Are k-forms tensors or tensor fields?

  • Thread starter logarithmic
  • Start date
  • Tags Tags
    Definition
1 reply · 2K views · last reply
logarithmic
Messages
103
Reaction score
0
Can someone explain whether, by definition, a k-form is a tensor, or a tensor field.

In "Tensor Analysis on Manifolds" (Bishop and Goldberg), it says: "A differential k-form is a [tex]C^\infty[/tex] skew-symmetric covariant tensor field of degree k (type (0,k))" [note: 0 refers to the dual space, and k refers to the original space], would suggest a k-form is a function like [tex]T: M \to T^0_kM[/tex], where M is a manifold and [tex]T^0_kM[/tex] is the set of type (0,k) tensors.

However, a 1-form df, is a map of the form [tex]df: T_pM \to \mathbb{R}[/tex] ([tex]T_pM[/tex] is the tangent space of M at p), which isn't what the above definition gives. It's consistent with the definition on wikipedia, which treats k-forms as tensors rather than tensor fields: http://en.wikipedia.org/wiki/Differential_form#Intrinsic_definitions

Adding to my confusion is some notes I have which says the space of k-forms at [tex]p\in M[/tex] is the vector space [tex]\Lambda^k(T_pM)[/tex], which is the set of functions [tex]T: (T_pM)^k \to \mathbb{R}[/tex]. Yet it says that k-form fields, are called k-forms. It also then in defining the pullback uses an expression [tex]\omega(V_1,\dots, V_k)[/tex], where the [tex]V_i[/tex] are vector fields. How can you apply a vector field to a k-form, when its domain is a point on a manifold (or k-tuple of elements from a tangent space, under the other definition)? Is this just loose notation to actually mean the value of the vector field, i.e., [tex]V_i(p)[/tex] for some p in M, which would actually be an element of the tangent space, instead of [tex]V_i[/tex]. In which case the definition of a k-form being a tensor, rather than a tensor field was used.

So can someone clear up all these conflicting usages and definitions of a k-form for me?
 
You must be prepared that it can mean both. Being aware of that always look at the context. It is much like when we say: "let's calculate the derivative". It is from the context that you have to deduce whether it is about the derivative at a given point or the derivative as a function. Usually, when you are paying attention to the context, there is only one way in which a given reasoning makes sense.
 
Last edited: