1: %% version of sprocl.tex modified for CPT'01 proceedings, August 12, 2001
2: %%
3: %%UNIX --- UPDATED ON 13/8/97
4: %====================================================================%
5: % sprocl.tex 27-Feb-1995 %
6: % This latex file rewritten from various sources for use in the %
7: % preparation of the standard proceedings Volume, latest version %
8: % by Susan Hezlet with acknowledgments to Lukas Nellen. %
9: % Some changes are due to David Cassel. %
10: %====================================================================%
11:
12: \documentstyle[sprocl]{article}
13: %\setcounter{page}{190}
14:
15: \font\eightrm=cmr8
16:
17: %\input{psfig}
18:
19: \bibliographystyle{unsrt} %for BibTeX - sorted numerical labels by
20: %order of first citation.
21:
22: \arraycolsep1.5pt
23:
24: % A useful Journal macro
25: \def\Journal#1#2#3#4{{#1} {\bf #2}, #3 (#4)}
26:
27: % Some useful journal names
28: \def\NPB{{\it Nucl. Phys.} B}
29: \def\PLB{{\it Phys. Lett.} B}
30: \def\PRL{\it Phys. Rev. Lett.}
31: \def\PRD{{\it Phys. Rev.} D}
32: \def\ZPC{{\it Z. Phys.} C}
33: \def\fr#1#2{{{#1} \over {#2}}}
34: \def\gsim{\mathrel{\rlap{\lower4pt\hbox{\hskip1pt$\sim$}}
35: \raise1pt\hbox{$>$}}}
36: % Some other macros used in the sample text
37: \def\st{\scriptstyle}
38: \def\sst{\scriptscriptstyle}
39: \def\mco{\multicolumn}
40: \def\epp{\epsilon^{\prime}}
41: \def\vep{\varepsilon}
42: \def\ra{\rightarrow}
43: \def\ppg{\pi^+\pi^-\gamma}
44: \def\vp{{\bf p}}
45: \def\ko{K^0}
46: \def\kb{\bar{K^0}}
47: \def\al{\alpha}
48: \def\ab{\bar{\alpha}}
49: \def\be{\begin{equation}}
50: \def\ee{\end{equation}}
51: \def\bea{\begin{eqnarray}}
52: \def\eea{\end{eqnarray}}
53: \def\CPbar{\hbox{{\rm CP}\hskip-1.80em{/}}}%temp replacemt due to no font
54:
55: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
56: %%BEGINNING OF TEXT
57: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
58:
59: \begin{document}
60:
61: \title{SOME CONSIDERATIONS REGARDING LORENTZ-VIOLATING THEORIES}
62:
63: \author{RALF LEHNERT}
64:
65: \address{Physics Department, Indiana University\\
66: Bloomington, IN 47405, U.S.A.
67: \\E-mail: rlehnert@indiana.edu}
68:
69: \maketitle\abstracts{
70: We investigate the compatibility
71: of Lorentz-violating quantum field theories
72: with the requirements
73: of causality and stability.
74: A general renormalizable model
75: for free massive fermions
76: indicates that these requirements are satisfied
77: at low energies provided
78: the couplings controlling the breaking are small.
79: However, for high energies
80: either microcausality or energy positivity
81: or both are violated in some observer frame.
82: We find evidence that this difficulty can be avoided
83: if the model is interpreted
84: as a sub-Planckian approximation
85: originating from a nonlocal theory
86: with spontaneous Lorentz violation.
87: The present study thereby supports the validity
88: of the standard-model extension
89: as the low-energy limit
90: of any realistic string theory
91: that exhibits spontaneous Lorentz breaking.}
92:
93: \section{Introduction}
94:
95: \noindent
96: From a theoretical point of view,
97: the minimal SU(3)$\times$SU(2)$\times$U(1) standard model
98: leaves unresolved a variety of issues.
99: It is therefore believed to be the low-energy limit
100: of an underlying framework
101: that also includes a quantum description of gravity.
102: On the other hand,
103: the standard model is phenomenologically successful,
104: so observable effects
105: from the presumed underlying physics
106: must be minuscule.
107: It then becomes an interesting challenge
108: to identify possible experimental signals
109: from such a fundamental theory
110: accessible with present techniques.
111:
112: A candidate signal of this type
113: is the violation CPT and Lorentz invariance:
114: In conventional renormalizable
115: gauge theories including the standard model,
116: these two symmetries are linked
117: by CPT theorem\cite{pauli} and hold exactly.
118: In contrast,
119: attempts to construct an underlying framework
120: often involve ingredients
121: that bypass the CPT theorem.
122: For example,
123: string (M) theory is known to admit
124: spontaneous Lorentz and CPT violation.\cite{str}
125: Other frameworks can also lead to similar
126: low-energy effects.\cite{klink,ch}
127:
128: For the microscopic description
129: of possible observable signals at presently accessible energy scales,
130: an extension of the minimal standard model of particle physics
131: has been developed.\cite{ck}
132: This standard-model extension
133: has provided the basis for numerous experimental
134: investigations discussed during this meeting and elsewhere,
135: which constrain CPT and Lorentz violation.\cite{exp}
136:
137: In this talk, we study the fundamental properties
138: of causality and stability
139: in the context of the Lorentz-violating standard-model extension.
140: These two properties appear essential for realistic theories,
141: for it would be difficult to make meaningful experimental predictions
142: without either causality or stability.
143: In particular,
144: it is of interest whether these two requirements
145: constrain the parameter space and the range of validity
146: of the standard-model extension,
147: and whether insight into the underlying theory can be gained.
148: Although the calculations presented here are carried out
149: for free massive fermions,
150: we expect that most of our results can be straightforwardly generalized
151: to the other sectors of the standard-model extension.
152:
153:
154: \section{Framework}
155:
156: \noindent
157: The general Lorentz-violating Lagrangian
158: for a single spin-$\frac{1}{2}$ fermion
159: \cite{ck}
160: can be cast into a variety of forms.
161: One such form
162: reminiscent of the ordinary Dirac Lagrangian
163: and emphasizing the derivative structure
164: is \cite{kl99}
165: \begin{equation}
166: {\cal L} = \frac{1}{2}{\it i}\overline{\psi}
167: {\Gamma}^{\nu}\hspace{-.15cm}
168: \stackrel{\;\leftrightarrow}
169: {\partial}_{\!\nu}\hspace{-.1cm}{\psi}
170: -\overline{\psi}M{\psi}
171: \label{lagr}
172: \quad ,
173: \end{equation}
174: where
175:
176: \begin{equation}
177: {\Gamma}^{\nu}:={\gamma}^{\nu}+c^{\mu \nu}
178: {\gamma}_{\mu}+d^{\mu \nu}{\gamma}_{5}
179: {\gamma}_{\mu}+e^{\nu}+if^{\nu}{\gamma}_{5}
180: +\frac{1}{2}g^{\lambda \mu \nu}
181: {\sigma}_{\lambda \mu}
182: \label{Gam}
183: \quad ,
184: \end{equation}
185: and
186:
187: \begin{equation}
188: M:=m+a_{\mu}{\gamma}^{\mu}+b_{\mu}{\gamma}_{5}
189: {\gamma}^{\mu}+\frac{1}{2}H^{\mu \nu}
190: {\sigma}_{\mu \nu}
191: \label{M}
192: \quad .
193: \end{equation}
194: The gamma matrices $\{1, \gamma_5,\gamma^{\mu},
195: \gamma_5\gamma^{\mu}, \sigma^{\mu \nu}\}$
196: have conventional properties,
197: and the signature of the Minkowski metric
198: $\eta_{\mu\nu}$ is $-2$.
199: The extent of Lorentz violation
200: is described by the parameters
201: $a_{\mu}$, $b_{\mu}$, $c_{\mu \nu}$,
202: $d_{\mu \nu}$, $e_{\mu}$, $f_{\mu}$,
203: $g_{\mu \nu \lambda}$ and $H_{\mu \nu}$.
204: As a consequence
205: of the presumed hermiticity
206: of the Lagrangian,
207: all these coefficients
208: are real,
209: with
210: $c_{\mu \nu}$ and $d_{\mu \nu}$ traceless,
211: $g_{\mu \nu \lambda}$ antisymmetric
212: in its first two indices
213: and $H_{\mu \nu}$ antisymmetric.
214: While all the parameters
215: violate Lorentz invariance,
216: only
217: $a_{\mu}$, $b_{\mu}$, $e_{\mu}$, $f_{\mu}$
218: and $g_{\mu \nu \lambda}$
219: break CPT symmetry as well.
220:
221: Since no departures from Lorentz symmetry
222: have been observed to date,
223: all Lorentz-breaking parameters
224: must be minuscule in a certain class
225: of observer inertial frames
226: called {\it concordant frames},
227: and the Earth must move nonrelativistically
228: with respect to these frames.
