Could not include topos theory - contents
Real numbers object
Recall that it is possible to define an internalization of the set of natural numbers, called a natural numbers object (NNO), in any cartesian monoidal category (a category with finite products). In particular, the notion makes sense in a topos. But a topos supports intuitionistic higher-order logic, so once we have an NNO, it is also possible to repeat the usual construction of the integers, the rationals, and then finally the real numbers; we thus obtain an internalization of in any topos with an NNO.
More generally, we can define a real numbers object (RNO) in any category with sufficient structure (somewhere between a cartesian monoidal category and a topos). Then we can prove that an RNO exists in any topos with an NNO (and in some other situations).
Let be a Heyting category. (This means, in particular, that we can interpret full first-order intuitionistic logic using the stack semantics.)
A commutative ring object in is an object equipped with morphisms , , , , and (where is the terminal object of and is the product operation in ) that make certain diagrams commute. (These diagrams may be found at ring object, in principle, although right now they're not there.)
Given a commutative ring object in , we define a binary relation on (that is a subobject of ) as
written in the internal language of . Then is a (Heyting) field object if is a tight apartness relation; that is if the following axioms (in the internal language) hold:
- irreflexivity: ,
- symmetry: (which can actually be proved using ),
- comparison: ,
- tightness: .
A (linearly) ordered field object in is a field object equipped with a binary relation such that the following axioms hold:
- strong irreflexivity: ,
- strong connectedness: ,
- transitivity: ,
- respecting addition: ,
- respecting multiplication: .
Given an ordered field object in , any object in , and subobjects and of , we say that is a Dedekind cut in (parametrised by ) if the following axioms hold:
- lower bound: ,
- upper bound: ,
- downward roundedness: ,
- upward roundedness: ,
- upward openness: ,
- downward openness: ,
- locatedness: ,
- separation: .
An ordered field object in is Dedekind complete if, given any object of and any Dedekind cut in parametrised by , there exists a morphism such that
Finally, a real numbers object in is a Dedekind-complete ordered field object.
In the last requirement, of Dedekind completeness, we postulate (under certain conditions) the existence of a morphism satisfying certain properties.
This morphism is in fact unique.
Here is an explicit ‘external’ proof:
Suppose that both satisfy the required properties, and consider , which is a subobject of . By the strong connectedness of , is contained in (factors through) , which is the union of and .
Now consider , and let be a generalised element of . If belongs to (factors through) , or equivalently belongs to , it follows that belongs to . Thus, also belongs to , or equivalently belongs to . By the strong irreflexivity of , this is contained in ; by the irreflexivity of , this is contained in (as a subobject of ). Since every that belongs to belongs to , is contained in (and so equals as a subobject) .
Similarly (either by swapping with or by using instead of ), is also . Therefore, is . By the tightness of , .
Here is an ‘internal’ proof, to be interpreted in the stack semantics of :
Suppose that both satisfy the required properties, and suppose that . By the strong connectedness of , or .
Now suppose that . It follows that belongs to , so . By the strong irreflexivity of , ; by the irreflexivity of , we have a contradiction.
Similarly (either by swapping with or by using instead of ), is also false. Therefore, is false. By the tightness of , .
In the definition of a Heyting field object, all of the axioms except the last are coherent and therefore make sense in any coherent category.
An object satisfying all but the last axiom of a field object is precisely a local ring object (so in particular an RNO is a local ring object).
It would be nice to say that a Heyting category with an RNO must have an NNO; after all, is contained in . However, my only argument is impredicative; although I don’t know a specific example, there could be a Π-pretopos with an RNO but no NNO. However, the argument works for a geometric Heyting category or a topos. (In light of the constructions below, the existence of an RNO is therefore equivalent to the existence of an NNO in a topos.)
If is an RNO in an infinitary Heyting category or topos, then there is unique subobject of that is both a sub-rig object of and an NNO under the operations and .
We usually speak of the RNO, if one exists. This is because any two RNOs in a Heyting category with an NNO are isomorphic, in an essentially unique way. (I can’t prove this without an NNO, although the previous theorem shows that we often have one.)
If and are both RNOs in a Heyting category with an NNO, then there is a unique isomorphism from to that preserves the structures on them (, , , , , ).
We can construct a real numbers object in the following cases (presumably among others):
- in a topos with an NNO;
- in a -pretopos with an NNO and weak countable choice;
- in a -pretopos with an NNO and subset collection.
(Actually, (1) is a special case of (3), but the usual construction in (1) is not available in (3). Thus, we still have three different constructions to consider.)
In a topos with an NNO
Let be an elementary topos with a natural numbers object . The Dedekind real numbers object of is the object of all Dedekind cuts. To be more precise, we will need to make some auxiliary definitions.
We first construct an integers object as follows. Let be the kernel pair of the addition map , and let be the product projections. We define to be the coequalizer of . A similar construction yields a rational numbers object .
We denote by the power object of in . A Dedekind cut is a generalized element of , satisfying the following conditions, expressed in the Mitchell–Bénabou language of and interpreted under Kripke–Joyal semantics:
The relation on is the order relation constructed in the usual way. We define to be the subobject of consisting of all Dedekind cuts as defined above.
In a -pretopos with
Summary: construct a Cauchy real numbers object and use (weak countable choice) to prove that it is an RNO.
Note that any Boolean topos with an NNO satisfies , so in all we have three different constructions available in that case.
In a -pretopos with an NNO and subset collection
Summary: modify the construction of a Cauchy real numbers object to use multi-valued Cauchy sequences.
The real numbers object in Set is the real line, the usual set of (located Dedekind) real numbers. Note that this is a theorem of constructive mathematics, as long as we assume that is an elementary topos with an NNO (or more generally a Π-pretopos with NNO and either WCC or subset collection).
In sheaves on a topological space
Let be a topological space, and its category of sheaves. It is well-known that is a Grothendieck topos (and so, a fortiori, an elementary topos), and the constant sheaf functor preserves finite limits and has the global section functor as a right adjoint. (Hence, and are the components of a geometric morphism .) The following claims are essentially immediate:
If is the set of natural numbers, then must be an NNO in , since has a right adjoint.
If is the set of integers, then is an integers object in (as defined above), since preserves finite limits and colimits.
Similarly, if is the set of rational numbers, then is a rational numbers object in .
Thus, for every topological space , the topos has a Dedekind real numbers object . Naïvely one might expect to be isomorphic to the constant sheaf , where is the classical set of real numbers, but this turns out not to be the case. Instead, we have a rather more remarkable result:
A Dedekind real numbers object in the topos is isomorphic to the sheaf of real-valued continuous functions on .
This allows us to define various further constructions on in internal terms in ; for example, a vector bundle over is an internal projective -module.
It is also possible to define the notion of a Cauchy real number object and construct one in any -pretopos with an NNO, but as the internal logic in general lacks weak countable choice, these are usually inequivalent. (There is also potentially a difference between the classical Cauchy RNO and the modulated Cauchy RNO; see definitions at Cauchy real number, to be interpreted in the stack semantics.)
Section D4.7 of
Discussion in homotopy type theory is in section 11 of of