Ancestor Virasoro Constraints in CohFT

• Ancestor Virasoro constraints are systems of differential equations that annihilate the total ancestor potential in cohomological field theories, providing a half-Virasoro algebra structure.
• They are constructed via methods such as loop-space quantization, period field expansions, and residue proofs, linking FJRW theory, singularity theory, and topological recursion.
• Semisimplicity and mirror symmetry proofs reduce complex ancestor invariants to one-dimensional models, verifying string, dilaton, and higher-order recursion relations.

Ancestor Virasoro constraints are systems of differential equations that annihilate the total ancestor potential of a cohomological field theory, singularity theory, or related enumerative structure. In the Fan–Jarvis–Ruan–Witten theory of an admissible Landau–Ginzburg pair $(W,G)$, they take the form

$$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$

where the operators $L_k^{(W,G)}$ realize a half-Virasoro algebra on the formal Fock space of ancestor variables (He et al., 2021). Closely related formulations appear for homogeneous CohFTs with vacuum, where Guo–Zhang propose the ancestor Virasoro conjecture for calibrated theories (Guo et al., 26 Feb 2025), in singularity theory, where Milanov identifies $N$ copies of Virasoro constraints with the Eynard–Orantin recursion for the total ancestor potential (Milanov, 2012), and in topological recursion for arbitrary spectral curves, where a direct residue proof yields $L_mA=0$ for all $m\ge -1$ (Guo et al., 27 Jul 2025). Across these settings, the constraints organize higher-genus ancestor invariants into a rigid representation-theoretic structure.

1. Total ancestor potential and ancestor correlators

For an admissible Landau–Ginzburg pair $(W,G)$, the FJRW theory defines a CohFT

$$\Lambda^{\mathrm{FJRW}}_{g,k}:H_{W,G}^{\otimes k}\longrightarrow H^*(\overline{\mathcal M}_{g,k},\mathbb C),$$

on the state space $H_{W,G}$. Its ancestor invariants are

$$\big\langle \tau_{m_1}(d_{i_1})\dots\tau_{m_k}(d_{i_k})\big\rangle^{(W,G)}_{g,k} = \int_{[\overline{\mathcal M}_{g,k}]^{\rm vir}} \prod_{j=1}^k\psi_j^{m_j}\; \Lambda^{\mathrm{FJRW}}_{g,k}(d_{i_1}\otimes\cdots\otimes d_{i_k}),$$

and the total ancestor potential is the formal exponential generating series

$$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$0

It is therefore a formal function of the variables $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$1 that packages all FJRW ancestor invariants in all genera (He et al., 2021).

In the more general framework of a CohFT with vacuum, one starts with a finite-dimensional complex vector space $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$2, multilinear classes

$$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$3

and a vacuum vector $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$4 satisfying the forgetful pull-back identity

$$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$5

After choosing an $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$6-calibration and a $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$7-calibration, the total ancestor potential is written as

$$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$8

with $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$9 the generating series of ancestor correlators in variables $L_k^{(W,G)}$0 (Guo et al., 26 Feb 2025). The common feature is that ancestor potentials are genus-graded partition functions whose coefficients are $L_k^{(W,G)}$1-class correlators on moduli spaces.

The same packaging principle appears in topological recursion. For a spectral curve $L_k^{(W,G)}$2, the Eynard–Orantin correlators $L_k^{(W,G)}$3 are expanded in the basis $L_k^{(W,G)}$4 near branch points, and the corresponding coefficients define ancestor correlators

$L_k^{(W,G)}$5

The ancestor partition function is then

$L_k^{(W,G)}$6

so the ancestor formalism extends well beyond geometric Gromov–Witten or FJRW theories (Guo et al., 27 Jul 2025).

2. Virasoro operators and loop-space quantization

In FJRW theory, the operator construction is formulated on the infinite-dimensional symplectic vector space

$L_k^{(W,G)}$7

with Darboux coordinates

$L_k^{(W,G)}$8

Quadratic Hamiltonians are quantized by

$L_k^{(W,G)}$9

With grading operator

$N$0

one sets

$N$1

After quantization and the dilaton shift $N$2, one obtains explicit quadratic differential operators in the $N$3, and these operators satisfy

$N$4

This is the half-Virasoro algebra relevant to the ancestor potential (He et al., 2021).

