Kirkwood–Dirac Quasiprobability Distribution

• The Kirkwood–Dirac quasiprobability distribution is a complex, normalized representation that generalizes classical joint probabilities for pairs of noncommuting observables.
• It provides rigorous structural characterizations and efficient experimental protocols for certifying quantum nonclassicality through weak measurements and shadow techniques.
• Its inherent negativity and complex values underpin quantum resource diagnostics, advancing applications in metrology, thermodynamics, and temporal quantum correlations.

The Kirkwood–Dirac (KD) quasiprobability distribution is a quasi-probability representation of quantum states that generalizes classical joint probability to the context of pairs (or sequences) of noncommuting observables or measurement bases. The KD distribution is normalized, possesses the correct marginals, but allows for entrywise negativity and complex values. This nonclassicality is directly linked to essential quantum phenomena including contextuality, enhanced quantum metrology, anomalous weak values, quantum thermodynamics, information scrambling, and foundational features such as temporal and spatial correlations. The KD framework supports rigorous structural characterizations, efficient experimental certification protocols, and unifies the analysis of state and process nonclassicality across both static and temporal domains.

1. Definition and Mathematical Structure

For a $d$-dimensional Hilbert space and orthonormal bases $\mathcal{A} = \{ |a_i\rangle \}$ and $\mathcal{B} = \{ |b_j\rangle \}$, the KD distribution of a quantum state $\rho$ is given by: $$KD_{ij}(\rho) = \langle b_j | a_i \rangle \langle a_i | \rho | b_j \rangle$$ Equivalently, for projective measurements, $$KD_{ij}(\rho) = \operatorname{Tr}[\rho\,|a_i\rangle\langle a_i|\,|b_j\rangle\langle b_j|]$$. For pure states $|\psi\rangle$, $$KD_{ij}(|\psi\rangle\langle\psi|) = \langle b_j|a_i\rangle \langle a_i|\psi\rangle \langle\psi|b_j\rangle$$.

The distribution is normalized, i.e., $\sum_{i,j} KD_{ij}(\rho) = 1$, and the marginals provide the Born probabilities: $\sum_{j} KD_{ij} = \langle a_i| \rho |a_i\rangle$, $\mathcal{A} = \{ |a_i\rangle \}$0 (Arvidsson-Shukur et al., 2024, Cai et al., 14 Mar 2026, Spriet et al., 3 Nov 2025). This representation is generally complex-valued and can have negative real parts. Only for simultaneously diagonalizable bases does the KD distribution reduce to a bona fide classical joint probability table.

The construction extends to sequences of projectors (multi-time or multi-observable contexts) as $\mathcal{A} = \{ |a_i\rangle \}$1, which forms the basis for temporal generalizations (Jia et al., 8 Jan 2026, Halpern et al., 2017).

2. Convex Geometry and KD-Classical States

The set of states with nonnegative KD distributions is a closed convex subset of the full state space. The extremal KD-positive (KD-classical) states always include the eigenprojectors of each basis. In prime dimension with the DFT as the transition matrix, the only pure KD-classical states are the basis vectors of $\mathcal{A} = \{ |a_i\rangle \}$2 and $\mathcal{A} = \{ |a_i\rangle \}$3, and their convex hull exhausts all KD-classical mixed states (Langrenez et al., 2023, Cai et al., 14 Mar 2026, Xu, 2022). For general ($\mathcal{A} = \{ |a_i\rangle \}$4-dimensional) orthonormal bases, KD-classical pure states are characterized by a global phase-matching and amplitude-consistency structure: for mutually unbiased bases (MUBs), a pure state is KD-classical if and only if $\mathcal{A} = \{ |a_i\rangle \}$5, with $\mathcal{A} = \{ |a_i\rangle \}$6 the numbers of nonvanishing $\mathcal{A} = \{ |a_i\rangle \}$7 and $\mathcal{A} = \{ |a_i\rangle \}$8 amplitudes, respectively (Xu, 2022).

When the transition unitary between $\mathcal{A} = \{ |a_i\rangle \}$9 and $\mathcal{B} = \{ |b_j\rangle \}$0 is generic (off-diagonal elements are all nonzero), the universal upper bound $\mathcal{B} = \{ |b_j\rangle \}$1 applies: all pure states with $\mathcal{B} = \{ |b_j\rangle \}$2 are KD-nonclassical (Bievre, 2021, Langrenez et al., 2024). In small dimensions and for certain non-DFT bases, mixed KD-positive states may exist that are not convex combinations of pure KD-positive states, and their geometry can feature higher-dimensional faces as explicit examples show in spin-1 systems (Langrenez et al., 2023, Langrenez et al., 2024).

3. Nonclassicality, Negativity, and Certification

A state is KD-nonclassical if its KD distribution has negative or imaginary entries. The extent of nonclassicality is quantified by the total KD-variation $\mathcal{B} = \{ |b_j\rangle \}$3 (Tan et al., 2024, Arvidsson-Shukur et al., 2024, Langrenez et al., 2024). A value strictly greater than zero indicates genuine nonclassicality; for pure states, $\mathcal{B} = \{ |b_j\rangle \}$4 if and only if the state is KD-classical (Langrenez et al., 2024).

