Strong Heegaard Invariants in Floer Homology

• Strong Heegaard Invariant is a functorial assignment of algebraic data to 3-manifolds, ensuring coherence and naturality across diagram moves.
• The framework relies on four axioms—functoriality, commutativity, continuity, and handleswap invariance—to guarantee path-independent and canonical invariants.
• This invariant underpins developments in Heegaard, bordered, involutive, and equivariant Floer homology by providing explicit canonical maps and distinguished elements.

Searching arXiv for papers on strong Heegaard invariants, naturality in Heegaard Floer homology, and related developments. A strong Heegaard invariant is a functorial assignment of algebraic data to 3-manifolds, or more generally to objects and morphisms in the mapping class groupoid of surfaces, that satisfies strict invariance properties under changes of description arising from Heegaard moves or surface diffeomorphisms. In Heegaard Floer theory, this notion upgrades invariance “up to isomorphism” to naturality, coherence, and functoriality, so that one obtains concrete groups for based 3-manifolds, links, and balanced sutured manifolds, together with canonical maps induced by diffeomorphisms and canonical identifications of distinguished elements such as contact classes (Juhász et al., 2012).

1. Definition and conceptual role

Heegaard Floer homology was originally formulated as an invariant of a 3-manifold only up to isomorphism: one chooses a Heegaard diagram, forms the corresponding Floer groups, and then compares different diagrammatic choices by isomorphisms. The framework of strong Heegaard invariants addresses the insufficiency of this “up to isomorphism” perspective for mapping class group actions, cobordism maps, and distinguished elements.

A weak Heegaard invariant assigns an object to each isotopy diagram in a diagram graph and assigns isomorphisms along the basic moves, with inverses for those moves. This suffices to say that the resulting algebraic object is well defined up to isomorphism. A strong Heegaard invariant adds further structure: the assignment must satisfy functoriality on each move type, commutativity on distinguished rectangles, continuity for diffeomorphisms isotopic to the identity, and handleswap invariance. In the formulation quoted in the naturality literature, this is codified as Definition 2.32.

The strengthening is categorical and geometric at the same time. It is categorical because one requires coherence for compositions of diagram moves. It is geometric because the allowed moves are the standard moves relating Heegaard diagrams: isotopies, handleslides, stabilizations, and diffeomorphisms. This suggests that “strong” refers not to a larger numerical invariant, but to a stricter descent property from diagrammatic data to manifold-level data.

2. Diagram categories and the axioms

The relevant diagrammatic framework is the graph or category $G(S)$ whose vertices are isotopy diagrams in a class $S$ and whose edges correspond to the four basic move types: isotopies, handleslides, stabilizations, and diffeomorphisms. In this setting, Heegaard Floer homology is first regarded as an invariant of abstract multi-pointed Heegaard diagrams, and the problem is to show that the induced identifications do not depend on choices of paths through $G(S)$ (Juhász et al., 2012).

The four axioms that distinguish strong from weak invariants are as follows.

Axiom	Content
Functoriality	Restrictions to each move type are functors
Commutativity	For each distinguished rectangle, $F(g)F(e)=F(h)F(f)$
Continuity	If a diffeomorphism is isotopic to the identity, the induced map is $\mathrm{Id}$
Handleswap invariance	For a handleswap triangle, $F(g)\circ F(f)\circ F(e)=\operatorname{Id}$

Distinguished rectangles are commutative squares built from pairs of move types, for example an $\alpha$-move commuting with a $\beta$-move. The commutativity condition requires equality of the two resulting composite isomorphisms. Handleswap invariance is the additional local relation needed to control elementary loops built from handleslides and diffeomorphisms.

A common misconception is that naturality follows formally once invariance under the individual moves has been shown. The strong-invariant framework isolates the extra coherence needed to rule out monodromy in the space of diagrams. In that sense, the axioms are not decorative; they are the mechanism by which diagram-level constructions descend to canonical manifold-level invariants.

3. Vanishing monodromy and canonical groups

The central technical problem is monodromy. Because many different sequences of diagram moves connect two Heegaard diagrams of the same manifold, one must show that the induced isomorphism is independent of the chosen path. The key result in this direction is Theorem 2.38: if $F$ is a strong Heegaard invariant for a class of sutured manifolds, then any two paths in the graph of diagrams of a given manifold connecting the same start and end diagrams induce the same isomorphism (Juhász et al., 2012).

This zero-monodromy statement is the heart of functoriality and naturality. Its proof uses Morse theory and Cerf theory to analyze the “space of Heegaard diagrams,” identifies simple generators for the relevant fundamental groupoid, and reduces all possible loops to distinguished rectangles and handleswap triangles. The strong-invariant axioms are precisely what is required to force trivial holonomy around those generators.

Once path-independence is available, one obtains a transitive system of groups and canonical isomorphisms. The actual group associated to a manifold is then defined by a colimit construction over the diagram category. In the language of the naturality paper, Heegaard Floer groups become concrete groups assigned to based 3-manifolds, and based diffeomorphisms induce canonical isomorphisms. The same framework applies to link Floer homology and sutured Floer homology.

This is also the point at which the mapping class group enters naturally. Vanishing monodromy ensures that the action of the mapping class group, or more generally of the relevant fundamental groupoid of diagram space, is well defined and isotopy invariant. A plausible implication is that any refinement of Heegaard Floer theory that needs canonical chain-level identifications must either fit this strong-invariant pattern or provide an equivalent replacement for it.

4. Bordered Floer theory and categorical representations

In bordered Heegaard Floer homology, the strong-invariant idea takes a particularly explicit categorical form. The paper “Combinatorial Proofs in Bordered Heegaard Floer homology” states that bordered Floer theory gives a linear-categorical representation of the strongly-based mapping class groupoid, and proves this combinatorially for the hat version of Heegaard Floer homology (Zhan, 2014).