For homogeneous CohFTs with vacuum, the corresponding operators are constructed from the infinitesimal symplectic operators

$N$5

and then quantized to

$N$6

In local coordinates $N$7, they take the general quadratic form

$N$8

where the coefficients are universal polynomials in the Euler operator and the calibration data (Guo et al., 26 Feb 2025).

In topological recursion, the explicit operators involve the spectral-curve data more transparently. For each $N$9, one has an operator

$L_mA=0$0

with first- and second-order differential terms determined by the coefficients $L_mA=0$1, the first $L_mA=0$2-matrix coefficient $L_mA=0$3, and the vacuum shift $L_mA=0$4. These operators also satisfy

$L_mA=0$5

In this formulation, the Virasoro algebra is read directly from the recursion data of the spectral curve (Guo et al., 27 Jul 2025).

A distinct but related singularity-theoretic realization appears in Milanov’s construction. There one introduces fields

$L_mA=0$6

expands them as

$L_mA=0$7

and obtains $L_mA=0$8 independent copies of the Virasoro algebra,

$L_mA=0$9

all with central charge $m\ge -1$0 (Milanov, 2012).

3. The constraint equations and their lowest consequences

The basic ancestor Virasoro statement in FJRW theory is the conjecture that for every admissible pair $m\ge -1$1,

$m\ge -1$2

Expanding in $m\ge -1$3 and in the variables $m\ge -1$4, this becomes an infinite system of linear relations among ancestor invariants in every genus and with every number of marked points (He et al., 2021).

The first two constraints recover standard structural equations. In the FJRW formulation,

$m\ge -1$5

which is the string equation, while $m\ge -1$6 gives a homogeneity equation of the form

$m\ge -1$7

In genus $m\ge -1$8, one may ignore $m\ge -1$9 and the second-order terms in $(W,G)$0, and the resulting modified Virasoro operators yield pure linear recursion among the primary genus-$(W,G)$1 correlators (He et al., 2021).

In the CohFT-with-vacuum setting, the ancestor Virasoro conjecture is formulated under a homogeneity condition on the vacuum and on the $(W,G)$2- and $(W,G)$3-calibrations: $(W,G)$4

$(W,G)$5

Under these assumptions, the conjecture states

$(W,G)$6

Guo–Zhang verify the genus-$(W,G)$7 part and deduce a simplified form of the genus-$(W,G)$8 part for arbitrary CohFTs (Guo et al., 26 Feb 2025).

At genus $(W,G)$9, the higher-insertion structure collapses dramatically: for $$\Lambda^{\mathrm{FJRW}}_{g,k}:H_{W,G}^{\otimes k}\longrightarrow H^*(\overline{\mathcal M}_{g,k},\mathbb C),$$0, the condition $$\Lambda^{\mathrm{FJRW}}_{g,k}:H_{W,G}^{\otimes k}\longrightarrow H^*(\overline{\mathcal M}_{g,k},\mathbb C),$$1 is equivalent to the trace identity

$$\Lambda^{\mathrm{FJRW}}_{g,k}:H_{W,G}^{\otimes k}\longrightarrow H^*(\overline{\mathcal M}_{g,k},\mathbb C),$$2

This reduction is significant because it replaces a formally infinite system of genus-$$\Lambda^{\mathrm{FJRW}}_{g,k}:H_{W,G}^{\otimes k}\longrightarrow H^*(\overline{\mathcal M}_{g,k},\mathbb C),$$3 differential constraints by a finite supertrace expression (Guo et al., 26 Feb 2025).

The topological-recursion version has parallel lowest equations. Writing

$$\Lambda^{\mathrm{FJRW}}_{g,k}:H_{W,G}^{\otimes k}\longrightarrow H^*(\overline{\mathcal M}_{g,k},\mathbb C),$$4

the $$\Lambda^{\mathrm{FJRW}}_{g,k}:H_{W,G}^{\otimes k}\longrightarrow H^*(\overline{\mathcal M}_{g,k},\mathbb C),$$5, $$\Lambda^{\mathrm{FJRW}}_{g,k}:H_{W,G}^{\otimes k}\longrightarrow H^*(\overline{\mathcal M}_{g,k},\mathbb C),$$6 equation yields the string equation

