home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky sci.physics:21597 sci.math:17313
- Newsgroups: sci.physics,sci.math
- Path: sparky!uunet!enterpoop.mit.edu!galois!riesz!jbaez
- From: jbaez@riesz.mit.edu (John C. Baez)
- Subject: Re: Can space-time intersect itself?
- Message-ID: <1992Dec19.213520.19051@galois.mit.edu>
- Sender: news@galois.mit.edu
- Nntp-Posting-Host: riesz
- Organization: MIT Department of Mathematics, Cambridge, MA
- References: <402@moene.indiv.nluug.nl> <1992Dec15.065413.26861@galois.mit.edu> <mcirvin.724718726@husc8>
- Date: Sat, 19 Dec 92 21:35:20 GMT
- Lines: 46
-
-
- Matt McIrvin writes:
- >A difference between "can be embedded" and "is embedded"
- >is that the former doesn't *require* the space to intersect
- >itself, whereas the latter well might! I can embed a Klein bottle
- >in R^3...
-
- Really? You should publish this shocking result immediately. :-)
-
- >but if I do it requires that the Klein bottle intersect
- >itself.
-
- Right. This is an immersion, not an embedding. One can't embed a compact
- nonorientable n-manifold in R^{n+1}, as it turns out, and the Klein
- bottle is nonorientable. In fact, one can show that one can't embed a compact
- nonorientable n-manifold in a simply connected (n+1)-manifold. The
- proof relies on the following rough intuition (which I won't bother to
- state precisely, so it won't be literally true): the complement of a
- compact embedded n-manifold in a simply connected (n+1)-manifold must
- consist of two parts, but if the embedded n-manifold is nonorientable,
- one must be able to get around from one part to the other, a
- contradiction. (Think about how the outside and the inside of the Klein
- bottle are really on the same side.)
-
- Here it is crucial that the n-manifold be without boundary. There is no
- problem embedding the Moebius strip in R^3; it's a compact nonorientable
- 2-manifold *with boundary*.
-
- >(By the way, is the word "embed" right in this context? Is what
- >you do to a Klein bottle in R^3 an embedding, or just an immersion?
- >Are all immersions embeddings? An immersed manifold certainly can
- >intersect itself, or, more properly, its image under the immersion
- >can.)
-
- An embedding is a 1-1 immersion, or loosely one with no
- self-intersections. An immersion is a smooth map from one smooth
- manifold to another whose differential has maximal rank at each point.
- Or, to put it loosely, it's 1-1 on each tangent space, and thus (one can
- show) locally looks like an embedding, but globally might fold around
- back on itself.
-
- The great advantage of being a mathematician over being a physicist is
- that mathematicians are the ones who get to define how all mathematical
- terms are used, and thus are always able to trip up physicists with
- terminology! :-)
-
-