Parity-Odd Matter Trispectrum Insights

• Parity-odd matter trispectrum is defined as the imaginary component of the four-point matter density function, encoding handedness via pseudoscalar contractions.
• It originates from primordial inflationary mechanisms, including axion–U(1) systems and dynamical Chern–Simons gravity, yielding measurable parity-violating signatures.
• Detection methods involve dedicated trispectrum estimators and compressed POP spectra, with significant implications for galaxy surveys, CMB lensing, and intrinsic alignment studies.

A parity-odd matter trispectrum is the parity-changing component of the connected four-point function of the matter density contrast in Fourier space. In the notation used for the matter field $\delta_m$,

$$\langle\delta_m(\vec{k}_1,\tau)\delta_m(\vec{k}_2,\tau)\delta_m(\vec{k}_3,\tau)\delta_m(\vec{k}_4,\tau)\rangle_c =(2\pi)^3\delta_D(\vec{k}_{1234})\,T_m(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$

with $$\vec{k}_{1234}\equiv \vec{k}_1+\vec{k}_2+\vec{k}_3+\vec{k}_4$$ (Azyzy et al., 7 Oct 2025). For real density fields, parity acts in Fourier space as

$$T_m(-\vec{k}_1,-\vec{k}_2,-\vec{k}_3,-\vec{k}_4,\tau)=T_m^*(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$

so one may write

$T_m=T_{m,+}+i\,T_{m,-},$

where $T_{m,+}$ is parity-even and $T_{m,-}$ is parity-odd (Azyzy et al., 7 Oct 2025). Under translational invariance, rotational invariance, and parity-even late-time dynamics, the trispectrum is the lowest scalar correlator that can carry parity-odd information; the power spectrum and bispectrum cannot do so in expectation (Coulton et al., 2023). In practice, parity oddness is the handed component of the tetrahedron formed by the four wavevectors, and it is encoded by pseudoscalar momentum contractions such as $\mathbf{k}_1\cdot(\mathbf{k}_2\times\mathbf{k}_3)$ (Jamieson et al., 2024).

1. Definition and kinematic structure

A standard decomposition of a scalar trispectrum factors out the unique parity-odd rotational invariant built from three independent wavevectors. In one formulation,

$$\langle \delta(\mathbf{k}_1)\cdots\delta(\mathbf{k}_4)\rangle = \delta_D(\mathbf{k}_1+\mathbf{k}_2+\mathbf{k}_3+\mathbf{k}_4) \left[\tau_+(k_1,k_2,k_3,k_4,K,K')+i\,(\mathbf{k}_1\cdot(\mathbf{k}_2\times\mathbf{k}_3))\,\tau_-(k_1,k_2,k_3,k_4,K,K')\right],$$

with $K\equiv |\mathbf{k}_1+\mathbf{k}_2|$ and $$\langle\delta_m(\vec{k}_1,\tau)\delta_m(\vec{k}_2,\tau)\delta_m(\vec{k}_3,\tau)\delta_m(\vec{k}_4,\tau)\rangle_c =(2\pi)^3\delta_D(\vec{k}_{1234})\,T_m(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$0 (Coulton et al., 2023). In this language, $$\langle\delta_m(\vec{k}_1,\tau)\delta_m(\vec{k}_2,\tau)\delta_m(\vec{k}_3,\tau)\delta_m(\vec{k}_4,\tau)\rangle_c =(2\pi)^3\delta_D(\vec{k}_{1234})\,T_m(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$1 is parity even and $$\langle\delta_m(\vec{k}_1,\tau)\delta_m(\vec{k}_2,\tau)\delta_m(\vec{k}_3,\tau)\delta_m(\vec{k}_4,\tau)\rangle_c =(2\pi)^3\delta_D(\vec{k}_{1234})\,T_m(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$2 is parity odd.

The geometric content is specific to four-point order. For scalar Fourier-space correlators of real fields, parity oddness requires a pseudoscalar, and in three dimensions the irreducible pseudoscalar is the scalar triple product $$\langle\delta_m(\vec{k}_1,\tau)\delta_m(\vec{k}_2,\tau)\delta_m(\vec{k}_3,\tau)\delta_m(\vec{k}_4,\tau)\rangle_c =(2\pi)^3\delta_D(\vec{k}_{1234})\,T_m(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$3 (Coulton et al., 2023). Two-point functions depend only on a single magnitude, and three-point functions are constrained by momentum conservation to closed triangles, so neither carries an independent handedness. A tetrahedron, by contrast, distinguishes left-handed from right-handed configurations. This is why a parity-odd matter trispectrum is naturally interpreted as a handedness-sensitive four-point statistic (Jamieson et al., 2024).