229: Throughout this talk,
230: we shall work under the assumption
231: that the size of the Lorentz violation
232: is such that,
233: in a concordant frame,
234: a hermitian hamiltonian can be found
235: and the dispersion relation still exhibits
236: two positive- and two negative-valued roots,\cite{kl01,kljmp}
237: paralleling the conventional Dirac case.
238: The lagrangian can then be canonically quantized
239: such that the energy is positive definite.\cite{kl01}
240:
241:
242:
243:
244: \section{Microcausality and Stability}
245:
246: \noindent
247: A quantum field theory is microcausal
248: if any two local observables
249: with spacelike separation
250: can be measured independently.
251: This is guaranteed
252: if any two local,
253: spatially separated operators
254: commute.
255: In the present case,
256: such local operators
257: are fermion bilinears
258: and the above condition
259: is satisfied if
260: \begin{equation}
261: iS(x-x^{\prime})=\{\psi(x),
262: \overline{\psi}(x^{\prime})\}=0
263: \quad,
264: \qquad(x-x^{\prime})^2<0
265: \label{anticom}
266: \end{equation}
267: holds.
268: Note that the anticommutator
269: function $S(x-x^{\prime})$ only depends
270: on the coordinate differences
271: due to translational invariance.
272:
273: To determine
274: an integral representation
275: for the $S(z)$,
276: we insert the plane-wave expansion
277: of the field operators\cite{kl01}
278: into the anticommutator and use the generalization
279: of the conventional spinor projectors.\cite{kljmp}
280: This gives the following expression:
281: \begin{equation}
282: S(z)=\int_{C}\fr{d^4\lambda}{(2\pi)^4}
283: e^{-i\lambda\cdot z}
284: \fr{{\rm cof}(\Gamma_{\mu}\lambda^{\mu}-M)}
285: {\det(\Gamma_{\mu}\lambda^{\mu}-M)}
286: \label{pwexp2}
287: \quad,
288: \end{equation}
289: where $C$ is the usual contour
290: encircling all poles in clockwise direction,
291: and ${\rm cof}(\cdot)$ denotes the matrix of cofactors.
292: Notice that $\lambda^{\mu}$
293: can be replaced by $i\partial^{\mu}$ in the numerator of the integrand.
294: It is then possible to pull the cofactor matrix outside the integral,
295: because the contour $C$ can be deformed
296: such that the integrand is analytic in a neighborhood of $C$.\cite{sg}
297: We obtain for the anticommutator function
298: \begin{equation}
299: S(z)=
300: {\rm cof}(\Gamma^{\mu} i\partial_{\mu}-M)
301: \int_{C} \fr{d^4\lambda} {(2\pi)^4}
302: \fr{e^{-i\lambda\cdot z}} {\det(\Gamma^{\mu}\lambda_{\mu}-M)}
303: \quad.
304: \label{ffgreen}
305: \end{equation}
306:
307: Next, we study $S(z)$ outside the lightcone.
308: We can take advantage
309: of observer Lorentz invariance
310: and boost to a frame such that
311: $z^{\mu}=(0,\vec{z})$.
312: To make further progress,
313: it is necessary to investigate the pole structure
314: of the integrand.
315: Due to the above observer transformation
316: we may no longer assume to be working in a concordant frame.
317: In particular,
318: it may not be possible to find a hermitian hamiltonian,
319: so that complex eigenenergies may occur.
320: Since the eigenenergies determine the location of the poles
321: of the integrand, the contour $C$ may fail to encircle them all.
322: Thus, the case where a hermitian hamiltonian
323: (and therefore real eigenenergies)
324: exist in all frames
325: has to be distinguished.
326: We consider this case first.
327:
328: A sufficient condition for the hermiticity of the hamiltonian
329: in all observer frames is that the derivative structure
330: of lagrangian (\ref{lagr}) is the conventional one,
331: {\it i.e.,} $\Gamma^{\mu}=\gamma^{\mu}$.
332: Then, all four roots
333: $E_{(j)}(\vec{\lambda})$, $j=1,\ldots,4$,
334: of the dispersion relation appearing in the denominator
335: of the integrand in Eq.\ (\ref{ffgreen})
336: are real.
337: In this case,
338: the contour integration can be directly performed.\cite{kl01}
339: This argument confirms microcausality
340: for the case $\Gamma^{\mu}=\gamma^{\mu}$.
341:
342: In cases
343: when there exist observer frames
344: that fail to admit the definition
345: of a hermitian hamiltonian,
346: the above line of reasoning cannot be employed,
347: and microcausality may break down.
348: For example,
349: consider a model with $c_{00}$ parameter only.
