Reverse every arrow and the theorem still holds
The opposite category has the same objects and all arrows reversed. Any statement provable from the axioms therefore has a dual, obtained by reversing arrows, and the dual is automatically a theorem. This is why projective and injective, kernel and cokernel, product and coproduct, limit and colimit come in pairs, and why half the proofs in the subject can be omitted. What duality does not do is transport constructions or examples.
Learning objectives
- Define the opposite category and the dual of a statement.
- List the standard dual pairs in homological algebra.
- Apply the duality principle to obtain a theorem for free.
- Explain why duality does not transport constructions.
Section 01The principle
Cop has the objects of C and Cop(A, B) = C(B, A). Composition reverses. Since the axioms of a category are self-dual, Cop is a category, and (Cop)op = C.
If a statement is provable for all categories from the axioms, so is its dual. No separate proof is required — the dual proof is the original proof read in Cop. This is a metatheorem about the logic, not a construction inside any one category.
Section 02The dual pairs
| Notion | Dual |
|---|---|
| Monomorphism | Epimorphism |
| Kernel | Cokernel |
| Product | Coproduct (direct sum) |
| Pullback | Pushout |
| Limit | Colimit |
| Projective | Injective |
| Free | Cofree |
| Projective resolution | Injective resolution |
| Left derived functor | Right derived functor |
| Right exact | Left exact |
| Initial object | Terminal object |
| Left adjoint | Right adjoint |
Some notions are their own duals: isomorphism, zero object, biproduct, and the axioms of an abelian category as a whole. That last point is why the theory is so symmetric — the opposite of an abelian category is abelian.
Section 03Where duality stops
Duality operates on statements, not on constructions or examples. ModΛop is a perfectly good abelian category, but it is not a category of modules over any ring, so nothing concrete transports.
Theorems
Every result about projectives yields a result about injectives with no further work.
Constructions
Free modules dualise to cofree modules only because a separate construction was found. There is no dual of a basis.
Existence
Enough projectives is immediate; enough injectives requires a proof. Projective covers may not exist even though injective hulls always do.
‘By duality’ is valid when invoking a dual theorem. It is not valid for asserting that a dual object exists, or that a dual construction behaves the same way. The asymmetry between projective covers and injective hulls is the standard cautionary example.
ReferenceFrequently asked questions
Is the opposite of a module category a module category?
Almost never. It is abelian, but by the Freyd–Mitchell theorem it embeds in a module category rather than being one. This is precisely why duality gives theorems but not models.
Why is an abelian category self-dual?
Because each axiom is either self-dual or paired with its dual in the list: existence of kernels pairs with cokernels, and the condition that every mono is a kernel pairs with every epi being a cokernel. Reversing arrows permutes the axioms among themselves.
Does duality apply to statements about specific objects?
Only if the statement is expressed purely in arrow terms. A claim about ℤ as an abelian group has no dual, because ℤ is not defined by a universal property that survives arrow reversal.
NavigateContinue in this stream
Curated next steps from this page. The site also surfaces algorithmically related reading below.
ProvenanceSources and further reading
This page is an original KEVOS explanatory article. It presents the underlying mathematics — definitions, algorithms, complexity results and selection criteria — in KEVOS editorial voice. No text is reproduced from any copyrighted source. Where numerical tables are relevant, KEVOS links to live authoritative databases rather than republishing static values.