betaFeedback

WorksPhysics and Astronomy

Microcausality Bounds on Dispersion Laws

Abstract

Microcausality holds in any state, including those that spontaneously break Lorentz invariance, and requires the poles (and other singularities) of the retarded Green's function to satisfy Im ω≤∣Im k∣\mathrm{Im}\,ω\le |\mathrm{Im}\,k| throughout the complex plane. Building on the work of Heller, Serantes, Spaliński and Withers, we study its consequences for dispersion relations that are linear at small kk, like Goldstone bosons, assuming tree-level, non-dissipative dynamics. We show that microcausality is equivalent to requiring that k±ω(k)k \pm ω(k) be typically real functions. This yields two-sided bounds on all higher-order coefficients of ω(k)ω(k) in units of its radius of convergence RR, the scale at which the linear Goldstone description breaks down. The optimal constraints are characterised by an infinite hierarchy of Toeplitz determinants. We derive analogous bounds on ω2(k)ω^2(k), whose radius of convergence R~\tilde R marks the scale at which new states must appear. The bounds shrink to zero as the sound speed csc_s approaches 1: for a fixed higher-derivative correction, both RR and R~\tilde R become parametrically smaller than naive dimensional analysis suggests. In particular, in the EFT of Inflation all higher-derivative quadratic operators must vanish as cs→1c_s \to 1. We test the bounds in several models: a U(1)U(1) superfluid, light in a dielectric medium, photon propagation in a strong magnetic field, colour-flavour-locked quark matter, and a homogeneous solid. Finally, we extend the bounds to gapped modes.

Prepared for submission to JHEP

Introduction

Our understanding of quantum field theory (QFT) relies to a large extent on Lorentz invariance. In most physical systems, however, Lorentz invariance is spontaneously broken by the state. In cosmology there is always a preferred frame, the one in which matter is at rest; in condensed matter, and more generally in any QFT at finite temperature or chemical potential, the medium defines a preferred frame. We stress that we are concerned with the spontaneous breaking of Lorentz invariance, and not in a fundamental breaking of this symmetry.

Without Lorentz invariance QFT is considerably richer, and correspondingly less understood. In this paper we address what is arguably the simplest question one can ask in this setting. We assume that boosts are broken, i.e. there is a preferred frame, and we study the dispersion relation ω(k)\omega(k) of a weakly coupled excitation, i.e. a pole of the retarded Green’s function. If Lorentz invariance were linearly realised, the answer would be unique: ω2=k2+m2\omega^2 = k^2 + m^2. Without it, ω(k)\omega(k) is a priori an arbitrary function. Microcausality—the vanishing of commutators of local operators at spacelike separation—is a property of the operator algebra and holds whatever the state, including states that break Lorentz invariance [1, 2]. Our goal is to understand what constraints microcausality imposes on ω(k)\omega(k).

Microcausality is mapped to analyticity in Fourier space. In particular, as we will review, the poles of the Green’s functions, whose location is described by the dispersion relation ω(k)\omega(k) must satisfy

Im⁡ω(k)≤∣Im⁡k∣.(1)\operatorname{Im}\omega(k) \leq\lvert\operatorname{Im} k\rvert. \tag*{(1)}

Heller, Serantes, Spaliński and Withers realised that this condition, imposed throughout the complex kk-plane, is very constraining [3, 4]. Applying it to hydrodynamic modes, they derived rigorous two-sided bounds on all the coefficients of their small-kk expansion, in units of its radius of convergence.