350: The anticommutator function for this model
351: is given explicitly by
352: \begin{equation}
353: S(z)=(i\zeta\gamma^0\partial^0
354: -i\gamma^j\partial^j+m)
355: \fr{1}{4\pi\zeta r}
356: \fr{\partial}{\partial r}[\Theta(w^2) J_0(m\sqrt{w^2})]
357: \quad,
358: \label{prop}
359: \end{equation}
360: where
361: $\zeta=1+c_{00}$,
362: $r=|\vec{z}|$,
363: $w^2=(z^0/\zeta)^2-\vec{z}^2$,
364: $\Theta$ denotes the Heaviside step function
365: and $J_0(y)$ is the zeroth-order Bessel function.
366: It follows
367: that the anticommutator function $S(z)$
368: vanishes only in the region defined by
369: $z^0<(1+c_{00})|\vec{z}|$.
370: The propagation of signals therefore could occur with maximal speed
371: $1/(1+c_{00})$.
372: For negative values of $c_{00}$,
373: this exceeds 1 and hence violates microcausality.
374:
375: The question arises,
376: at which energy scale
377: this breakdown of microcausility occurs.
378: To this end,
379: it is useful to introduce a definition
380: of the velocity of a particle valid for arbitrary 3-momenta.
381: Even in the conventional Lorentz- and CPT-symmetric case,
382: the notion of a quantum velocity operator is nontrivial.
383: The issue is further complicated in the present context.\cite{ck}
384: Here, we consider the group velocity
385: defined for a monochromatic wave
386: in terms of the dispersion relation.
387: This choice seems appropriate for a variety of reasons.\cite{kl01}
388: Insight about the scale $\tilde{M}$ of microcausality violations
389: can then be gained by determining the value of the 3-momentum
390: at which the the group velocity reaches 1.
391: Analyses for a variety of parameter combinations
392: yield
393: \begin{equation}
394: \tilde{M}\gsim {\cal O}(M_P)
395: \quad .
396: \label{mscale1}
397: \end{equation}
398: Here, we have assumed that the parameters
399: $c_{\mu\nu}$, $d_{\mu\nu}$, $e_{\mu}$,
400: $f_{\mu}$ and $g_{\mu\nu\lambda}$
401: are of order $m/M_P$,
402: where $M_P$ denotes the Planck scale.
403:
404: We mention in passing
405: that the conclusion of microcausality breakdown
406: at ${\cal O}(M_P)$ may be invalid
407: if the $c_{\mu\nu}$ coefficient is nonzero.
408: For example,
409: in the above model with only a coupling $c_{00}<0$,
410: one can show that $\tilde{M}\gsim{\cal O}(\sqrt{mM_P})$.
411: The effect of the $c_{\mu\nu}$ parameter on the dispersion relation
412: is special for the following reason:
413: The general spinorial and derivative structure
414: of the associated quadratic field term
415: is identical to the conventional Dirac kinetic term.
416: Thus, it is a first-order correction
417: to an existing zeroth-order term.
418: None of the other Lorentz-violating couplings
419: exhibits this feature.
420:
421: The above analysis shows
422: that the standard-model extension
423: can develop problems when the symmetry-breaking scale
424: is approached.
425: This should not come as a surprise
426: because the effects of the presumed underlying theory
427: are likely to be no longer negligible at these energies.
428: However,
429: given the impracticality
430: of achieving Planck-scale momenta
431: in the laboratory,
432: the issue of microcausality breakdown
433: is largely unimportant at the level of the standard-model extension.
434:
435: Another important ingredient
436: for realistic field theories is the requirement of stability.
437: A field theory is stable if the energy is positive definite
438: {\it in all observer frames}.
439: This implies that the 4-momenta of all one-particle states
440: in a particular frame must be timelike or lightlike
441: with nonnegative 0th component.
442: Only under this last condition,
443: does energy positivity become an observer-invariant notion.
444: This is satisfied in the conventional Dirac case.
445:
446: In the present context,
447: the energy is positive definite in concordant frames.
448: In these frames,
449: the dispersion relation
450: has still two positive- and two negative-valued roots,
451: which yield positive particle energies
452: after the usual reinterpreta-tion.\cite{kl01}
453: However,
454: contrary to the conventional case,
455: these energies are in some instances
456: 0th components of spacelike 4-momenta.
457: As a result,
458: energy positivity becomes observer-dependent.
459:
460: As an example,
461: consider a model that has all Lorentz-violating
462: parameters except $b_{\mu}$ set to zero.
