1: \documentclass[12pt]{article}

2: \usepackage{a4wide}

3: \usepackage{amssymb}

4: \usepackage{graphicx}

5: %\voffset-1cm

6: \begin{document}

7: {\renewcommand{\thefootnote}{\fnsymbol{footnote}}

8: \hfill  AEI--2005--025\\ 

9: \medskip

10: \hfill gr--qc/0505057\\

11: \medskip

12: \begin{center}

13: {\LARGE  Elements of Loop Quantum Cosmology}\\

14: \vspace{1.5em}

15: Martin Bojowald\footnote{e-mail address: {\tt mabo@aei.mpg.de}}

16: \\

17: \vspace{0.5em}

18: Max-Planck-Institut f\"ur Gravitationsphysik, Albert-Einstein-Institut,\\

19: Am M\"uhlenberg 1, D-14476 Potsdam, Germany

20: \vspace{1.5em}

21: \end{center}

22: }

23: 

24: \setcounter{footnote}{0}

25: 

26: \newcommand{\lP}{\ell_{\rm P}}

27: \newcommand{\md}{{\rm d}}

28: \newcommand{\sgn}{{\rm sgn}}

29: 

30: \begin{abstract}

31:   The expansion of our universe, when followed backward in time,

32:   implies that it emerged from a phase of huge density, the big bang.

33:   These stages are so extreme that classical general relativity

34:   combined with matter theories is not able to describe them properly,

35:   and one has to refer to quantum gravity. A complete quantization of

36:   gravity has not yet been developed, but there are many results about

37:   key properties to be expected. When applied to cosmology, a

38:   consistent picture of the early universe arises which is free of the

39:   classical pathologies and has implications for the generation of

40:   structure which are potentially observable in the near future.

41: \end{abstract}

42: 

43: \section{Introduction} 

44: 

45: General relativity provides us with an extremely successful

46: description of the structure of our universe on large scales, with

47: many confirmations by macroscopic experiments and so far no conflict

48: with observations.  The resulting picture, when applied to early

49: stages of cosmology, suggests that the universe had a beginning a

50: finite time ago, at a point where space, matter, and also time itself

51: were created. Thus, it does not even make sense to ask what was there

52: before since ``before'' does not exist at all. At very early stages,

53: space was small such that there were huge energy densities to be

54: diluted in the later expansion of the universe that is still

55: experienced today. In order to explain also the structure that we see

56: in the form of galaxies in the correct statistical distribution, the

57: universe not only needs to expand but do so in an accelerated manner,

58: a so-called inflationary period, in its early stages. With this

59: additional input, usually by introducing inflation with exponential

60: acceleration \cite{Guth,NewInfl,InflAS} lasting long enough to expand

61: the scale factor $a(t)$, the radius of the universe at a given time

62: $t$, by a ratio $a_{\rm final}/a_{\rm initial}>e^{60}$. The resulting

63: seeds for structure after the inflationary phase can be observed in

64: the anisotropy spectrum of the cosmic microwave background (most

65: recently of the WMAP satellite \cite{WMAPParam}), which agrees well

66: with theoretical predictions over a large range of scales.

67: 

68: Nonetheless, there are problems remaining with the overall picture.

69: The beginning was extremely violent with conditions such as diverging

70: energy densities and tidal forces under which no theory can prevail.

71: This is also true for general relativity itself which led to this

72: conclusion in the first place: according to the singularity theorems

73: any solution to general relativity, under reasonable conditions on the

74: form of matter, must have a singularity in the past or

75: future \cite{SingTheo}. There, space degenerates, e.g.\ to a single

76: point in cosmology, and energy densities and tidal forces diverge.

77: From the observed expansion of our current universe one can conclude

78: that according to general relativity there must have been such a

79: singularity in the past (which does not rule out further possible

80: singularities in the future). This is exactly what is usually referred

81: to as the ``beginning'' of the universe, but from the discussion it is

82: clear that the singularity does not so much present a boundary to the

83: universe as a boundary to the classical theory: The theory predicts

84: conditions under which it has to break down and is thus incomplete.

85: Here it is important that the singularity in fact lies only a finite

86: time in the past rather than an infinite distance away, which could be

87: acceptable. A definitive conclusion about a possible beginning can

88: therefore be reached only if a more complete theory is found which is

89: able to describe these very early stages meaningfully.

90: 

91: Physically, one can understand the inevitable presence of

92: singularities in general relativity by the characteristic property of

93: classical gravitation being always attractive. In the backward

94: evolution in which the universe contracts, there is, once matter has

95: collapsed to a certain size, simply no repulsive force strong enough

96: to prevent the total collapse into a singularity. A similar behavior

97: happens when not all the matter in the universe but only that in a

98: given region collapses to a small size, leading to the formation of

99: black holes which also are singular.

100: 

101: This is the main problem which has to be resolved before one can call

102: our picture of the universe complete. Moreover, there are other

103: problematic issues in what we described so far. Inflation has to be

104: introduced into the picture, which currently is done by assuming a

105: special field, the inflaton, in addition to the matter we know. In

106: contrast to other matter, its properties must be very exotic so as to

107: ensure accelerated expansion which with Einstein's equations is

108: possible only if there is negative pressure. This is achieved by

109: choosing a special potential and initial conditions for the inflaton,

110: but there is no fundamental explanation of the nature of the inflaton

111: and its properties. Finally, there are some details in the anisotropy

112: spectrum which are hard to bring in agreement with theoretical models.

113: In particular, there seems to be less structure on large scales than

114: expected, referred to as a loss of power.

115: 

116: \section{Classical Cosmology}

117: 

118: In classical cosmology one usually assumes space to be homogeneous and

119: isotropic, which is an excellent approximation on large scales today.

120: The metric of space is then solely determined by the scale factor

121: $a(t)$ which gives the size of the universe at any given time $t$. The

122: function $a(t)$ describes the expansion or contraction of space in a

123: way dictated by the Friedmann equation \cite{Friedmann}

124: \begin{equation} \label{Friedmann}

125:   \left(\frac{\dot{a}}{a}\right)^2=\frac{8\pi}{3}G\rho(a)

126: \end{equation}

127: which is the reduction of Einstein's equations under the assumption of

128: isotropy. In this equation, $G$ is the gravitational constant and

129: $\rho(a)$ the energy density of whatever matter we have in the

130: universe. Once the matter content is chosen and $\rho(a)$ is known,

131: one can solve the Friedmann equation in order to obtain $a(t)$.

132: 

133: As an example we consider the case of radiation which can be

134: described phenomenologically by the energy density $\rho(a)\propto

135: a^{-4}$. This is only a phenomenological description since it ignores

136: the fundamental formulation of electrodynamics of the Maxwell field.

137: Instead of using the Maxwell Hamiltonian in order to define the energy

138: density, which would complicate the situation by introducing the

139: electromagnetic fields with new field equations coupled to the

140: Friedmann equation, one uses the fact that on large scales the energy

141: density of radiation is diluted by the expansion and in addition

142: red-shifted. This leads to a behavior proportional $a^{-3}$ from

143: dilution times $a^{-1}$ from redshift. In this example we then

144: solve the Friedmann equation $\dot{a}\propto a^{-1}$

145: by $a(t)\propto\sqrt{t-t_0}$ with a constant of integration $t_0$.

146: This demonstrates the occurrence of singularities: For any solution

147: there is a time $t=t_0$ where the size of space vanishes and the

148: energy density $\rho(a(t_0))$ diverges. At this point not only the

149: matter system becomes unphysical, but also the gravitational evolution

150: breaks down: When the right hand side of (\ref{Friedmann}) diverges at

151: some time $t_0$, we cannot follow the evolution further by setting up

152: an initial value problem there and integrating the equation. We can

153: thus only learn that there is a singularity in the classical theory,

154: but do not obtain any information as to what is happening there and

155: beyond. These are the two related but not identical features of a

156: singularity: energy densities diverge and the evolution breaks down.

