© | Dror Bar-Natan: Classes: 2004-05: Math 1300Y - Topology:

(91)

Next: Jules Henri Poincare Previous: Class Notes for Tuesday March 29, 2005

Dreams on the (Co)Homology of Manifolds

This document in PDF: ManifoldDreams.pdf

The Context. Let $M$ be an $n$-dimensional manifold (a topological space that locally looks like ${\mathbb{R}}^n$) and let $R$ be a ring.

Dream 1   The following should be a list of spaces and their duals; there ought to be no (co)homology beyond that list:

$$\xymatrix{ H_0(M) & H_1(M) & H_2(M) & \cdots & H_{n-1}(M) & H_n(M) \\ H^0(M) & H^1(M) & H^2(M) & \cdots & H^{n-1}(M) & H^n(M) }$$

Dream 2   There should be an ``intersection pairing''

$$H_k\times H_l \to H_{k+l-n},$$

induced from the intersection pairing of submanifolds which ought to satisfy $\partial(\sigma\cap\lambda) = (\partial\sigma)\cap\lambda + \sigma\cap(\partial\lambda)$.

Dream 3   In particular, there should be a pairing

$$H_k\times H_{n-k} \to H_0=R,$$

so with some further optimism, $H_k$ ought to be the same as $(H_{n-k})^\star=H^{n-k}$. (And why not call that ``Poincaré Duality''?)

Dream 4   $H_n$ should be $R$ (and hence $H^n$ should be $R$ as well).

Dream 5   There should be a ``cap product''

$$\cap:H_k\times H^l \to H_{k-l}.$$

Dream 6   There should be a ``cup product''

$$\cup:H^k\times H^l \to H^{k+l},$$

and so $H^\star:=\oplus_k H^k$ ought to be a ring!

Jules Henri Poincaré

(from http://www-groups.dcs.st-and.ac.uk/~history/Mathematicians/Poincare.html)

Warning. Dreams are based on reality. Often, distorted reality.

The generation of this document was assisted by L^ATEX2HTML.