A coherence theorem for pseudonatural transformations

From MaRDI portal



Abstract: We prove coherence theorems for bicategories, pseudofunctors and pseudonatural transformations. These theorems boil down to proving the coherence of some free (4,2)-categories. In the case of bicategories and pseudofunctors, existing rewriting techniques based on Squier's Theorem allow us to conclude. In the case of pseudonatural transformations this approach only proves the coherence of part of the structure, and we use a new rewriting result to conclude. To this end, we introduce the notions of white-categories and partial coherence.


This paper contains coherence theorems expressed in terms of free \((n,p)\)-categories and the following terminology. An \((n,p)\)-category for this author is a strict \(n\)-category in which, for all \(k>p\), every \(k\)-cell is invertible. The notion of computad \(\Sigma\) appropriate to this structure is called an \((n,p)\)-polygraph (following terminology of Albert Burroni rather than the reviewer). The free \((n,p)\)-category on \(\Sigma\) is denoted by \(\Sigma^{*(p)}\). The \(2\)-dimensional notion of sesquicategory is generalized to dimension \(n\) and called an \(n\)-white-category. The word used for rewrite termination and confluence is convergence. The word used for existence of \((m+1)\)-cells between parallel pairs of \(m\)-cells is \(m\)-coherent. There is a \((4,2)\)-polygraph \(\mathbf{Assoc}\) encoding the structure of category with an up-to-isomorphism associative tensor product satisfying the Mac Lane pentagon. Mac Lane's coherence theorem for the structure becomes: \(\mathbf{Assoc}^{*(2)}\) is \(3\)-coherent. Coherence for bicategories becomes the result that, for any set \(\mathbf{C}\), the \(4\)-polygraph \(\mathbf{BiCat[C]}\) is \(3\)-convergent and the free \((4, 2)\)-category \(\mathbf{BiCat[C]}^{*(2)}\) is \(3\)-coherent. Coherence for pseudofunctors becomes the result that, for any function \(\mathbf{f} : \mathbf{C}\to \mathbf{D}\), the \(4\)-polygraph \(\mathbf{PFonct[f]}\) is \(3\)-convergent and the free \((4,2)\)-category \(\mathbf{PFonct[f]}^{*(2)}\) is \(3\)-coherent. Yet a coherence of this precise form for pseudonatural transformations is spoilt by the fact that \(\mathbf{PNTrans[f,g]}\) is not \(3\)-confluent. The problem is resolved by observing that the so-called \(2\)-Squier condition of depth \(2\) is satisfied.











This page was built for publication: A coherence theorem for pseudonatural transformations

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q502628)