Blog Archive

Saturday, May 2, 2026

Connected Objects in Extensive Categories

Connected Objects in Extensive Categories

Let C be a locally small extensive category that possesses a terminal object 1.

For an object XC, the following three conditions are equivalent:

  1. The hom-functor HomC(X,):CSet preserves finite coproducts.

  2. The canonical map

    HomC(X,1)+HomC(X,1)HomC(X,1+1)

    is a bijection (which forces X to be non-initial).

  3. X is not initial and cannot be written as a coproduct XX1+X2 with both X1,X2 non-initial.

An object satisfying (1)(3) is called connected. Examples: Top, Sch, Grpd and any topos (e.g. Quiver, GSet).

Remark

Fundamental groupoid and path-connectedness

Consider the fundamental groupoid functor

Π1:TopGrpd.

A topological space X is path-connected if and only if Π1(X) is a connected groupoid (equivalently, a connected object in Grpd). Indeed, connectedness of Π1(X) means exactly that there is a morphism (homotopy class of a path) between any two objects (points) — i.e., X is path-connected.


Proof of the equivalences

We assume throughout that C is extensive and has a terminal object 1. Coproduct injections are denoted i1:AA+B and i2:BA+B.

(1)(3)

Suppose Hom(X,) preserves finite coproducts.
Then Hom(X,0) (empty coproduct), so X is not initial.

Assume, for contradiction, that XX1+X2 with X1,X20.
The identity idX corresponds under the isomorphism XX1+X2 to a morphism f:XX1+X2.
Since Hom(X,) preserves the coproduct,

Hom(X,X1+X2)Hom(X,X1)Hom(X,X2),

the morphism f must belong to exactly one of the two summands.
Without loss of generality assume it belongs to the first; thus there exists u:XX1 such that f=i1u.

Now pull back the coproduct inclusions along f:

P1X1i1XfX1+X2i2P2X2

Extensivity gives XP1+P2. Because f factors through i1, the top square is a pullback of an isomorphism along i1, hence P1X. By disjointness of the coproduct, the pullback of i1 and i2 is the initial object 0; consequently P20. Thus XX+0, forcing X20 (by comparing the two decompositions), contradicting the assumption that X2 is non-initial.

Hence no non-trivial decomposition exists — i.e. condition (3) holds.

(3)(1)

Assume X is non-initial and cannot be split into two non-initial summands.
We show the natural map

jJHom(X,Aj)Hom(X,jJAj)

is a bijection for every finite family (Aj). By induction it suffices to treat binary coproducts A+B.

Take an arbitrary f:XA+B.
Pull back the injections along f:

X1Ai1XfA+Bi2X2B

Extensivity yields XX1+X2.

By condition (3), one of X1,X2 must be initial.
If X20, then f factors uniquely through i1:AA+B (the map from 0 is unique and the universal property of the coproduct collapses onto the remaining summand).
Hence every morphism XA+B lands entirely in one summand, which is precisely the statement that

Hom(X,A)Hom(X,B)Hom(X,A+B)

is an isomorphism. Together with Hom(X,0) (true for any non-initial X because 0 is strict in an extensive category), we obtain preservation of all finite coproducts, i.e. (1).

(3)(2)

Decompositions of X are in bijection with morphisms into 1+1:

  • Decomposition map: If XX1+X2, send it to the copairing X11, X21, yielding f:X1+1.

  • Map decomposition: Given f:X1+1, pull back 1:11+1 and 2:11+1 to obtain P1,P2 with XP1+P2.

These constructions are mutually inverse, establishing

{decompositions XX1+X2}Hom(X,1+1).

The two summands on the left of the canonical map

Hom(X,1)+Hom(X,1)Hom(X,1+1)

correspond exactly to the trivial decompositions XX+0 and X0+X.

  • If X satisfies (3), every decomposition is trivial, so the map is bijective and X is non-initial, giving (2).

  • If (2) holds, the bijection forces every map X1+1 to come from a trivial decomposition, hence X has no non-trivial decomposition. Moreover X cannot be initial (otherwise left side would be 1+1, right side a singleton). Thus (3) holds.


Image of a connected object is connected

Assume now that C has an (epi, mono) factorization system.

Proposition. If X is connected and f:XY is any morphism, then Im(f) is connected.

Proof.

Assume Im(f) is disconnected, then there exists a non-trivial map k:Im(f)1+1

Then ke:X1+1 is a nontrivial map as well, contradiction. ◻

Popular Posts