463: The dispersion relation for this model is given by
464: \begin{equation}
465: (\lambda^2-b^2-m^2)^2+4b^2\lambda^2 -4(b\cdot\lambda)^2=0
466: \quad .
467: \label{bdisp}
468: \end{equation}
469: It is straightforward to show
470: that observer frames in which
471: $b_{\mu}=(b_0,0,0,b_3)$ and ${b_3}^2>m^2+|b^{\mu}b_{\mu}|$
472: can always be chosen.
473: In such a frame,
474: the spacelike 4-vectors
475: ${\lambda^{\mu}}_{\pm}=(0,0,0,p_{\pm})$
476: satisfy the dispersion relation (\ref{bdisp}).
477: Here, the real quantities $p_{\pm}$
478: are defined by
479: \begin{equation}
480: {p_{\pm}}^2=(2{b_3}^2+b^2-m^2)
481: \pm\sqrt{(2{b_3}^2+b^2-m^2)^2-(m^2+b^2)^2}
482: \quad .
483: \label{moment}
484: \end{equation}
485: Moreover, the existence of these spacelike solution
486: remains unaffected,
487: when a nonzero $a_{\mu}$ coefficient is included.
488:
489: The instabilities resulting from these spacelike
490: 4-momenta are most transparent
491: for sufficiently boosted observers:
492: It is always possible to convert a spacelike vector
493: with a positive 0th component to one with a negative 0th component
494: by an appropriate observer Lorentz transformation.
495: In the present case,
496: this means that there exist otherwise acceptable observer frames
497: in which a single root of the dispersion relation
498: involves both positive and negative energies
499: for varying 3-momenta.
500: In such observer frames,
501: the canonical quantization procedure fails.
502:
503: In concordant frames, the energy is positive definite.
504: However, the physics is independent of the observer,
505: so the appearance of negative energies in a boosted frame
506: must also lead to instabilities in the concordant frames.
507: The above discussions implies that these instabilities
508: can only be associated with the spacelike momenta
509: satisfying the dispersion relation.
510: To illustrate this,
511: let us introduce a U(1) gauge interaction for the moment
512: because in the free fermion model
513: the particle number is conserved.
514: As an example,
515: consider the following process in a concordant frame:
516: A high-energy fermion emits a virtual photon,
517: which then decays into a fermion-antifermion pair.
518: We can write this as
519: \begin{equation}
520: f_{+1}\longrightarrow f_{+1} +f_{+1}+\bar{f}_{-1}
521: \quad ,
522: \label{decay}
523: \end{equation}
524: where $f$ and $\bar{f}$ denote fermions and antifermions,
525: respectively,
526: and the subscript labels the helicity state.
527: In ordinary QED,
528: such a process is kinematically forbidden
529: even though both the U(1) charge and angular momentum are conserved.
530: However it can occur in the present context
531: if the incoming fermion
532: has an appropriate spacelike 4-momentum.\cite{kl01}
533: Thus, there exist unstable single-particle states.
534:
535: The scale $\tilde{M}$ of the 3-momentum
536: at which spacelike 4-momenta occur
537: can be calculated explicitly
538: for various parameter combination.
539: We find that
540: \begin{equation}
541: \tilde{M} \gsim {\cal O}(M_P)
542: \quad ,
543: \label{bscale}
544: \end{equation}
545: where we have assumed that the derivative-coupling coefficients
546: have the same suppression as in the microcausality case,
547: and the remaining parameters $a_{\mu}$, $b_{\mu}$ and $H_{\mu\nu}$
548: are of order $m^2/M_P$.
549: This estimate shows that the instabilities
550: appear only for Planck-scale 4-momenta in a concordant frame.
551: The corresponding negative energies
552: occur only for Planck-boosted observers
553: relative to this frame.
554: Since the Earth moves nonrelativistically with
555: respect to concordant frames,
556: the model maintains stability
557: for all experimentally attainable physical momenta
558: and in all experimentally attainable observer frames.
559:
560: As for microcausality,
561: the presence of a $c_{\mu\nu}$ parameter
562: can invalidate (\ref{bscale}).
563: For example,
564: a model with a positive $c_{00}$ coefficient only,
565: exhibits spacelike momenta at a scale of order
566: $\sqrt{mM_P}$.
567: It follows that when microcausality and stability
568: are imposed on a model with a $c_{\mu\nu}$ coupling,
569: effects from the presumed underlying theory
570: are likely to become non-negligible
571: already at energies close to the geometric mean of $M_P$ and $m$.
572: As this scale is within reach of some experiments,
573: a theoretical analysis of such effects
574: may require high-energy corrections to the standard-model extension.
