1: %\documentclass[prl,preprint,draft,amsmath,amssymb]{revtex4}
2: \documentclass[prl,twocolumn,amsmath,amssymb]{revtex4}
3: %\usepackage{graphicx}
4:
5: %\newcommand{\vectr}[1]{\mathbf{#1}}
6: %\newcommand{\dprod}[2]{\bar#1\cdot {#2}\,}
7:
8: \begin{document}
9:
10: \title{Very Special Relativity}
11:
12: \date{Jan 26, 2006}
13:
14: \author{Andrew G. Cohen}
15: \email{cohen@bu.edu}
16: \author{Sheldon L. Glashow}
17: \email{slg@bu.edu}
18: \affiliation{Physics Department, Boston University\\
19: Boston, MA 02215, USA}
20:
21: \begin{abstract}
22: By Very Special Relativity (VSR) we mean descriptions of nature
23: whose space-time symmetries are certain proper subgroups of the
24: Poincar\'e group. These subgroups contain space-time translations
25: together with at least a 2-parameter subgroup of the Lorentz group
26: isomorphic to that generated by $K_{x}+J_{y}$ and $K_{y}-J_{x}$. We
27: find that VSR implies special relativity (SR) in the context of
28: local quantum field theory or of $CP$ conservation. Absent both of
29: these added hypotheses, VSR provides a simulacrum of SR for which
30: most of the consequences of Lorentz invariance remain wholly or
31: essentially intact, and for which many sensitive searches for
32: departures from Lorentz invariance must fail. Several feasible
33: experiments are discussed for which Lorentz-violating effects in VSR
34: may be detectable.
35: \end{abstract}
36:
37: \maketitle
38:
39: % \section{Introduction}
40: % \label{sec:introduction}
41:
42:
43: Special relativity (SR) is based on the hypothesis that the laws of
44: physics share many of the symmetries of Maxwell's equations. Whereas
45: the maximal symmetry group of Maxwell's equations is the 15-parameter
46: conformal group $SU(2,4)$, the existence of particles with mass (and
47: the known violations of $P$ and $T$) constrains space-time symmetry to
48: be no greater than the Poincar\'e group (the connected component of
49: the Lorentz group along with space-time translations). The special
50: theory of relativity identifies this group as the symmetry of nature.
51:
52: Although no decisive departure from exact Lorentz invariance has yet
53: been detected, ever more sensitive searches should be and are being
54: carried out. A perturbative framework has been developed to
55: investigate a certain class of departures from Lorentz invariance. For
56: example, Coleman and Glashow\cite{Coleman:1998ti,Coleman:1997xq}
57: consider the case of space-time translations along with exact
58: rotational symmetry in the rest frame of the cosmic background
59: radiation, but allow small departures from boost invariance in this
60: frame. Perturbative departures from Lorentz-invariance are then
61: readily parametrized in terms of a fixed time-like 4-vector or
62: `spurion.' Others\cite{Colladay:1996iz,Colladay:1998fq} consider the
63: introduction into the Lagrangian of more general spurion-mediated
64: perturbations (sometimes referred to as `expectation values of Lorentz
65: tensors following spontaneous Lorentz breaking.').
66:
67: In this note we pursue a different approach to the possible failure of
68: Lorentz symmetry. We ask whether the exact symmetry group of nature
69: may be isomorphic to a proper subgroup of the Poincar\'e group. To
70: preserve energy and momentum conservation, we consider only those
71: subgroups that include space-time translations along with a proper
72: subgroup of the Lorentz group. Up to isomorphism and possible
73: discrete elements, there are four such distinct subgroups with
74: 1-parameter, three with 2-parameters, five with 3-parameters, and just
75: one with 4-parameters.
76:
77: For reasons soon to be explained, we restrict our attention to those
78: subgroups of the Lorentz group containing the generators $T_{1} \equiv
79: K_{x}+J_{y}$ and $T_{2} = K_{y} - J_{x}$, where $\mathbf{J}$ and
80: $\mathbf{K}$ are the generators of rotations and boosts
81: respectively. These commuting generators form a group, $T(2)$, which
82: is isomorphic to the group of translations in the plane. Three larger
83: subgroups of the Lorentz group are obtained by adjoining one or two
84: generators to $T(2)$. Each of them has a natural action on the plane:
85: the addition of $J_{z}$ yields a group isomorphic to the
86: three-parameter group of Euclidean motions, $E(2)$; the addition of
87: $K_{z}$ yields one isomorphic to the three-parameter group of
88: orientation-preserving similarity transformations, or homotheties,
89: $HOM(2)$; and lastly, the addition of both $J_{z}$ and $ K_{z}$ yields
90: one isomorphic to the four-parameter similitude group,
91: $SIM(2)$\footnote{There is a further one-parameter family of
92: three-dimensional groups in which we adjoin the generator $K_{z}+a
93: J_{z}$ for any real, non-zero $a$. We do not consider these here.}.