$$\Lambda^{\mathrm{FJRW}}_{g,k}:H_{W,G}^{\otimes k}\longrightarrow H^*(\overline{\mathcal M}_{g,k},\mathbb C),$$7

while the $$\Lambda^{\mathrm{FJRW}}_{g,k}:H_{W,G}^{\otimes k}\longrightarrow H^*(\overline{\mathcal M}_{g,k},\mathbb C),$$8, $$\Lambda^{\mathrm{FJRW}}_{g,k}:H_{W,G}^{\otimes k}\longrightarrow H^*(\overline{\mathcal M}_{g,k},\mathbb C),$$9 equation gives the dilaton equation

$H_{W,G}$0

These formulas make explicit that the ancestor Virasoro hierarchy begins with the expected string and dilaton constraints and then continues to higher nontrivial recursions (Guo et al., 27 Jul 2025).

4. Proven cases and proof mechanisms

He–Shen prove the FJRW ancestor Virasoro conjecture in three principal classes of examples. The first class consists of invertible polynomials with maximal symmetry group $H_{W,G}$1. In this setting, Berglund–Hübsch–Krawitz mirror symmetry and the Polishchuk–Vaintrob / Saito–Givental identification imply that the FJRW CohFT of $H_{W,G}$2 is equivalent to the Saito–Givental semisimple CohFT of the mirror singularity $H_{W,G}$3. Since Givental–Teleman implies that any semisimple CohFT satisfies Virasoro constraints, one obtains $H_{W,G}$4 (He et al., 2021).

The second class is formed by selected two-variable examples $H_{W,G}$5 with minimal group $H_{W,G}$6. Here the proof proceeds by direct semisimplicity analysis in genus $H_{W,G}$7: one computes the quantum multiplication by the Euler vector field and verifies that it has distinct eigenvalues. Givental’s result then yields the full constraints (He et al., 2021).

The third class consists of Calabi–Yau type polynomials. When the central charge satisfies $H_{W,G}$8, a degree count shows that the only possible nonzero correlators satisfy the homogeneity condition needed for $H_{W,G}$9; string and dilaton hold formally, and higher $$\big\langle \tau_{m_1}(d_{i_1})\dots\tau_{m_k}(d_{i_k})\big\rangle^{(W,G)}_{g,k} = \int_{[\overline{\mathcal M}_{g,k}]^{\rm vir}} \prod_{j=1}^k\psi_j^{m_j}\; \Lambda^{\mathrm{FJRW}}_{g,k}(d_{i_1}\otimes\cdots\otimes d_{i_k}),$$0 follow by grading symmetry. In the elliptic-curve case $$\big\langle \tau_{m_1}(d_{i_1})\dots\tau_{m_k}(d_{i_k})\big\rangle^{(W,G)}_{g,k} = \int_{[\overline{\mathcal M}_{g,k}]^{\rm vir}} \prod_{j=1}^k\psi_j^{m_j}\; \Lambda^{\mathrm{FJRW}}_{g,k}(d_{i_1}\otimes\cdots\otimes d_{i_k}),$$1, the proof uses the Landau–Ginzburg/Calabi–Yau correspondence of Li–Shen–Zhou: under a holomorphic Cayley transform, the ancestor Gromov–Witten potential of the elliptic curve is carried to the FJRW ancestor potential, and since the Gromov–Witten side satisfies Virasoro and the transform commutes with the Virasoro operators, the FJRW side does as well (He et al., 2021).

