A classification theorem for homotopy commutative 𝐻-spaces with finitely generated mod 2 cohomology rings (Q3978697)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 12738
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | A classification theorem for homotopy commutative 𝐻-spaces with finitely generated mod 2 cohomology rings |
scientific article; zbMATH DE number 12738 |
Statements
A classification theorem for homotopy commutative 𝐻-spaces with finitely generated mod 2 cohomology rings (English)
0 references
25 June 1992
0 references
homotopy commutative \(H\)-space
0 references
Steenrod algebra
0 references
ladder Toda brackets
0 references
secondary operations
0 references
Dyer-Lashof operations
0 references
mod 2 cohomology
0 references
0.8738718628883362
0 references
0.8715649843215942
0 references
0.8668815493583679
0 references
0.8615679740905762
0 references
0.8594363927841187
0 references
One of the most dramatic theorems in the study of \(H\)-spaces is Hubbuck's Torus Theorem -- a homotopy commutative \(H\)-space which is a finite complex has the homotopy type of a product of circles. The proof uses \(K\)-theory and work of Browder to establish that the universal cover of such an \(H\)-space must be acyclic. The condition of finiteness is essential to the proof. Suppose all spaces are localized at the prime 2. A natural generalization of finiteness is to require that the mod 2 cohomology of the homotopy commutative \(H\)-space be finitely generated as an algebra. The main theorem of this paper proves a no less dramatic generalization -- a 2-connected, homotopy commutative \(H\)-space with coassociative mod 2 cohomology which is finitely generated is acyclic. This gives a classification theorem -- all such \(H\)-spaces are homotopy equivalent to products of the Eilenberg-Mac Lane spaces, \(S^1\), \(K(\mathbb{Z}/2^r\mathbb{Z},1)\), and \(K(\mathbb{Z},2) \). The proof depends on a careful consideration of the module of indecomposables of the mod 2 cohomology of the \(H\)-space over the Steenrod algebra. The tools for analysis are the ladder Toda brackets of J. P. Lin and secondary operations controlled by the \(h_n\)-deviation. A space \(X\) is an \(H_n\)-space if there is a \(\mathbb{Z}/2\mathbb{Z}\)-equivariant mapping \(\theta^X_n: S^n \times X \times X\to X\), where the \(\mathbb{Z}/2\mathbb{Z}\)-action on \(S^n \times X \times X\) is given by the antipodal map and the switch map. The mapping \(\theta^X_n\) is required to satisfy \(\theta^X_n(w;x,*)=\theta^X_n(w;*,x)=x\). An \(H_0\)-space is an \(H\)-space, and an \(H_1\)-space is a homotopy commutative \(H\)-space. Given a mapping \(f : X \to K\) of \(H_n\)-spaces, suppose that the following diagram commutes up to \(\mathbb{Z}/2\mathbb{Z}\)-equivariant homotopy: \[ \begin{matrix} S^{n-1}\times X\times X &\xrightarrow{\theta^X_{n-1}} &X\\ ^{1\times f\times f}\downarrow && \downarrow ^{f} \\ S^{n-1} \times K\times K & \xrightarrow{\theta^K_{n-1}} & K. \end{matrix} \] The \(h_n\)-deviation is the obstruction to extending the diagram with \(n\) for \(n - 1\). The \(H_n\)-space structure gives an \(H_{n-1}\)-space structure by restricting to the equator. The \(h_n\)-deviation lies in \([X \wedge X,\Omega^nK]\) and when \(K\) is an Eilenberg-Mac Lane space, \(h_n(f)\) is a cohomology class. In another paper [J. Pure Appl. Algebra 73, 181--201 (1991; Zbl 0729.55013)], the author has related the \(h_n\)-deviation to Dyer-Lashof operations. The author uses this relation in the presence of a secondary operation associated to a relation in the Steenrod algebra, and a cohomology class represented by an \(h_{n-1}\)-map. The \(h_n\)-deviation detects a non-zero Dyer-Lashof operation when the \(h_n\)-deviation is non-zero. This leads to an argument like Browder's `infinite implications' producing indecomposables in forbidden degrees, thus eliminating the original class. The author proves some interesting corollaries of his Main Theorem: If \(X\) is a simply-connected \(H\)-space that is homotopy commutative and has mod 2 cohomology finitely generated as an algebra, then the integral cohomology of \(X\) is 2-torsion free. He also constructs examples out of simply-connected finite loop spaces and their 3-connective covers of loop spaces with finitely generated mod 2 cohomology algebras which cannot have a homotopy commutative multiplication.
0 references