94:
95: We refer to any scheme whose space-time symmetries consist of
96: translations along with any one of the Lorentz subgroups described
97: above as Very Special Relativity (VSR). We shall see that the four VSR
98: avatars thus defined have quite different character. Nevertheless,
99: they all share the following remarkable defining property: that the
100: incorporation of either $P$, $T$ or $CP$ enlarges these subgroups to
101: the full Lorentz group. Conjugation by one of these discrete
102: transformations treats boosts and rotations oppositely, thereby
103: extending the group to allow boosts and rotations in the $x$--$y$
104: plane independently. Further commutation leads to the remaining
105: $z$-boost and $z$-rotation. The group $T(2)$ is the smallest subgroup
106: of the Lorentz group with this property, the only others being those
107: containing $T(2)$, hence our focus. It follows that Lorentz-violating
108: effects in VSR are absent for theories conserving any one of these
109: three discrete symmetries (and perhaps, that Lorentz-violating effects
110: in VSR are necessarily small because $CP$ violating effects are
111: small.)
112:
113: In previous approaches, the breaking of Lorentz symmetry was expressed
114: in terms of local operators incorporating one or more invariant
115: tensors, or spurions. In the case of $SO(3)$ symmetry, the spurion
116: takes the form of a time-like 4-vector, and the lowest dimension
117: operators involving it affect both particle propagation and the
118: kinematics of particle decays. The limits on such departures from SR
119: in this model are exceptionally strong. For example, the mere
120: observation of ultra-high energy cosmic rays places an upper bound of
121: $10^{-23}$ on one dimensionless measure of Lorentz
122: violation\cite{Coleman:1998ti}, while an analysis of neutrino data
123: bounds flavor-dependent Lorentz violation in the neutrino sector to
124: less than $10^{-25}$\cite{Battistoni:2005gy}.
125:
126: We may attempt to apply a similar spurion strategy to the four VSR
127: groups described above. The smallest group $T(2)$ admits many possible
128: invariant tensors. The simplest of these whose little group is no
129: greater than $T(2)$ is the antisymmetric two-index tensor\footnote{For
130: a finite-dimensional representation of the Lorentz group labeled by
131: two non-negative half-integers $(n,m)$, such an invariant tensor
132: corresponds to the state with weight $\vert{}n,-m\rangle{}$. To
133: avoid invariance under the larger group $E(2)$ $n$ must differ from
134: $m$.}
135: \begin{equation*}
136: F =
137: \begin{pmatrix}
138: 0& 1& 0& 0 \\
139: -1& 0& 0& -1 \\
140: 0& 0& 0& 0 \\
141: 0 &1& 0& 0
142: \end{pmatrix}.
143: \end{equation*}
144: ($F$ may be thought of as the field-strength for a zero frequency
145: electromagnetic wave with linear polarization in the $x$-direction.)
146:
147: The group $E(2)$ admits the 4-vector $n=(1,0,0,1)$ as an invariant
148: tensor. (This is also an invariant tensor for $T(2)$, but one which
149: preserves rotations about the $z$-axis as well.) The existence of
150: invariant tensors for the VSR groups $T(2)$ and $E(2)$ allows the
151: construction of Lorentz-violating local operators that, among other
152: things, affect the propagation of particles, much as in the $O(3)$
153: case. However unlike that case, these new operators necessarily
154: violate $P$, $CP$ and $T$.
155:
156: The remaining two VSR groups $HOM(2)$ and $SIM(2)$ are entirely
157: different in this regard. There are no invariant tensors for these
158: cases\footnote{For an irreducible representation labeled by two
159: half-integers $(n,m)$ the only state annihilated by $T_{1}$ and
160: $T_{2}$ is $\vert n,-m\rangle$. But this is an eigenstate of $K_{z}$
161: with eigenvalue $n+m$ which vanishes only for $n=m=0$.}. No local
162: Lorentz symmetry-breaking operator preserving either of these groups
163: exists and there is no obvious local, perturbative description of
164: their departures from SR. Consequently, spurions cannot access
165: scenarios in which the symmetry group of nature is $HOM(2)$ or
166: $SIM(2)$.
