My question concerns Section 2 of the article "The Simplicial Model of Univalent Foundations (after Voevodsky)" (https://arxiv.org/pdf/1211.2851.pdf).

Let $\alpha$ be a strongly inaccessible cardinal.

Let $f : X \to Y$ be a map of simplicial sets.

  1. We say that $f$ is well-ordered if it is equipped with a well-ordering of $Y_x := f_n^{-1}(x)$ for each simplex $x\in X_n$.

  2. We say that $f$ is $\alpha$-small if $\left\lvert{Y_x}\right\rvert < \alpha$ for every simplex $x$.

Let $f: X \to Y$ and $g : Z \to Y$ be well-ordered simplicial maps. A morphism $f \to g$ is a fiber-preserving simplicial map $h : X \to Z$ such that $h_n : f_n^{-1}(y) \to g_n^{-1}(y)$ is order-preserving for every natural number $n$ and every $y\in Y_n$.

Define the functor $\mathcal{U}_{\alpha} : \mathbf{sSet}^{\text{op}} \to \mathbf{Set}$ so that $\mathcal{U}_{\alpha}(X)$ consists of all isomorphism classes of $\alpha$-small well-ordered Kan fibrations $Y\to X$. Also, let $$ \mathrm{U}_{\alpha} = \mathcal{U}_{\alpha}\circ \mathcal{Y}^{\text{op}} : \varDelta^{\text{op}} \to \mathbf{Set} $$ where $\mathcal{Y} :\varDelta \to \mathbf{sSet}$ denote the Yoneda embedding.

The authors say that if $\beta <\alpha$ is also inaccessible, then the unique map $\mathrm{U}_{\beta} \to 1$ is $\alpha$-small (p. 23, near the bottom). This amounts to saying that the set

$$ \left(\mathrm{U}_{\beta}\right)_n = \mathcal{U}_{\beta}(\Delta[n]) $$ has cardinality $<\alpha$ for each $n$.

I am unable, however, to see a set-theoretic justification for this. I'd be grateful if someone could provide one!


$\mathcal{U}_\beta(\Delta[n])$ is the set of isomorphism classes of $\beta$-small well-ordered fibrations over $\Delta[n]$. Such a thing is uniquely determined by an isomorphism class of $\beta$-small well-orderings over each element of $\Delta[n]$, together with the face and degeneracy maps between them. There are $\beta$ isomorphism classes of $\beta$-small well-orderings, and there are countably many elements of $\Delta[n]$. Since $\beta \cdot \aleph_0 = \beta$, there are $\beta$ choices of these fibers. Now for every face and degeneracy map, the domain and codomain are $\beta$-small, hence (since $\beta$ is inaccessible) there are $<\beta$ possible maps. Since there are countably many face and degeneracy maps, there are no more than $\beta \cdot \aleph_0 = \beta$ choices of the face and degeneracy maps. Thus there are overall $\beta$ choices, and $\beta<\alpha$.

  • $\begingroup$ To clarify, do you mean the face and degeneracy maps between the well-ordered fibers? $\endgroup$ – CuriousKid7 Mar 30 '20 at 1:26
  • $\begingroup$ Yes, that's what I mean. $\endgroup$ – Mike Shulman Mar 31 '20 at 4:25

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.