List of Implications
The following 93 implications and equivalences are available*.
abelian is equivalent to additive and equalizers and coequalizers and mono-regular and epi-regular
by definition- abelian implies Malcev
additive is equivalent to preadditive and finite products
by definitionadditive implies disjoint finite coproducts
If is a morphism that factors through and , then and , so .additive and pullbacks and subobject classifier implies trivial
see MSE/4086192binary products and equalizers implies pullbacks
The pullback of and is the equalizer of .binary products and pullbacks implies equalizers
The equalizer of is the pullback of with the diagonal .binary products and inhabited implies connected
For any two objects we have the zig-zag .cartesian closed and finite coproducts implies distributive
Each functor is left adjoint and hence preserves finite coproducts (in fact, all colimits).cartesian closed and coproducts implies infinitary distributive
Each functor is left adjoint and hence preserves coproducts (in fact, all colimits).cartesian closed implies finite products
by definitioncartesian closed and initial object implies strict initial object
See the nLab.complete implies finitely complete and filtered limits and wide pullbacks and connected limits
trivialcomplete is equivalent to products and equalizers
Mac Lane, V.2, Cor. 2connected limits is equivalent to wide pullbacks and equalizers
countable products implies finite products
trivialdiscrete implies essentially discrete and locally small and skeletal
trivialdisjoint coproducts is equivalent to coproducts and disjoint finite coproducts
easydisjoint finite coproducts implies finite coproducts
by definitiondisjoint finite coproducts and thin implies trivial
For every object the two inclusions must be equal, so their equalizer is , but also since the coproduct is disjoint. Hence .distributive implies finite products and finite coproducts
by definitiondistributive implies strict initial object
See the nLab.distributive and exact filtered colimits and coproducts implies infinitary distributive
Each functor preserves finite coproducts and filtered colimits, hence all coproducts.elementary topos is equivalent to cartesian closed and finitely complete and subobject classifier
by definitionelementary topos implies finitely cocomplete and disjoint finite coproducts and epi-regular
Mac Lane & Moerdijk, Cor. IV.5.4, Cor. IV.10.5, Thm. 4.7.8.elementary topos and locally essentially small implies well-copowered
This follows from Mac Lane & Moerdijk, Theorem IV.7.8 (and Prop. I.3.1).equalizers implies Cauchy complete
If is an idempotent, then the equalizer of provides a splitting of .equalizers and countable products implies sequential limits
Mac Lane, V.2, Prop. 3. The proof can easily be adapted to this case. Namely, the limit of is the equalizer of two suitable endomorphisms of .essentially discrete is equivalent to thin and groupoid
trivialessentially discrete implies locally essentially small and connected limits
trivialessentially discrete and connected implies trivial
trivialessentially finite and finite products implies thin
Mac Lane, V.2, Prop. 3. The proof can easily be adapted to this case.essentially finite implies essentially small
trivialessentially small implies well-powered and well-copowered and locally essentially small
trivialessentially small and products implies thin
Mac Lane, V.2, Prop. 3. The proof works for any category with products.essentially small and thin and complete implies cocomplete
The supremum of a subset in a (small) partial order is the infimum of the set of upper bounds.essentially small and thin and complete and infinitary distributive implies cartesian closed
This is an application of the adjoint functor theorem. Specifically, if is a complete lattice in which always holds, then the functor is a left adjoint because it preserves all suprema.exact filtered colimits implies filtered colimits and finitely complete
by definitionfiltered limits implies sequential limits
trivialfinitary algebraic implies locally finitely presentable
Adamek-Rosicky, Cor. 3.7finite implies small and essentially finite
trivialfinite products is equivalent to terminal object and binary products
The non-trivial direction follows since finite products can be constructed recursively via .finite products and filtered limits implies products
The product is the filtered limit of the finite partial products where ranges over the finite subsets of .finite products and sequential limits implies countable products
If is an infinite sequence of objects, then their product is the limit of the sequence .finitely complete is equivalent to finite products and equalizers
Mac Lane, V.2, Cor. 1Grothendieck abelian is equivalent to abelian and coproducts and generator and exact filtered colimits
by definitionGrothendieck abelian implies locally presentable
See Deriving Auslander's formula, Cor. 5.2, or Sheafifiable homotopy model categories, Prop. 3.10.Grothendieck abelian implies cogenerator
Kashiwara-Schapira, Thm. 9.6.3Grothendieck abelian and self-dual implies trivial
This follows since the dual of a non-trivial Grothendieck abelian category cannot be Grothendieck abelian. See Peter Freyd, Abelian categories, p. 116.Grothendieck topos is equivalent to elementary topos and coproducts and generator and locally essentially small
Mac Lane & Moerdijk, Appendix, Prop. 4.4Grothendieck topos implies locally presentable and cogenerator