For mixed states, tight witnesses of nonclassicality are provided by convex-roof extensions both of the support uncertainty and of the total nonpositivity, with the latter being a faithful identifier for the convex hull of pure KD-positive states (Langrenez et al., 2024). Moment inequalities—specifically, violations of Hankel-matrix positivity conditions built from sums of KD entries' powers—also serve as experimentally tractable witnesses of KD-nonpositivity, operationally efficient for randomized (shadow) measurement protocols without the need for full tomography (Chakrabarty et al., 9 Jun 2025).

The existence of negative KD entries is a universal indicator of quantum-resource features: for any resource theory (coherence, entanglement, magic, etc.), a suitable choice of incompatible measurements ensures that resourceful states yield negative KD values, and the total negativity can be related to the geometric Frobenius distance to the classical (resource-free) set (Tan et al., 2024).

4. Operational Roles in Quantum Information and Thermodynamics

The KD distribution underlies the statistical structure of weak measurement outcomes, conditional expectations, and post-selected weak values. When used as a quasi-probabilistic underpinning, only the KD representation has the property that its conditional expectation (given one measurement context) coincides with the mean-square optimal predictor of an observable as a function of that context (Spriet et al., 3 Nov 2025). This structural feature uniquely distinguishes the KD distribution among all Born-compatible quasiprobabilities.

Negative or complex KD values are essential for quantum-advantage protocols in metrology: the presence of KD negativity is required for unbounded Fisher information in post-selected weak-value metrology (Cai et al., 14 Mar 2026, Arvidsson-Shukur et al., 2024, Arvidsson-Shukur et al., 2020). In quantum thermodynamics, the KD distribution describes work and energy fluctuations beyond two-point measurement protocols, enabling the recovery of interference from energy path histories and satisfying fluctuation theorems (e.g., Crooks or Jarzynski equalities) even for coherent initial states (Hernández-Gómez et al., 2024, Pezzutto et al., 10 Mar 2025).

The framework extends naturally to temporal processes, yielding temporal KD distributions whose negativity diagnoses nonclassicality in spacetime—central to the study of temporal quantum correlations, causal modeling, and Leggett–Garg inequalities (Jia et al., 8 Jan 2026).

5. Experimental Measurement and Certification Protocols

KD distributions may be directly reconstructed experimentally via weak measurement, interferometric schemes, or circuits based on Bargmann invariants (Wagner et al., 2023, Hernández-Gómez et al., 2024). Weak coupling to an ancilla, coherent control protocols, and shadow-assisted polynomial estimators permit efficient access to both the real and imaginary KD statistics; these methods often require sample complexity only polynomial in $\mathcal{B} = \{ |b_j\rangle \}$5 for practical low order moment witnesses (Chakrabarty et al., 9 Jun 2025). Direct KD certification thus circumvents the exponential overheads of full state tomography in moderate dimensions, and enables in-situ resource witnessing in contemporary platforms including superconducting circuits and nitrogen-vacancy centers (Pezzutto et al., 10 Mar 2025, Hernández-Gómez et al., 2024).

6. Distinctive Properties and Hierarchies Among Quasiprobabilities

The KD distribution is unique among all Born-compatible quasiprobabilities (those with correct marginals for two observables) in yielding the best-predictor conditional expectations and informational completeness for state reconstruction (Spriet et al., 3 Nov 2025, Umekawa et al., 2023). Only the KD distribution expresses the quantum joint statistics strictly in terms of the eigenvalues of the input observables, respecting their possible measurement outcomes (Umekawa et al., 2023). KD, Wigner, and generalized phase-space distributions form a strict hierarchy: classical joint probabilities $\mathcal{B} = \{ |b_j\rangle \}$6 KD distributions $\mathcal{B} = \{ |b_j\rangle \}$7 postquantum quasiprobabilities, with universal bounds (e.g., $\mathcal{B} = \{ |b_j\rangle \}$8 and $\mathcal{B} = \{ |b_j\rangle \}$9) separating these sets (Liu et al., 12 Apr 2025).

7. Extensions, Temporal Structure, and Quantum Chaos

The KD formalism supports several generalizations. Extended KD distributions for sequences of noncommuting projectors naturally encompass temporal quantum phenomena. In the multi-time domain, the right, left, and doubled temporal KD quasiprobabilities provide a unified operational foundation linking effective process-tensor, superdensity, and pseudo-density operator frameworks (Jia et al., 8 Jan 2026). The KD view of temporal processes provides explicit schemes for temporal Bloch tomography.

A prominent application is the out-of-time-ordered correlator (OTOC) in quantum chaos, which admits a quasiprobability description as an extended KD distribution. Here, the fine structure and temporal symmetry-breaking in the KD distribution track the onset and dynamics of information scrambling, and distinguish chaotic from integrable behavior (Halpern et al., 2017).

---

The KD distribution, through its precise algebraic structure, convex geometries, natural fit with quantum measurement theory, and operational certification protocols, has become a central analytical and experimental tool for diagnosing and leveraging nonclassical resources in quantum science (Cai et al., 14 Mar 2026, Arvidsson-Shukur et al., 2024, Spriet et al., 3 Nov 2025).