Acausality of Massive Charged Spin 2 Fields
The recent claim in hep-th/0302225 that, contrary to all previous work, massive charged s=2 fields propagate causally is false.
arXiv subjects
Publications and source records attributed to S. Deser.
The recent claim in hep-th/0302225 that, contrary to all previous work, massive charged s=2 fields propagate causally is false.
We define and compute the energy of higher curvature gravity theories in arbitrary dimensions. Generically, these theories admit constant curvature vacua (even in the absence of an explicit cosmological constant), and asymptotically constant curvature solutions with non-trivial energy properties. For concreteness, we study quadratic curvature models in detail. Among them, the one whose action is the square of the traceless Ricci tensor always has zero energy, unlike conformal (Weyl) gravity. We also study the string-inspired Einstein-Gauss-Bonnet model and show that both its flat and Anti-de-Sitter vacua are stable.
I provide a very brief sketch of some of Dirac's interests and work in gravity, particularly his Hamiltonian formulation of Einstein's theory and its relation to his earlier research.
We analyze large gauge invariance in combined nonabelian and thermal QFT and their physical consequences for D=3 effective actions. After briefly reviewing the structure of bundles and large gauge transformations that arise in non-simply connected 3-manifolds and gauge groups, we discuss their connections to Chern-Simons terms and Wilson-Polyakov loops. We then provide an invariant characterization of the ``abelian'' fluxes encountered in explicit computations of finite temperature effective actions. In particular we relate, and provide explicit realizations of, these fluxes to a topological index that measures the obstruction to global diagonalization of the loops around compactified time. We also explore the fate of, and exhibit some everywhere smooth, large transformations for non-vanishing index in various topologies.
We define the notion of energy, and compute its values, for gravitational systems involving terms quadratic in curvature. While our construction parallels that of ordinary Einstein gravity, there are significant differences both conceptually and concretely. In particular, for D=4, all purely quadratic models admit vacua of arbitrary constant curvature. Their energies, including that of conformal (Weyl) gravity, necessarily vanish in asymptotically flat spaces. Instead, they are proportional to that of the Abbott-Deser (AD) energy expression in constant curvature backgrounds and therefore also proportional to the mass parameter in the corresponding Schwarzschild-(Anti) de Sitter geometries. Combined Einstein-quadratic curvature systems reflect the above results: Absent a cosmological constant term, the only vacuum is flat space, with the usual (ADM) energy and no explicit contributions from the quadratic parts. If there is a Lambda term, then the vacuum is also unique with that Lambda value, and the energy is just the sum of the separate contributions from Einstein and quadratic parts to the AD expression . Finally, we discuss the effects on energy definition of both higher curvature terms and higher dimension.
In three dimensions, there are two distinct mass-generating mechanisms for gauge fields: adding the usual Proca/Pauli-Fierz, or the more esoteric Chern-Simons (CS), terms. Here we analyze the three-term models where both types are present, and their various limits. Surprisingly, in the tensor case, these seemingly innocuous systems are physically unacceptable. If the sign of the Einstein term is ``wrong'' as is in fact required in the CS case, then the excitation masses are always complex; with the usual sign, there is a (known) region of the two mass parameters where reality is restored, but instead we show that a ghost problem arises, while, for the ``pure mass'' two-term system without an Einstein action, complex masses are unavoidable. This contrasts with the smooth behavior of the corresponding vector models. Separately, we show that the ``partial masslessness'' exhibited by (plain) massive spin-2 models in de Sitter backgrounds is formally shared by the three-term system: it also enjoys a reduced local gauge invariance when this mass parameter is tuned to the cosmological constant.
I review recent work on massive higher (s>1) spins in constant curvature (deSitter) spaces. Some of the novel properties that emerge are: partial masslessness and new local gauge invariances, unitarily forbidden ranges of mass, correlation between fermions/bosons and nagative/positive cosmological constant Lambda and finally the consistency requirement that in the limit of infinite spin towers, Lambda must tend to zero.
I review briefly, primarily for relativists, a series of recent results, obtained with A. Waldron, on the novel behavior of massive higher (s>1) spin systems in constant curvature backgrounds. We find that the cosmological constant Lambda, together with the mass parameter, define a "phase plane" in which partially massless gauge invariant lines separate allowed regions from forbidden, non-unitary, ones. These lines represent short multiplet systems, with missing lower helicities, removed by novel local gauge invariances, and (despite having nonvanishing m) propagating on the light cone. In the limit of an infinite tower of these higher spin bosons and fermions, unitarity requires Lambda to vanish.
Massive spin 2 theories in flat or cosmological ($Λ\ne 0$) backgrounds are subject to discontinuities as the masses tend to zero. We show and explain physically why their Newtonian limits do not inherit this behaviour. On the other hand, conventional ``Newtonian cosmology'', where $Λ$ is a constant source of the potential, displays discontinuities: e.g. for any finite range, $Λ$ can be totally removed.
This lecture consists of two parts. The first is a (totally unsystematic) survey of some of the high points in the evolution of gravity and its successors, primarily in the course of the past century. The second summarizes some new work on surprising properties of higher $(> 1)$ spin fields in cosmological backgrounds: the presence of $Ł$ gives rise to discrete sets of massive models endowed with gauge invariances, that divide the ($m^2, Ł$) plane into unitary and non-unitary phases. The unitary region common to fermions and bosons shrinks to flat space ($ Ł\to 0 $) as their spins increase.
We show explicitly that all partially and strictly massless fields with spins s<=3 in (A)dS have null propagation. Assuming that this property holds also for s>3, we derive the mass-cosmological constant tunings required to yield all higher spin partially massless theories. As s increases, the unitarily allowed region for massive spins is squeezed around Λ=0, so that an infinite tower of massive particles forces vanishing Λ. We also speculate on the relevance of this result to string theory and supergravity in (A)dS backgrounds.
We analyze the physics of massive spin 2 fields in (A)dS backgrounds and exhibit that: The theory is stable only for masses m^2 >= 2Λ/3, where the conserved energy associated with the background timelike Killing vector is positive, while the instability for m^2<2Λ/3 is traceable to the helicity 0 energy. The stable, unitary, partially massless theory at m^2=2Λ/3 describes 4 propagating degrees of freedom, corresponding to helicities (+/-2,+/-1) but contains no 0 helicity excitation.
Massive spin s>=3/2 fields can become partially massless in cosmological backgrounds. In the plane spanned by m^2 and Λ, there are lines where new gauge invariances permit intermediate sets of higher helicities, rather than the usual flat space extremes of all 2s+1 massive or just 2 massless helicities. These gauge lines divide the (m^2,Λ) plane into unitarily allowed or forbidden intermediate regions where all 2s+1 massive helicities propagate but lower helicity states can have negative norms. We derive these consequences for s=3/2,2 by studying both their canonical (anti)commutators and the transmutation of massive constraints to partially massless Bianchi identities. For s=2, a Hamiltonian analysis exhibits the absence of zero helicity modes in the partially massless sector. For s=5/2,3 we derive Bianchi identities and their accompanying gauge invariances for the various partially massless theories with propagating helicities (+/-5/2,+/-3/2) and (+/-3,+/-2), (+/-3,+/-2,+/-1), respectively. Of these, only the s=3 models are unitary. To these ends, we also provide the half integer generalization of the integer spin wave operators of Lichnerowicz. Partial masslessness applies to all higher spins in (A)dS as seen by their degree of freedom counts. Finally a derivation of massive d=4 constraints by dimensional reduction from their d=5 massless Bianchi identity ancestors is given.
We extend to its spin~3/2 supersymmetric partner the very recent demonstration that the massless limit of massive spin~2 exchange amplitudes can be made continuous in background AdS spaces, in contrast to the known flat space discontinuities for both systems. In an AdS background, unlike spin~2 where the limit m\to0 is the massless one, spin~3/2 ``masslessness'' requires m\to\sqrt{-Λ/3}, the supergravity value tuning the mass and cosmological constant that uniquely provides gauge invariance and two helicities. We find that continuity of the spin~3/2--mediated exchange amplitude can be regained in two ``massless'' limits m\to0 and m\to\sqrt{-Λ/3}; only the latter corresponds to cosmological supergravity.
The (m^2,Λ) plane of spin s>1 massive fields in (A)dS backgrounds is shown to consist of separate phases, divided by lines of novel ``partially massless'' gauge theories that successively remove helicities, starting from the lowest, 0 or +/-(1/2). The norms of the excluded states flip as the gauge lines are crossed and only the region containing the massive Minkowski theory is unitary. The partially massless gauge theories are unitary or not, depending on the ordering of the gauge lines. This ``level splitting'' of massless Minkowski gauge theories is specific to non-zero Λ.
We discuss some properties of the intrinsically nonlinear Clebsch decomposition of a vector field into three scalars in d=3. In particular, we note and account for the incompleteness of this parameterization when attempting to use it in variational principles involving Maxwell and Chern-Simons actions. Similarities with string decomposition of metrics and their actions are also pointed out.
We study the general non-minimally coupled charged massive spin 3/2 model both for its low energy phenomenological properties and for its unitarity, causality and degrees of freedom behaviour. When the model is viewed as an effective theory, its parameters (after ensuring the correct excitation count) are related to physical characteristics, such as the magnetic moment g factor, by means of low energy theorems. We also provide the corresponding higher spin generalisation. Separately, we consider both low and high energy unitarity, as well as the causality aspects of our models. None (including truncated N=2 supergravity) is free of the minimal model's acausality.
We compute the tree level 4-particle bosonic scattering amplitudes in D=11 supergravity. By construction, they are part of a linearized supersymmetry-, coordinate- and 3-form gauge-invariant. While this on-shell invariant is nonlocal, suitable SUSY-preserving differentiations turn it into a local one with correct dimension to provide a natural lowest (two-loop) order counterterm candidate. Its existence shows explicitly that no symmetries protect this ultimate supergravity from the nonrenormalizability of its lower-dimensional counterparts.