Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

So I've been thinking about the definition of categories as just arrows with a defined composition (i.e. without objects). I understand this is silly, but it's fun and I have a question about it: functors are easy to define in this setting; they're just maps that preserve the composition. But how does one discuss natural transformations in this setting, since we don't have objects around to "index" them by? I suppose we could index by identity arrows, but this feels wrong somehow, and I'm sure one of you smart people has a better answer.

share|cite|improve this question
up vote 7 down vote accepted

A natural transformation of functors $\mathbb{C} \to \mathbb{D}$ is the same thing as a functor $\mathbb{2} \times \mathbb{C} \to \mathbb{D}$, where $\mathbb{2} = \{ 0 \to 1 \}$. The domain of the natural transformation is the restriction to $\{ 0 \} \times \mathbb{C}$, and the codomain is the restriction to $\{ 1 \} \times \mathbb{D}$.

share|cite|improve this answer
How does composition of natural transformations work when viewed this way? – Shay Aug 1 '14 at 20:22
Use the category $\mathbb{3} = \{ 0 \to 1 \to 2 \}$. (Exercise.) – Zhen Lin Aug 1 '14 at 20:23
Oh my goodness, why did I never realize that a natural transformation looks like a... for lack of a better term, pre-homotopy? facepalm – Malice Vidrine Aug 1 '14 at 20:29
I like to think that is because the definition of homotopy is wrong! But unfortunately exponential objects in $\mathbf{Top}$ are a bit complicated to describe (assuming they even exist). – Zhen Lin Aug 1 '14 at 20:31
See also… – Martin Brandenburg Aug 1 '14 at 20:59

Identity arrows are it. For an identity $x$ and functors $F,G$, the component of a natural transformation $\sigma:F\to G$ at $x$ is going to be a morphism for which $F(x)$ and $G(x)$ are left and right identities, respectively.

Keep in mind in the normal definition of a category, there's a bijection between objects and their identities; indexing with either one gives you essentially the same natural transformation even before we consider tossing out the objects.

share|cite|improve this answer
That is correct, but it is just an immediate reformulation (using the equivalence between the two definitions of a category). – Martin Brandenburg Aug 1 '14 at 21:01
True enough, and that was how I hoped it would be read. The way that the original poster thinks "feels wrong" works in a perfectly adequate and straightforward way. – Malice Vidrine Aug 1 '14 at 21:13
(Naturally, Zhen Lin's answer is much more interesting.) – Malice Vidrine Aug 1 '14 at 21:31

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.