How to go about proving theorems about category theory?

94 Views Asked by At
  1. Prove that the functor $V : Cat → Set$ that sends a category $\mathcal{C}$ to its set of arrows, is represented by the category ${0 \rightarrow 1}$ with two objects and a single non identity morphism between them; do not forget to prove that the bijection $V(\mathcal{C}) \cong Hom_{Cat}({0 \rightarrow 1}, \mathcal{C})$ you found is natural!"

  2. Prove that there exists a category of relations, where the objects are all sets, and given sets $A, B$, the set of morphisms $A \rightarrow B$ is the powerset of $A \times B$. Given a relation $R \in \mathcal{P}(A \times B)$ and a relation $S \in \mathcal{P}(B \times C)$, define the composition $S \circ R \in \mathcal{P}(A \times C)$ as the subset $\{(a,c)\in A \times C \ | \ \exists \ b \in B$ such that $(a,b)\in R, (b,c)\in S \}$. Check that this defines an associative composition operation, i.e. $(S \circ R) \circ T = S \circ (R \circ T)$ whenever it makes sense. Given this composition law, there is only one possible choice for what is the identity relation $I \subseteq A \times A$: which one?

These are the types of questions I am tying to solve. I start by listing all the facts I know, the lemmas I can use, the properties of natural transformations, functors etc.

So for the 1st question, $V$ sends a category to its set of morphisms , and we want to prove that $V(\mathcal{C}) \cong Hom_{Cat}(0 \rightarrow 1, \mathcal{C} )$, the $Hom$-functor that maps $\mathcal{C}$ to the set of morphisms between $0 \rightarrow 1$ and $\mathcal{C}$. In $0 \rightarrow 1$, we have only two morphisms, identity. I know that we need to prove the square from $V(\mathcal{C})$ to $Hom(0 \rightarrow 1,\mathcal{C})$, to $V(\mathcal{D})$ to $Hom(0 \rightarrow 1,\mathcal{D})$ commutes, that the natural transformation is an isomorphism(has a well defined inverse)

$$ \require{AMScd} \begin{CD} V(\mathcal{C}) @>a^{\mathcal{C}} >> Hom(0 \rightarrow 1, \mathcal{C})\\ @V {V(F)} VV @VV {Hom(0 \rightarrow 1,F)} V \\ V(\mathcal{D}) @>a^{\mathcal{D}} >> Hom(0 \rightarrow 1, \mathcal{D})\\ \end{CD} $$

And after that, I am totally lost.

I will gladly accept any general tips on approaching these types of questions as well as more specific tips, especially for the 2nd question

1

There are 1 best solutions below

3
On

It looks to me like you have some misunderstandigs about the definitions. So maybe it will help you when I describe the components of the natural transformation explicitly, you can then try to check that it is indeed natural.

I will denote the category $a \to b$ with two objects and one non-identity arrow by $\mathbf 2$. The category has exactly two objects $a$ and $b$, two identity arrows $id_a$ and $id_b$ and a single non-identity arrow $a\to b$, I give it the name $f$. The composition operations in $\mathbf 2$ are forced, for example $f\circ id_a$ must be $f$ because $\mathbf 2$ is a category. You sometimes write $\{0,1\}$ or even $(0,1)$ for that category, which looks a bit weird to me.

What data does a functor $F:\mathbf 2\to \mathcal C$ contain? The functor $F$ must send the arrow $f$ to some arrow $F(f)$ in $\mathcal C$. The action on the remaining objects and arrows is forced. We must have that $F(a)$ and $F(b)$ must be the domain and codomain of $F(f)$ and that $F(id_a)$ and $F(id_b)$ are the identities of domain and codomain. Let me write $\mathcal C_1$ for the set of arrows in $\mathcal C$, i.e. $\mathcal C_1 = V(\mathcal C)$ in your notation. We see that there is a bijection $$\mathcal C_1 = Cat(\mathbf 2,\mathcal C)$$ Given an arrow $g:x\to y$ in $\mathcal C$, we can define a functor $F:\mathbf 2\to \mathcal C$ by setting $F(a) = x, F(b)=y,F(f) = g$, $F(id_a) =id_x$ and $F(id_b) = id_y$. Conversely, given a functor $F:\mathbf 2\to \mathcal C$ we get an arrow by evaluating $F$ at $f$. The arrow is $F(f) : F(a) \to F(b)$. You can check that the two functions are inverse to each other. I have described the component of the natural isomorphism at $\mathcal C$ for you, your job is now to check that it is natural. This means that given a functor $G:\mathcal C \to \mathcal D$, you have to check that the corresponding square of functions commutes. Remember that $Cat(\mathbf 2,G)$ acts by post-composition on functors. I hope this is enough to get you started. If not, write me a comment! :)