In this paper we study the consequences of (1) for dispersion relations that are linear at small momentum,

ω(k)=±(csk+∑n=1∞α2n+1k2n+1),∣k∣<R,(2)\omega(k) = \pm\left(c_s k + \sum_{n=1}^{\infty}\alpha_{2n+1}k^{2n+1}\right), \qquad\lvert k\rvert< R , \tag*{(2)}

with RR the radius of convergence. This applies to Goldstone bosons1 (phonons in superfluids and solids, the inflaton fluctuation in the effective field theory (EFT) of Inflation [6], the scalar mode of dark energy [7, 8, 9] . . . ) but also, for instance, to photons travelling in a medium. The coefficients α2n+1\alpha_{2n+1} are encoded in the higher-derivative operators of the EFT. As in [3, 4], we neglect loops in the EFT, so that at low energy the singularities of the Green’s function are only poles and ω(k)\omega(k) is analytic in a disc around k=0k = 0. This is the case, for instance, when the UV completion is weakly coupled. Unlike [3, 4], we restrict to non-dissipative dynamics and assume that ω(k)\omega(k) is real for real kk.

In Section 2.2, to guide the discussion, we present a simple UV complete model of a U(1)U(1) superfluid. In Section 2.3 we show that condition (1) is equivalent to the requirement that the two functions k±ω(k)k \pm\omega(k) be typically real [10] in the disc ∣k∣<R|k| < R, i.e. that their imaginary part has the same sign as Im⁡k\operatorname{Im} k. The Taylor coefficients of typically real functions obey universal bounds: each coefficient satisfies the Rogosinski bounds, while the complete set of constraints on the sequence of coefficients is characterised by an infinite hierarchy of Toeplitz determinants (see Appendix § 8). (Related tools from geometric function theory have recently found applications in the SS-matrix bootstrap [11, 12].) The allowed region for the dimensionless coefficients α2n+1R2n\alpha_{2n+1}R^{2n} is obtained by intersecting the constraints of the two functions k±ω(k)k \pm\omega(k), eq. (37), and its projections, computed numerically, are shown in Figure 6.

Figure 6. The Rogosinski-Toeplitz bounds on a3a_3, a5a_5, a7a_7 and a9a_9. Each coloured region corresponds to the N+1N+1 truncation of the Toeplitz determinant. For N=2N=2, the purple region describes −(1−cs)K2∩(1+cs)K2-(1-c_s)K_2\cap(1+c_s)K_2, shown (also in purple) in Figure 5. As N→∞N\to\infty, the bounds converge towards the optimal region (37) shown in red. The grey regions are the bounds given by [3] (light grey) and [4] (dark grey). The values of a3a_3, a5a_5, a7a_7 and a9a_9 for the U(1)U(1) superfluid studied in § 2.2 are also shown as the black curve from cs=0c_s=0 to 1.

Rather than as bounds on the coefficients at fixed RR, these results should be seen as upper bounds on RR once some coefficients of the series are known: beyond this bound the series cannot satisfy microcausality (1), so something must happen for ∣k∣≤R|k| \le R to save the day. (This is analogous to how unitarity bounds are used to estimate the scale of new physics.) In all the examples we study, ω(k)\omega(k) is microcausal up to the radius of convergence of the series. However, one can also consider situations in which microcausality, eq. (1), holds only in a smaller disc. For instance, some approximation may break down, so that the system apparently loses microcausality before the radius of convergence is reached; or a pole may “hide” on a second sheet of the Green function and become unphysical (see the discussion in [13, 14]). In these cases our results apply with RR the maximum radius within which microcausality holds. For simplicity we will keep calling RR the radius of convergence, but it should always be understood as the maximum radius for which the series (2) is microcausal.

The bounds become particularly interesting as cs→1c_s \to1 (Section 2.3.3), where the allowed region shrinks to zero linearly in 1−cs1-c_s:

−(2n+1)(1−cs)≤α2n+1R2n≤κ2n+1(1−cs),(3)-(2n+1)(1-c_s) \le\alpha_{2n+1}R^{2n} \le\kappa_{2n+1}(1-c_s), \tag*{(3)}

with κ2n+1>0\kappa_{2n+1} > 0 defined in eq. (29). That microcausality forbids any higher-order correction for cs=1c_s = 1 was shown in [3], and in [15] it was shown that a dispersion relation at order k3k^3 is microcausal only in a window of momenta that shrinks as cs→1c_s \to1.

