A Understand (k,l) Tensors in Gen. Relativity

GeoffFB
Messages
4
Reaction score
0
TL;DR Summary
Type (k,l) tensors, dual vectors and vectors.
In both Wald and Carroll, a type (k,l) tensor has k dual vectors and l vectors, yet a (1,0) tensor is a vector and a (0,1) tensor is a dual vector. I must be missing something simple. Please explain.
 
Physics news on Phys.org
GeoffFB said:
In both Wald and Carroll, a type (k,l) tensor has k dual vectors and l vectors
No, it doesn't. A type (k, 1) tensor can be contracted with k dual vectors and l vectors to produce a scalar. That does not mean the k dual vectors and l vectors are "part of" the tensor.

GeoffFB said:
a (1,0) tensor is a vector and a (0,1) tensor is a dual vector.
Yes, because a (1, 0) tensor can be contracted with 1 dual vector to produce a scalar, so it's a vector; and a (0, 1) tensor can be contracted with 1 vector to produce a scalar, so it's a dual vector.
 
GeoffFB said:
Summary:: Type (k,l) tensors, dual vectors and vectors.

In both Wald and Carroll, a type (k,l) tensor has k dual vectors and l vectors, yet a (1,0) tensor is a vector and a (0,1) tensor is a dual vector. I must be missing something simple. Please explain.
A tensor, as they define it, is a multilinear map, it has k dual vectors and l vectors as input. So a (0,1) tensor will be a linear map that has a vector as an input i.e. it will be a dual vector.
 
  • Like
Likes vanhees71 and cianfa72
GeoffFB said:
In both Wald and Carroll
You are evidently misinterpreting statements from these sources, but unless you give specific quotes and where they are from, it's impossible to tell exactly what you are misinterpreting.
 
Sorry!
General Relativity by Robert M. Wald, Page 20.
Spacetime and Geometry by Sean M. Carroll, Page 21.
 
GeoffFB said:
General Relativity by Robert M. Wald, Page 20.
What on this page led you to believe what you said in the OP?
 
PeterDonis said:
No, it doesn't. A type (k, 1) tensor can be contracted with k dual vectors and l vectors to produce a scalar. That does not mean the k dual vectors and l vectors are "part of" the tensor.Yes, because a (1, 0) tensor can be contracted with 1 dual vector to produce a scalar, so it's a vector; and a (0, 1) tensor can be contracted with 1 vector to produce a scalar, so it's a dual vector.
Thanks. Now I understand.
 
PeterDonis said:
What on this page led you to believe what you said in the OP?
I was confused by thinking that type (0, 1) tensor meant (k, l) tensor, not understanding about the mapping to R.
 
Back
Top