From BV-BFV Quantization to Reshetikhin-Turaev Invariants

This paper proposes a program to bridge perturbative BV-BFV quantization of Chern-Simons theory with non-perturbative Reshetikhin-Turaev invariants by conjecturing their equivalence as extended topological quantum field theories, mediated by factorization homology, derived character stacks, and $\mathbb{E}_n$-Koszul duality.

The Gist

Imagine you are trying to understand the shape of a mysterious, invisible mountain range. You have two very different teams of explorers trying to map it, but they are speaking completely different languages and using different tools.

Team A (The Perturbative Explorers) uses a high-powered microscope. They zoom in on tiny, specific spots on the mountain (called "flat connections"). They take measurements, draw tiny diagrams, and write down long, complicated formulas. However, these formulas are like a recipe that goes on forever; if you try to add up all the ingredients, the numbers get infinitely huge and break the calculator. They can only describe the mountain locally, spot by spot, and their map is full of "what if" scenarios that never quite add up to a whole picture. This is the BV-BFV quantization approach.

Team B (The Non-Perturbative Explorers) uses a giant, magical net. Instead of looking at tiny spots, they throw the net over the entire mountain at once. They don't care about the tiny details of the rocks; they care about the overall shape, the loops, and the knots. They produce a perfect, finished map of the whole mountain instantly. This is the Reshetikhin–Turaev (RT) approach.

The Problem:
For decades, mathematicians have known that these two teams are looking at the same mountain. Team A's messy, infinite formulas seem to be the "ingredients" that make up Team B's perfect map. But no one has been able to prove exactly how to turn Team A's broken, infinite recipe into Team B's perfect map. It's like knowing that flour, eggs, and sugar make a cake, but not knowing how to bake them together without the kitchen exploding.

The Author's Big Idea:
Nima Moshayedi proposes a new way to connect these two teams. He suggests we stop trying to fix the broken recipe (Team A) by adding more ingredients. Instead, he wants to build a translation bridge using a new kind of geometry.

Here is the analogy for his solution:

1. The Common Ground: The "Shape of the Mountain"

Both teams are actually standing on the same piece of land, called the Derived Character Stack.

• Think of this as the "soul" of the mountain. It's not just the physical rocks; it's a magical, multi-layered version of the mountain that remembers every possible way you could walk on it, every time you got stuck, and every loop you made.
• Team A sees this land as a place where you can take tiny steps (perturbations).
• Team B sees this land as a place where you can weave a net (algebraic structures).

2. The Bridge: "Factorization Homology"

This is the author's main tool. Imagine you have a bag of LEGOs (the local rules Team A discovered).

• Factorization Homology is a magical machine that takes those LEGOs and snaps them together to build a giant castle (the global map Team B has).
• The machine works by saying: "If I know the rules for how one brick connects to another, I can figure out how to build the whole city."
• Moshayedi argues that if you feed Team A's local rules into this machine, it automatically spits out Team B's perfect map. You don't need to manually fix the infinite recipe; the machine does the heavy lifting by organizing the pieces correctly.

3. The Secret Glue: "Koszul Duality"

Why does the machine work? The author introduces a concept called Koszul Duality.

• Think of this as a Rosetta Stone for math.
• Team A speaks "Local Language" (tiny, messy details).
• Team B speaks "Global Language" (big, clean patterns).
• Koszul Duality is the dictionary that translates between them. It says: "The messiness you see in the small details is actually just the shadow of the big, clean pattern."
• It explains that the "broken" infinite series from Team A isn't actually broken; it's just a local view of a global structure that Team B already sees. The "glue" that holds the mountain together (the non-perturbative data) is hidden in the way these local pieces talk to each other.

4. The "Stokes" Phenomenon: The Tunneling

Sometimes, the mountain has hidden tunnels connecting different peaks.

• Team A's microscope can't see the tunnels; it only sees the peaks.
• Team B's net catches the whole mountain, including the tunnels.
• Moshayedi suggests that the "glue" (Koszul Duality) is actually the mathematical description of these tunnels. When Team A's numbers start to go crazy (diverge), it's because they are trying to describe a tunnel they can't see. The "magic" of the bridge is realizing that these divergences are just the algebraic signature of the tunnels connecting the peaks.