For homogeneous CohFTs with vacuum, Guo–Zhang establish several complementary results. The genus-$$\big\langle \tau_{m_1}(d_{i_1})\dots\tau_{m_k}(d_{i_k})\big\rangle^{(W,G)}_{g,k} = \int_{[\overline{\mathcal M}_{g,k}]^{\rm vir}} \prod_{j=1}^k\psi_j^{m_j}\; \Lambda^{\mathrm{FJRW}}_{g,k}(d_{i_1}\otimes\cdots\otimes d_{i_k}),$$2 part of the conjecture is a formal consequence of topological recursion relations in genus $$\big\langle \tau_{m_1}(d_{i_1})\dots\tau_{m_k}(d_{i_k})\big\rangle^{(W,G)}_{g,k} = \int_{[\overline{\mathcal M}_{g,k}]^{\rm vir}} \prod_{j=1}^k\psi_j^{m_j}\; \Lambda^{\mathrm{FJRW}}_{g,k}(d_{i_1}\otimes\cdots\otimes d_{i_k}),$$3, namely WDVV together with string and dilaton, and of the cone property $$\big\langle \tau_{m_1}(d_{i_1})\dots\tau_{m_k}(d_{i_k})\big\rangle^{(W,G)}_{g,k} = \int_{[\overline{\mathcal M}_{g,k}]^{\rm vir}} \prod_{j=1}^k\psi_j^{m_j}\; \Lambda^{\mathrm{FJRW}}_{g,k}(d_{i_1}\otimes\cdots\otimes d_{i_k}),$$4 for the graph of $$\big\langle \tau_{m_1}(d_{i_1})\dots\tau_{m_k}(d_{i_k})\big\rangle^{(W,G)}_{g,k} = \int_{[\overline{\mathcal M}_{g,k}]^{\rm vir}} \prod_{j=1}^k\psi_j^{m_j}\; \Lambda^{\mathrm{FJRW}}_{g,k}(d_{i_1}\otimes\cdots\otimes d_{i_k}),$$5. In the semisimple case, the full higher-genus statement follows from the Givental–Teleman reconstruction

$$\big\langle \tau_{m_1}(d_{i_1})\dots\tau_{m_k}(d_{i_k})\big\rangle^{(W,G)}_{g,k} = \int_{[\overline{\mathcal M}_{g,k}]^{\rm vir}} \prod_{j=1}^k\psi_j^{m_j}\; \Lambda^{\mathrm{FJRW}}_{g,k}(d_{i_1}\otimes\cdots\otimes d_{i_k}),$$6

where $$\big\langle \tau_{m_1}(d_{i_1})\dots\tau_{m_k}(d_{i_k})\big\rangle^{(W,G)}_{g,k} = \int_{[\overline{\mathcal M}_{g,k}]^{\rm vir}} \prod_{j=1}^k\psi_j^{m_j}\; \Lambda^{\mathrm{FJRW}}_{g,k}(d_{i_1}\otimes\cdots\otimes d_{i_k}),$$7 is the Witten–Kontsevich $$\big\langle \tau_{m_1}(d_{i_1})\dots\tau_{m_k}(d_{i_k})\big\rangle^{(W,G)}_{g,k} = \int_{[\overline{\mathcal M}_{g,k}]^{\rm vir}} \prod_{j=1}^k\psi_j^{m_j}\; \Lambda^{\mathrm{FJRW}}_{g,k}(d_{i_1}\otimes\cdots\otimes d_{i_k}),$$8-function, $$\big\langle \tau_{m_1}(d_{i_1})\dots\tau_{m_k}(d_{i_k})\big\rangle^{(W,G)}_{g,k} = \int_{[\overline{\mathcal M}_{g,k}]^{\rm vir}} \prod_{j=1}^k\psi_j^{m_j}\; \Lambda^{\mathrm{FJRW}}_{g,k}(d_{i_1}\otimes\cdots\otimes d_{i_k}),$$9 is the homogeneous $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$00-matrix, and $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$01 is the translation by the $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$02-vector. Since the Virasoro constraints hold for each $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$03, they survive the $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$04- and $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$05-conjugations (Guo et al., 26 Feb 2025).

These proofs exhibit a recurrent mechanism: semisimplicity reduces ancestor Virasoro constraints to the one-dimensional KdV model through Givental’s quantization formalism. A plausible implication is that semisimplicity is not merely a technical convenience but a structural regime in which Virasoro symmetries become directly reconstructible from calibration data and canonical coordinates.

5. Relations with mirror symmetry, topological recursion, and singularity theory

In the invertible FJRW case with maximal group, the ancestor constraints are tied to mirror symmetry in a precise sense: the FJRW theory is identified with the Saito–Givental $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$06-model of the transpose polynomial $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$07, and Virasoro annihilation follows from semisimplicity on the mirror side (He et al., 2021). For elliptic curves, the use of quasi-modular forms and the Cayley transform converts Gromov–Witten Virasoro constraints into FJRW Virasoro constraints; the map preserving Virasoro is identified with the comparison between quasi-modular expansions around a cusp (He et al., 2021).

