MathOverflow is a question and answer site for professional mathematicians. Join them; it only takes a minute:

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

Given the Chromatic Convergence Theorem, can we state this globally as some convergence in the category of categories? That is, we have the subcategories of finite spectra in each stable homotopy category of $E(n)$-localized spectra, say $\mathcal{L}_n^{fin}$. Is it correct that there should be "inclusion" functors $\mathcal{L}_n^{fin}\to\mathcal{L}_{n+1}^{fin}$? Can we say that if we take some kind of limit (maybe this should be phrased in terms of quasi-categories or something?) we get the homotopy category of finite spectra?

share|cite|improve this question
Of course assume we've already localized at a prime $p$. – Jon Beardsley May 25 '12 at 20:40
up vote 4 down vote accepted

You have inclusion and localization functors. The inclusion functors go in the direction you indicate (any $E(n)$-local spectrum is also $E(n+1)$-local), and the localization functors are left adjoints, goint in the oposite direction. In your heuristics, I would consider the localization functors because of the following reason. Given a spectrum $X$ the units of the adjunctions give rise to a tower

$$\cdots\rightarrow L_{E(n+1)}X\longrightarrow L_{E(n)}X\rightarrow\cdots\rightarrow L_{E(1)}X\longrightarrow L_{E(0)}X=X_{\mathbb{Q}},$$

and the chromatic convergence theorem says that for $X$ a $p$-local finite spectrum, $X$ can be computed as the following homotopy inverse limit

$$X=\operatorname{holim}_n L_{E(n)}X.$$

Now we can use this imput for the computation of mapping spaces between two $p$-local finite spectra $X$ and $Y$,

$$\operatorname{Map}(Y,X)=\operatorname{Map}(Y,\operatorname{holim}_nL_{E(n)}X)=\operatorname{holim}_n\operatorname{Map}(Y,L_{E(n)}X). $$

If $\mathcal{L}_{n}^{fin}$ denotes the category of $E(n)$-local $p$-local finite spectra, simplicially enriched via mapping spaces (e.g. using the hammock localization), then the sequence of localization functors

$$\cdots\rightarrow \mathcal{L}_{n+1}^{fin}\longrightarrow \mathcal{L}_{n}^{fin}\rightarrow\cdots\rightarrow \mathcal{L}_{1}^{fin}\longrightarrow \mathcal{L}_{0}^{fin}$$

can be regarded as a sequence of simplicial functors which are the identity on objects. Given two objects $X$ and $Y$, the formula above shows that the homotopy limit at the level of mapping spaces recovers morphisms in the category of $p$-local spectra.

I guess you would now like to say that this means that $\operatorname{holim}_n\mathcal{L}_{n}^{fin}$ is weakly equivalent to the (hammock localization of) the category of $p$-local spectra, where the homotopy limit is taken in the model category of simplicial categories. I think this is indeed true, but do not have at hand any reference to confirm it.

share|cite|improve this answer

As Fernando showed, you do have a functor from the category of finite spectra to an appropriate limit of the categories of $E(n)$-local finite spectra, and this recovers the mapping space between any pair of finite spectra: $$ {\rm Map}(Y,X) \simeq {\rm holim}_n {\rm Map}(L_{E(n)} Y, L_{E(n)} X). $$ However, this can't be made into an equivalence of categories because some non-finite objects show up as limits of finite objects.

For each $n$, choose a finite type $n$ complex $X_n$ whose $n$'th Morava $K$-theory is nontrivial (hence the $m$'th Morava $K$-theory is nontrivial for all $m \geq n$). Being of type $n$ implies that $L_{E(k)} X_n \simeq *$ for $k < n$. We therefore get a system of objects $$ Y_n = L_{E(n)} (X_0 \vee X_1 \vee \dots \vee X_n) $$ together with natural weak equivalences $L_{E(n-1)} Y_n \to Y_{n-1}$ induced by projecting off the factor $X_n$. This assembles into an object in an appropriate homotopy limit category.

However, there is no finite complex $Y$ such that $L_{E(n)} Y \simeq Y_n$ for all $n$, because the number of cells necessary to build $Y_n$ grows in an unbounded fashion. You can check this by observing that the $n$'th Morava $K$-theory of $Y_n$ has rank at least $n$ because each wedge factor contributes rank at least 1.

share|cite|improve this answer
Thanks Tyler, that's a really nice counterexample. This may not be such a bad thing. That is, I'd really be interested in knowing more about this limit of categories. For instance, what do its localizing subcategories look like? We know the structure of localizing subcategories for each category in the sequence. Also, can we get a more general "chromatic convergence" type situation by just looking at a limit of the E(n) local subcats, and not even their finite objects. – Jon Beardsley May 31 '12 at 15:08

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.