The same point may be phrased in terms of reality. For a real scalar field $$\langle\delta_m(\vec{k}_1,\tau)\delta_m(\vec{k}_2,\tau)\delta_m(\vec{k}_3,\tau)\delta_m(\vec{k}_4,\tau)\rangle_c =(2\pi)^3\delta_D(\vec{k}_{1234})\,T_m(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$4, $$\langle\delta_m(\vec{k}_1,\tau)\delta_m(\vec{k}_2,\tau)\delta_m(\vec{k}_3,\tau)\delta_m(\vec{k}_4,\tau)\rangle_c =(2\pi)^3\delta_D(\vec{k}_{1234})\,T_m(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$5, so the Fourier-space parity-even part is real and the parity-odd part is imaginary (Jamieson et al., 2024). Several primordial calculations use this fact directly and identify the parity-even and parity-odd parts of the scalar trispectrum with the real and imaginary parts of the reduced four-point function (Fujita et al., 2023).

2. Primordial inflationary origins

A large fraction of the literature treats the parity-odd matter trispectrum as the late-time descendant of a primordial parity-odd scalar or curvature trispectrum generated during inflation. One explicit mechanism is a spectator axion–$$\langle\delta_m(\vec{k}_1,\tau)\delta_m(\vec{k}_2,\tau)\delta_m(\vec{k}_3,\tau)\delta_m(\vec{k}_4,\tau)\rangle_c =(2\pi)^3\delta_D(\vec{k}_{1234})\,T_m(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$6 system with Chern–Simons coupling,

$$\langle\delta_m(\vec{k}_1,\tau)\delta_m(\vec{k}_2,\tau)\delta_m(\vec{k}_3,\tau)\delta_m(\vec{k}_4,\tau)\rangle_c =(2\pi)^3\delta_D(\vec{k}_{1234})\,T_m(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$7

for which only one gauge helicity is amplified. In that setup, the sourced curvature trispectrum $$\langle\delta_m(\vec{k}_1,\tau)\delta_m(\vec{k}_2,\tau)\delta_m(\vec{k}_3,\tau)\delta_m(\vec{k}_4,\tau)\rangle_c =(2\pi)^3\delta_D(\vec{k}_{1234})\,T_m(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$8 is parity violating, with the odd component encoded in $$\langle\delta_m(\vec{k}_1,\tau)\delta_m(\vec{k}_2,\tau)\delta_m(\vec{k}_3,\tau)\delta_m(\vec{k}_4,\tau)\rangle_c =(2\pi)^3\delta_D(\vec{k}_{1234})\,T_m(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$9. The paper finds that in exact equilateral configurations the ratio of parity-odd to parity-even can reach $$\vec{k}_{1234}\equiv \vec{k}_1+\vec{k}_2+\vec{k}_3+\vec{k}_4$$0, while in a quasi-equilateral configuration with $$\vec{k}_{1234}\equiv \vec{k}_1+\vec{k}_2+\vec{k}_3+\vec{k}_4$$1 and $$\vec{k}_{1234}\equiv \vec{k}_1+\vec{k}_2+\vec{k}_3+\vec{k}_4$$2, the parity-odd component can become comparable to or larger than the parity-even one over appreciable angular regions (Fujita et al., 2023).

Axion inflation with a Chern–Simons interaction $$\vec{k}_{1234}\equiv \vec{k}_1+\vec{k}_2+\vec{k}_3+\vec{k}_4$$3 gives a related one-loop scalar trispectrum in which one vector helicity is predominantly produced during inflation. In that calculation the parity-even piece is sizeable, while the parity-odd contribution is about one to two orders of magnitude smaller. The numerical study reports that the massive and massless vector fields give significant parity-even signals, while the parity-odd part is typically subdominant, with stronger suppression for heavier vectors (Niu et al., 2022).

