morphism
A morphism is the categorical word for an arrow — the abstraction of “a structure-respecting map from one thing to another.” In familiar categories morphisms really are functions of some kind (homomorphisms, continuous maps, linear maps), but the category-theoretic stance is to treat them as primitive: a morphism need not be a function on underlying elements at all. What is fixed is that each morphism f has a source object (domain) and a target object (codomain), written f : A -> B.
Several adjectives classify morphisms purely by how they compose, not by anything internal. A morphism f : A -> B is a monomorphism if it is left-cancellable: f ∘ g = f ∘ h forces g = h. It is an epimorphism if it is right-cancellable: g ∘ f = h ∘ f forces g = h. It is an isomorphism if it has a two-sided inverse g with g ∘ f = id_A and f ∘ g = id_B. An endomorphism is a morphism A -> A, and an automorphism is an invertible endomorphism.
A common trap: in Set, monomorphism means injective and epimorphism means surjective, and a map that is both is an isomorphism. But this last implication fails in general categories. In the category of rings, the inclusion Z -> Q is both monic and epic yet is not an isomorphism. So “mono and epi” is strictly weaker than “iso” outside nice settings, and the three adjectives must be kept conceptually distinct.
In Vect over R, a morphism R^3 -> R^2 is a linear map, represented by a 2-by-3 matrix; it is monic iff its kernel is 0 and epic iff its image is all of R^2. The matrix [1, 0, 0; 0, 1, 0] is epic but not monic.
Mono/epi for linear maps coincide with injective/surjective.