575: In the next section,
576: we discuss a possible type of such corrections.
577:
578: \section{High-Energy Effects}
579:
580: \noindent
581: The results from the previous section indicate
582: that quantum field theories of massive fermions
583: containing terms explicitly breaking Lorentz invariance
584: can develop difficulties with mircocausality or stability.
585: However, in concordant frames,
586: these difficulties primarily appear
587: as the Planck scale is approached.
588: The question arises,
589: whether there exist combinations of Lorentz-violating coefficients
590: that maintain both causality and stability.
591: Many parameter combinations
592: eliminate one of the two difficulties.
593: However,
594: we are unaware of any set of values of the couplings
595: $a_{\mu}$, $b_\mu$, $\ldots$, $H_{\mu\nu}$
596: that simultaneously guarantee microcausality and stability
597: at all energy scales.
598: Moreover,
599: it has been shown rigorously
600: that in {\it conventional} quantum field theory,
601: such a set of parameters would have to yield
602: the ordinary Lorentz-symmetric dispersion relation.\cite{be}
603: It is then likely that these Lorentz-breaking parameters
604: can be absorbed into a field redefinition
605: and remain unobservable.
606:
607: The Lorentz-violating standard-model extension
608: was developed following a top-down approach.
609: The original motivation was the possibility
610: of spontaneous Lorentz-symmetry breakdown
611: in an underlying framework such as strings.\cite{str}
612: Indeed, the standard-model extension includes
613: all Lorentz-violating, but observer-invariant, terms
614: compatible with renormalizability and the usual gauge structure.
615: It is thus the low-energy limit
616: of any potential spontaneous Lorentz breaking
617: in a more fundamental theory.
618: It is therefore not surprising
619: that difficulties develop as the Planck-scale is approached.
620: One would expect higher-order nonrenormalizable operators
621: to gain importance.
622: On the other hand,
623: the essential status of the requirements of causality and stability
624: suggests to adopt the inverse line of reasoning.
625: Such a bottom-up approach could provide
626: valuable insights into the nature of the underlying theory
627: at the Planck scale.
628:
629: The standard-model extension breaks Lorentz invariance explicitly.
630: However, a desirable feature of the fundamental theory
631: is spontaneous symmetry breaking.
632: One immediate advantage of this mechanism is that the dynamics
633: remains Lorentz covariant.
634: Therefore it does not come as a surprise
635: that such an underlying framework avoids at least some of
636: the difficulties plaguing more general models
637: involving Lorentz and CPT violation.
638: For instance,
639: one consequence of the spontaneous character
640: of the Lorentz violation is
641: that observer invariance is naturally maintained.
642: In the previous section,
643: this property has proved to be an important advantage.
644: In contrast,
645: if observer Lorentz invariance is imposed
646: in a theory with explicit Lorentz breaking,
647: an additional \it ad hoc \rm choice is required.
648:
649: Another effect of spontaneous Lorentz violation is
650: that the parameters
651: $a_{\mu}$, $b_\mu$, $\ldots$, $H_{\mu\nu}$
652: are only fixed at low energies.
653: As the Planck scale is approached,
654: they must be associated with dynamical fields.
655: A natural question is,
656: whether these fluctuations alone can simultaneously maintain
657: microcausality and stability.
658: This issue has been previously been discussed
659: in the context of a toy model.
660: It was shown that a satisfactory resolution
661: within the context of ordinary point-particle field theory
662: seems unlikely.\cite{kl01}
663: This is consistent with other ideas.\cite{be}
664: As expected,
665: ingredients beyond conventional quantum field theory
666: appear necessary.
667:
668: A class of theories with free-field terms
669: maintaining causality and stability
670: must contain terms beyond the ones in Eq.\ (\ref{lagr}).
671: The new terms have to be nonrenormalizable,
672: and in a realistic scenario with spontaneous Lorentz violation
673: they would correspond to higher-order nonrenormalizable operators
674: correcting the standard-model extension at energies
675: determined by the Planck scale.
676:
677: The first step is to investigate
678: whether any type of dispersion relation
679: can satisfy all the requirements for consistency.
680: In a concordant frame,
681: such a dispersion relation
682: would reproduce the physics of Eq.\ (\ref{lagr})
683: for small 3-momenta
684: but would avoid group velocities exceeding 1 and spacelike 4-momenta
685: for large 3-momenta.
686: These requirements could be implemented
687: by combining the Lorentz- and CPT-breaking parameters
688: with a suitable factor suppressing them only at large 3-momenta.