The objects of the strongly-based mapping class groupoid are pointed matched circles, and the morphisms are isotopy classes of basepoint-preserving diffeomorphisms. The algebraic assignment is:

• to an object $Z$, a strand dg-algebra $S$0;
• to a morphism $S$1, a bimodule $S$2.

Composition is expressed by box tensor product: $S$3

The mapping class groupoid is generated by arcslides, and the proof proceeds by verifying all relations among these generators: involution, triangle, commutativity, and pentagons. The key theorem is that the homotopy type of a tensor product corresponding to a factorization of $S$4 into arcslides does not depend on the choice of factorization. In particular,

$S$5

The significance of this result is twofold. First, it gives a completely combinatorial proof of invariance for the hat theory in the bordered setting. Second, it exhibits strong Heegaard invariance as a representation-theoretic structure rather than merely a consistency condition on moves. This suggests a broader viewpoint in which strong invariance is the low-dimensional topological manifestation of categorical functoriality.

5. Refined theories built from strong naturality

Several refinements of Heegaard Floer theory rely explicitly on strong naturality. In “Involutive Heegaard Floer homology,” conjugation symmetry on Heegaard Floer complexes is used to define a new three-manifold invariant by forming the mapping cone of $S$6, where $S$7 is obtained by composing conjugation with the naturality equivalences among Heegaard diagrams (Hendricks et al., 2015). The chain-level construction is

$S$8

and the resulting homology $S$9 is a topological invariant of the pair $G(S)$0. The paper states that the construction crucially relies on naturality and strong Heegaard invariants, because the chain complexes for varying Heegaard diagrams must form a transitive system up to homotopy.

A closely related development occurs in $G(S)$1-equivariant Heegaard Floer cohomology for branched double covers of knots. The paper “$G(S)$2-equivariant Heegaard Floer cohomology of knots in $G(S)$3 as a strong Heegaard invariant” proves that $G(S)$4 is a strong Heegaard invariant and that it can be computed from arbitrary knot Heegaard diagrams rather than only from bridge diagrams (Kang, 2018). The proof adapts the Juhász–Thurston framework by verifying functoriality, commutativity, continuity, and handleswap invariance for the equivariant theory.

The same paper constructs a transverse knot invariant

$G(S)$5

and shows that it refines both the LOSS invariant and the $G(S)$6-equivariant contact class. This illustrates a recurrent pattern: once strong Heegaard invariance is available, diagram-dependent constructions can be promoted to canonical objects and distinguished classes in refined Floer theories.

6. Contact invariants, flavor dependence, and current significance

Strong Heegaard invariance is especially relevant for contact topology because distinguished elements such as the Heegaard Floer contact invariant require canonical identifications across diagrams. For a contact structure $G(S)$7 on a closed, oriented 3-manifold $G(S)$8, Ozsváth and Szabó constructed contact invariants

$G(S)$9

related by the natural map

$F(g)F(e)=F(h)F(f)$0

These constructions are invariant under contact isotopy and natural under contact $F(g)F(e)=F(h)F(f)$1-surgery, and they satisfy functoriality properties with respect to symplectic cobordism (Cavallo et al., 30 Apr 2026).

The 2026 paper “The hat and plus version of the Heegaard Floer contact invariant are not equivalent” gives the first example in the literature of contact structures for which the hat invariant is nonzero while the plus invariant vanishes. More precisely, there exist infinite families of Seifert fibered spaces, including Brieskorn spheres such as $F(g)F(e)=F(h)F(f)$2, which admit tight, zero-twisting contact structures $F(g)F(e)=F(h)F(f)$3 such that

$F(g)F(e)=F(h)F(f)$4

(Cavallo et al., 30 Apr 2026).

In the context described there, the distinction is traced to the image of $F(g)F(e)=F(h)F(f)$5 lying in the kernel of $F(g)F(e)=F(h)F(f)$6 for certain negative-definite, non-$F(g)F(e)=F(h)F(f)$7-space graphs. The paper further states that for negative-definite graphs and small Seifert fibered spaces that are not $F(g)F(e)=F(h)F(f)$8-spaces, the nonvanishing of $F(g)F(e)=F(h)F(f)$9 detects tightness, while $\mathrm{Id}$0 detects fillability. This shows that the choice of flavor matters for classification and for the role played by strong Heegaard invariants in contact topology.

A common misconception is that once a distinguished class is natural, all standard flavors of the theory encode the same geometric information. The nonequivalence of $\mathrm{Id}$1 and $\mathrm{Id}$2 shows that strong invariance does not collapse flavor-dependent structure. Rather, it provides the coherent framework within which such differences become mathematically meaningful.

7. Interpretation and research use

Within Heegaard Floer theory, a strong Heegaard invariant is best understood as a descent mechanism from diagrammatics to topology. It supplies sufficient conditions for an invariant of multi-pointed Heegaard diagrams to descend to a natural invariant of 3-manifolds, links, or sutured manifolds. The sufficient conditions are exactly the assignment to diagrams, the move maps, and the four strong axioms.

Its practical role is equally important. It permits the construction of canonical groups, canonical maps under diffeomorphisms, mapping class group actions, and canonical distinguished elements. In bordered Floer theory it yields a linear-categorical representation of the strongly-based mapping class groupoid. In involutive and equivariant theories it enables chain-level enhancements that would otherwise depend on arbitrary choices. In contact topology it underwrites the use of Floer-theoretic contact classes as invariants of contact structures rather than of presentations.

This suggests a unifying interpretation: strong Heegaard invariance is the coherence layer of Heegaard Floer theory. It does not replace the analytic or combinatorial content of the theory, but it organizes that content so that constructions made on Heegaard diagrams become canonical topological data.