Bousfield-Kan Completion and Very Large Groups

Jaime Benabent Guerrero, Ramón Flores · 2025

Read Bousfield-Kan Completion and Very Large Groups by Jaime Benabent Guerrero, Ramón Flores (2025) 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-2511-02582.txt.

Introduction

Jaime Benabent Guerrero and Ramón Flores work in algebraic topology on the Bousfield-Kan completion. It is a standard procedure that completes a space relative to a chosen homology theory. Some spaces are called R-bad when this completion behaves pathologically for the ring R in question, and recognising them matters because such methods silently fail on such spaces. The authors establish criteria for deciding when a space is R-bad, and apply them to closed surfaces, the compact two-dimensional manifolds such as spheres and tori.

Catalog record

CREATORJaime Benabent Guerrero, Ramón Flores
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE28 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-02582

Browse the full library · About the Tower of Babel

Full text

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

Bousfield-Kan Completion and Very Large Groups Jaime Benabent Guerrero, Ramón Flores Abstract . We establish new criteria for the R R -badness of a space and apply it to the case of closed surfaces. 1. Introduction The Bousfield–Kan R -completion, developed in [ 8 ] , provides a unifying framework for understanding how spaces behave with respect to a chosen ring of coefficients R , generalizing both p -completion and rationalization within a single homotopical setting. When R=\mathbb{Z}/p , the R -completion isolates the p -primary information of a space, capturing precisely those features detected by mod- p (co)homology. This construction has played a fundamental role in homotopy theory, notably in the study of classifying spaces of finite and profinite groups, in the analysis of p -local homotopy types, and in the characterization of maps inducing mod- p homology equivalences. It is also central to the proof of the Sullivan conjecture and to the homotopy classification of finite H -spaces. Moreover, the theory provides a geometric analogue of p -adic completion for groups, linking the R -completion of a space with the p -adic completion of its fundamental group in the nilpotent case. When R is a subring of \mathbb{Q} , such as \mathbb{Z}[1/p] or \mathbb{Q} itself, R -completion is related to rational localization. In these cases, the process annihilates torsion phenomena invisible to R -homology and produces a more algebraic homotopy type that can often be described via differential graded Lie algebras or minimal models. This rational completion framework underlies rational homotopy theory, where the \mathbb{Q} -completion of a simply connected space encodes its entire homotopy type in purely algebraic terms. Furthermore, R -completion for subrings of \mathbb{Q} provides the essential bridge in the arithmetic fracture square, connecting rational and p -adic information and allowing global homotopy types to be reconstructed from their local components. In a broader sense, the R -completion formalism offers a conceptual mechanism for studying localizing phenomena simultaneously across arithmetic and homological contexts. It has profound applications to mapping spaces, localizations of ring spectra, and the interplay between algebraic and topological completions. Whether isolating p -primary structure or extracting rational information, R -completion remains one of the most powerful tools for decomposing complex homotopy types into algebraically manageable local parts. However, unlike Bousfield homological localization, the Bousfield–Kan completion functor is not always idempotent on spaces; that is, applying it twice does not necessarily yield the same result as applying it once. From the beginning, Bousfield-Kan sought conditions under which the completion functor becomes idempotent, and many of these are listed in their foundational monograph. It is precisely on such “good” spaces that completion behaves predictably. Yet there remain many spaces for which idempotency is not known, and explicit examples of spaces that fail it —the so-called R -bad spaces— are rare and of particular interest. This paper addresses that gap by establishing a general criterion for detecting R -badness. Our main result (Theorem 3.3 ) states that if X is a connected space whose second integral homology group H_{2}(X;\mathbb{Z}) is countable and whose fundamental group admits a surjection onto a non-commutative free group (a very large