4

If I understand correctly, $\Delta$-complex on a space $X$ is defined to be a collection $\Delta(X)$ of cont. funtions $\sigma:\Delta^n\to X$ such that:

1) restriction of $\sigma$ to any face is in $\Delta(X)$

2) restriction of $\sigma$ to the interior is injective

3) $A \subseteq X$ is open if every $\sigma^{-1}(A)$ is open

Can anyone explain me how a singular complex $S(X)$ can be regarded as a $\Delta$-complex considering the condition (2)? I have read some books but I still don't get it. Thank you for your help.

Shaun
  • 44,997

1 Answers1

6

The singular complex $S(X)$ is not $Δ$-complex structure on space $X$, but it is a completely new space constructed as a geometric realization of an abstract $Δ$-structure given by singular simplices of original space $X$.

Example: Take $X = \{*\}$ the one-point space. For each $n$ there is only one singular $n$-simplex – the constant one. The singular complex $Δ(X)$ looks as follows. It has one $n$-cell for each $n$. Each of them represents the corresponding $n$-simplex. So the $0$-skeleton of $Δ(X)$ is just a point. In $1$-skeleton, we glue a line segment to the point with both endings. In $2$-skeleton, we glue a border of a triangle to the line segment. Each segment of the border is glued to the only line in $1$-skeleton. The first one is glued forwards, the second one backwards and the third one forwards. We continue in this manner with all $n$-cells. So $Δ(X)$ of the one-point space is actually infinite-dimensional $Δ$-complex with one cell in each dimension.

user87690
  • 9,133
  • If I'm not mistaking, the point has singular simplex the final simplicial set $\Delta[0]$ whose geometric realisation is the point. I think you forgot in your description that each $n$-simplex with $n>0$ is actually degenerate. – Pece Aug 23 '14 at 20:52
  • @Pece: Yes, the singular simplices are degenerate in $X$, but not in $S(X)$ by definition. We are not giving a $Δ$-structure to $X$, but constructing a $Δ$-complex, simplicial chain complex of which is singular chain complex of $X$. – user87690 Aug 23 '14 at 21:08
  • 1
    Ok maybe we just have a conflict of definitions. For me, the singular simplicial complex of $X$ is the simplicial set $SX$ defined in every degree by $(SX)n = \hom{\mathsf{Top}}(\Delta^{\mathrm{top}}_n,X)$ where $\Delta^{\mathrm{top}}_n$ is the topological simplex of dimension $n$. The geometric realization $|-|$ is the left adjoint of $S$ : in particular, $|\Delta[n]| = \Delta^{\mathrm{top}}_n$. Then, we have $| S(\ast) | = | \Delta[0] | = \Delta^{\mathrm{top}}_0 = \ast$. – Pece Aug 23 '14 at 21:25
  • Yes, we have. The singular complex I'm talking about is a certain $Δ$-complex. It is defined on page 108 in Hatcher's book so one can regard singular chain complex of a topological space as simplicial chain complex of its singular complex (the word “complex” is used in two different meanings). – user87690 Aug 24 '14 at 08:24
  • @user87690 In 1-skeleton, we glue a line segment to the point with both endings. What does this sentence mean? Could you explain it to me? – Brooks Feb 21 '16 at 09:18
  • @Brooks: The sentence describes how the 1-skeleton of the $Δ$-structure $Δ(X)$ we are building looks like. There is only one 1-simplex, geometric realization of which is a line segment, and its border, which consits of the two endpoints, is attached to the 0-skeleton in the only possible way – the endpoints are attached to the only point in 0-skeleton we have. – user87690 Feb 24 '16 at 20:27