A distinct route uses dynamical Chern-Simons gravity. There, inflationary gravitons become chiral, and graviton exchange between two scalar pairs imprints chirality onto the scalar four-point function. In the collapsed limit, the curvature trispectrum takes the form

$$\vec{k}_{1234}\equiv \vec{k}_1+\vec{k}_2+\vec{k}_3+\vec{k}_4$$4

and the ratio of odd to even pieces becomes

$$\vec{k}_{1234}\equiv \vec{k}_1+\vec{k}_2+\vec{k}_3+\vec{k}_4$$5

so the angular dependence directly reflects the spin-2 exchange structure (Creque-Sarbinowski et al., 2023).

Not all parity-violating inflationary theories produce a nonzero scalar parity-odd trispectrum under standard assumptions. In dynamical Chern-Simons gravity treated through parity-odd interaction vertices and general-relativistic bulk propagators, the scalar parity-odd trispectrum vanishes for Bunch-Davies initial conditions, both for the graviton-mediated exchange diagram and for the contact diagram. A nonzero signal arises only for non-Bunch-Davies scalar initial states (Orlando et al., 22 May 2025). This establishes that explicit parity-odd vertices are not, by themselves, sufficient.

3. From primordial curvature to the intrinsic matter trispectrum

The primordial-to-matter mapping is conceptually standard. In one common formulation,

$$\vec{k}_{1234}\equiv \vec{k}_1+\vec{k}_2+\vec{k}_3+\vec{k}_4$$6

so a primordial curvature trispectrum propagates as

$$\vec{k}_{1234}\equiv \vec{k}_1+\vec{k}_2+\vec{k}_3+\vec{k}_4$$7

up to late-time gravitationally induced trispectrum contributions and biasing effects for galaxies (Fujita et al., 2023). In linear theory the same statement is written as

$$\vec{k}_{1234}\equiv \vec{k}_1+\vec{k}_2+\vec{k}_3+\vec{k}_4$$8

with

$$\vec{k}_{1234}\equiv \vec{k}_1+\vec{k}_2+\vec{k}_3+\vec{k}_4$$9

(Azyzy et al., 7 Oct 2025).

Once nonlinear structure formation is included, the parity-odd matter trispectrum is modified but not erased. Numerical $$T_m(-\vec{k}_1,-\vec{k}_2,-\vec{k}_3,-\vec{k}_4,\tau)=T_m^*(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$0-body simulations with parity-violating initial conditions show that the realization-averaged power spectrum, bispectrum, halo mass function, and matter PDF are not affected by the modifications to the initial conditions, while the parity-odd trispectrum of matter shows distinct signatures proportional to the parity-violating parameter $$T_m(-\vec{k}_1,-\vec{k}_2,-\vec{k}_3,-\vec{k}_4,\tau)=T_m^*(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$1 (Coulton et al., 2023). At $$T_m(-\vec{k}_1,-\vec{k}_2,-\vec{k}_3,-\vec{k}_4,\tau)=T_m^*(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$2, using 500 simulations with $$T_m(-\vec{k}_1,-\vec{k}_2,-\vec{k}_3,-\vec{k}_4,\tau)=T_m^*(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$3 and 500 with $$T_m(-\vec{k}_1,-\vec{k}_2,-\vec{k}_3,-\vec{k}_4,\tau)=T_m^*(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$4, the parity-odd matter trispectrum is strongly detected, whereas the matter power spectrum, halo power spectrum, matter PDF, halo mass function, matter bispectrum, and parity-even matter trispectrum remain null in the odd-$$T_m(-\vec{k}_1,-\vec{k}_2,-\vec{k}_3,-\vec{k}_4,\tau)=T_m^*(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$5 difference (Coulton et al., 2023).

The nonlinear evolution has now been computed to one-loop order within the effective field theory of large-scale structure. In that treatment,

$$T_m(-\vec{k}_1,-\vec{k}_2,-\vec{k}_3,-\vec{k}_4,\tau)=T_m^*(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$6

and the paper demonstrates the structure of infrared divergence cancellations, as required by the equivalence principle (Azyzy et al., 7 Oct 2025). The universal ultraviolet counterterm has the form

$$T_m(-\vec{k}_1,-\vec{k}_2,-\vec{k}_3,-\vec{k}_4,\tau)=T_m^*(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$7

with the same effective sound-speed coefficient structure familiar from the power spectrum renormalization (Azyzy et al., 7 Oct 2025). For the specific primordial template studied there, nonlinear clustering suppresses the parity-odd signal on short scales in squeezed configurations and produces both BAO damping and a pronounced BAO phase shift (Azyzy et al., 7 Oct 2025).