167:
168: The situation for these groups is much like that of $CPT$
169: in the context of Lorentz-invariant local quantum field theory: all
170: local operators preserving Lorentz invariance preserve a larger
171: symmetry (Lorentz plus $CPT$). Here, all local operators preserving
172: $SIM(2)$ (or $HOM(2)$) also preserve a larger symmetry
173: (Lorentz). Nevertheless it is easy to construct non-local amplitudes
174: that violate Lorentz invariance while respecting $SIM(2)$ (or
175: $HOM(2)$).
176:
177: One way to do this makes use of the non-invariant null vector $n\equiv
178: (1,0,0,1)$. This vector is invariant under $T_{1}, T_{2}$
179: transformations and $z$-axis rotations, but transforms as $n \to
180: e^{\phi}n$ under boosts in the $z$-direction. Consequently, ratios of
181: dot-products of this vector with kinematic variables (such as momenta)
182: are invariant under $SIM(2)$ or $HOM(2)$ but not under all Lorentz
183: transformations. For example, the amplitude for the two body decay of
184: a spinless particle at rest can depend on the 4-momenta of the decay
185: products, $p_{1} \text{ and } p_{2}$. The ratio $(p_{1}\cdot
186: n)/(p_{2}\cdot n)$ is then an invariant, and thus the amplitude for
187: the decay may depend on the direction of the decay products relative
188: to the VSR-preferred direction (nominally, the $z$ axis).
189:
190: Because VSR includes space-time translations, particle states may be
191: labeled by their 4-momenta. For $SIM(2)$ and $HOM(2)$, the only
192: invariant that can be constructed from the 4-momentum of a massive
193: particle is the mass itself, just as in SR. Therefore, all positive
194: energy time-like momenta of fixed length are equivalent under $SIM(2)$
195: or $HOM(2)$ transformations. (A given time-like momentum may be
196: transformed to the rest frame by three successive transformations: a
197: $T_{1}$ transformation with parameter $-p_{x}/(E-p_{z})$; a $T_{2}$
198: transformation with parameter $-p_{y}/(E-p_{z})$; and a boost in the
199: $z$-direction with parameter $e^\phi = (E-p_{z})/M$.) This result
200: implies that many of the elementary consequences of SR, such as
201: time-dilation, the law of velocity addition, the existence of a
202: center-of-mass frame, and a universal and isotropic maximal attainable
203: velocity hold in these variants of VSR. Indeed, invariance under
204: $HOM(2)$, rather than (as is often taught) the Lorentz group, is both
205: necessary and sufficient to ensure that the speed of light is the same
206: for all observers, and \emph{inter alia}, to explain the null result
207: to the Michelson-Morley experiment and its more sensitive successors.
208:
209: Nature seems well described by the Standard Model, a Lorentz-invariant
210: local quantum field theory in which the existence of three fermion
211: families is necessary for $CP$ violation. As a result, $CP$ violating
212: effects are usually small and are nearly absent in all flavor-diagonal
213: processes. In this context, we note that the failure to detect the
214: neutron electric dipole moment shows that $\bar{\theta} < 10^{-10}$,
215: while arguments suggest that $\bar{\theta}$ may be considerably
216: smaller.
217:
218: Were $CP$ an exact symmetry, VSR would imply SR. Consequently,
219: because $CP$ violating effects in nature are small, we expect VSR
220: departures from SR to be correspondingly small. However, such
221: departures may be a dominant effect in processes for which $CP$
222: violation is significant. For example, in the decay $K_L\rightarrow
223: \pi^++\pi^-$, the pions need not be isotropically distributed in the
224: kaon rest frame. Their directions could be correlated to the
225: VSR-preferred direction. This effect is likely to be tiny for
226: the spurion-inaccessible variants of VSR, because $CP$ violation in
227: kaon decay is predominantly indirect (propagator dominated) and as we
228: have noted, particle propagation is unaffected for $SIM(2)$ or
229: $HOM(2)$.
230:
231:
232: For the spurion-inaccessible variants of VSR, observable departures
233: from SR might be found in studies of the $CP$-violating decays of
234: neutral $B$ mesons (such as $B_0\rightarrow J/\Psi+ K_s$), where $CP$
235: violation is largely direct and significant angular correlations may
236: be present. A straightforward search for departures from SR can be
237: performed without knowing the times of individual events, so long as
238: the VSR preferred direction $\vec n$ is not coincident with the
239: Earth's polar axis. In that case, we anticipate an angular
240: distribution (in the $B$ rest frame) depending on the angle between a
241: decay product and the polar axis.
242:
243:
244: VSR may have radical consequences for neutrino physics. Neutrinos are
245: now known to have mass. Several mechanisms have been contrived to
246: remedy the absence of neutrino mass in the Standard Model. All of
247: these invoke new particles or new interactions. In the `Dirac'
248: picture, lepton number is conserved with neutrinos acquiring mass via
249: (anomalously small) Yukawa couplings to sterile $SU(2)$-singlet
250: neutrinos. In the `Majorana' picture, lepton number is
251: violated. Neutrino masses result from a seesaw mechanism involving
252: heavy sterile states, or via dimension-6 operators resulting from
253: unspecified new interactions.
254:
255: In VSR, neutrino mass has a natural origin. Lepton-number conserving
256: neutrino masses, although not Lorentz invariant, are VSR invariant.
257: There is no guarantee that neutrino masses have a VSR origin,
258: but if so their sizes may be an indication of the magnitude of
259: Lorentz-violating effects in other sectors. For example VSR allows for
260: an (anisotropic!) electric dipole moment for charged leptons. $SU(2)$
261: invariance may then relate such dipole moments to neutrino masses:
262: $d_{\text{lepton}} \sim (m_{\nu}/2m_{l})^{2}(e/2m_{l})$. For the
263: electron and for $m_{\nu}^{2}\simeq 10^{-4}\text{ eV}^{2}$ this is the
264: same size as the current experimental
265: sensitivity\cite{Regan:2002ta}. We leave detailed explication of these
266: and related matters to a subsequent publication.
267:
268: A VSR origin of neutrino masses requires no additional states and need
269: not introduce lepton number violation\footnote{The massless and
270: massive unitary representations of $SIM(2)$ and $HOM(2)$ are
271: one-dimensional, unlike those of the Poincar\'e group.}. This is a
272: significant departure from conventional notions. However, because all
273: observable neutrino phenomena involve ultra-relativistic neutrinos
274: ($\gamma >> 1$), neutrino phenomenology is virtually identical to that
275: of the usual scenarios: the neutrino helicity will differ
276: significantly (but unobservably) from $-1/2$, but only in a narrow
277: cone about the preferred axis with opening angle $\sim 1/\gamma$; and
278: neutrinoless double beta decay is forbidden (and therefore also
279: unobservable) by lepton number conservation.
280:
281: Previous authors\cite{Coleman:1998ti,Coleman:1997xq,Colladay:1998fq,%
282: Colladay:1996iz} have noted that spurion-mediated Lorentz violation
283: can lead to two varieties of potentially observable Lorentz-violating
284: effects: $CPT$ conserving or $CPT$ violating. The same is true for VSR
285: in its spurion-accessible variants, $T(2)$ and $E(2)$. However, this
286: is not necessarily the case for VSR in its $SIM(2)$ avatar. The $CPT$
287: operation is equivalent to a \emph{complex} VSR transformation: a
288: rotation about the $z$-axis by $\pi$ along with an imaginary boost by
289: the same amount in the $z$-direction. Thus for amplitudes satisfying
290: appropriate analyticity properties, $CPT$ follows from $SIM(2)$. (This
291: argument is similar to the canonical proof of $CPT$ invariance in
292: Lorentz invariant theories). While complex $HOM(2)$ can reverse the
293: sign of any given 4-vector, $CPT$ invariance is not implied because
294: the necessary transformation is momentum-dependent.
295:
296: Our paper initiates an exploration of the possibility that the many
297: empirical successes of special relativity need not demand Lorentz
298: invariance of the underlying theoretical framework. Could the Lorentz
299: invariant Standard Model emerge as an effective theory from a more
300: fundamental scheme, perhaps operative at the Planck scale, that is VSR
301: (but not SR) invariant? Such a scheme, as we have noted, cannot be a
302: precisely local quantum field theory and its effects, especially for
303: the case of the spurion-inaccessible variants of VSR, are difficult to
304: estimate.
305:
306: \begin{acknowledgments}
307: AGC was supported in part by the Department of Energy under grant
308: no. DE-FG02-01ER-40676; SLG by the National Science Foundation under
309: grant no. NSF-PHY 0099529.
310: \end{acknowledgments}
311:
312: \bibliography{vsr}
313:
314:
315: \end{document}
316:
317:
318:
319:
320: %%% Local Variables:
321: %%% mode: latex
322: %%% TeX-master: t
323: %%% End:
324:
325: