So, I read the John Baez essay "Lectures on n-categories and cohomology" and I understand the notion of a (-1)-category" and a (-2)-category" and how to derive them. However, I'm not totally clear on what a (-1)-morphism is.

At nLab at http://ncatlab.org/nlab/show/k-morphism, I found:

For the purposes of negative thinking, it may be useful to recognise that every ∞-category has a (−1)-morphism, which is the source and target of every object.

I have an idea of how this works, but I'm not sure if it's correct. Basically, my thinking is that if $A$ and $B$ are objects in a 1-category $C$, and $f$ and $g$ are both parallel morphisms that map $A\to B$, and we take $S = hom_C(A,B)$ then we get a 0-category (a set) that contains $f$ and $g$ as elements (and their identities as morphisms). Both $hom_S(f,g)$ and $hom_S(f,g)$ each give us an empty (-1)-category (obviously isomorphic). Then $hom_S(f,f)$ and $hom_S(g,g)$ each give us a (-1)-category with a single object (also obviously isomorphic). Now, we treat the isomorphic (-1)-categories as equivalent, so there are only the two: the empty one named $False$; and, the non-empty one named $True$. So, since all non-empty (-1)-categories are equivalent, we decide that their morphisms are also equivalent.

So, here's my question that I'm not sure about:

  • Question 1: Am I correct in my understanding that the identity morphism of an object in a (-1)-category is a (-1)-morphism? It's a 0-morphism if the (-1)-category were treated as a 1-category, but if we take into account that the (-1)-category is derived from some higher category, then does that effectively make the single 1-morphism in the (-1)-category a (-1) morphism in the higher-category?

This seems like it would make sense of the assertion that the single (-1)-morphism is the source and target of every 0-morphism/object.

I also came across another passage meant to explain the (-1)-morphism that threw me off a little bit:

Also note that every k-morphism has k+1 identity (k+1)-morphisms, which just happen to all be the same (which can be made a result of the Eckmann–Hilton argument). Thus, the (−1)-morphism has 0 identity 0-morphisms, so we don't need any object. (This confused me once.)
  • Question 2: Does this mean that an empty category still has a (-1)-morphism? If so, where does it come from?
  • Question 3: Does this mean that a 1-morphism has 2 identity 2-morphisms? My understanding of an "identity" necessitates uniqueness. Having 2 identities for the same thing seems to cause a contradiction. How can I reconcile this?
  1. An object is always a 0-morphism, so the identity morphism of an object is a 1-morphism.
  2. Yes, the empty category has a $(-1)$-morphism. The convention is that the free $\infty$-category generated by a $(-1)$-morphism is the empty $\infty$-category, and so every $\infty$-category has a unique $(-1)$-morphism. Note that this accords with the homotopy hypothesis: it is commonly accepted that a $(-1)$-cell should be the empty space, because this yields reduced homology in a natural way.
  3. Higher identity morphisms are unique in an $\infty$-category with strict interchange laws, as noted in the quote. In principle they could be different (but equivalent): an $n$-cube can be degenerate in at least $n$ different ways, if we respect orientation.

For an example, take the $\infty$ category $Kom$, i.e. the nerve of the category of chain complexes over a field. A $-1$ morphism is a map from one chain complex to another chain complex of degree $-1$.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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