The scale RR has a transparent physical meaning: it is the momentum at which the description in terms of a linear dispersion relation plus small corrections breaks down. This does not necessarily require new states. In the U(1)U(1) superfluid of Section 2.2, for cs<1/2c_s < 1/\sqrt{2}, RR is set by the collision of the two Goldstone branches ±ω(k)\pm\omega(k), a branch point of ω2\omega^2. In physical terms, 1/R1/R is the so-called healing length of the superfluid, above which the Bogoliubov spectrum crosses over to the non-relativistic one, ω≃k2/2M\omega\simeq k^2/2M. New degrees of freedom appear instead at the radius of convergence of ω2(k)\omega^2(k) that we denote R~\widetilde{R}. This shows that the scales RR and R~\widetilde{R} have a different physical meaning.

Therefore, in Section 2.4 we derive analogous bounds on the coefficients α~2n\widetilde{\alpha}_{2n} of ω2(k)\omega^2(k), in units of R~\widetilde{R}. In this case the microcausality condition is not linear in the function, and we only obtain necessary conditions: a csc_s-independent bound, ∣α~2n∣R~2(n−1)≤1|\widetilde{\alpha}_{2n}|\widetilde{R}^{2(n-1)} \le1 (Section 2.4.1), and a csc_s-dependent one, which follows from the typical-realness of (k2−ω2)/k(k^2-\omega^2)/k and again shrinks to zero as cs→1c_s \to1 (§ 2.4.2). Together, our bounds, once a higher-derivative term is known, allow to set rigorous upper bounds on the two scales RR and R~\widetilde{R}.

In § 3 we test the bounds in several models with spontaneously broken Lorentz invariance: light in a medium described by a Drude-Lorentz model; a photon in a strong homogeneous magnetic field, whose radius of convergence is set by the pair-production threshold; the octet and singlet Goldstones of colour-flavour-locked quark matter, for which the radius of convergence is determined numerically with the Domb-Sykes method (Appendix § 10). A model of a homogeneous solid is studied in Appendix § 11.

The implications for cosmology are briefly discussed in § 4. In particular, the bounds (3) imply that, at the level of the quadratic action, the EFT of Inflation is forced to become Lorentz invariant as cs→1c_s \to1, as conjectured in [16]. The same conclusions apply to the EFT of Dark Energy. In § 5 we extend the bounds on ω2(k)\omega^2(k) to gapped modes. We conclude in § 6 with many open directions.

Microcausality bounds

Our starting point is the retarded two-point function: given a bosonic, local, and gauge-invariant operator O\mathcal{O}, the retarded Green’s function is defined as

G(t,x⃗;t′,x⃗′)=iθ(t−t′)⟨[O(t,x⃗),O(t′,x⃗′)]⟩.(4)G(t,\vec{x};t',\vec{x}') = i\theta(t-t')\langle[\mathcal{O}(t,\vec{x}),\mathcal{O}(t',\vec{x}')]\rangle. \tag*{(4)}

As evident from (4), the retarded two-point function has support in the future light cone i.e. t≥∣x⃗∣t \ge|\vec{x}|. Although Lorentz symmetry is spontaneously broken in the theories that we will consider in this paper, the notion of microcausality still exists. Indeed, microcausality is an operatorial statement and thus is valid for any state (or density matrix), including Lorentz-breaking ones (see, for example, [1]).

For states with time-translational invariance2 and spatial homogeneity, the retarded two-point function will depend only on the differences t−t′t-t' and x⃗−x⃗′\vec{x}-\vec{x}', and we can work with the retarded two-point function in Fourier space:

G~(ω,k⃗)=∫ddx G(0;t,x⃗)ei(ωt−k⃗⋅x⃗).(5)\widetilde{G}(\omega,\vec{k}) = \int\mathrm{d}^{d}x\,G(0;t,\vec{x})e^{i(\omega t-\vec{k}\cdot\vec{x})}. \tag*{(5)}

Under the assumption that the operator O\mathcal{O} is Hermitian, G(t,x⃗)G(t,\vec{x}) will be real, and its Fourier transform G~(ω,k⃗)\widetilde{G}(\omega,\vec{k}) satisfies the reality condition:

G~∗(ω,k⃗)=G~(−ω∗,−k⃗∗).(6)\widetilde{G}^{*}(\omega,\vec{k}) = \widetilde{G}(-\omega^{*},-\vec{k}^{*}). \tag*{(6)}

Taking ω=ωR+iωI\omega=\omega_R+i\omega_I and k⃗=k⃗R+ik⃗I\vec{k}=\vec{k}_R+i\vec{k}_I, (5) becomes

G~(ω,k⃗)=∫ddx G(0;t,x⃗)ei(ωRt−k⃗R⋅x⃗)e−(ωIt−k⃗I⋅x⃗)(7)\widetilde{G}(\omega,\vec{k}) = \int\mathrm{d}^{d}x\,G(0;t,\vec{x})e^{i(\omega_R t-\vec{k}_R\cdot\vec{x})}e^{-(\omega_I t-\vec{k}_I\cdot\vec{x})} \tag*{(7)}

and, given the support of G(t,x⃗)G(t,\vec{x}) and the assumption that G(t,x⃗)G(t,\vec{x}) is polynomially bounded3, (7) converges in the region

Im⁡ω>∣Im⁡k⃗∣.(8)\operatorname{Im}\omega> \lvert\operatorname{Im}\vec{k}\rvert. \tag*{(8)}

As a consequence, we can conclude that G~(ω,k⃗)\widetilde{G}(\omega,\vec{k}) must be analytic within this region.

For simplicity, we further restrict to systems with isotropy, although many of our results can be straightforwardly generalised to the anisotropic case. Due to rotational symmetry, G(t,x⃗)G(t,\vec{x}) will only depend on x⃗\vec{x} through x2x^2 and, accordingly, G~(ω,k⃗)\widetilde{G}(\omega,\vec{k}) will only depend on k⃗\vec{k} through k2k^2. Thus one can reduce to study the retarded two-point function in terms of the two complex variables ω\omega and kk4 with an additional symmetry:

G~(ω,−k)=G~(ω,k).(9)\widetilde{G}(\omega,-k)=\widetilde{G}(\omega,k). \tag*{(9)}

Following (8), G~(ω,k)\widetilde{G}(\omega,k) will be analytic within the region

Im⁡ω>∣Im⁡k∣.(10)\operatorname{Im}\omega> \lvert\operatorname{Im}k\rvert. \tag*{(10)}

A more general and rigorous discussion on the properties of the retarded two-point function when boosts are broken can be found in [17].