689: This factor must be essentially constant at small 3-momenta
690: and must overwhelm polynomial powers at large 3-momenta.
691: Since the the size of 3-momenta
692: is frame dependent,
693: it is to be expected that a suitable factor
694: would also be frame-dependent
695: and hence involve Lorentz- and CPT-violating coefficients.
696:
697: To make further progress,
698: it is useful to consider explicit examples.
699: To simplify the discussion,
700: the masses and the Lorentz-violating parameters
701: are taken to be of order 1
702: in appropriate units.
703: This makes it possible to focus on resolving
704: the problems of stability and causality
705: at Planck-scale energies in a concordant frame
706: without the complications introduced by the hierarchy of scales.
707:
708: Consider a model with a negative $c_{00}$ parameter only.
709: As discussed in the previous section,
710: this model violates microcausality at high energies.
711: The replacement
712: $c_{00}\rightarrow c_{00}\exp (c_{00}{\lambda_0}^2)$
713: in the dispersion relation
714: has been shown to result in subluminal group velocities
715: for all 3-momenta
716: without introducing instabilities.\cite{kl01}
717: In an arbitrary frame,
718: this modification takes the form
719: \begin{equation}
720: c_{\mu\nu}\rightarrow c_{\mu\nu}
721: \exp (c_{\mu\nu}\lambda^{\mu}\lambda^{\nu})
722: \quad ,
723: \label{crepl}
724: \end{equation}
725: establishing observer invariance
726: of the resulting dispersion relation.
727: It can also be shown
728: that introducing similar exponential suppression factors
729: in models with instabilities
730: can also resolve this problem
731: while maintaining subluminal group velocities.
732:
733: The above demonstrations prove that stable and causal
734: dispersion relations
735: violating Lorentz and CPT symmetry can exist.
736: The occurrence of transcendental functions of the 4-momenta
737: corresponds to derivative couplings of arbitrary order
738: in the lagrangian.
739: A satisfactory framework incorporating Lorentz and CPT violation
740: appears necessarily to be nonlocal in this sense.
741: Although it is in principle conceivable that
742: a model with explicit Lorentz breaking might satisfy the requirements
743: of causality and stability,
744: it would appear somewhat contrived to
745: implement both the necessary observer Lorentz invariance
746: and nonlocal couplings by hand.
747: On the other hand,
748: one can see that spontaneous Lorentz and CPT violation
749: in a nonlocal theory
750: can naturally yield the desired ingredients
751: for stability and causality at all scales.
752:
753: It would be interesting to identify theories
754: from which these dispersion relations emerge naturally.
755: A promising candidate for this type of framework is string theory.
756: It provided the original motivation for the construction
757: of the standard-model extension.
758: Moreover, strings are known to admit spontaneous Lorentz
759: violation and they have nonlocal interaction.
760: A complete treatment of this question would be desirable,
761: but is hampered by the absence
762: of a satisfactory realistic string theory.
763: Instead, we consider the field theory
764: of the open bosonic string as an example
765: and show that its structure is compatible
766: with dispersion relations of the desired type.
767:
768: The open bosonic string has no fermion modes.
769: We will therefore consider the scalar tachyon.
770: The relevant quadratic terms of the lagrangian for the tachyon
771: in the presence of Lorentz violation are given by:\cite{str}
772: \begin{eqnarray}
773: {\cal L} &\supset &
774: \fr{1}{2} \partial_\mu \phi \partial^\mu \phi
775: + (\alpha^{\prime -1} + k_0) \phi^2
776: + \ldots
777: + k_1 \langle B_{\mu\nu} \rangle \partial^\mu \phi \partial^\nu \phi
778: \nonumber\\
779: &&\qquad
780: + \ldots + k_2 \langle D_{\mu\nu\rho\sigma} \rangle
781: \partial^\mu \phi \partial^\nu \phi \partial^\rho \phi
782: \partial^\sigma \phi
783: + \ldots
784: \quad .
785: \label{string}
786: \end{eqnarray}
787: Here,
788: the scalar parameters $k_0$, $k_1$, $k_2$, $\ldots$
789: are determined by the theory,
790: but their specific values are irrelevant
791: for the present considerations.
792: Each ellipsis represents quadratic terms involving
793: vacuum expectation values of other tensors
794: and terms with powers of $\lambda^2$.
795:
796: For a plane-wave tachyon solution,
797: the structure of the dispersion relation
798: resulting from lagrangian (\ref{string})
799: indeed exhibits features
800: needed to maintain causality and stability.
801: For example,
802: it contains all terms of the dispersion relation that results
803: from the replacement (\ref{crepl}) in the $c_{00}$ model,
804: as the reader is invited to verify.
805:
806: We emphasize that the purpose of the above discussion
807: is only to provide an outline indicating how
808: a satisfactory dispersion relation for Lorentz violation
809: could emerge in the context of string theory.
810: In particular,
811: we do not claim that the tachyon itself
812: must \it necessarily \rm obey such a relation,
813: although it is conceivable that it does.\cite{str}
814: Here,
815: the tachyon dispersion relation is used
816: merely as an illustration to display explicitly
817: the appearance of nonlocal couplings in string theory
818: that could be appropriate for
819: a stable and causal theory with spontaneous Lorentz violation.
820: This type of coupling is generic both for other fields in
821: the open bosonic string
822: and for fields in other string theories,
823: including ones with fermions.
824:
825:
826:
827: \section*{References}
828: \begin{thebibliography}{99}
829:
830: \bibitem{pauli}
831: J.S.\ Bell,
832: Birmingham University thesis (1954);
833: G.\ L\"uders, Det.\ Kong.\ Danske Videnskabernes
834: Selskab Mat.fysiske Meddelelser {\bf 28}, 5 (1954);
835: W.\ Pauli, in
836: {\it Niels Bohr and the Development of Physics},
837: edited by W.\ Pauli
838: (McGraw-Hill, New York, 1955).
839:
840: \bibitem{str}
841: V.A.\ Kosteleck\'y and S.\ Samuel,
842: Phys.\ Rev.\ D {\bf 39}, 683 (1989);
843: V.A.\ Kosteleck\'y and R.\ Potting,
844: Nucl.\ Phys.\ B {\bf 359}, 545 (1991);
845: Phys.\ Lett.\ B {\bf 381}, 89 (1996).
846:
847: \bibitem{klink}
848: F.R.\ Klinkhamer,
849: Nucl.\ Phys.\ B {\bf 578}, 277 (2000).
850:
851: \bibitem{ch}
852: S.M.\ Carroll {\it et al.},
853: Phys.\ Rev.\ Lett.\ {\bf 87}, 141601 (2001).
854:
855: \bibitem{ck}
856: D.\ Colladay and V.A.\ Kosteleck\'y,
857: Phys.\ Rev.\ D {\bf 55}, 6760 (1997);
858: Phys.\ Rev.\ D {\bf 58}, 116002 (1998).
859:
860: \bibitem{exp}
861: Reviews of various experimental approaches
862: can be found, for example, in these proceedings
863: and V.A.\ Kosteleck\'y, ed.,
864: {\it CPT and Lorentz Symmetry}
865: (World Scientific, Singapore, 1999).
866:
867: \bibitem{kl99}
868: V.A.\ Kosteleck\'y and C.D.\ Lane
869: Phys.\ Rev.\ D {\bf 60}, 116010 (1999);
870: J.\ Math.\ Phys. {bf 40}, 6245 (1999).
871:
872: \bibitem{kl01}
873: V.A.\ Kosteleck\'y and R.\ Lehnert,
874: Phys.\ Rev.\ D {\bf 63}, 65008 (2001).
875:
876: \bibitem{kljmp}
877: V.A.\ Kosteleck\'y and R.\ Lehnert,
878: in preparation.
879:
880: \bibitem{sg}
881: See, e.g.,
882: G.\ Sansone and J.\ Gerretsen,
883: {\it Lectures on the Theory of Functions of a Complex Variable, Vol.1},
884: Noordhoff, Groningen, 1960.
885:
886: \bibitem{be}
887: H.J.\ Borchers and D.\ Buchholz,
888: Commun.\ Math.\ Phys.\ {\bf 97},
889: 169 (1985);
890: J.\ Bros, H.\ Epstein and V.\ Glaser,
891: Nuovo Cimento {\bf 31},
892: 1265 (1964).
893:
894:
895: \bibitem{ksobscs}
896: V.A.\ Kosteleck\'y and S.\ Samuel,
897: Phys.\ Lett.\ B {\bf 207}, 169 (1988);
898: Nucl.\ Phys.\ {\bf B336}, 263 (1990);
899: Phys.\ Rev.\ D {\bf 42}, 1289 (1990);
900: Phys.\ Rev.\ Lett.\ {\bf 64}, 2238 (1990);
901: Phys.\ Rev.\ Lett. {\bf 66}, 1811 (1991).
902:
903: \end{thebibliography}
904:
905: \end{document}
906:
907: %LEHNERT%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
908: %% End of sprocl.tex
909: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
910: