MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I want to check that the homotopy category of cochain complexes of an idempotent splitting, preadditive category is idempotent splitting.

Let $a\xleftarrow{e}{}a$ be an idempotent chain map up to chain homotopy, $e^2\sim e$; that is, there exists maps $a_{i-1}\xleftarrow{h_i}a_i$ with $e_i^2-e_i=h_{i+1}d_i+d_{i-1}h_i$. Assuming that idempotents split in the underlying category, how can I construct a chain complex $im\left(e\right)$ with chain maps $im\left(e\right){\xleftarrow{p}\atop \xrightarrow[i]{}}a$ such that $ip\sim e$ and $pi\sim1_{im\left(e\right)}$?

share|cite|improve this question
Also, you probably mean "cochain", not "chain". – darij grinberg Jan 23 '11 at 23:41
up vote 5 down vote accepted

Let me re-denote your chain complex $a$ by $C$. You can define a chain complex $D$ as the mapping telescope of the infinite sequence $$ \cdots\overset e\to \quad C \quad \overset e\to \quad C \quad \overset e\to \cdots $$ This can be constructed as follows: Form the homotopy coequalizer of the pair of maps $$ 1,S_a: \bigoplus_{\Bbb N} \quad C \quad \to \quad \bigoplus_{\Bbb N} \quad C $$ where $1$ is the identity and $S_a$ is given by applying $a$ and then shifting by one unit to the right in the index. (The homotopy coequalizer is gotten from this diagram by replacing the target $\oplus_{\Bbb N} C$ with its cylinder $\oplus_{\Bbb N} C \otimes I$ and forming the coequalizer of the two inclusions given by $1$ and $S_a$ on each end.)

The effect of this construction is to homotopically invert the map $e$, giving you a model for $C[e^{-1}]$. There is an evident inclusion $i: C \to D$. There is a map $D \to C$ which is defined on the $k$-th summand using the map $e^{\circ k}$. This will do what you want it to.

share|cite|improve this answer
$\otimes$? $I$? What do we know about our category? – darij grinberg Jan 24 '11 at 0:15
Good point. Does Eitan's category have mapping cylinders and can one form colimits? I don't actually know what a "chain homotopy category" is. Maybe he means the homotopy category of a category of chain complexes. What are these complexes defined over? – John Klein Jan 24 '11 at 0:37
But don't you assume that there are countable coproducts (or co-powers) in your argument? In that case it's a general fact about such pre-additive categories proved with essentially the same 'mapping telescope argument', see Freyd, Splitting homotopy idempotents, in: Proceedings of the Conference on Categorical Algebra, La Jolla, CA, 1965, Springer, New York, 1966, pp. 173–176. – Theo Buehler Jan 24 '11 at 2:30
Also, it really doesn't work without countable sums? Can we prove it? – darij grinberg Jan 24 '11 at 9:10
@darij: $C \otimes I$ is the mapping cylinder over the identity $C \to C$, see Weibel Ch. 1.6 for an explicit description of the complex. The notation $- \otimes I$ is a standard abuse from homotopical algebra. – Theo Buehler Jan 24 '11 at 10:49

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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