Goldstone dispersion relation

References

  1. [1]S. Dubovsky, A. Nicolis, E. Trincherini and G. Villadoro, Microcausality in curved space-time, Phys. Rev. D 77 (2008) 084016 [0709.1483].DOI
  2. [2]G. Cintia, F. Piazza and S. Ramos, Modified microcausality from perturbation theory, Phys. Rev. D 112 (2025) 085015 [2504.16992].arxiv.org/abs/2504.16992
  3. [3]M.P. Heller, A. Serantes, M. Spaliński and B. Withers, Rigorous Bounds on Transport from Causality, Phys. Rev. Lett. 130 (2023) 261601 [2212.07434].DOI
  4. [4]M.P. Heller, A. Serantes, M. Spaliński and B. Withers, The space of transport coefficients allowed by causality, Nature Phys. 20 (2024) 1948 [2305.07703].DOI
  5. [5]H.B. Nielsen and S. Chadha, On How to Count Goldstone Bosons, Nucl. Phys. B 105 (1976) 445.DOI
  6. [6]C. Cheung, P. Creminelli, A.L. Fitzpatrick, J. Kaplan and L. Senatore, The Effective Field Theory of Inflation, JHEP 03 (2008) 014 [0709.0293].DOI
  7. [7]P. Creminelli, G. D’Amico, J. Norena and F. Vernizzi, The Effective Theory of Quintessence: the w<-1 Side Unveiled, JCAP 02 (2009) 018 [0811.0827].
  8. [8]G. Gubitosi, F. Piazza and F. Vernizzi, The Effective Field Theory of Dark Energy, JCAP 02 (2013) 032 [1210.0201].DOI
  9. [9]J. Gleyzes, D. Langlois, F. Piazza and F. Vernizzi, Healthy theories beyond Horndeski, Phys. Rev. Lett. 114 (2015) 211101 [1404.6495].DOI
  10. [10]W. Rogosinski, Über positive harmonische Entwicklungen und typisch-reelle Potenzreihen, Mathematische Zeitschrift 35 (1932) 93.DOI
  11. [11]P. Haldar, A. Sinha and A. Zahed, Quantum field theory and the Bieberbach conjecture, SciPost Phys. 11 (2021) 002 [2103.12108].DOI
  12. [12]P. Raman and A. Sinha, QFT, EFT and GFT, JHEP 12 (2021) 203 [2107.06559].
  13. [13]L. Gavassino, How acausal equations emerge from causal dynamics, 2604.07031.
  14. [14]R. Brants, Causal UV completions of relativistic hydrodynamics, 2605.21377.arxiv.org/abs/2605.21377
  15. [15]L. Hui, A. Nicolis, A. Podo and S. Zhou, Microcausality without Lorentz invariance, JHEP 07 (2025) 188 [2502.04215].DOI
  16. [16]D. Baumann, D. Green, H. Lee and R.A. Porto, Signs of Analyticity in Single-Field Inflation, Phys. Rev. D 93 (2016) 023523 [1502.07304].DOI
  17. [17]P. Creminelli, A. Longo, B. Salehian and A. Zahed, Analyticity and positivity of Green’s functions without Lorentz, JHEP 06 (2026) 135 [2512.10843].arxiv.org/abs/2512.10843
  18. [18]B. Bellazzini, J. Elias Miró, R. Rattazzi, M. Riembau and F. Riva, Positive moments for scattering amplitudes, Phys. Rev. D 104 (2021) 036006 [2011.00037].DOI
  19. [19]P. Creminelli, M. Delladio, O. Janssen, A. Longo and L. Senatore, Non-analyticity of the S-matrix with spontaneously broken Lorentz invariance, JHEP 06 (2024) 201 [2312.08441].
  20. [20]L. Hui, I. Kourkoulou, A. Nicolis, A. Podo and S. Zhou, S-matrix positivity without Lorentz invariance: a case study, JHEP 04 (2024) 145 [2312.08440].
  21. [21]A. Esposito, A. Nicolis and R. Penco, Effective Field Theories for Material Media, 2607.06666.arxiv.org/abs/2607.06666
  22. [22]K. Schmüdgen, The Moment Problem, vol. 277 of Graduate Texts in Mathematics, Springer, Cham, 1 ed. (2017), 10.1007/978-3-319-64546-9.DOI
  23. [23]J.D. Jackson, Classical Electrodynamics, Wiley (1998).
  24. [24]W. Heisenberg and H. Euler, Consequences of Dirac’s theory of positrons, Z. Phys. 98 (1936) 714 [physics/0605038].
