25
$\begingroup$

Consider the inclusion $i_n: \varDelta_{\leq n} \hookrightarrow \varDelta$. This induces a truncation functor $tr_n:\mathbf{sSet}\to\mathbf{sSet}_{\leq n}$, which has a left and a right adjoint, $tr_n^L$ and $tr_n^R$, obtained via Kan extension. This allows us to define the $n$-(co)skeleton of a simplical set $X$: $$ sk_n X := tr_n^L tr_n X, \quad cosk_n X := tr_n^Rtr_n X. $$ It is easy to see that these functors form an adjunction.

$sk_n$ is analogous to forgetting about cells of dimension greater than $n$ in a CW complex, i.e. taking the n-skeleton. What is the dual analogy (if it exists) for the coskeleton of a simplical set $X$?

$\endgroup$
2
  • 2
    $\begingroup$ You glue higher dimensional cells along each boundary for dimension greater than $n$. When your simplicial set is a Kan complex, this is equivalent to killing higher homotopy groups (again greater than $n$). $\endgroup$
    – user40276
    Jun 28, 2016 at 22:52
  • $\begingroup$ Also check out the “Formal definition” section of en.wikipedia.org/wiki/Hypercovering. Except the third paragraph is probably slightly wrong as one need to include the augmentation map into the diagram of which the limit is taken. Also I suspect that cosk sk isn’t necessary; just cosk will do. $\endgroup$ Nov 4, 2022 at 7:53

3 Answers 3

26
$\begingroup$

It should be added that if $X$ is a Kan complex, then $X\to \mathrm{cosk}_n X$ computes a model for the $(n-1)$-truncation of the homotopy type of $X$. That's a fact you can easily read off using elementary definition of homotopy groups of a Kan complex, together with the bijection $$ \{ \Delta^k \to \mathrm{cosk}_n X \} \Leftrightarrow \{ \mathrm{sk}_n \Delta^k \to X\}. $$

$\endgroup$
14
$\begingroup$

A simplicial set $X$ is $k$-coskeletal iff the following condition holds:

a simplex of dimension $\geq k$ is present iff all of its $(k-1)$-dimensional faces are present in $X$.

A standard example is the usual nerve of a small category, which is readily seen to be $2$-coskeletal. (Similarly, the nerve of an $n$-category is $(n+1)$-coskeletal).

$\endgroup$
3
  • 4
    $\begingroup$ Your last point is not correct: the nerve of a category of 2-coskeletal, and the nerve of an n-category should be (n+1)-coskeletal. See eg ncatlab.org/nlab/show/simplicial+skeleton#examples and references there. $\endgroup$
    – David Roberts
    Aug 16, 2018 at 7:28
  • 10
    $\begingroup$ Phrasing @David's comment as a fuzzy slogan: In a category, not every triangle commutes, but every tetrahedron (or higher structure) of morphisms made from commutative triangles commutes. $\endgroup$ Nov 3, 2018 at 21:37
  • 1
    $\begingroup$ Thanks @davidroberts, now fixed I hope. $\endgroup$ Nov 11, 2019 at 16:19
9
$\begingroup$

For a simplicial set $X$ you can see the $n$-th skeleton to have no $m$-simplices other than the degenerate ones for $m>n$.(As you say)

For $m\leq n$ both the $n$-skeleton and the $n$-coskeleton have the same $m$-simplices as $X$.

And to answer your question for $m>n$ the $n$-coskeleton has $m$-simplices ALL possible configurations of a $m+1$ number of $(m-1)$-simplices in $\mathrm{cosk}_nX$ that agree in the appropriate $m-2$-faces to form the shape of a standard m-simplex.

If I say it in a lower dimension it will sound less tongue-twisty and easier to imagine it geometrically, which I think is what you are looking for: Take $n=2$

  • For $m=3$ the 3-simplices of the 2-coskeleton are all four possible triangles in $X$ that you can arrange in a tetrahedral shape.

  • For $m=4$ the 4-simplices are all five possible tetrahedra in $\mathrm{cosk}_2X$ that form the shape of a standard 4-simplex (pentachoron).

So in this sense, the coskeleton is the largest simplicial set containing X and the skeleton the smallest.

$\endgroup$
1
  • $\begingroup$ Perhaps I would rephrase the first dot point as: For $m=3$ the 3-simplices of the 2-coskeleton correspond to all possible choices of four triangles in $X$ which can be arranged into a tetrahedral shape.$^\dagger$ and finish at the end with: $\dagger$ I.e. to a map on the $2$-truncations $(\Delta^3)^{(2)}\to X_\bullet^{(2)}$, which is determined by where the $4$ non-degenerate $2$-simplicies are mapped, such that these form a tetrahedron. $\endgroup$
    – Skies burn
    Feb 25, 2020 at 0:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.