Talk:Supplement I (Ex)

From Manifold Atlas
Jump to: navigation, search

Before stating the exercise, we recall the definition of \Sigma^m, and explain why it is indeed a simplicial complex.

Definition 0.1. The simplicial complex \Sigma^m is defined to have k-simplices equal to the set of (m-k)-simplices of \partial \Delta^{m+1}, denoted by \sigma^* for \sigma\in\partial\Delta^{m+1}. The face relations are given by \sigma^* \leq \tau^* iff \sigma \geq \tau.

Lemma 0.2. This is a simplicial complex.

Proof. The following is an equivalent definition of \Sigma^m:

  • Its vertices are the m-dimensional faces in \partial\Delta^{m+1}.
  • A collection of vertices of \Sigma^m -- i.e. m-dimensional faces in \partial\Delta^{m+1} -- is a simplex of \Sigma^m iff the intersecion of all the faces is non-trivial.

This is an equivalent definition because of the following special ('miraculous'!) property of \partial\Delta^{m+1}:

  • The (m-k)-simplices are precisely the k-fold intersections of m-simplices.

This equivalent description makes it clear that \Sigma^m is a simplicial complex, since a subcollection of a non-trivially intersecting collection is again non-trivially intersecting.

\square

Remark 0.3. The first description of \Sigma^m gives us a 'map' of simplicial complexes

\displaystyle  \tau \mapsto \tau^* \;\colon\; \partial\Delta^{m+1} \longrightarrow \Sigma^m .

This is of course not a real map of simplicial complexes, since it sends k-simplices to (m-k)-simplices. However, it does induce a map of barycentric subdivisions

\displaystyle  (\tau_0 < \tau_1 < \cdots < \tau_k) \mapsto (\tau_0^* > \tau_1^* > \cdots > \tau_k^*) \;\colon\; (\partial\Delta^{m+1})^\prime \longrightarrow (\Sigma^m)^\prime ,

which is an isomorphism of simplicial complexes. So we may view subcomplexes of (\partial\Delta^{m+1})^\prime as living in (\Sigma^m)^\prime via this isomorphism.

Exercise 0.4. Let K be a subcomplex of \partial\Delta^{m+1}, and consider the barycentric subdivision K^\prime as a subcomplex of (\Sigma^m)^\prime. Show that

\displaystyle  D(\sigma ,K) = \sigma^* \cap K^\prime .

Solution 0.5. First consider what happens if we view everything as living inside (\partial\Delta^{m+1})^\prime:

  • The barycentric subdivision K^\prime is the subcomplex with k-simplices \lbrace \tau_0 < \tau_1 < \cdots < \tau_k \rbrace with \tau_i a simplex of K for each i.
  • The dual complex D(\sigma ,K) is by definition the subcomplex of K^\prime whose k-simplices are \lbrace \tau_0 < \tau_1 < \cdots < \tau_k \rbrace with \tau_i a simplex of K and \tau_i \geq \sigma for each i.
  • Finally, \sigma is the subcomplex with k-simplices \lbrace \tau_0 < \tau_1 < \cdots < \tau_k \rbrace with \tau_i \leq \sigma for all i.

So we clearly have D(\sigma ,K) \neq \sigma\cap K^\prime \subseteq (\partial\Delta^{m+1})^\prime, since we have different conditions \tau_i \geq \sigma and \tau_i \leq \sigma on the LHS and RHS respectively. This is solved by passing from (\partial\Delta^{m+1})^\prime to (\Sigma^m)^\prime and replacing \sigma by \sigma^*. Considered as subcomplexes of (\Sigma^m)^\prime, we have:

  • K^\prime has k-simplices \lbrace \tau_0^* > \tau_1^* > \cdots > \tau_k^* \rbrace with \tau_i a simplex of K for each i.
  • D(\sigma ,K) has k-simplices \lbrace \tau_0^* > \tau_1^* > \cdots > \tau_k^* \rbrace with \tau_i a simplex of K and \tau_i \geq \sigma, equivalently \tau_i^* \leq \sigma^*, for all i.
  • \sigma^* has k-simplices \lbrace \tau_0^* > \tau_1^* > \cdots > \tau_k^* \rbrace with \tau_i^* \leq \sigma^* for all i.

Hence,

\displaystyle  D(\sigma ,K) = \sigma^* \cap K^\prime \subseteq (\Sigma^m)^\prime .

Remark 0.6. Once we know that the dual cell decomposition of \partial\Delta^{m+1} is actually a simplicial complex, in other words that \Sigma^m is a simplicial complex, then the argument is very simple and formal. The main geometric input to the argument is the special property of \partial\Delta^{m+1} mentioned above, that (m-k)-simplices are precisely the k-fold intersections of m-simplices.

Personal tools
Namespaces
Variants
Actions
Navigation
Interaction
Toolbox