The definition of a natural numbers object a priori only allows for recursively defined morphisms in which the next value Φ(s(n)) depends only on the previous value Φ(n). In many cases, however, we would also like to use n itself to define Φ(s(n)). This can be done in categories with finite products:
Lemma 1.
Let (N,z,s) be a natural numbers object in a category with finite products. Then for every a:1→X and every g:N×X→X there exists a unique morphism Φ:N→X such that Φ(z)=a and Φ(s(n))=g(n,Φ(n)).
Here, we use element notation to simplify the exposition. For example, the equation Φ(s(n))=g(n,Φ(n)) means that the following diagram commutes:
N(idN,Φ)↓⏐N×XsgN↓⏐ΦX
Proof. Define the morphism b:1→N×X by b:=(z,a) and the morphism h:N×X→N×X by h(n,x):=(n,g(n,x)). By the universal property of (N,z,s), there is a unique morphism Ψ:N→N×X such that:
Ψ(z)=b
Ψ(s(n))=h(Ψ(n))
Write Ψ(n)=(Ψ0(n),Ψ1(n)), where Ψ0:N→N and Ψ1:N→X. The two equations above then become:
Ψ0(z)=z
Ψ0(s(n))=Ψ0(n)
Ψ1(z)=a
Ψ1(s(n))=g(Ψ0(n),Ψ1(n))
The uniqueness in the universal property of (N,z,s) implies Ψ0=idN. Therefore, Φ:=Ψ1 is the unique morphism Φ:N→X satisfying Φ(z)=a and Φ(s(n))=g(n,Φ(n)). □
The next result appears in Johnstone, Part A, Lemma 2.5.5. Our proof is slightly more concise because we have extracted Lemma 1.
Lemma 2.
Let (N,z,s) be a natural numbers object in a category with finite products. Then
1zNsN
is a coproduct cocone. Thus, N≅1⊔N.
Proof. Let a:1→X and b:N→X be morphisms. We need to show that there is a unique morphism c:N→X satisfying c(z)=a and c(s(n))=b(n). This follows by applying Lemma 1 to the morphism g:N×X→X defined by g(n,x):=b(n). □
Next, we will check when the terminal object 1 itself is a natural numbers object. In that case, z:1→1 and s:1→1 are necessarily equal to id1.
Lemma 3.
Let 1 be a terminal object in a category. Then (1,id1,id1) is a natural numbers object if and only if for every endomorphism g:X→X and every morphism a:1→X we have g∘a=a. If the category has finite products, (1,id1,id1) is a parametrized natural numbers object if and only if g=idX for every endomorphism g:X→X, i.e. the category is one-way.
Proof. The first statement is immediate from the definition of a natural numbers object. For the second, (1,id1,id1) is a parametrized natural numbers object if and only if, for all morphisms f:A→X and all endomorphisms g:X→X, there is a unique morphism Φ:A→X such that Φ∘idA=f and Φ∘idA=g∘Φ. These equations simplify to Φ=f and f=g∘f. Since this must hold for every f:A→X, we must have g=idX (by the Yoneda Lemma or by a direct argument). □
Next, we prove a partial converse to the result that countably distributive categories have a parametrized natural numbers object.
Lemma 4.
Let C be a category with finite products, countable copowers denoted ⊗, and a parametrized natural numbers object 1zNsN. Then there is an isomorphism N≅N⊗1, and for every object A the natural morphism
α:N⊗A→A×(N⊗1)
is an isomorphism.
Proof. We will use element notation extensively. In particular, for every element a∈A and n∈N, there is an element n⊗a∈N⊗A, formally defined by the nth coproduct inclusion. The morphism α is defined by
α(n⊗a)=(a,n⊗1).
In any category with a terminal object and countable copowers, we can construct the non-parametrized NNO N⊗1 with the element 0⊗1∈N⊗1 and the map
s:N⊗1→N⊗1,s(n⊗1):=(n+1)⊗1.
See here for a detailed proof. Since, by assumption, 1zNsN is a parametrized NNO, it is also a non-parametrized NNO and is therefore isomorphic to the one just constructed. We may assume without loss of generality that they are equal and hence work with N=N⊗1.
Next, apply the parametrized universal property of the NNO to the diagram
AfN⊗AgN⊗A
defined by f(a):=0⊗a and g(n⊗a):=(n+1)⊗a. It gives a morphism
Φ:A×N→N⊗A
satisfying
Φ(a,0⊗1)=0⊗a,Φ(a,s(m))=g(Φ(a,m)).
For m:=n⊗1∈N, where n∈N, the second equation becomes
Φ(a,(n+1)⊗1)=g(Φ(a,n⊗1)).
By induction on n∈N, it follows that
Φ(a,n⊗1)=n⊗a,
which is exactly the statement that Φ∘α=idN⊗A.
It remains to prove α∘Φ=idA×N. We first observe that
α∘g=(idA×s)∘α
as morphisms N⊗A→A×(N⊗1). Indeed, for every n⊗a∈N⊗A,
α(g(n⊗a))=α((n+1)⊗a)=(a,(n+1)⊗1),
while
(idA×s)(α(n⊗a))=(idA×s)(a,n⊗1)=(a,(n+1)⊗1).
The universal property applied to the diagram
A(idA,z)A×NidA×sA×N
shows that there is a unique morphism Ψ:A×N→A×N satisfying
Ψ∘(idA,z)=(idA,z)
and
Ψ∘(idA×s)=(idA×s)∘Ψ,
namely idA×N. Thus, it suffices to verify that α∘Φ:A×N→A×N satisfies these two equations. We have
α(Φ(a,z))=α(0⊗a)=(a,0⊗1)=(a,z),
and hence α∘Φ∘(idA,z)=(idA,z). Moreover,
α∘Φ∘(idA×s)=α∘g∘Φ=(idA×s)∘α∘Φ.
This finishes the proof. □
Remark. Actually, the mentioned result and Lemma 4 can be combined into an equivalent characterization as follows: In a category with finite products and countable copowers, the NNO (which exists, see here) is a parametrized NNO if and only if for all objects A the canonical morphism
∐n∈NA=∐n∈N(A×1)→A×∐n∈N1
is an isomorphism. This is the precise connection to countable distributivity.
Context
This page is referenced by the following categories.