4. Estimation strategies and compressed statistics

A major practical development is the construction of dedicated parity-odd estimators. In Fourier space, one explicit matter-trispectrum estimator inserts the triple product directly,

$$T_m(-\vec{k}_1,-\vec{k}_2,-\vec{k}_3,-\vec{k}_4,\tau)=T_m^*(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$8

which makes the estimator sensitive to tetrahedral orientation and nulls parity-even contributions (Coulton et al., 2023). The same work derives an FFT-based implementation and shows that the estimator vanishes if $$T_m(-\vec{k}_1,-\vec{k}_2,-\vec{k}_3,-\vec{k}_4,\tau)=T_m^*(\vec{k}_1,\vec{k}_2,\vec{k}_3,\vec{k}_4,\tau),$$9 or $T_m=T_{m,+}+i\,T_{m,-},$0, reflecting its antisymmetry properties (Coulton et al., 2023).

A complementary line of work compresses the parity-odd trispectrum into power spectra of composite fields. The “Parity-Odd Power” spectra are defined from nonlinear combinations involving gradients, curls, and filtering functions, and they are explicit weighted compressions of the parity-odd trispectrum shape $T_m=T_{m,+}+i\,T_{m,-},$1 (Jamieson et al., 2024). Two basic constructions are introduced: a vector POP spectrum sensitive to soft diagonal limits, and a scalar POP spectrum sensitive to soft side limits. In controlled simulations with an injected parity-odd primordial trispectrum, both POP spectra agree with semi-analytic predictions, Gaussian simulations give POP spectra consistent with zero, and the correlation matrices are dominated by the diagonal, with off-diagonal entries below about $T_m=T_{m,+}+i\,T_{m,-},$2 (Jamieson et al., 2024).

These approaches serve different purposes. The direct trispectrum estimator retains the full tetrahedral information and is the natural statistic for model-specific searches. The POP spectra are one-dimensional compressions with simpler covariance and lower computational cost. This suggests a division between discovery-oriented compressed statistics and full-shape inference with parity-odd trispectrum estimators.

5. Intrinsic matter, intrinsic galaxy, and observed galaxy trispectra

A central conceptual distinction is between intrinsic parity violation of the matter field and apparent parity violation of the observed galaxy field.

Object	Parity status	Origin
Intrinsic matter trispectrum	Parity invariant in standard GR scalar structure formation	Late-time gravitational dynamics are parity even (Paul et al., 2024)
Intrinsic galaxy trispectrum	Parity invariant under standard local, parity-respecting bias	Inherits parity-even intrinsic matter statistics (Paul et al., 2024)
Observed galaxy trispectrum	Can acquire a parity-odd component	Relativistic redshift-space and lightcone projection effects (Paul et al., 2024)

The observed galaxy overdensity contains a Newtonian piece and a relativistic piece,

$T_m=T_{m,+}+i\,T_{m,-},$3

with

$T_m=T_{m,+}+i\,T_{m,-},$4

(Paul et al., 2024). The Newtonian terms contain only even powers of $T_m=T_{m,+}+i\,T_{m,-},$5 and preserve parity, while the leading relativistic terms are odd in the line-of-sight cosine and generate an imaginary parity-odd observed trispectrum (Paul et al., 2024). The paper argues that this apparent parity violation can be significant on intermediate linear scales and dominant around the equality scale and above (Paul et al., 2024).

This distinction matters directly for interpretation. A measured parity-odd galaxy four-point signal does not automatically imply intrinsic parity violation of the underlying matter field. It may reflect observer-dependent projection effects. At the same time, several primordial models were developed precisely to interpret the large parity-odd trispectrum signal extracted from BOSS galaxy-clustering data, and the rolling spectator axion–$T_m=T_{m,+}+i\,T_{m,-},$6 model was explicitly proposed as a possible primordial explanation (Fujita et al., 2023).

6. Projections to the CMB, lensing, and other observables

The parity-odd matter trispectrum is not confined to three-dimensional large-scale structure analyses. It projects into several complementary observables, though often with nontrivial suppression.