The Goal

Moshayedi isn't claiming to have finished the map yet. He is proposing a roadmap.
He says: "Let's stop trying to fix the broken recipe. Instead, let's build this translation machine (Factorization Homology) and use this dictionary (Koszul Duality) to show that Team A and Team B are actually describing the exact same thing, just from different angles."

If he is right, it means that the gap between "messy, local physics" and "perfect, global topology" isn't a wall that can't be crossed. It's just a matter of learning the right language to translate between them.

In short:

• Team A = The microscopes (messy, local, infinite).
• Team B = The nets (perfect, global, exact).
• The Paper = A proposal to build a bridge using "Factorization Homology" (the LEGO machine) and "Koszul Duality" (the dictionary) to prove they are looking at the same mountain.

Technical Summary

1. Problem Statement

The paper addresses a fundamental gap in mathematical physics and topology: the relationship between two distinct constructions of invariants for 3-manifolds arising from Chern–Simons (CS) theory:

• Perturbative Side (BV-BFV): The Batalin–Vilkovisky (BV) and Batalin–Fradkin–Vilkovisky (BFV) formalism (developed by Cattaneo, Mnev, and Reshetikhin) provides a rigorous framework for the perturbative quantization of gauge theories on manifolds with boundaries. It yields invariants as formal power series in $\hbar$ (asymptotic expansions around flat connections).
• Non-Perturbative Side (Reshetikhin–Turaev): The Reshetikhin–Turaev (RT) construction produces exact, non-perturbative topological invariants using the algebraic data of Modular Tensor Categories (MTCs), typically derived from quantum groups $U_q(\mathfrak{g})$ at roots of unity.

The Core Question: How can the perturbative BV-BFV data be identified with the exact RT invariants? Standard approaches rely on Borel resummation and resurgence theory to bridge the divergent perturbative series to the exact answer. This paper proposes an alternative, purely algebraic and categorical bridge that avoids analytic resummation difficulties.

2. Methodology and Framework

The author proposes a program to bridge these frameworks by passing through Factorization Homology and Derived Algebraic Geometry (DAG). The strategy relies on four conceptual layers:

1. Classical Geometry (Layer 1): Both frameworks are grounded in the derived character stack $\text{Loc}_G(\Sigma)$, the moduli space of flat $G$-connections on a manifold $\Sigma$.

   • The BV-BFV formalism sees this as a phase space with a shifted symplectic structure.
   • The RT construction accesses it via representation theory.
   • Key Insight: The Pantev–Toën–Vaquié–Vezzosi (PTVV) theorem establishes that $\text{Loc}_G(\Sigma)$ carries a canonical $(2-d)$-shifted symplectic structure, where $d$ is the dimension of $\Sigma$. This unifies the BV structure ($d=3$, $-1$-shifted), the BFV structure ($d=2$, $0$-shifted), and the boundary condition structure ($d=1$, $1$-shifted).

2. Deformation Quantization (Layer 2): The paper posits that the quantization of these shifted symplectic stacks yields the quantum group $U_q(\mathfrak{g})$ and its representation category $\text{Rep}_q(G)$.

   • The deformation quantization of the 1-shifted stack $[G/G]$ (the character stack of a circle) produces the braided monoidal category $\text{Rep}_q(G)$.

3. Factorization Homology (Layer 3): This is the computational engine. Factorization homology integrates an $E_n$-algebra over an $n$-manifold.

   • The category $\text{Rep}_q(G)$ is an $E_2$-algebra (braided monoidal).
   • Integrating this $E_2$-algebra over a surface $\Sigma$ yields the "quantum character variety," which, at roots of unity, collapses to the RT state space (distinguished object in the category of Hilbert spaces).

4. Koszul Duality and Resurgence (Layer 4): The paper introduces $E_n$-Koszul duality as the algebraic mechanism to reconstruct global non-perturbative data from local perturbative data. It conjectures that the Stokes phenomena (tunneling between flat connections) in the analytic resurgence picture correspond to transition maps in the algebraic Koszul reconstruction.

3. Key Contributions and Conjectures

The paper formulates a precise program consisting of seven conjectures that, if proven, would establish a natural equivalence between the two theories.

