# Learning tensors from scratch

1. Nov 30, 2006

### aeroboyo

Hello,

I'd like to learn about tensors so i can start learning about special relativity. I understand nothing right now about tensors, except that they mean different things to mathematicians and to physicists, which is where my confusion begins!

Should i learn about the modern way of doing tensors without coordinates, or should i learn the 'classic method'. I'm very confused. Anyone know any good websites with basic lectures or notes on tensors? Also, does any1 happen to have msn so i might be able to ask a few questions in real time as they prop up? Thank you.

2. Nov 30, 2006

### Daverz

3. Nov 30, 2006

### aeroboyo

wow that online book is complicated! I guess i've been living on the moon all my life.

4. Nov 30, 2006

### Daverz

5. Nov 30, 2006

### aeroboyo

one of those online books makes more sense... is an understanding of linear algebra is necessary to begin learning tensor calculus? I'm a 1st year engineering student just now and the linear algebra course is in the 2nd semester so i've never done any linear algebra before. I guess you could say i'm like a baby in a swimming pool wearing inflatable arm bands.

6. Nov 30, 2006

### Daverz

Yes, ideas like linear independence and change of basis are important.

7. Nov 30, 2006

### mathwonk

tensors are just a language for describing multilinear functions. e.g. dot products are tensors. because they are bilinear.

8. Dec 1, 2006

### kesh

you can start learning special relativity without tensors. tensors are essential for general relativity

9. Dec 1, 2006

### aeroboyo

is this tensor 'language' covered in what is called 'tensor analysis' in some maths methods books?

10. Dec 1, 2006

### mathwonk

you can start learning by reading what i said.

11. Dec 1, 2006

### Daverz

Many math methods books cover Cartesian tensors, tensors in (flat) Euclidean space. They may also introduce tensors in non-linear coordinate systems (curvilinear coordinates), but still in flat space, in which concepts like the Christoffel symbols arise for the first time. This is the case for Boas (judging from the TOC).

On a curved space or surface, linear coordinate systems cannot be introduced at all, we must use curvilinear coordinates.

The full mathematics of GR in all its technical glory is "tensor analysis on manifolds", where a manifold is the n-dimensional generalization of the concept of a surface.

12. Dec 1, 2006

### aeroboyo

So i'd assume that Frankels Geometry of Physics is preety much the 'technical glory' of General Relativity. It covers manifolds in the 1st chapter, tensor fields on manifolds in the 2nd chapter... and gets more insanely complicated as it goes on... finally covering einsteins field equations in chapter 11. I think i'm right in saying that another way to describe 'tensor analysis on manifolds' is 'differential geometry'. I think it might take me 10 years to work through Frankels text!

Last edited: Dec 1, 2006
13. Dec 2, 2006

### PerennialII

14. Dec 6, 2006

### Chris Hillman

Do what I say, not what I do...

That book is about much more than the background you need for classical gtr. I think you are worrying to much about all the stuff you think you have to read. It might help to be told that bits of knowledge are highly "superadditive" in mathematical subjects: when you have mastered n techniques, and then master another one, your knowledge and ability has increased by some factor much greater than $$1+1/n$$

aeroboy, I actually think you should stop posting for a few weeks to that you can focus on your reading...

15. Dec 10, 2006

### andytoh

The best book on tensors and manifolds for general relativity is "Tensors and Manifolds with Applications to Relativity" by Wasserman, 1992. I am typing out the solutions to all the exercises right now. The download link below are my solutions (with the questions typed out as well) to the exercises in Chapters 1 and 2:

Last edited: Jan 22, 2008
16. Dec 10, 2006

### Daverz

Wasserman is for students with abstract algebra, advanced linear algebra, and topology under their belt.

17. Dec 10, 2006

### andytoh

which makes it all the more interesting when we tie it with general relativity.