nLab
preserved limit

Preservation of limits

Idea

If J:IC is a diagram and x is its limit in C, then we may naïvely say that this limit is preserved by a functor F:CD if F(x) is the limit of the composite diagram IJCFD. However, it is not enough to state this at the level of objects; we also need to impose some coherence conditions, preserving the entire universal cone. Furthermore, we can use a trick involving the Yoneda embedding to get a meaningful condition even if J has no limit in C at all.

Definitions

Let J:IC be a diagram and let F:CD be a functor.

Recall (see limit) that a cone over J in C may be defined as an object x of C together with a natural transformation to J from the composite const x I:I!1{x}C, where 1 is the terminal category. Then a terminal object in the category of these cones (if it exists) is a limit of J in C. Thus, a limit consists of an object x and a natural transformation η:const x IJ.

The functor F preserves the limit (x,η) if (F(x),Fη) is a limit of the functor FJ in D. (Here, Fη:const F(x) IFJ is a whiskering.)

Dually, F preserves a colimit of J if F op:C opD op preserves it as a limit of J op:I opC op.

If F preserves all limits or colimits of a given type (i.e. over a given category I), we simply say that F preserves that sort of limit (e.g. F preserves products, F preserves equalizers, etc.).

Examples

Let I be the empty category, so that a limit of the unique functor J:IC is a terminal object 1. Then F preserves this terminal object if and only if F(1) is a terminal object of D.

Let I be the discrete category 2, so that J picks out two objects a and b of C and the limit of J is a product a×b of a and b. Note that this product comes equipped with product projections π:a×ba and ρ:a×bb. Then F preserves this product if and only if F(a×b) is a product of F(a) and F(b) and furthermore the product projections are F(π) and F(ρ).

A functor that preserves all small limits in C that exist is called a continuous functor. Usually this term is only used when C has all small limits, i.e. is a complete category.

Preservation of weighted limits

Analogously, an enriched functor between enriched categories may preserve weighted limits. Are there any tricky points that we should mention?

Preservation of limits that don't exist

Sometimes we want to say that a functor F:CD preserves a limit that does not actually exist in C. For instance, a finitely continuous functor is usually defined as one that preserves all finite limits. If C is a finitely complete category, then this is fine; such a functor is called left exact. But what if C does not have all finite limits?

If C and D are locally small, then we can use the Yoneda lemma to turn the question into one involving categories that do have the required limits (and in fact have all limits), the presheaf categories [C op,Set] and [D op,Set]. (For colimits, use [C,Set] and [D,Set]; for V-enriched categories, use [C op,V] and [D op,V], which will work if V is complete.)

The left Kan extension of the composite CFDYon[D op,Set] along the Yoneda embedding CYon[C op,Set] (which always exists) is a functor from [C op,Set] to [D op,Set], which may be written as F (alluding to the bimodule nature of profunctors). A diagram J:IC becomes a diagram IJCYon[C op,Set] in [C op,Set], where it has a limit. If F preserves this limit, then we say that F preserves the hypothetical limit of J.

Since the Yoneda embedding preserves and reflects all limits, if J has a limit in C, then this condition is equivalent to the condition that F preserve it in the ordinary sense, but in general it is stronger than requiring that F preserve the limit only if it exists in C.

Finishing the motivating example, a flat functor may be defined as one that preserves all finite limits, whether or not they exist.

Revised on September 4, 2012 16:06:09 by Stephan Müller? (141.20.52.69)