A new theoretical mathematics preprint proposes an explicit way to describe Dwyer–Kan hammock localization in the setting of ∞-categories, the paper’s abstract framework for objects and their maps. It builds the construction from all finite zig-zag shapes and shows that the resulting mapping animae—the formal objects that record connections between pairs of objects—agree, by equivalence, with both localization mapping animae and hammock animae. The result brings three mathematical descriptions into a common framework.
At its core, the paper asks how to make a concise, conceptual model of localization: a process built around a selected class of maps. In the paper’s notation, the input is an abstract ∞-category equipped with a wide subcategory W, and the target is the localization at W.
A model built from every finite zig-zag
The central construction defines the anima of 1-simplices by a colimit, a mathematical assembly, over all finite zig-zag shapes. The practical idea is that the model does not begin with one fixed arrow pattern: it gathers the finite ways of moving through the structure and packages them into one object.
The authors then show that this zig-zag construction is an explicit Segal anima realizing the Segalification of the relative Rezk nerve. In plain language, the operation turns a nerve-based description into a form where composition is represented directly, while the zig-zag construction makes that form explicit.
This is a methods result about abstract mathematical constructions, not a study based on recruited participants, specimens or observations. The paper compares localization models within the relative ∞-categorical framework rather than reporting measurements from an empirical comparison group.
Three descriptions, one localization
The paper’s main equivalence identifies the mapping animae of the zig-zag construction with the mapping animae of the localization and with the hammock construction. Read simply, the claim is that the formal record of connections between two objects is the same in each of these descriptions, under the equivalences established by the paper.
It further states that the ∞-category associated with the zig-zag anima is equivalent to the localization at W. That matters because the zig-zag object is not being offered only as a convenient diagram: under the paper’s hypotheses, it presents the intended localized ∞-category.
The comparison with Dwyer–Kan’s hammock localization is described as fully faithful and essentially surjective at the level of Segal animae. In broad terms, the comparison preserves the relevant structure and reaches the objects represented on the other side, up to the equivalence used in this framework.
When the long routes collapse
The paper’s most concrete shortcut comes from its fraction theorem. If a relative ∞-category supports left fractions, the hammock mapping animae reduce to zig-zags of shape [1,−1]. Right-fraction support gives [−1,1], while two-sided fraction support gives [−1,1,−1]. The signs record the directions in the corresponding short zig-zag shapes.
These are conditional formulas, not a universal replacement for hammock constructions. Without the relevant support hypotheses, hammock animae may involve zig-zags of arbitrary length and can be difficult to analyze. The theorem therefore tells mathematicians when the shorter description is available, rather than removing the need for longer constructions in every case.
Fraction support also gives a criterion for the relative Rezk nerve itself. Supporting left, right or two-sided fractions makes the nerve Segal; with the additional two-out-of-three assumptions, completeness is equivalent to saturation. The paper thus ties the shape of its mapping objects to structural properties of the nerve.
The conditions are not merely abstract labels
Two results spell out settings where the required hypotheses are available. The preprint says that every relative ∞-category admitting CLF supports left fractions, making CLF a sufficient route to the short left-fraction mapping formula.
On the right-fraction side, every ∞-category of fibrant objects supports right fractions, according to the paper. This supplies the short right-fraction description for that class, subject to the axioms that define an ∞-category of fibrant objects.
A bridge across two hammock models
The authors also connect their construction with Mazel-Gee’s hammock localization. Levelwise colimits of Mazel-Gee’s model agree with the zig-zag construction up to a map of Segal animae that is fully faithful and essentially surjective.
Read together, the comparisons make the preprint a unifying account of several localization languages: the relative Rezk nerve, the explicit zig-zag anima, Dwyer–Kan’s hammock model and Mazel-Gee’s construction. The claim is about mathematical equivalence within the stated framework, not about one method being empirically faster or more accurate than another.
What the proofs and shortcuts leave open
The proof architecture is broad but not entirely model-independent. The paper says its existing proofs of Lemmas 3.19 and 3.27, in Subsections 3.3 and 3.4, rely critically on quasicategory combinatorics, while the remaining sections use formal ∞-categorical properties.
That distinction sets a boundary around the paper’s conceptual ambition. The framework unifies the constructions at the level of the stated theorems, but not every proof ingredient has been detached from the specific combinatorics of quasicategories.
The paper’s uncertainty is mathematical rather than statistical: its conclusions are theorem-level statements conditional on the relative ∞-category framework and, for the short formulas, on the relevant fraction-support hypotheses. No empirical validation or statistical estimate is part of the supplied evidence.
One question remains explicitly unsettled in the supplied analysis: how the paper’s fraction-support condition relates to Cisinski’s calculus of fractions. The relationship is not established in either direction, so the preprint should not be read as resolving that comparison.
A further open direction concerns arbitrary-length hammock animae when the support hypotheses are unavailable. The supplied analysis says that these cases still need further treatment, underscoring that the short formulas apply to a qualified setting rather than replacing every hammock description.
A mathematical preprint, with disclosed revision support
The document supplied for review is an arXiv version 1 preprint dated 20 Aug 2026. Its conclusions are presented through definitions, theorem statements and proof arguments within the paper’s mathematical framework.
The acknowledgments report support for K.A. from JSPS KAKENHI Grant Number 24KJ1443 and identify B.C. as an associate member of CRC 1785 Generalised Motivic Methods in Geometry.
The authors also disclose that coding agents, named as Claude and Codex, were used in manuscript revision, including a simpler proof of Lemma 3.22. The disclosure concerns the revision process; the paper remains a proof-based theoretical study.
Paper data and sources
Original title: Hammock localization via Segal animae
Authors: Kensuke Arakawa, Bastiaan Cnossen
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-20
DOI: Not available
Original paper · Full text