Beyond Wigner: Non-Invertible Symmetries Preserve Probabilities

1. The Problem

The paper addresses a fundamental tension between two pillars of modern quantum theory:

• Wigner's Theorem (1931): States that any symmetry transformation preserving transition amplitudes (and thus probabilities) in a quantum system must be implemented by a unitary (or anti-unitary) operator acting on a fixed Hilbert space. Crucially, unitary operators are invertible.
• Non-Invertible Symmetries: Recent developments in Quantum Field Theory (QFT) and Condensed Matter Physics have identified "generalized" or "categorical" symmetries described by fusion categories. Unlike group symmetries, these symmetries are generated by topological defects that are non-invertible under fusion (e.g., $D \otimes D = 1 \oplus \eta$ in the Ising model).

The Conflict: If non-invertible symmetries act as linear operators on a fixed Hilbert space, they cannot be unitary (as they map states to zero or mix sectors in a way that violates invertibility). This seemingly contradicts the requirement that physical symmetries must preserve probabilities. The central question is: How can non-invertible symmetries preserve quantum probabilities without violating Wigner's theorem?

2. Methodology

The authors resolve this paradox by redefining the mathematical framework in which these symmetries act. Their methodology involves:

• Twisted Sectors: Instead of acting on a single Hilbert space $\mathcal{H}$, the authors propose that symmetry defects act on a collection of twisted sector Hilbert spaces ($\mathcal{H}_X$), where $X$ represents a topological defect "puncturing" the spatial manifold (e.g., a cylinder).
• Transition Channels: When a defect $A$ acts on a state in a twisted sector $\mathcal{H}_X$, it can transition the system to a different twisted sector $\mathcal{H}_Y$. The authors introduce the concept of transition channels (morphisms in the fusion category) $\phi: A \otimes X \to Y \otimes A$ that mediate these transitions.
• Tube Algebra and Generalized Charges: They utilize the tube category (or tube algebra) associated with the fusion category $\mathcal{C}$. A "generalized charge" is defined as a linear functor $U$ from the tube category to the category of Hilbert spaces. This functor maps the algebraic structure of defects to linear operators between the various twisted Hilbert spaces.
• Unitarity Constraint: The authors rigorously demonstrate that for the probability-preserving property to hold, the underlying fusion category $\mathcal{C}$ must be unitary (possessing a specific $^\dagger$-structure compatible with the fusion rules).

3. Key Contributions

A. The Categorical Probability Preservation (CPP) Theorem

The core theoretical contribution is Theorem 2 (CPP). It states that for a unitary fusion category $\mathcal{C}$, the action of a symmetry defect $A$ on a twisted sector $\mathcal{H}_X$ does not preserve the inner product on $\mathcal{H}_X$ alone. However, if one considers the sum over all possible outgoing twisted sectors $S$ and a specific basis of transition channels $\{e^S_i\}$, the total map becomes an isometry.

Mathematically, for any state $|\Phi\rangle, |\Psi\rangle \in \mathcal{H}_X$:
$$\sum_{S, i} \langle U(\frac{A}{S}{X} e^S_i) \Phi, U(\frac{A}{S}{X} e^S_i) \Psi \rangle = \langle \Phi, \Psi \rangle$$
where $U(\frac{A}{S}{X} e^S_i)$ is the operator mapping $\mathcal{H}_X \to \mathcal{H}_S$.

B. Interpretation as Quantum Channels

The authors show that non-invertible symmetries naturally act as trace-preserving quantum channels (completely positive maps).

• The action of a defect $A$ on a density matrix $\rho$ is decomposed into a sum over Kraus operators corresponding to the transition channels.
• The completeness relation of these Kraus operators ensures that the total probability is conserved ($\text{Tr}[\mathcal{E}(\rho)] = \text{Tr}[\rho]$).
• This resolves the "non-invertible" issue: the symmetry is not a unitary operator on a single space, but a unitary embedding into a larger, direct-sum Hilbert space $\mathcal{H}_{total} = \bigoplus_S \mathcal{H}_S$.

C. Necessity of Unitarity

The paper proves that the CPP theorem fails if the fusion category is not unitary. Using the Yang-Lee category (a non-unitary counterpart to the Fibonacci category) as a counter-example, they show that no basis of transition channels exists that satisfies the isometry condition. This provides a physical justification for why physical symmetries must be described by unitary categories.

4. Results and Examples

The authors validate their theory through explicit calculations in several examples:

• Critical Ising CFT: They recover the known Kramers-Wannier duality. The non-invertible defect $D$ maps the spin field $\sigma$ (in the untwisted sector) to the disorder operator $\mu$ (in the $\eta$-twisted sector). The probability is conserved by summing over the outcomes in the enlarged space.
• Rep($S_3$): They analyze the unitary fusion category of representations of the permutation group $S_3$. They explicitly construct the transition channels for the non-invertible defect $\pi$ and calculate the transition probabilities between different twisted sectors, verifying the isometry condition.
• Fibonacci Category: They demonstrate the construction for the Fibonacci category (objects $1, \tau$ with $\tau \otimes \tau = 1 \oplus \tau$), showing how the golden ratio dimensions naturally appear in the transition probabilities.
• Yang-Lee (Counter-Example): They show that for the non-unitary Yang-Lee category, the quantum dimension is negative ($d_\tau = -1/\phi$), leading to a violation of the probability conservation condition, confirming the necessity of the unitary structure.
• Higher Dimensions: The framework is generalized to $d$-dimensional spacetime using fusion $(d-1)$-categories, where defects act on local operators via linking, preserving the probability structure in higher dimensions.

5. Significance

• Resolution of a Theoretical Paradox: The paper definitively answers whether non-invertible symmetries are compatible with quantum mechanics. The answer is yes, provided one expands the Hilbert space to include all twisted sectors and treats the symmetry action as a quantum channel rather than a simple unitary operator.
• Physical Justification for Unitarity: It provides a direct physical argument (probability conservation) for why symmetry categories in physical theories must be unitary, complementing previous mathematical arguments based on reflection positivity.
• Scattering and S-Matrix: The results have immediate applications to scattering theory involving non-invertible defects. The transition channels identified in the paper encode the possible scattering outcomes of operators, offering a new tool for calculating S-matrices in theories with generalized symmetries.
• Unified Framework: It unifies the treatment of invertible and non-invertible symmetries under the umbrella of "generalized charges" acting on twisted sectors, bridging the gap between topological quantum field theory and quantum information theory (quantum channels).

In summary, the paper establishes that non-invertible symmetries preserve probabilities by acting as isometries between distinct twisted Hilbert spaces, effectively realizing the symmetry as a unitary process in an enlarged configuration space.