home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky sci.math:10497 sci.physics:13291
- Path: sparky!uunet!dtix!darwin.sura.net!mips!sdd.hp.com!cs.utexas.edu!qt.cs.utexas.edu!yale.edu!yale!gumby!destroyer!ubc-cs!unixg.ubc.ca!ramsay
- From: ramsay@unixg.ubc.ca (Keith Ramsay)
- Newsgroups: sci.math,sci.physics
- Subject: Re: tensors: How about 3rd, 4th rank?
- Message-ID: <1992Aug21.234025.2455@unixg.ubc.ca>
- Date: 21 Aug 92 23:40:25 GMT
- References: <5134@tuegate.tue.nl> <3djygrk@rpi.edu> <1992Aug21.191559.20170@news2.cis.umn.edu>
- Sender: news@unixg.ubc.ca (Usenet News Maintenance)
- Organization: University of British Columbia, Vancouver, B.C., Canada
- Lines: 44
- Nntp-Posting-Host: unixg.ubc.ca
-
- In article <1992Aug21.191559.20170@news2.cis.umn.edu>
- benzvi@leibniz.geom.umn.edu (benzvi) writes:
- >... given two vector spaces over a common scalar field,
- >V and W, V(tensor)W is defined as the set of elements
- >v(tensor)w (v in V, w in W).
-
- Not exactly. V tensor W is *spanned* by those elements.
-
- For example, if V and W are both 3-dimensional vector spaces, with
- basis {i,j,k}, then V tensor W is a 9=3*3 dimensional space with basis
- {i tensor i, i tensor j,...,k tensor k}. The elements of the form v
- tensor w are just the elements of the form (ai+bj+ck) tensor
- (di+ej+fk)=ad(i tensor i)+ae(i tensor j)+...+cf(k tensor k). This (6
- dimensional surface) misses such things as (i tensor i)+(j tensor j).
-
- I recommend that one think of tensors in terms of the multi-linearity
- properties being utilized, rather than thinking exclusively in terms
- of coordinates and how the components transform as the coordinates are
- changed.
-
- An example of a natural higher order tensor is the (Riemann) curvature
- tensor. Curvature is manifested by noticing that vectors which are
- "parallel-transported" around closed loops don't come back the same.
- (Try this on a globe. Take a pencil, and put it on the equator
- pointing east. Move it eastward around the equator 90 degrees. Now
- move it up to the north pole, without changing the orientation. Now
- slide it similarly along a meridian back to the original place. It is
- no longer pointing east. This is because the globe is curved.)
-
- There is a corresponding curvature tensor. To specify a (tiny) closed
- (parallelogram-like) loop takes two vectors, and the initial vector to
- be transported around it is another one. The "result" is a vector
- indicating by how much the vector has been perturbed by going around
- the loop. One can show that the perturbation is linear in all three
- "input" vectors (if we take the correct limit of "tinyness" of our
- closed loop). Hence we have a tensor of type (3,1) (or is it (1,3)?).
-
- The fact that it is linear in each of the components is the key thing
- as far as its being a tensor. I hope this helps to dispel some of the
- mystery. Now go read some books. :-)
-
- Keith Ramsay The measure of our intellectual capacity is the
- ramsay@unixg.ubc.ca capacity to feel less and less satisfied with our
- answers to better and better problems.-C.W.Churchman
-