

From here by picking basis one forms and vectors you can arrive at the formula you had. So in short it is the induced connection on the cotangent bundle of the Levi-Civita connection.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .