I am reading "An introduction to Spacetime and Geometry" by Sean Carroll, and on page 21 section 1.6, it is mentioned that the dual of a vector is a type (0,1) tensor.

The definition given of the page for a tensor is as follows (1):

\begin{equation}\tag{1} T:T^*_p\underbrace{\times...\times}_{k \text{ times}}T^*_p\times T_p\underbrace{\times...\times}_{l \text{ times}}T_p\mapsto\mathbb{R} \end{equation}

In (1), $T_p$ stands for the tangent space at a point of a manifold $M$, and $T^*_p$ the dual to the same space. $\times$ denotes the Cartesian product.

By this definition (1), a $(k,l)$ tensor would have $k$ dual vectors and $l$ regular vectors.

However, the textbook claims that a dual vector can be thought of as a type (0,1) tensor and a regular one a type (1,0). This seems to be in contradiction of (1), where dual vectors appear $k$-times and regular ones $l$-times.

Which leads to my question:

Have I misinterperated the definition? If not, why are the indices of dual and regular vectors switched for a type $(k,l)$ tensor?


A dual vector is an element of the dual space $T_p^*$. This means it is a map from $T_p$ to $\mathbb{R}$. In general, a $(k,l)$ tensor is a map which takes $k$ dual vectors and $l$ vectors as inputs, and outputs a real number. A dual vector is hence a $(0,1)$ tensor. Your statement that a $(k,l)$ tensor has $k$ dual vectors and $l$ vectors is imprecise, and is the source of the confusion.


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.