[Paper Review] Chains(R) does not admit a geometrically meaningful properadic homotopy Frobenius algebra structure

Theo Johnson-Freyd|arXiv (Cornell University)|Aug 15, 2013

Homotopy and Cohomology in Algebraic Topology3 references3 citations

TL;DR

This paper proves that Chains(ℝ) does not admit a geometrically meaningful properadic homotopy Frobenius algebra structure, despite supporting a unique quasilocal dioperadic action. While a cofibrant resolution of the dioperad for nonunital shifted Frobenius algebras acts on Chains(ℝ) in a homotopically unique, geometrically natural way, the same structure fails to extend to a properadic resolution due to a homological obstruction involving the non-exactness of multiplication by −1/12 in the quasilocality category.

ABSTRACT

The embedding Chains(R) into Cochains(R) as the compactly supported cochains might lead one to expect Chains(R) to carry a nonunital commutative Frobenius algebra structure, up to a degree shift and some homotopic weakening of the axioms. We prove that under reasonable "locality" conditions, a cofibrant resolution of the dioperad controlling nonunital shifted-Frobenius algebras does act on Chains(R), and in a homotopically-unique way. But we prove that this action does not extend to a homotopy Frobenius action at the level of properads or props. This gives an example of a geometrically meaningful algebraic structure on homology that does not lift in a geometrically meaningful way to the chain level.

Motivation & Objective

• To determine whether the homology of ℝ, which carries a nonunital shifted Frobenius algebra structure, lifts to a geometrically meaningful chain-level structure.
• To investigate whether such a lift exists at the level of properads, which allow graph-like compositions, as opposed to dioperads with only tree-like compositions.
• To clarify the distinction between dioperadic and properadic resolutions in the context of geometrically meaningful algebraic structures on chain complexes.
• To resolve a contradiction with a prior claim in [Wil07] that such a properadic structure should exist on Chains(ℝ).
• To establish that not all geometrically meaningful structures on homology extend to the chain level, even when homotopy transfer theory guarantees a formal lift.

Proposed method

• Introduce the notion of quasilocality as a minimal geometricity condition, ensuring operations extend canonically from compactly supported chains to non-compactly supported ones.
• Define the dioperad and properad governing nonunital shifted Frobenius algebras (Frob₁), and prove its Koszulity to construct a small cofibrant replacement.
• Construct a contractible space of quasilocal, translation-invariant actions of the dioperadic resolution on Chains(ℝ), showing uniqueness and geometric naturality.
• Use homotopy transfer theory to lift the homological Frobenius structure from H•(ℝ) to Chains(ℝ) via the dioperadic resolution.
• Analyze the obstruction to extending the dioperadic action to a properadic one by examining the action of the generator on Thom forms and the non-exactness of multiplication by −1/12 in the quasilocality category.
• Demonstrate that any attempt to force the standard multiplication via intersection of chains leads to non-quasilocality, violating geometric meaning.

Experimental results

Research questions

• RQ1 Does Chains(ℝ) admit a geometrically meaningful properadic homotopy Frobenius algebra structure, given that it supports such a structure at the dioperadic level?
• RQ2 Can the homotopy Frobenius structure on H•(ℝ) be lifted to Chains(ℝ) in a way that respects locality and geometric composition rules?
• RQ3 Why does the properadic resolution fail to act on Chains(ℝ) despite the existence of a unique quasilocal dioperadic action?
• RQ4 What is the role of the non-exactness of multiplication by −1/12 in obstructing properadic extensions?
• RQ5 How does the failure of the properadic lift illustrate a fundamental limitation in lifting algebraic structures from homology to chains?

Key findings

• A cofibrant resolution of the dioperad controlling nonunital shifted Frobenius algebras acts on Chains(ℝ) in a unique, quasilocal, and translation-invariant way.
• This dioperadic action extends the strictly associative and commutative multiplication (degree −1) and cocommutative comultiplication (degree 0) on compactly supported de Rham forms.
• The properadic resolution of the same structure does not act on Chains(ℝ) in a quasilocal way that induces both the comultiplication and the standard multiplication on homology.
• The obstruction arises because multiplication by −1/12 is not exact in the quasilocality category, preventing a well-defined homomorphism from the properadic resolution to the endomorphism properad.
• Any attempt to force the standard multiplication via intersection of chains results in a non-quasilocally defined action, violating the geometricity condition.
• The result resolves a contradiction with [Wil07] by showing that properadic lifts are not geometrically meaningful, even when dioperadic lifts are.