# Why is an object not defined as identity morphism?

I've seen that there was a single-sorted definition of a category. In some ways, it seems more understandable than the original definition.

I don't know much about category theory. But I would like to know how each definition is useful.

• The two-sorted definition corresponds to how most people think of categories, and how they talk about categories. For example, we speak of the categories of sets (not of functions), of groups (not of group homomorphisms), of topological spaces (not of continuous maps), etc. (If I remember correctly, Ehresmann did write about the categories of functions, of homomorphisms, of continuous maps, etc., but that never caught on.) – Andreas Blass Jul 19 at 16:39
• One can identify an object with its identity morphism. Then we have a category consisting of morphisms that can be composed. Usually the composition is thought of one function for the whole category. However, the composition function can be broken up into a performing compositions, one such place for each object. In higher order category theory there may be 2-morphisms between the usual morphisms. They may also be places or performing compositions. With the appropriate additional axioms higher order category theory fits into this scheme. – Jay Kangel Jul 19 at 17:52
• My impression is that the single-sorted POV is slightly more widespread in the context of groupoids (considered as small categories where every morphism is iso) but I am not a specialist and would be happy to be corrected here by others. However, one often ends up introducing the "unit space" of the groupoid which is the "set of objects", so even then two sorts seem to emerge – Yemon Choi Jul 19 at 18:20
• Another area where the single sorted definition is more widespread is when working with strict $n$-categories or strict $\infty$-categories. Working with the single sorted definition allows to compose arrows of different dimension without writing iterated identities everywhere. It makes the manipulation of expressions a little more bearable, so lots of paper use it (where the "lots" need to be taken relatively as there is not that many paper on that topics anyway) – Simon Henry Jul 19 at 19:54
• Actually I would argue that the homsets-definition (ncatlab.org/nlab/show/category#AFamilyOfCollectionsOfMorphisms) is the one that corresponds to how more people think of categories in practice. – Mike Shulman Jul 21 at 23:26