home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math
- Path: sparky!uunet!europa.asd.contel.com!paladin.american.edu!howland.reston.ans.net!spool.mu.edu!enterpoop.mit.edu!galois!riesz!jbaez
- From: jbaez@riesz.mit.edu (John C. Baez)
- Subject: Re: wanted: diff. criterium for pureness of tensors.
- Message-ID: <1993Jan11.203344.11891@galois.mit.edu>
- Sender: news@galois.mit.edu
- Nntp-Posting-Host: riesz
- Organization: MIT Department of Mathematics, Cambridge, MA
- References: <1993Jan11.115730.42655@urz.unibas.ch>
- Date: Mon, 11 Jan 93 20:33:44 GMT
- Lines: 42
-
- In article <1993Jan11.115730.42655@urz.unibas.ch> kullmann@urz.unibas.ch (Peter Kullmann) writes:
- >
- >Is there a differentiable criterium for the pureness *) of tensors in the
- >outer tensor algebra over a real vector space? I.e. a function
- >
- > f:Alt(k,V) ---> R differentiable
- >
- >such that if s \in Alt(k,V): f(s) = 0 iff s is pure.
- >
- >This has of course to do with Grassmanians, and what I actually want is to
- >have Grassmann (n,k) as a level surface in Alt(k,R^n). I doubt it very much
- >that this is possible, so an argument for the contrary would be welcomed as
- >well:-)
-
- Such a function exists in the complex case (vector spaces over
- the complex numbers) - I don't know if you are asking the real case to
- be deliberately difficult, and I don't know if the real case is actually
- harder, but I'll tell you how to do it for the complex case - from some
- lecture notes by Guillemin.
-
- What you call pure tensors he calls "decomposable." Pick a basis e_i
- for V and the dual basis e^i for V*, and let e^I for any multiindex I =
- (i_1,...,i_k) be the exterior (or wedge, or outer) product e^{i_1} x ...
- x e^{i_k}; here I'm using x for exterior product. Then s is
- decomposable iff for all I, iota(e^I)s x s = 0. Here iota(v)s is the
- interior product of v (in the exterior algebra over V*) and s (in the
- exterior algebra over V).
-
- The conditions iota(e^I)s x s = 0 are a bunch of quadratic equations in
- the components of s. So we have found a *bunch* of functions that *all*
- vanish precisely on the decomposable tensors. If you really need *one*
- function that vanishes precisely on the decomposable tensors, take all
- these functions, take the squares of their absolute values, and add them
- up.
-
- Feel free to ask if some of the above doesn't make sense. In
- particular, the interior product of something in the exterior algebra
- over V* with something in the exterior algebra over V is a notion that's
- not very commonly used, though it's an obvious generalization of the
- interior product of something in V* with something in the exterior
- algebra over V.
-
-