1
$\begingroup$

There is

a unique simplicial set with a unique non-degenerate simplex in each dimension, (updated) and such that all faces of the non-degenerate simplex are non-degenerate.

Does it have a name, and what can be said about it ? What is it geometrically ? Can one think of it as a "fat" point ?

I have seen it referred to as the dunce's hat but I cannot find a reference. An explicit construction is as follows: $ S_n $ is the set of equivalence relations on $\{0,1,\dotsc,n\}$ such that each equivalence class is an interval (i.e. of form $[a,a+1,\dotsc,b]$ for some $a$, $b$). In another notation, $S_n := \bigsqcup_{m\leq n} \operatorname{Hom}_\text{surj}(n,m)$ is the set of order-preserving surjections from $n$ to $m$, $m\leq n$ (i.e. $\operatorname{Hom}_\text{surj}(n,m)$ is the set of all order-preserving surjections from the linear order $n$ to linear order $m$).

(Update). By the answer of Dmitri Pavlov $S$ is weakly contractible. A comment by Reid Barton suggested $S$ is something like an "ordered" classifying space (BG) for the standard representation of the group of automorphisms of a dense linear order. Namely, $$S_n := Hom_{orders} (n^\leq, \Bbb Q^\leq)/Aut(\Bbb Q^\leq)$$ That is, (a representative of) an $n$-simplex is an ordered tuple $x_0\leq ... \leq x_n $, and two tuples $(x_0\leq ... \leq x_n)$ and $(y_0\leq...\leq y_n)$ are considered equal iff there is an order preserving map $g:\Bbb Q\to \Bbb Q$ such that $y_i=gx_i$,$0\leq i \leq n$. (Such a $g$ exists iff both tuples give rise to the same equivalence relation $i\approx j$ iff $x_i=x_j$.)

Does this construction have a name ? What is a reference and correct terminology for the classifying spaces of group representations (actions) ?

A related question:

let $\mathrm{Eq}_\bullet$ be the simplicial set where $\mathrm{Eq}_n$ is the set of equivalence relations on $\{0,1,\dotsc,n\}$. Equivalently, it is something like a classifying space for the standard representation of the symmetric group of an infinite set $X$: $$Eq_n := Hom_{Sets}(n, X)/Aut(X)$$

Where can I read about this simplicial set ?

Note it classifies equivalence relations in the following sense: to give an equivalence relation on a set $X$ is the same as to give a morphism to $\mathrm{Eq}_\bullet$ from the simplicial set represented by $X$.

$\endgroup$
11
  • 9
    $\begingroup$ I’m not sure the specification “simplicial set with a unique non-degenerate simplex in each dimension” does define a simplicial set. From your description I guess you mean “[…] such that every face is non-degenerate”, or equivalently, “extended from a semisimplicial set”. $\endgroup$
    – Zhen Lin
    Commented Jan 30, 2023 at 13:26
  • $\begingroup$ Do you mean such a simplicial set is non-unique ? Could you give a counterexample, as I fail to see one ? In my example the unique non-degenerate simplex does seem to have the stronger property you mention. $\endgroup$
    – user494312
    Commented Jan 30, 2023 at 16:36
  • 4
    $\begingroup$ What are the faces of the $3$-simplex? Are they all equal to the non-degenerate $2$-simplex? There are a few degenerate $2$-simplices; what if some faces are instead degenerate? If every face of every $n$-simplex is just the $0$-simplex (with appropriate degeneracies applied), it looks to me like a wedge of spheres. With other choices for faces, you get other spaces. $\endgroup$ Commented Jan 30, 2023 at 17:09
  • $\begingroup$ Isn't the nerve of the groupoid $B(\mathbb{Z}/2)$ a simplicial set with a single nondegenerate simplex in each dimension (the one with the nonidentity element of $\mathbb{Z}/2$ on all its "short" edges)? It doesn't sound like the one you are describing is the same. $\endgroup$ Commented Jan 30, 2023 at 20:44
  • $\begingroup$ On second thought maybe it is the same: label edge $i$ with the nonidentity element if you put $i$ and $i+1$ in different equivalence relations in your description. $\endgroup$ Commented Jan 30, 2023 at 20:46

1 Answer 1

1
$\begingroup$

As already pointed out in the comments, such a simplicial set is highly nonunique. For example, in addition to the simplicial set $S$ described in the second paragraph one could take the wedge of simplicial spheres, with one sphere for every dimension.

If we concentrate our attention on the simplicial set $S$ described in the second paragraph, it is easy to show that $S$ is weakly contractible. For example, it has a single vertex, therefore it is connected and by writing down an explicit presentation for the fundamental group we see that the fundamental group is trivial. Next, computing the chain complex of normalized simplicial chains on $S$ with coefficients in an abelian group $A$, we get the chain complex $$A←A←A←A←⋯,$$ where the differentials alternate between zero and identity on $A$. This chain complex is contractible, which proves that $S$ is weakly contractible.

$\endgroup$
5
  • $\begingroup$ In fact, I am interested in the set $S$ itself. If one requires the stronger property suggested by Zhen Lin "in each dimension there is a unique non-degenerate simplex, and, moreover, all its faces are non-degenerate', it does seem to describe a simplicial set uniquely. $\endgroup$
    – user494312
    Commented Jan 31, 2023 at 9:48
  • $\begingroup$ If we restrict the dimension, we get a simplicial sphere, right ? Namely, the following property describes a simplicial sphere: "in each dimension <N there is a unique non-degenerate simplex, and, moreover, all its faces are non-degenerate; in each dimension $\geq N$ each simplex is degenerate". $\endgroup$
    – user494312
    Commented Jan 31, 2023 at 9:51
  • $\begingroup$ I don't think so. As was suggested to you, for $N = 2$ you get the topologist's dunce hat. $\endgroup$
    – Zhen Lin
    Commented Jan 31, 2023 at 10:09
  • $\begingroup$ In odd dimensions you get up to homotopy a sphere and in even dimensions you get something contractible. $\endgroup$
    – IJL
    Commented Jan 31, 2023 at 11:44
  • $\begingroup$ @user494312: As already pointed out in the comments, the claim about truncated simplicial sets is only true in odd dimensions. You can easily prove this using the technique explained in the answer, applying it to the map from S to the simplicial n-sphere. $\endgroup$ Commented Jan 31, 2023 at 20:29

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .