Solution

ID: past-exam-of-the-mathematics-course-of-the-university-of-cambridge/2005/iii/paper-26/3/solution

For a locally small category , an object , and a functor , the covariant form of the Yoneda lemma is the natural bijection
For define , where . For , functoriality gives , so this is a natural transformation. Conversely, naturality of at gives . The two constructions are inverse. They are natural in , because a natural transformation commutes with evaluation. They are also natural in : an arrow induces by precomposition, and the resulting map corresponds to . Applying the result to the opposite category gives for .
Now let be a small category, and let . Finite products in a category in exist pointwise; the terminal functor is the constant singleton functor. Define the exponential of covariant set-valued functors by
Smallness ensures that these natural transformations form a set. For , define at by
Precomposition respects identities and composition, so this defines a functor. Its evaluation map of an exponential object is
To check naturality, for both sides give : one side uses naturality of , and the other uses the defining action of .
Here is the full universal property. Given , define by
For , naturality of makes this commute with , so is a natural transformation. Replacing by is the same as precomposing by , proving naturality in . Conversely, gives . Starting with recovers it immediately. Starting with , its naturality at yields
so it too is recovered. The formulas are natural in and , giving the required adjunction . Therefore the covariant functor category is cartesian closed. This proof supplies the actual exponential and does not need a colimit presentation of its arguments.
Finally suppose itself is a small category which is a Cartesian closed category. In its presheaf category, the same construction, applied to , gives
The first isomorphism uses the product's universal property: a map into is precisely a pair of maps into and . The next is the Yoneda lemma, and the last is the exponential adjunction in . Each step is natural in , , and . Under these identifications, evaluation corresponds to of the evaluation . Hence the Yoneda embedding preserves exponentials, with the natural identification .

New to topics? Read the docs here!