The Leibniz adjunction in homotopy type theory, with an application to simplicial type theory

Tom de Jong, Nicolai Kraus, Axel Ljungström · 2026

Read The Leibniz adjunction in homotopy type theory, with an application to simplicial type theory by Tom de Jong, Nicolai Kraus, Axel Ljungström (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2601-21843.txt.

Introduction

Tom de Jong, Nicolai Kraus, and Axel Ljungström work in homotopy type theory, a foundation for mathematics where types behave like spaces. Simplicial type theory extends it so that types also carry directed morphisms, arrows with a direction. A Segal type is one where those arrows can be composed.

Defining such composition usually drags in endless higher coherence conditions. The authors show that all those coherences can be stated and derived. This holds if the theory is set up in one way. It is homotopy type theory with a postulated interval. Their tool is the Leibniz adjunction named in the title. The 2026 preprint then applies the machinery to simplicial type theory, giving formal mathematics a cleaner way to handle directed structures.

Catalog record

CREATORTom de Jong, Nicolai Kraus, Axel Ljungström
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE50 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2601-21843

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

The Leibniz adjunction in homotopy type theory, with an application to simplicial type theory Tom de Jong, Nicolai Kraus, Axel Ljungström Abstract . Simplicial type theory extends homotopy type theory and equips types with a notion of directed morphisms. A Segal type is defined to be a type in which these directed morphisms can be composed. We show that all higher coherences can be stated and derived if simplicial type theory is taken to be homotopy type theory with a postulated interval type. In technical terms, this means that if a type has unique fillers for ( 2 , 1 ) (2,1) -horns, it has unique fillers for all inner ( n , k ) (n,k) -horns. This generalizes a result of Riehl and Shulman for the case n = 3 n=3 , k ∈ { 1 , 2 } k\in\{1,2\} . Our main technical tool is the Leibniz adjunction: the pushout-product is left adjoint to the pullback-hom in the wild category of types. While this adjunction is well known for ordinary categories, it is much more involved for higher categories, and the fact that it can be proved for the wild category of types (a higher category without stated higher coherences) is non-trivial. We make profitable use of the equivalence between the wild category of maps and that of families. We have formalized the results in Cubical Agda. 1. Introduction [ RS17 ] have introduced simplicial type theory (STT) as a synthetic (bespoke) framework for higher categories [ RS17 ] . It extends homotopy type theory, where types are equipped with invertible (higher) paths, making them behave like \infty -groupoids rather than categories. Therefore, STT additionally equips types with a notion of directed morphisms via an interval. We naturally wish to compose these directed morphisms, and those types where this is possible are referred to as Segal types . Only having composition is not enough for a well behaved theory of higher categories: we want composition to be associative and for identities to be neutral with respect to composition. For ordinary 1-categories, that would be sufficient, since equations are mere properties, but for higher categories, equations carry genuine data and there are (infinitely many) additional coherences to be fulfilled. A simple example illustrating this is the requirement that the different ways of rewriting {\id}\circ(f\circ\id) , illustrated below, are in fact the same, up to a higher morphism. {\lx@inpgf@ignorespaces{\id}\circ(f\circ\id)} {\lx@inpgf@ignorespaces({\id}\circ f)\circ\id} {\lx@inpgf@ignorespaces f\circ\id} {\lx@inpgf@ignorespaces{\id}\circ f} {\lx@inpgf@ignorespaces f} Informally, the first main result of this paper ( Theorem 4.9 ) states that for Segal types, all higher coherences can be derived. Simplicial techniques give us a precise and combinatorial framework for stating higher coherences. The 2-simplex \Delta^{2} is pictured as a filled-in triangle with directed edges {\lx@inpgf@ignorespaces 1} {\lx@inpgf@ignorespaces 0} {\lx@inpgf@ignorespaces 2} \scriptstyle{\lx@inpgf@ignorespaces 12} \scriptstyle{\lx@inpgf@ignorespaces 