Let $f,g\colon X\rightarrow Y$ be maps between spaces.

I am looking for criteria for the equalizer of $f$ and $g$ to be a homotopy equalizer and I am happy to get answers for whatever model category of spaces you prefer.

As the equalizer of $f$ and $g$ can be written as the pullback of $(f,g)\colon X\rightarrow Y\times Y$ along the diagonal $Y\rightarrow Y\times Y$, it is a homotopy equalizer if $(f,g)$ is a fibration, which unfortunately doesn't happen very often.

I feel there should be something weaker which is more likely to happen. An indication for this is the following thought. If $k\colon Z\rightarrow Y$ and $l\colon W\rightarrow Y$, the pullback of $k$ and $l$ is the equalizer of the maps from $Z\times W$ to $Y$ given by projecting to one factor first and then using $k$ respectively $l$. The map analogous to $(f,g)$ is then $k\times l$. But $k\times l$ is often not a fibration even if the pullback is a homotopy pullback, e.g. if only one of the maps is a fibration and the other is not.

  • 3
    $\begingroup$ Write the homotopy equalizer $hE(f,g)$ as the homotopy pullback of $X\xrightarrow{\Delta} X\times X\xleftarrow{(p,q)} N(f,g)$ Where $N(f,g)$ is the homotopy pullback of $f$ and $g$ and $(p,q):N(f,g)\rightarrow X\times X$ are the canonical maps. When $f$ or $g$ is a fibration then $N(f,g)$ is homotopy equivalent to the strict pull back $P(f,g)$ of the two maps. When both $f$ and $g$ are fibrations, then the map $(p,q):N(f,g)\rightarrow X\times X$ is a fibration so therefore $hE(f,g)$ is homotopy equivalent to the strict pullback of the above and this is precisely the equalizer $E(f,g)$. $\endgroup$
    – Tyrone
    Jun 24, 2016 at 16:40
  • $\begingroup$ This is an answer. $\endgroup$
    – Effy
    Jun 28, 2016 at 16:47
  • $\begingroup$ @Tyrone Why is the map $(p,q): N(f,g) \rightarrow X \times X$ a fibration? Take $X = Y$ and $f,g$ to be identity then the $(p,q)$ is just the diagonal map which is not a fibration. $\endgroup$ May 24, 2018 at 11:21
  • $\begingroup$ $N(f,g)$ is the homotopy pullback, not strict pullback. $\endgroup$
    – Tyrone
    May 24, 2018 at 11:43
  • $\begingroup$ I still don't understand; Assume model category is right proper, then if either of $f$ or $g$ is a fibration then homotopy pullback is same (weakly equivalent) as strict pullback $\endgroup$ May 24, 2018 at 11:52

1 Answer 1


It is not sufficient that $f$ and $g$ are both fibrations.

Here is a counterexample in the category $\mathbf{Cat}$. The counterexample works both in the Thomason model structure (which gives a model category of spaces) as well as the canonical model structure.

Let $I$ be the "walking isomorphism": the category with two objects that is equivalent to $1$.

Let $F$ and $G$ be the two automorphisms of $I$. These are isomorphisms, and thus (acyclic) fibrations.

Since $I$ is contractible, every homotopy equalizer of $F$ and $G$ is also a contractible category.

The strict equalizer of $F$ and $G$ is the empty category.


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.