Milanov’s work places ancestor Virasoro constraints inside the Eynard–Orantin formalism of singularity theory. Starting from a miniversal unfolding $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$08 with primitive form, one defines period vectors $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$09 and phase forms governing the propagators $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$10. The total ancestor potential $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$11 is then acted on by operators $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$12 extracted from the expansion of the fields $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$13, and the main theorem states that

$$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$14

Moreover, this system is exactly equivalent to the local Eynard–Orantin recursion built from the same period data (Milanov, 2012). The equivalence furnishes a dictionary between recursive residue calculus and Virasoro representation theory.

The 2025 topological-recursion formulation removes the restriction to semisimple singularity theory and works for an arbitrary spectral curve $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$15. The crucial residue identity

$$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$16

is shown term-by-term to be equivalent to $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$17 (Guo et al., 27 Jul 2025). In this setting, the Virasoro constraints arise directly from the recursion kernel residue rather than from Givental–Teleman reconstruction.

The Airy curve

$$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$18

provides the basic example. Here the nonzero correlators are the usual $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$19-class integrals on $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$20, the ancestor potential begins with explicitly listed low-order terms, and the operator $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$21 reproduces the Witten–Kontsevich Virasoro constraints (Guo et al., 27 Jul 2025). This identifies the classical point-target theory as the local model for the general ancestor Virasoro formalism in topological recursion.

6. Generalizations, examples, and conceptual distinctions

A central distinction is between ancestor and descendent Virasoro constraints. The ancestor version acts on the total ancestor potential, typically after a dilaton shift or its analogue. In a CohFT with vacuum, the $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$22-calibration explicitly plays the role of the dilaton shift in the absence of a flat unit (Guo et al., 26 Feb 2025). In topological recursion, the ancestor variables $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$23 can be related to descendant-type variables $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$24 by a Miura-type change of variables, and one then recovers descendant Virasoro operators from the ancestor ones (Guo et al., 27 Jul 2025). This rules out the common conflation of ancestor and descendant constraints as merely notational variants.

A second distinction concerns the algebra itself. In FJRW theory, the operators form a half-Virasoro algebra indexed by $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$25 (He et al., 2021). In homogeneous CohFT with vacuum and in topological recursion, the operators satisfy the standard commutation relations $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$26 for $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$27 (Guo et al., 26 Feb 2025, Guo et al., 27 Jul 2025). In Milanov’s singularity-theory formulation, one obtains $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$28 independent copies of the Virasoro algebra, each of central charge $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$29, one for each branch point or canonical sector (Milanov, 2012). These are structurally related but not identical realizations.

Guo–Zhang also provide two explicit applications of the general CohFT-with-vacuum formalism. For the deformed negative $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$30-spin theory $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$31, one constructs a homogeneous CohFT on $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$32, writes the $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$33-matrix and $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$34-vector in closed form, and obtains

$$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$35

This generalizes BGW-type Virasoro constraints to arbitrary $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$36 (Guo et al., 26 Feb 2025). For the semisimple CohFT $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$37 associated with extended Grothendieck’s dessins d’enfants, explicit $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$38 and $$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$39 calibrations produce a descendent potential whose specialization recovers the dessins-counting generating function of Kazarian–Zograf (Guo et al., 26 Feb 2025).

A further generalization appears in Drinfeld–Sokolov hierarchies. There the DS tau-function can be written as the ancestor potential of a semisimple Frobenius manifold, and the Virasoro symmetries of the hierarchy translate into constraints

$$L_k^{(W,G)}\,\mathcal A_{W,G}(\mathbf t;\hbar)=0,\qquad k\ge -1,$$40

In the point-target case this reproduces the Witten–Kontsevich theorem, while similarity equations select the Brezin–Gross–Witten solution and other special solutions characterized by ordinary differential equations of Painlevé type (Liu et al., 2019).

Taken together, these developments show that ancestor Virasoro constraints form a unifying structure across FJRW theory, semisimple and calibrated CohFTs, singularity theory, topological recursion, and integrable hierarchies. What varies from setting to setting is the realization of the operators—via loop-space quantization, period fields, calibration data, or recursion residues—rather than the underlying principle that the total ancestor potential is characterized by Virasoro annihilation.