Technical Summary: Simple Homotopy Types of Even Dimensional Manifolds

Authors: Csaba Nagy, John Nicholson, and Mark Powell
Subject: Geometric Topology, Algebraic K-Theory, Surgery Theory

1. Problem Statement

The paper addresses a fundamental gap in manifold topology: the classification of manifolds up to simple homotopy equivalence versus homotopy equivalence.

• Context: While the classification of 3-dimensional lens spaces up to homotopy equivalence was established early on (distinguishing them from homeomorphism via Reidemeister torsion), the existence of closed manifolds in even dimensions ($n \ge 4$) that are homotopy equivalent but not simple homotopy equivalent was an open problem.
• The Gap: Previous constructions of such examples relied on odd-dimensional manifolds. However, the product of two odd-dimensional manifolds that are homotopy equivalent is always simple homotopy equivalent (due to the vanishing Euler characteristic). Thus, standard product constructions failed to generate even-dimensional counterexamples.
• Goal: To construct the first infinite families of closed, connected, orientable even-dimensional manifolds ($n \ge 4$) that are homotopy equivalent but not simple homotopy equivalent, and to characterize the "simple homotopy manifold set" (the set of such manifolds modulo simple homotopy equivalence) for these examples.

2. Methodology

The authors employ a sophisticated blend of geometric topology and algebraic number theory.

A. Geometric Framework: Surgery and Cobordism

1. Simple Homotopy Manifold Sets: They define three sets for a manifold $M$:
   • $M^h_s(M)$: Manifolds homotopy equivalent to $M$ modulo simple homotopy equivalence.
   • $M^{hCob}_s(M)$: Manifolds $h$-cobordant to $M$ modulo simple homotopy equivalence.
   • $M^h_{s,hCob}(M)$: Manifolds homotopy equivalent to $M$ modulo the equivalence relation generated by simple homotopy equivalence and $h$-cobordism.
2. Theoretical Characterization (Theorem B): Using the s-cobordism theorem and the surgery exact sequence, they characterize these sets in terms of:
   • The Whitehead group $Wh(\pi_1(M))$ equipped with an involution.
   • The surgery obstruction map $\sigma_s$.
   • The action of homotopy automorphisms $hAut(M)$.
   • Specifically, they relate these sets to subquotients of $Wh(\pi_1(M))$ involving Tate cohomology groups $\hat{H}^{n+1}(C_2; Wh(\pi_1(M)))$.

B. Specific Examples: $S^1 \times L$

The authors focus on manifolds of the form $M = S^1 \times L$, where $L$ is an $(n-1)$-dimensional lens space with fundamental group $C_m$ (cyclic of order $m$).

• Fundamental Group: $\pi_1(M) \cong C_\infty \times C_m$.
• Key Insight: They prove that for these manifolds, the map $\psi$ in the Ranicki-Rothenberg exact sequence is surjective. This simplifies the characterization of the manifold sets to orbit sets of the action of $hAut(M)$ on specific subgroups of $Wh(C_\infty \times C_m)$.

C. Algebraic Machinery: K-Theory and Number Theory

To compute the size of these orbit sets, the authors decompose the Whitehead group using the Bass-Heller-Swan decomposition:
$$Wh(C_\infty \times C_m) \cong Wh(C_m) \oplus \tilde{K}_0(\mathbb{Z}C_m) \oplus NK_1(\mathbb{Z}C_m)^2$$
The analysis reduces to studying the involution on $\tilde{K}_0(\mathbb{Z}C_m)$, which is isomorphic to the locally free class group $C(\mathbb{Z}C_m)$.

• Decomposition: $C(\mathbb{Z}C_m)$ is analyzed via a short exact sequence involving the kernel group $D(\mathbb{Z}C_m)$ and the ideal class groups of cyclotomic fields $C(\mathbb{Z}[\zeta_d])$.
• Number Theoretic Tools: The proofs rely heavily on:
  • Class numbers ($h_m$) and their "minus parts" ($h_m^-$) of cyclotomic fields.
  • Divisibility properties of class numbers (results by Masley-Montgomery, Horie, etc.).
  • Growth rates of class numbers (super-exponential growth).
  • Kernel groups $D(\mathbb{Z}C_m)$ and their involutions, analyzed using idelic approaches and unit groups.

