Can wall-crossing formulas be derived purely from the spectral dualities without topological recursion?

This is a genuinely deep question about the logical architecture of modern mathematical physics. Let me work through what the archive shows.

The short answer

No — at least not in the direction you might hope. The preprints show that **wall-crossing formulas are now being derived *using* topological recursion and spectral duality together, not independently from spectral duality alone. The spectral data (the underlying geometry) determines the wall-crossing behavior, but the machinery to extract it involves recursion.

What the archive reveals

Wall-crossing formulas as outputs of duality + recursion:**

The cleanest recent result comes from a reformulation perspective. In 4d $\mathcal{N}=2$ theories, Kontsevich-Soibelman wall-crossing formulas can be recast as identities for symmetric quivers that encode 3d boundary dualities [AG-2025.06-900]. The authors show that *both sides* of the wall-crossing identity are captured by quiver structures — but crucially, they use wall-crossing *type* formulas involving trees of unlinkings to express closed Donaldson-Thomas invariants in terms of open ones [AG-2025.06-900]. This is a restatement of wall-crossing, not a derivation from pure spectral data.

Why recursion is integral, not accidental:

The convergence of three frameworks—topological recursion, $x$-$y$ duality, and wall-crossing—suggests they're aspects of one structure, not independent paths. Symplectic duality itself is now understood *as* a composition of $x$-$y$ dualities *within* log topological recursion [AG-2024.05-860]. The free energy differences $F_g - F_g^\vee$ (measuring how the spectrum changes under swapping roles of $x$ and $y$) are computed via residue formulas that depend on singularity structure—information that topological recursion is designed to extract [AG-2025.02-1029].

For open topological strings and 3D-5D wall-crossing, the partition function is derived *via* log topological recursion applied to the spectral curve; wall-crossing appears as a *consequence* when you vary parameters [AG-2025.10-794].

The deeper structure:

Spectral duality alone gives you a map between geometric data—but wall-crossing is a *computational problem*: you need to count or sum something (BPS states, invariants, correlators). Topological recursion is the algorithm that solves that counting from spectral input. The recent papers show this isn't a crutch; it's how the mathematics is actually organized. Even in degenerate and irregular regimes where classical topological recursion breaks down, the resolution comes via generalizations of duality within a *broader* recursion framework [AG-2024.08-554].

Why this matters

If wall-crossing could be derived purely from spectral duality without recursion, it would mean the wall-crossing formulas were "already visible" in the geometry. Instead, the archive suggests wall-crossing requires *computing with* the spectrum—which is what recursion does. This is actually more powerful: it means the same recursion apparatus works across many different wall-crossing scenarios (finite chambers, wild quivers, open/closed duality) once you input the right spectral data.

---

The preprints don't directly address whether a hypothetical "pure duality" approach exists; they show the working mathematics integrates recursion as essential.