Let $C$ be an arbitrary category, $X,Y\in Obj(C)$ and $$f:X\to Y,$$ $$g:X\to Y.$$ How to define coincidence of such morphisms?
- Anybody can ask a question
- Anybody can answer
- The best answers are voted up and rise to the top
That is, just as an abstract group has a primitive binary operation satisfying certain "equalities" but does not further define what constitutes "equality" of group elements, an abstract category is defined in terms of objects and morphisms (arrows) between them, without proposing any definition of equality of morphisms (or objects) beyond saying certain properties stated as equalities are satisfied.
In this context "primitive" has the connotation of basic concepts in terms of which all the other concepts may be defined. For example, equality of two functors might be expressed in terms of the equal assignments of objects and morphisms.
Added: If one wishes to be "bare bones" about the primitives, objects can be dispensed with in favor of doing everything in terms of morphisms, something that Eilenberg and MacLane noted in their original 1945 exposition.