For "locally presentable" see Prop. 3.4.16 in Handbook of Categorical Algebra Vol. 3. For "cogenerator" see the nLab.groupoid implies self-dual and mono-regular and pullbacks and filtered limits and left cancellative and well-powered
easygroupoid and equalizers implies thin
The equalizer of any parallel pair must be an isomorphism, so .groupoid and binary products and inhabited implies trivial
Let be an inhabited groupoid with binary products. Then it is connected, so we may assume for a group with unique object . But then , so there are such that , is bijective. From here it is an easy exercise to deduce .groupoid and initial object implies trivial
easyinfinitary distributive implies finite products and coproducts
by definitioninfinitary distributive implies distributive
trivialleft cancellative and coequalizers implies thin
If are two parallel morphisms, then their coequalizer is a regular epimorphism, but also a monomorphism by assumption, so it must be an isomorphism. But this means that .left cancellative and initial object implies strict initial object
It suffices to prove that in general any monomorphism into an initial object is an isomorphism. If is the unique morphism, then since is initial. But then is a split epimorphism and a monomorphism, hence an isomorphism.left cancellative implies Cauchy complete
Any idempotent monomorphism must be the identity and therefore splits.left cancellative and right cancellative and balanced implies groupoid
triviallocally finitely presentable implies locally presentable
Locally finitely presentable categories are by definition the locally -presentable categories.locally finitely presentable implies exact filtered colimits
Special case of Adamek-Rosicky, Prop. 1.59 with .locally finitely presentable implies locally ℵ₁-presentable
triviallocally presentable implies locally essentially small and well-powered and well-copowered and complete and cocomplete and generator
For the non-trivial conclusions see Adamek-Rosicky, Thm. 1.20, Cor. 1.28, Rem. 1.56, Thm. 1.58.locally presentable and self-dual implies thin
This follows from Adamek-Rosicky, Thm. 1.64.locally small implies locally essentially small
triviallocally ℵ₁-presentable implies locally presentable
trivialMalcev implies finitely complete
by definitionmono-regular implies balanced
Any regular monomorphism that is an epimorphism must be an isomorphism.pointed is equivalent to zero morphisms and initial object
easypointed and cartesian closed implies trivial
We have for every object .preadditive implies locally essentially small and zero morphisms
trivialpreadditive and finite coproducts implies finite products
Mac Lane, VIII.2., Theorem 2products implies finite products and countable products
trivialpullbacks and terminal object implies binary products
If is a terminal object, then .right cancellative and initial object implies strict initial object
Let be a morphism. Let be the unique morphism. It is an epimorphism by assumption. Also, since is initial. But then is a split monomorphism and an epimorphism, hence an isomorphism.small implies locally small and essentially small
trivialsplit abelian implies abelian
by definitionstrict initial object implies initial object
by definitionstrict initial object and pointed implies trivial
If is the zero object, then for every object the unique morphism is an isomorphism by assumption.subobject classifier and locally essentially small implies well-powered
Mac Lane & Moerdijk, Prop. I.3.1subobject classifier implies finitely complete and mono-regular
The first part holds by convention, and the second part: any monomorphism is the equalizer of \chi_U,\chi_X : X \to \Omega$.terminal object implies connected
If denotes the terminal object, then for any two objects we have the zig-zag .thin implies equalizers and left cancellative
Any two parallel morphisms are equal, so their equalizer is the identity, and every morphism is a monomorphism as well.thin and finitely complete implies Malcev
In a thin category, every subobject of containing is already .trivial implies finitary algebraic and Grothendieck topos and split abelian and self-dual and essentially discrete and essentially finite
trivialwide pullbacks and terminal object implies complete
See the nLab.wide pullbacks is equivalent to pullbacks and filtered limits
To prove , a wide pullback can be constructed as a filtered limit of finite pullbacks, and finite pullbacks can be reduced to binary products (the empty-indexed pullback always exists). Conversely, assume that wide pullbacks exist in . For every object then the slice category has wide pullbacks and a terminal object, hence is complete. Since a filtered limit can be finally reduced to such a slice, we are done.zero morphisms and inhabited implies connected
trivial
*Deductions from these implications are automatically incorporated into each category whenever applicable. For instance, if a category is identified as complete, the property of having a terminal object is automatically inferred and added.
Moreover, implications are automatically dualized when the corresponding dual properties exist. For example, the statement that finitely complete categories with filtered limits are complete automatically implies that finitely cocomplete categories with filtered colimits are cocomplete. Similarly, if a category is self-dual and, for example, complete, it is automatically inferred to be cocomplete as well. Use this link to see all implications, including the dualizations.