3. Key Contributions and Results

A. Existence of New Examples (Theorem A)

The paper constructs the first examples of closed, connected, orientable $n$-manifolds ($n \ge 4$ even) that are homotopy equivalent but not simple homotopy equivalent.

• Construction: $M = S^1 \times L$ where $L$ is a lens space with $\pi_1(L) = C_m$.
• Condition: If $m$ is not square-free (e.g., $m=4$), the set $M^h_s(M)$ is infinite.
• Smoothness: For $n > 4$, these examples can be taken to be smooth.

B. Quantitative Classification (Theorems C, D, E)

The authors provide a complete classification of the size of the simple homotopy manifold sets for $M = S^1 \times L$ based on the integer $m$:

1. Triviality ($|Set| = 1$):

   • $|M^h_s(M)| = 1$ if and only if $m \in \{2, 3, 5, 6, 7, 10, 11, 13, 14, 17, 19\}$.
   • $|M^{hCob}_s(M)| = 1$ if and only if $m \in \{2, 3, 5, 6, 7, 10, 11, 13, 14, 15, 17, 19, 29\}$.
   • For $M^h_{s,hCob}(M)$, there are infinitely many $m$ for which the set is trivial.

2. Finiteness vs. Infiniteness:

   • The sets are infinite if and only if $m$ is not square-free.
   • This is driven by the Nil group $NK_1(\mathbb{Z}C_m)$, which is non-zero (and infinite) precisely when $m$ is not square-free.

3. Asymptotic Behavior:

   • As $m \to \infty$, the size of the sets $|M^h_s(M)|$ and $|M^{hCob}_s(M)|$ grows super-exponentially.
   • For the set modulo $h$-cobordism ($M^h_{s,hCob}$), the size is always finite, but the supremum of the sizes over $n$ grows exponentially with $m$.

C. Algebraic Results on $\tilde{K}_0(\mathbb{Z}C_m)$

A significant portion of the paper is dedicated to proving new results about the involution on $\tilde{K}_0(\mathbb{Z}C_m)$:

• They determine exactly when the subgroups $\{x \in \tilde{K}_0 \mid x = -x\}$ and $\{x - \bar{x} \mid x \in \tilde{K}_0\}$ are trivial.
• They establish lower bounds on the growth of these groups using the odd part of the class number $h_m^-$.
• They analyze the kernel group $D(\mathbb{Z}C_m)$ and its interaction with the involution, providing explicit computations for specific $m$.

4. Significance

• Resolution of an Open Problem: The paper definitively answers the question of whether homotopy equivalence implies simple homotopy equivalence for even-dimensional manifolds, showing it does not, and providing the first infinite families of counterexamples.
• Bridging Fields: It creates a deep link between high-dimensional manifold topology and algebraic number theory. The classification of topological objects is reduced to the arithmetic properties of cyclotomic fields (class numbers).
• Refinement of Classification: By distinguishing between $h$-cobordant and non-$h$-cobordant manifolds, the authors provide a nuanced view of the "simple homotopy manifold set," showing that even within the same $h$-cobordism class, there can be infinitely many distinct simple homotopy types.
• Methodological Advancement: The work demonstrates the power of combining surgery theory with integral representation theory and the arithmetic of class groups to solve geometric problems. The detailed analysis of the involution on $\tilde{K}_0(\mathbb{Z}C_m)$ provides tools applicable to other problems in algebraic K-theory.

In summary, this paper establishes that for even-dimensional manifolds, the gap between homotopy equivalence and simple homotopy equivalence is vast and structurally rich, governed by the arithmetic of cyclotomic fields and the properties of the Whitehead group.