Homework 3
Problem 1
To any group
- Show there is a correspondence[1] between groups and one-object categories in which every arrow is invertible.
- Suppose
and are groups. Show that group morphisms correspond[2] to functors . - Recall that a permutation representation of
is a group morphism , where is the permutation group on a set . Show that permutation representations of correspond to functors . - Suppose
are group morphisms, with corresponding functors . Show there is a natural transformation if and only if and are conjugate, i.e., there is an element such that for all .
-
We have seen how to assign to each group
a one-object category whose arrows are labeled by the elements of (and with arrow composition corresponding to the group operation). By construction, the identity element in is sent to the identity arrow (on the unique object) in the category .Now suppose
is a category with a unique object and in which every arrow is invertible. Let be the set of arrows in , and define a binary operation on using the arrow composition in . (Note that every pair of arrows is composable, since all arrows in start and end at the unique object.) Since is a category, its composition law is associative and the identity arrow acts as a two-sided identity. It follows that the operation on is associative and the element corresponding to the identity arrow acts as two-sided identity. By assumption, every arrow in is invertible, which by construction of means that every element in has an inverse under the so-defined binary operation. Thus is a group.Finally, it is "clear" these two processes are mutually inverse. (Here we should technically make a bijection between isomorphism classes of groups and something analogous on the category side.)
-
Suppose
is a group morphism. We wish to define a functor . Each of the categories and has a unique object, so the object function of must map the unique object of to the unique object of . The set of arrows in is and the set of arrows in is , so on arrows we define by . By our construction of the categories and , the map sends the (unique) identity arrow in to the (unique) identity arrow in (because is a group morphism and hence ), and is compatible with composition (since is a group morphism and hence ). Thus, does indeed define a functor from to .Conversely, suppose
is a functor between one-object categories in which every arrow is invertible. Let and be the groups corresponding to and , respectively. Then the set of elements of (respectively, ) is the set of arrows in (respectively, ), so the arrow function of the functor is a map . The compatibility of the functor with arrow composition (and the fact that it sends the identity arrow to the identity arrow) directly translate to being a group morphism.Finally, it is once again "clear" that these two processes are mutually inverse. (Once again, this is not technically true. We are really sketching an equivalence between two specific categories.)
-
Suppose first that
is a group morphism. We define a functor as follows. On objects, maps the unique object of to the set . On arrows, maps each arrow in (corresponding to an element ) to the corresponding bijection . (Recall that the group consists of all set bijections from to itself.) Note that the identity element must be sent to the identity permutation , and that the group morphism property of ensures the functor is compatible with composition: . We have therefore indeed defined a functor .Conversely, suppose
is a functor. The object function of assigns to the unique object of a set . The arrow function of assigns to each arrow in (corresponding to an element ) an arrow , with identity arrow (corresponding to the identity element ) sent to the identity arrow . The compatibility of the functor with composition implies . This implies both that is has the group morphism property and also that every map is invertible (i.e., bijective), since every arrow in is invertible. Thus, and is a group morphism.As with the previous two parts, these two associations are "clearly" mutually inverse. (We're really asking for trouble at this point.)
-
First suppose there is a natural transformation
where are functors from to . For the sake of clarity, let us label the unique object in by and the unique object in by . Note that both and must map to . The natural transformation is then entirely defined by a single arrow in , i.e., by an element . The naturality condition requires that for every arrow in (corresponding to an element ) we have a commutative diagramIn other words, we must have
, i.e., . Since this must hold for all , this exactly says and are conjugate maps.Conversely, suppose
are conjugate group morphisms, so that there exists some such that for all . Let be the functors corresponding to and , respectively. Define a map by sending the unique object of to the arrow in corresponding to the element . The conjugate property exactly ensures that this defines a natural transformation from to , by the same diagram as above.
Problem 2
Let
- Show that
is the object function of a functor . - Show that
is the object function of a functor . - For each set
let be the evaluation map, defined by . Show that these maps are the components of a natural transformation , where is the identity functor on .
-
Let's denote the functor we are creating by
. It is already defined on objects, so we need to define it on arrows. To that end, let be a set map. We can define a set map by sending each map (which is what the elements of the set are) to the map (which is an element of the set ). In other words, is the "compose with " map. Observe that if is an identity arrow in , then is the "compose with " map, which is the identity on . (Sanity check: for any , the composition is the same set map .)It remains to check that this map on arrows is compatible with composition, so suppose
and are two set maps. For any (i.e., set map ), observe that , where the second equality held by the associativity of function composition. Thus, we do indeed have and hence our arrow map is compatible with composition. We therefore have defined a functor , as desired. -
Define the functor
by sending each set to the set , and each set map to the set map . (At the level of elements this is simply .) It is straightforward to show this does indeed define a functor, with the verification being nearly identical to the steps used in part 1. -
Recall that a natural transformation
is the data consisting of a set map for every set , such that for every set map the diagram below commutes:In the case of our evaluation map
we are being asked to show that the diagram below is commutative:Let's chase some elements. Starting from the top-left corner, let
be any element of the set . Going across and down corresponds to , while going down and then across corresponds to . By the definition of function composition in , we always have , and so the diagram does indeed commute. Thus, evaluation does indeed define a natural transformation. (Did you ever really doubt it?)
Problem 3
Recall the notion of pullbacks, which for the sake of this exercise we will only consider in the category
Show that the functor which assigns to each diagram of the form
Note
You don't need to check every tiny detail for this one. Define the pullback as a functor (giving the maps on objects and arrows), and then explicitly define the set map that should be a natural bijection between the appropriate hom-sets.
Let
Recall that for each diagram in
the pullback consists of the set
Encoding the diagrams as objects in a functor category
To see how the pullback construction is an adjoint functor, we first encode such diagrams as elements of a functor category. Let