I am looking for references concerning the following facts, which I believe to be true:

In a Kan complex $K$, can I use "empty cubes" $\partial(\Delta^1)^{(n+1)}$ with the vertex $(0,0,\ldots,0)$ at the basepoint (under the obvious notation $0$ and $1$ for the two $0$-simplices of $\Delta^1$) as representatives for the simplicial homotopy groups of $K$?

More precisely, I would like to have an "algorithm" (i.e. a precise order in which to do the homotopies) to homotope the empty cube to to a single $n$-simplex with the boundary mapping to the basepoint. Moreover, I would also like to have the extension to the whole cube: If the empty cube can be filled then the homotopy class should be trivial, and I would like an extension of the homotopy giving the corresponding homotopy of the $n$-simplex to the basepoint.

Note: I want to do all of this in simplicial sets, so no cubical sets. This question is related but works with a different representation of homotopy classes (and also does not assume $K$ to be Kan). It does not answer my question in a satisfactory way.

Note 2: I am aware that one of the reasons we work with simplices is that for many things they behave "better" than cubes. Unfortunately, I would need what I detailed above for a nice application that I'm afraid would become horribly complicated if I were to use the standard definition of simplicial homotopy groups.


Yes. The map ∂◻^{n+1}→S^n that collapses the complement of a single n-dimensional face of ◻^{n+1} is a simplicial weak equivalence. This implies that the induced map Map(S^n,K)→Map(∂◻^{n+1},K) is a simplicial weak equivalence because both mapping spaces are derived (K is Kan and all simplicial sets are cofibrant).

For an explicit algorithm, consider the inclusions ∂◻^{n+1}→J and S^n→J, where J denotes the pointed join of ∂◻^{n+1} and S^n. These inclusions are acyclic cofibrations of simplicial sets. A map ∂◻^{n+1}→K can be extended to a map J→K by repeatedly filling in horns using the Kan property of K. The resulting map J→K can then be restricted to a map S^n→K, which yields the desired representative for an empty cube. The case of fillable cubes is treated in the same way.

| cite | improve this answer | |
  • $\begingroup$ Thanks! Would you have a reference where this is done/stated? Having something citable would be great! $\endgroup$ – Daniel Robert-Nicoud Mar 9 at 17:09
  • $\begingroup$ @DanielRobert-Nicoud: The primary reference is Proposition I.11.5 in Goerss and Jardine's Simplicial Homotopy Theory, which shows that the maps that are claimed to be weak equivalences are indeed weak equivalences. $\endgroup$ – Dmitri Pavlov Mar 9 at 21:23
  • $\begingroup$ Great, thank you! Time to go find my copy of the book! $\endgroup$ – Daniel Robert-Nicoud Mar 9 at 21:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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