  25. [25]F. Karbstein, Photon polarization tensor in a homogeneous magnetic or electric field, Phys. Rev. D 88 (2013) 085033 [1308.6184].DOI
  26. [26]A.K. Harding and D. Lai, Physics of Strongly Magnetized Neutron Stars, Rept. Prog. Phys. 69 (2006) 2631 [astro-ph/0606674].DOI
  27. [27]M.G. Alford, A. Schmitt, K. Rajagopal and T. Schäfer, Color superconductivity in dense quark matter, Rev. Mod. Phys. 80 (2008) 1455 [0709.4635].
  28. [28]K. Zarembo, Dispersion laws for Goldstone bosons in a color superconductor, Phys. Rev. D 62 (2000) 054003 [hep-ph/0002123].DOI
  29. [29]C. Domb and M.F. Sykes, On the susceptibility of a ferromagnetic above the curie point, Proceedings of the Royal Society of London. Series A, Mathematical and Physical Sciences 240 (1957) 214.DOI
  30. [30]N. Arkani-Hamed, P. Creminelli, S. Mukohyama and M. Zaldarriaga, Ghost inflation, JCAP 04 (2004) 001 [hep-th/0312100].
  31. [31]G.W. Horndeski, Second-order scalar-tensor field equations in a four-dimensional space, Int. J. Theor. Phys. 10 (1974) 363.DOI
  32. [32]A. Adams, N. Arkani-Hamed, S. Dubovsky, A. Nicolis and R. Rattazzi, Causality, analyticity and an IR obstruction to UV completion, JHEP 10 (2006) 014 [hep-th/0602178].DOI
  33. [33]E.H. Lieb and D.W. Robinson, The finite group velocity of quantum spin systems, Commun. Math. Phys. 28 (1972) 251.
  34. [34]S.D. Chowdhury, S.A. Hartnoll, A. Hebbar and R. Khondaker, Bootstrapping transport in the Drude-Kadanoff-Martin model, 2509.18255.arxiv.org/abs/2509.18255
  35. [35]S.D. Chowdhury, A. Hebbar and S.A. Hartnoll, Lieb-Robinson causality and non-Fermi liquids, Phys. Rev. B 113 (2026) 245125 [2510.18950].arxiv.org/abs/2510.18950
  36. [36]C. de Rham and S. Melville, Gravitational Rainbows: LIGO and Dark Energy at its Cutoff, Phys. Rev. Lett. 121 (2018) 221101 [1806.09417].DOI
  37. [37]P.L. Duren, Univalent Functions, vol. 259 of Grundlehren der mathematischen Wissenschaften, Springer-Verlag, New York (1983).
  38. [38]M.S. Robertson, On the theory of typically-real functions, Bulletin of the American Mathematical Society 41 (1935) 565.
  39. [39]L. Koczan and P. Zaprawa, “On typically real functions with n-fold symmetry.” 2006.
  40. [40]C. Carathéodory, Über den variabilitätsbereich der fourierschen konstanten von positiven harmonischen funktionen, Rendiconti del Circolo Matematico di Palermo 32 (1911) 193.DOI
  41. [41]J.B. Conway, Functions of One Complex Variable I, vol. 11 of Graduate Texts in Mathematics, Springer-Verlag, 2nd ed. (1978), 10.1007/978-1-4612-6313-5.DOI
  42. [42]M. Van Dyke, Analysis and improvement of perturbation series, Quarterly Journal of Mechanics and Applied Mathematics 27 (1974) 423.DOI
  43. [43]L. Alberte, M. Baggioli, V.C. Castillo and O. Pujolas, Elasticity bounds from Effective Field Theory, Phys. Rev. D 100 (2019) 065015 [1807.07474].DOI
  44. [44]A. Esposito, R. Krichevsky and A. Nicolis, Gravitational Mass Carried by Sound Waves, Phys. Rev. Lett. 122 (2019) 084501 [1807.08771].DOI
  45. [45]S. Endlich, A. Nicolis and J. Wang, Solid Inflation, JCAP 10 (2013) 011 [1210.0569].
  46. [46]A. Esposito, R. Krichevsky and A. Nicolis, Solidity without inhomogeneity: Perfectly homogeneous, weakly coupled, UV-complete solids, JHEP 11 (2020) 021 [2004.11386].DOI

Paper details

vlnc:W_h7dvzh49
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 10

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Physics and Astronomy

Claude (Anthropic)

View statement
Looking for related papers…
Contents