For the primary CMB temperature trispectrum on large angular scales, a parity-violating matter 4-point function projects inefficiently. When the parity-violating support is limited to scales $T_m=T_{m,+}+i\,T_{m,-},$7, the odd CMB trispectrum is suppressed by at least one factor of $T_m=T_{m,+}+i\,T_{m,-},$8, with $T_m=T_{m,+}+i\,T_{m,-},$9, so $T_{m,+}$0. A non-detection of parity violation in the CMB trispectrum is therefore not necessarily inconsistent with a genuine detection in the large-scale-structure 4PCF (Slepian et al., 20 Oct 2025).

CMB lensing is more sensitive because it probes a broad radial kernel rather than a thin shell. A general reduced lensing-trispectrum expression has been derived for arbitrary input matter trispectra, and for a parity-odd input trispectrum the angular signal is nonzero only when

$T_{m,+}$1

the exact parity-odd selection rule for the lensing trispectrum in that formalism (Greco et al., 21 May 2025). This provides a route for future CMB lensing experiments to probe parity violation in large-scale structure.

In the scalar CMB trispectrum literature, the same odd-$T_{m,+}$2 selection rule is a defining signature. A harmonic-space analysis of parity violation in the primordial scalar trispectrum found that the CMB trispectrum is nonvanishing in the $T_{m,+}$3 domain, and a Fisher calculation yielded a minimum detectable size of the dipolar coefficient $T_{m,+}$4 in a cosmic-variance-limited-level temperature survey (Shiraishi, 2016).

Tensorial tracers provide another complementary channel. Within the EFT description of intrinsic alignments, the parity-odd IA power spectrum probes the collapsed limit of the parity-odd primordial trispectrum, and in the $T_{m,+}$5-gauge model the renormalized observable takes the form

$T_{m,+}$6

with $T_{m,+}$7 for approximately scale-invariant initial conditions (Kurita et al., 10 Sep 2025). This is not a direct matter-trispectrum measurement, but it is a biased response to a specific collapsed-limit sector of the parity-odd primordial trispectrum.

7. Formal constraints, computational status, and unresolved issues

The modern theory of parity-odd scalar trispectra includes a strong structural constraint: under tree-level approximation, Bunch-Davies initial conditions, and exact scale invariance with IR convergence, the maximally-connected parts of wavefunction coefficients are purely real for massless scalar external states. Since parity-odd scalar correlators are purely imaginary, the tree-level parity-odd scalar four-point function is necessarily factorised and does not diverge when the total energy is conserved (Stefanyszyn et al., 2023). In the language of inflationary correlators, the parity-odd trispectrum is therefore built from lower-point building blocks rather than from a genuinely singular maximally-connected contribution (Stefanyszyn et al., 2023).

This general theorem clarifies several model-specific results. It is consistent with the finding that the scalar parity-odd trispectrum generated by dynamical Chern-Simons interaction vertices vanishes for Bunch-Davies initial conditions, and that non-Bunch-Davies scalar initial states are required to obtain a nonzero signal in that setup (Orlando et al., 22 May 2025). It also helps explain why model building often focuses either on helicity-asymmetric sourced sectors, such as axion–gauge systems, or on nonstandard initial states.

On the computational side, the full parity-violating trispectrum in axion inflation is technically demanding because it is a one-loop sourced object with very high-dimensional integrals. A recent analysis showed how to simplify the full one-loop sourced primordial curvature trispectrum by performing all angular integrations analytically and reducing the problem to low-dimensional radial integrals. This provides the first feasible route to evaluating the full parity-odd primordial trispectrum across momentum configurations and is presented as the computational bridge toward full late-time 4PCF predictions for DESI, Euclid, and Roman (Reinhard et al., 2024).

Several issues remain open. The relation between primordial parity-odd curvature trispectra, intrinsic matter trispectra, and observed galaxy trispectra requires simultaneous control of nonlinear evolution, galaxy bias, redshift-space distortions, relativistic projection effects, and survey window functions. The existing literature now contains explicit primordial mechanisms, nonlinear matter evolution, estimator constructions, and projection formalisms, but these ingredients are not yet unified into a single end-to-end survey analysis. This suggests that the parity-odd matter trispectrum is best viewed, at present, as a sharply defined theoretical observable with multiple realizations: intrinsic and primordial in some models, apparent and projection-induced in others, and complementary to CMB, lensing, and intrinsic-alignment probes.