group, see Definition 3.5 ), then X is R -bad for any R=\mathbb{Z}/p or any subring R\subseteq\mathbb{Q} . In particular, the classifying spaces of these groups are R -bad. This criterion, which uses and generalizes recent work of Ivanov-Mikhailov [ 13 ] showing that the wedge of two circles is R -bad, provides a systematic framework for producing new examples of non-idempotent completions. Our initial motivation arose from a question posed by A. Murillo [ 15 ] concerning whether compact orientable surfaces are “good” spaces with respect to rational completion. Surprisingly, even for such elementary geometric objects, only the sphere and the torus were known to be R -good, while the cases of bigger genus remained open. By observing that the fundamental groups of compact surfaces (with very few exceptions) are very large, and by combining this fact with the aforementioned badness criterion, we completely resolve Murillo’s question. We prove that all compact orientable surfaces other than the sphere and the torus are R -bad for R=\mathbb{Z}/p and for subrings of \mathbb{Q} , and that the same holds for all compact non-orientable surfaces of genus bigger than 3; the cases of genus 1 and 2 where previously known, while our methods do not fit in the case of genus 3, which remains unsolved. Moreover, our methods extend to a wide class of very large groups satisfying mild finiteness assumptions, thus offering a flexible and general approach to producing new examples of R -bad spaces. The paper is organized as follows. In Section 2, we recall the main definitions and results concerning Bousfield–Kan R -completion, together with a review of known examples of R -good and R -bad spaces. In Section 3 we prove our main theorem, establishing the criterion for R -badness, while in Section 4 this criterion is applied to the case of closed surfaces. Finally, Section 5 extends the arguments to broader families of very large groups such as general Artin groups, right-angled Artin groups, Bestvina–Brady groups, and graph braid groups, thereby illustrating the scope and utility of our criterion. 2. Bousfield-Kan R -Completion In this section, we recall the main definitions and results concerning the Bousfield-Kan R -completion and compile some currently known examples of R -good and R -bad spaces for different choices of a ring R . 2.1. R -completion In [ 8 ] is introduced, for any ring R , a R - completion in the category of spaces, i.e. an endofunctor R_{\infty}:\mathcal{S}\to\mathcal{S} together with a natural transformation \epsilon:1_{\mathcal{S}}\to R_{\infty} , called the coaugmentation, such that R_{\infty}\epsilon_{X}=\epsilon_{R_{\infty}} . This functor is defined, up to homotopy, for the following property: Proposition 2.2 . A morphism of spaces, f:X\to Y , is a R -homology equivalence if and only if the morphism given by the R -completion, R_{\infty}f:R_{\infty}X\rightarrow R_{\infty}Y , is a homotopy equivalence. This construction, which is performed by means of an inverse homotopy limit of a tower of spaces, has an important disadvantage with regard to its respective homological localization (see [ 7 ] ): it is not always idempotent, i.e. in general R_{\infty}X is not always homotopy equivalent to R_{\infty}R_{\infty}X . This motivated the following definition, with its corresponding proposition: Definition 2.3 . Let X be a space. (1) X is called R - complete if \epsilon_{X} is a homotopy equivalence. (2) X is called R - good if \epsilon_{X} is an R -homology equivalence. (3) X is called R - bad if it is not R -good. Proposition 2.4 . Let X be a space. The following conditions are equivalent: • X is R -good. • R_{\infty}X is R -good. • R_{\infty}X is R -complete. In particular, if we define inductively R^{0}_{\infty}:=1_{\mathcal{S}} and R^{n}_{\infty}:=R_{\infty}\circ R^{n-1}_{\infty} for n\geq 1 , for every space we either have an infinite chain R_{\infty}X\stackrel{{\scriptstyle\simeq}}{{\to}}R^{2}_{\infty}X\stackrel{{\scriptstyle\simeq}}{{\to}}R^{3}_{\infty}X\stackrel{{\scriptstyle\simeq}}{{\to}}\cdots of homotopy equivalences or we never have a homotopy equivalence in the infinite chain X\stackrel{{\scriptstyle\not\simeq}}{{\to}}R_{\infty}X\stackrel{{\scriptstyle\not\simeq}}{{\to}}R^{2}_{\infty}X\stackrel{{\scriptstyle\not\simeq}}{{\to}}R^{3}_{\infty}X\stackrel{{\scriptstyle\not\simeq}}{{\to}}\cdots. 2.5. The choice of the ring . Unless explicit mention against, the result of this paper are valid when R=\mathbb{Z}/p for p prime, or R is a subring of the rationals. The reason of this is that

Read the complete text