Fundamental theorem of topos theory

The fundamental theorem of topos theory states that the slice \mathbf{E} / X of a topos \mathbf{E} over any one of its objects X is itself a topos. Moreover, if there is a morphism f : A \rightarrow B in \mathbf{E} then there is a functor f^*: \mathbf{E} / B \rightarrow \mathbf{E} / A which preserves exponentials and the subobject classifier.

The pullback functor

For any morphism f in \mathbf{E} there is an associated "pullback functor"  f^* := - \mapsto f \times - \rightarrow f which is key in the proof of the theorem. For any other morphism g in \mathbf{E} which shares the same codomain as f, their product f \times g is the diagonal of their pullback square, and the morphism which goes from the domain of f \times g to the domain of f is opposite to g in the pullback square, so it is the pullback of g along f, which can be denoted as f^*g.

Note that a topos \mathbf{E} is isomorphic to the slice over its own terminal object, i.e. \mathbf{E} \cong \mathbf{E} / 1, so for any object A in \mathbf{E} there is a morphism f : A \rightarrow 1 and thereby a pullback functor f^* : \mathbf{E} \rightarrow \mathbf{E} / A, which is why any slice \mathbf{E} / A is also a topos.

For a given slice \mathbf{E} / B let X \over B denote an object of it, where X is an object of the base category. Then B^* is a functor which maps:  - \mapsto {B \times - \over B} . Now apply f^* to B^*. This yields

 f^* B^* : - \mapsto {B \times - \over B} \mapsto {{A \over B} \times {B \times - \over B} \over {A \over B}} \cong {\Big({A \times_B \, B \times - \over B}\Big) \over {A \over B}} \cong {A \times_B B \times - \over A} \cong {A \times - \over A} = A^*

so this is how the pullback functor f^* maps objects of \mathbf{E} / B to \mathbf{E} / A. Furthermore, note that any element C of the base topos is isomorphic to {1 \times C \over 1} = 1^* C, therefore if f : A \rightarrow 1 then f^*: 1^* \rightarrow A^* and f^* : C \mapsto A^* C so that f^* is indeed a functor from the base topos \mathbf{E} to its slice \mathbf{E} / A.

Logical interpretation

Consider a pair of ground formulas φ and ψ whose extensions [\_|\phi] and [\_|\psi] (where the underscore here denotes the null context) are objects of the base topos. Then φ implies ψ iff there is a monic from [\_|\phi] to [\_|\psi]. If these are the case then, by theorem, the formula ψ is true in the slice \mathbf{E} / [\_|\phi], because the terminal object [\_|\phi] \over [\_|\phi] of the slice factors through its extension [\_|\psi]. In logical terms, this could be expressed as

 \vdash \phi \rightarrow \psi \over \phi \vdash \psi

so that slicing \mathbf{E} by the extension of φ would correspond to assuming φ as a hypothesis. Then the theorem would say that making a logical assumption does not change the rules of topos logic.

References

This article is issued from Wikipedia - version of the Saturday, April 16, 2016. The text is available under the Creative Commons Attribution/Share Alike but additional terms may apply for the media files.