Alphabeta Math
TheoremStatement: Literature-sourcedProof: AI-adaptedSession-authored (Fable 5 assisted)precheck passaudited 2026-08-26
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

An end is a limit over the twisted arrow category, and a coend is a colimit over its opposite

Statement

Let T:Cop×CD be a functor and let π:Tw(C)Cop×C be the twisted arrow projection (The twisted arrow category and its projection to Cop×C).

Ends. The assignments ωξ with ξf:=T(1c,f)ωc for f:cc, and ξω with ωc:=ξ1c, are mutually inverse and give an isomorphism of categories Wd(T)Cone(Tπ) that leaves the vertex unchanged (Wedges and cowedges, and the categories they form, Constant diagrams, cones, cocones, and their morphisms). Consequently an end is the limit over the twisted arrow category: a wedge (e,ω) is an end of T exactly when the corresponding cone is a limit of Tπ (The end and the coend of a functor Cop×CD, Limits and colimits as terminal cones and initial cocones, with existence and uniqueness in their universal properties), either exists exactly when the other does, and then

cT(c,c)    limTw(C)Tπ

by the unique isomorphism compatible with every component (Any two limits, or any two colimits, of one diagram are uniquely isomorphic compatibly with their structure maps).

Coends. Let πsw:Tw(C)opCop×C (Opposite category Cop) send an object f:cc to the pair (c,c), with the domain and the codomain of f swapped, and send the morphism gf of Tw(C)op determined by (a,b):fg to the morphism (b,a):(d,d)(c,c). Then πsw is a functor, Cwd(T)Cocone(Tπsw) by mutually inverse assignments leaving the vertex unchanged, and a coend of T is exactly a colimit of Tπsw:

cT(c,c)    colimTw(C)opTπsw.

Facts & Assumptions

Given: A functor T:Cop×CD and the twisted arrow projection π.

[F1]

A wedge from d to T is a dinatural transformation from a constant functor to T: a family ωc:dT(c,c) with T(1c,f)ωc=T(f,1c)ωc for every f:cc; a cowedge from T to d is a family ρc:T(c,c)d with ρcT(f,1c)=ρcT(1c,f); a morphism of wedges is a morphism of the vertices commuting with every component (Wedges and cowedges, and the categories they form).

[F2]

A cone over D:JC with apex c is a family λj:cD(j) satisfying D(u)λj=λk for u:jk; a cocone under D is a family ρj:D(j)c satisfying ρkD(u)=ρj; a morphism of cones (c,λ)(c,λ) is h:cc with λjh=λj (Constant diagrams, cones, cocones, and their morphisms).

[F3]

The objects of Tw(C) are the morphisms of C, a morphism fg is a pair (a,b) with bfa=g for a:dc and b:cd, and the projection π sends f:cc to (c,c) and (a,b) to (a,b) (The twisted arrow category and its projection to Cop×C).

[F4]

A limit of D is a terminal cone: explicitly, for every cone (X,ξ) there exists a unique morphism u:XL such that λju=ξj for every j. A colimit is an initial cocone (Limits and colimits as terminal cones and initial cocones, with existence and uniqueness in their universal properties).

[F5]

An end of T is a terminal object of the category of wedges over T and a coend an initial object of the category of cowedges under T; in short, an end is a terminal wedge and a coend an initial cowedge (The end and the coend of a functor Cop×CD).

[F6]

The opposite category has the same objects and reverses every morphism: Cop(A,B)=C(B,A) (Opposite category Cop).

[L1]

If (L,λ) and (L,λ) are limits of one diagram D, there is a unique isomorphism u:LL satisfying λju=λj for every j; dually for colimits (Any two limits, or any two colimits, of one diagram are uniquely isomorphic compatibly with their structure maps).