01} \scriptstyle{\lx@inpgf@ignorespaces 02} while the (2,1)-horn \Lambda^{2}_{1} is obtained by omitting the filling and the hypotenuse: {\lx@inpgf@ignorespaces 1} {\lx@inpgf@ignorespaces 0} {\lx@inpgf@ignorespaces 2} \scriptstyle{\lx@inpgf@ignorespaces 12} \scriptstyle{\lx@inpgf@ignorespaces 01} The idea is that a map \Lambda^{2}_{1}\to X picks out two composable morphisms in X , by 01\mapsto f and 12\mapsto g , say, and if we can extend it to a map \Delta^{2}\to X , we can take their composition g\circ f to be the image of the edge 02 . The filling, a 2 -dimensional cell, then witnesses that the triangle commutes up to this higher cell. If such an extension is unique (in the homotopy type theory sense), then we say that X has unique fillers for \Lambda^{2}_{1} -horns, and we call X a Segal type . Of course, one can similarly consider higher-dimensional simplices \Delta^{n} and (inner or outer) horns \Lambda^{n}_{k} , for n\in\mathbb{N} and 0\leq k\leq n , which encode the higher coherences of the composition. Theorem 4.9 can then be phrased more precisely as: any Segal type has unique fillers for all such inner horns. This theorem generalizes [ RS17 , Proposition 5.12] of Riehl and Shulman where they consider n=3 and k\in\{1,2\} , and is a version of Lurie’s [ Lur09 , Corollary 2.3.2.2] in simplicial type theory. The second contribution lies in our approach and technical development. The idea, going back Joyal (cf. just before Proposition 7.6 in [ JT06 ] where [ Joy08 ] is also cited), is to reduce the situation to the abstract framework provided by the orthogonality and Leibniz calculus [ RV13 ] . Our main technical tool is then that the so-called pushout-product is left adjoint to the so-called pullback-hom ( Theorem 3.16 ) in the wild category of types. While this result is well known for 1-categories, a naive approach to proving this for the wild category of types quickly becomes intractable due to the involved bookkeeping of coherences. Thankfully, as a consequence of the univalence axiom, functions and dependent type families are equivalent. This yields an equivalence between the wild category of maps and the wild category of families, and we illustrate how the latter is a rather useful substitute for the former. 1.1. Foundational setup Instead of working with Riehl–Shulman’s simplicial type theory as introduced in [ RS17 ] , we follow the approach of [ GWB24 ] [ GWB24 , GWB25 ] (in line with [ PS25 ] [ PS25 ] ) and work in plain homotopy type theory with a postulated interval type. In contrast, in the Riehl–Shulman framework, the interval is a concept on the meta-level. The advantage of their approach is that one gets additional definitional equalities, but the advantage of the approach followed here is the possibility of formalizing everything in an established proof assistant like (Cubical) Agda (cf. Section 1.3 ). Compared to [ GWB24 , GWB25 ] , our assumptions on the interval type are rather minimal: we only require it to be a poset that forms a bounded distributive lattice. In particular, we do not require the order to be total, i.e. we do not require x\leq y\vee y\leq x for all x and y . Further, we do not assume the modalities of [ GWB24 , GWB25 ] . 1.2. Related work In addition to the works [ RS17 , GWB24 , GWB25 ] discussed above, we remark that various closure properties of left/right orthogonal maps (cf. Section 4.1 ) have previously been formalized in Agda: first in agda-unimath [ Bak25 ] by [ Bak25 ] , and later by [ Tot26 ] in [ Tot26 ] . We note that the former does not include that left orthogonal maps are closed under pushout-products ( Lemma 4.5 ), which we obtain as an application of the Leibniz adjunction. The latter does contain this result (as of 8 January 2026), but using a different proof method. 1.3. Formalization All our results are formalized using Cubical Agda [ VMA19 ] . Our formalization builds on the cubical library [ Com26 ] and type checks using Agda 2.8.0. The code and its HTML rendering are available at https://github.com/leibniz-stt/agda-formalization and https://leibniz-stt.github.io/README.html . The file Paper.html acts an interface between the formal development and the paper. In particular, (nearly) all environments in this paper are marked with a symbol which is a clickable link to the corresponding formalized statement in that file. 2. Preliminaries At times we find it convenient

Read the complete text