2
$\begingroup$

In Hovey, Model Categories, he says that given a category $\mathcal{C}$ with a small limits and colimits, if we form the category of pointed objects ,i.e. the category with objects as morphisms $*\rightarrow A$ where * is the terminal object, then this category $\mathcal{C}_*$also has all small limits and colimits.

To prove this, for any Functor $F:J\rightarrow \mathcal{C}_* $ where $J$ is small we compose it with the forgetful functor $\mathcal{C}_* \rightarrow \mathcal{C}$ and take the limit here. My question is how will this give a limit in $\mathcal{C}_*$? we will get an object in the category, but how do we get the arrow from $*$ to that object?

$\endgroup$
3
  • 1
    $\begingroup$ You can use the universal property of the limit $\lim (\pi\circ J)$ of $J\to {\mathcal C}_{\ast}\stackrel{\pi}{\to}{\mathcal C}$ to construct a unique morphism ${\ast}\to \lim (\pi\circ J)$ such that all projections $\lim (\pi\circ J)\to J(j)$ for $j\in J$ are morphisms in ${\mathcal C}_{\ast}$ w.r.t. this map. $\endgroup$
    – Hanno
    Feb 17 '15 at 10:21
  • $\begingroup$ Thanks, that works. Does this mean this statement will be true for any slice category of C? $\endgroup$
    – Arun Kumar
    Feb 17 '15 at 10:30
  • 1
    $\begingroup$ Is true for slices under an object. Of course, the dual statement is true for slices over an object. $\endgroup$
    – Zhen Lin
    Feb 17 '15 at 13:50
2
$\begingroup$

Here is how it works in general: If $\mathcal{C}$ is a category and $P$ is an object (in your case, a final object), then the forgetful functor $P/\mathcal{C} \to \mathcal{C}$ creates limits. In particular, if $\mathcal{C}$ is complete, then $P/\mathcal{C}$ is complete, too. To see this, consider a diagram $(P \to X_i)$ in $P/\mathcal{C}$ and consider a limit cone $(\lim_i X_i \to X_i)$. Since the $(P \to X_i)$ can be regarded as a cone, it lifts to a morphism $P \to \lim_i X_i$. Then one verifies without effort that $((P \to \lim_i X_i) \to (P \to X_i))$ is a limit cone in $P/\mathcal{C}$.

The case of colimits is more complicated, since the forgetful functor doesn't create colimits. However, if $\mathcal{C}$ is cocomplete, then $P/\mathcal{C}$ will be cocomplete: An initial object is $(\mathrm{id} : P \to P)$. The coproduct of a non-empty family of objects $(P \to X_i)$ is given by the wide pushout $\coprod_P X_i$ in $\mathcal{C}$ equipped with the morphism $P \to X_i \to \coprod_P X_i$ for some $i$ (it doesn't matter which one). The coequalizer of two morphisms $P \to (X \rightrightarrows Y)$ is the coequalizer $Y \to C$ of the underlying morphisms in $X \rightrightarrows Y$ equipped with the morphism $P \to Y \to C$. Coequalizers are created by $P/\mathcal{C} \to \mathcal{C}$.

$\endgroup$

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.