Proof

technique · direct
1.1

Let ω be a wedge with vertex X and put ξf:=T(1c,f)ωc for f:cc, so that ξf:XTπ(f). This family is a cone over Tπ: for a morphism (a,b):fg of Tw(C), with g:dd and bfa=g, functoriality of T gives T(a,b)T(1c,f)=T(a,bf)=T(1d,bf)T(a,1c), the wedge equation at a:dc gives T(a,1c)ωc=T(1d,a)ωd, and T(1d,bf)T(1d,a)=T(1d,bfa)=T(1d,g); hence Tπ(a,b)ξf=T(1d,g)ωd=ξg.

F1F2F3
1.2

Let ξ be a cone over Tπ with apex X and put ωc:=ξ1c. The pairs (1c,f):1cf and (f,1c):1cf are morphisms of Tw(C), since f1c1c=f and 1c1cf=f, so the cone equation at each of them gives T(1c,f)ωc=ξf and T(f,1c)ωc=ξf; the two left-hand sides are therefore equal and ω is a wedge.

F1F2F3
1.3

The assignment πsw is a functor: it sends the identity (1c,1c) of f to the identity of (c,c), and if (a,b):fg and (a,b):gh then their composite in Tw(C) is (aa,bb), whose image (bb,aa) is exactly the composite in Cop×C of the images (b,a) and (b,a), taken in the order these two morphisms compose in Tw(C)op.

F3F6
2.1

The two assignments are mutually inverse: from a wedge, ξ1c=T(1c,1c)ωc=ωc; from a cone, the family rebuilt in step 1.1 has T(1c,f)ξ1c=ξf by the first computation of step 1.2. A morphism h of the vertices satisfies ωch=ωc for every c exactly when it satisfies ξfh=ξf for every f, since each family determines the other by the displayed formulas. So Wd(T) and Cone(Tπ) are isomorphic categories over the identity on vertices, terminal objects correspond, and by [F5] and [F4] an end of T is precisely a limit of Tπ; [L1] then supplies the unique component-compatible isomorphism between any two such limits.

F4F5L1step 1.1step 1.2
3.1

For cowedges, put ρf:=σcT(f,1c) for a cowedge σ and f:cc, so ρf:Tπsw(f)=T(c,c)X. Given (a,b):fg, so a morphism gf of Tw(C)op sent by Tπsw to T(b,a), functoriality gives T(f,1c)T(b,a)=T(bf,a)=T(bf,1c)T(1d,a), the cowedge equation at bf:cd gives σcT(bf,1c)=σdT(1d,bf), and T(1d,bf)T(1d,a)=T(1d,bfa)=T(1d,g), so ρfT(b,a)=σdT(1d,g)=σdT(g,1d)=ρg by the cowedge equation at g; conversely σc:=ρ1c recovers a cowedge from a cocone through the two morphisms of step 1.2 read in Tw(C)op, and the two assignments are mutually inverse exactly as in step 2.1. Hence Cwd(T)Cocone(Tπsw), initial objects correspond, and a coend of T is a colimit of Tπsw over Tw(C)op, the integrand being read with domain and codomain swapped.

F3F5F6step 1.3step 2.1

Remarks

The swap in the coend clause is not cosmetic and is not a consequence of formal duality applied carelessly. Dualising D turns a cowedge under T into a wedge over T viewed in Dop, and the index category that then computes it is the opposite of the twisted arrow category, with the integrand evaluated at (c,c) rather than (c,c). Taking the colimit of Tπ over Tw(C)op instead — the same diagram whose limit is the end — gives a different object, and FALSE: under this page's convention a coend is the colimit of the same twisted-arrow diagram whose limit is the end computes both on a two-object category to show they differ.

No smallness hypothesis appears anywhere above: the two categories are isomorphic whatever the size of C, and the statement is about which universal objects exist, not about whether the index category is a set. The size hypothesis enters only when existence is to be deduced from completeness, which is Ends exist over a small index category in a complete target, and coends in a cocomplete one.

Depends on

Used by

Dependency tree · two levels

16 results within two dependency steps of this one, each drawn at its shortest distance from it. An arrow runs from a result to what uses it, so the chart reads left to right and ends at this result, which carries a heavier outline. Every node is a link to that result. Click elsewhere on the chart to enlarge it.

Sources