• Conjecture 6.5 (E2-Equivalence): The $E_2$-monoidal category of boundary conditions derived from BV-BFV quantization on a disk ($B^{(k)}_{CS}$) is equivalent to the ribbon category $\text{Rep}_q(G)$ (the modular tensor category used in RT).
• Conjecture 6.8 (Main Conjecture): There is a natural equivalence of $(3-2-1)$-extended Topological Quantum Field Theories (TQFTs) between the non-perturbative completion of the BV-BFV functor ($Z_{BV, np}$) and the RT functor ($Z_{RT}$).
  • This implies they assign the same categories to circles, the same vector spaces to surfaces, and the same linear maps to 3-cobordisms.
• Conjecture 6.11 (Quantization Agreement): The perturbative BV-BFV quantization of the cylinder $D^2 \times [0,1]$ agrees with the shifted deformation quantization of the 1-shifted symplectic stack $[G/G]$.
• Conjecture 6.14 (Quantized Lagrangian Correspondences): The Lagrangian morphism between the bulk character stack and the boundary character stack (derived from the BV-BFV axiom) quantizes to the linear map (partition function) assigned to the cobordism.
• Conjecture 7.4 (Koszul Reconstruction): The global category of CS boundary conditions is equivalent to the category of ind-coherent sheaves on the formal completion of the character stack: $B_{CS} \simeq \text{IndCoh}(\text{Loc}_G(M)^\wedge)$. This suggests the global non-perturbative category is "glued" from local perturbative data.
• Conjecture 8.3 (Resurgence–Koszul Correspondence): The Stokes data (alien derivatives) governing the resurgent structure of the perturbative series are identified with the transition maps (morphisms) in the Koszul duality reconstruction. This provides an algebraic interpretation of non-perturbative tunneling.
• Conjecture 6.16 (Cellular Compatibility): The cellular (simplicial) BV-BFV partition function is equivalent to the factorization homology colimit.

4. Results and Evidence

While the main conjectures remain open, the paper provides substantial evidence and consistency checks:

• Abelian Case ($G=U(1)$): The program is verified explicitly. The perturbative expansion is one-loop exact (no divergence), the character stack is smooth, and the BV-BFV sum over flat connections reproduces the RT invariant exactly without needing resurgence.
• Genus 0 and 1:
  • For $S^2$, the state space is 1-dimensional in all frameworks.
  • For $T^2$ (Torus) with $G=SU(2)$, the dimension of the state space is $k+1$ in all frameworks. The paper verifies that the $SL_2(\mathbb{Z})$ mapping class group representations (S and T matrices) match the Verlinde formula, the geometric quantization of the "pillowcase," and the factorization homology results.
• Lens Spaces and Seifert Manifolds: The paper reviews existing literature (e.g., Jeffrey, Gukov–Mariño–Putrov) showing that the asymptotic expansion of RT invariants matches the sum of perturbative BV-BFV contributions around flat connections.
• Resurgence Evidence: It highlights the work on the Poincaré homology sphere where the "alien derivatives" (Stokes constants) connecting different perturbative sectors precisely account for the difference between the Borel-resummed series and the exact RT invariant, supporting the Resurgence–Koszul correspondence.

5. Significance

This paper is significant for several reasons:

1. Unification of Frameworks: It proposes a unified geometric origin (the derived character stack with shifted symplectic structures) for both the perturbative (field-theoretic) and non-perturbative (algebraic/topological) approaches to Chern–Simons theory.
2. Algebraic Resolution of Divergence: It suggests that the "gap" between perturbative and non-perturbative physics is not primarily an analytic problem (requiring delicate Borel resummation) but an algebraic one. The passage from local to global data is achieved via Koszul duality and sheaf theory on derived stacks.
3. Extended TQFT: It elevates the comparison to the level of $(3-2-1)$-extended TQFTs, identifying the specific $E_2$-category assigned to the circle, which is a crucial step toward understanding the fully local structure of the theory.
4. Connection to Geometric Langlands: The paper situates the problem within the Geometric Langlands program, viewing the Koszul reconstruction of Chern–Simons boundary conditions as a 3-dimensional, quantized analogue of the Langlands correspondence (relating sheaves on character stacks to D-modules).
5. Roadmap for Future Research: By explicitly stating the technical gaps (e.g., the need for a derived BV-BFV formalism, the commutativity of semisimplification with factorization homology), the paper provides a clear roadmap for mathematicians and physicists to close the loop between perturbative quantum field theory and topological invariants.

In summary, Moshayedi's work offers a sophisticated, conjectural framework that reinterprets the Reshetikhin–Turaev invariants as the natural output of a derived-geometric quantization of the Chern–Simons character stack, mediated by factorization homology and Koszul duality.