157: 

158: One could think that the problem comes from too strong idealizations

159: such as symmetry assumptions or the phenomenological description of

160: matter. That this is not the case follows from the singularity

161: theorems which do not depend on these assumptions. One can also

162: illustrate the singularity problem with a field theoretic rather than

163: phenomenological description of matter. For simplicity we now assume

164: that matter is provided by a scalar $\phi$ whose energy density then

165: follows from the Hamiltonian

166: \begin{equation} \label{Hscalar}

167:   \rho(a)=a^{-3}H(a)= a^{-3}({\textstyle\frac{1}{2}}a^{-3}p_{\phi}^2+

168: a^3V(\phi))

169: \end{equation}

170: with the scalar momentum $p_{\phi}$ and potential $V(\phi)$. At small

171: scale factors $a$, there still is a diverging factor $a^{-3}$ in the

172: kinetic term which we recognized as being responsible for the

173: singularity before. Since this term dominates over the non-diverging

174: potential term, we still cannot escape the singularity by using this

175: more fundamental description of matter. This is true unless we manage

176: to arrange the evolution of the scalar in such a way that

177: $p_{\phi}\to0$ when $a\to0$ in just the right way for the kinetic term

178: not to diverge. This is difficult to arrange in general, but is

179: exactly what is attempted in slow-roll inflation (though with a

180: different motivation, and not necessarily all the way up to the

181: classical singularity).

182: 

183: For the evolution of $p_{\phi}$ we need the scalar equation of motion,

184: which can be derived from the Hamiltonian $H$ in (\ref{Hscalar}) via

185: $\dot{\phi}=\{\phi,H\}$ and $\dot{p}_{\phi}=\{p_{\phi},H\}$. This

186: results in the isotropic Klein--Gordon equation in a time-dependent

187: background determined by $a(t)$,

188: \begin{equation} \label{KG}

189:  \ddot{\phi}+3\dot{a}a^{-1}\dot{\phi}+V'(\phi)=0\,.

190: \end{equation}

191: In an expanding space with positive $\dot{a}$ the second term implies

192: friction such that, if we assume the potential $V'(\phi)$ to be flat

193: enough, $\phi$ will change only slowly (slow-roll). Thus, $\dot{\phi}$

194: and $p_{\phi}=a^3\dot{\phi}$ are small and at least for some time we

195: can ignore the kinetic term in the energy density. Moreover, since

196: $\phi$ changes only slowly we can regard the potential $V(\phi)$ as a

197: constant $\Lambda$ which again allows us to solve the Friedmann

198: equation with $\rho(a)=\Lambda$. The solution $a\propto

199: \exp(\sqrt{8\pi G\Lambda/3}\,t)$ is inflationary since $\ddot{a}>0$ and

200: non-singular: $a$ becomes zero only in the limit $t\to-\infty$.

201: 

202: Thus, we now have a mechanism to drive a phase of accelerated

203: expansion important for observations of structure. However, this

204: expansion must be long enough, which means that the phase of slowly

205: rolling $\phi$ must be long. This can be achieved only if the

206: potential is very flat and $\phi$ starts sufficiently far away from

207: its potential minimum. Flatness means that the ratio of $V(\phi_{\rm

208:   initial})$ and $\phi_{\rm initial}$ must be of the order $10^{-10}$,

209: while $\phi_{\rm initial}$ must be huge, of the order of the Planck

210: mass \cite{ChaoticInfl}. These assumptions are necessary for agreement

211: with observations, but are in need of more fundamental explanations.

212: 

213: Moreover, inflation alone does not solve the singularity

214: problem \cite{InflSing}. The non-singular solution we just

215: obtained was derived under the approximation that the kinetic term can

216: be ignored when $\dot{\phi}$ is small. This is true in a certain range

217: of $a$, depending on how small $\dot{\phi}$ really is, but never very

218: close to $a=0$. Eventually, even with slow-roll conditions, the

219: diverging $a^{-3}$ will dominate and lead to a singularity.

220: 

221: \section{Quantum Gravity}

222: 

223: For decades, quantum gravity has been expected to complete the

224: picture which is related to well-known properties of quantum mechanics

225: in the presence of a non-zero $\hbar$.

226: 

227: \subsection{Indications}

228: 

229: First, in analogy to the singularity problem in gravity, where

230: everything falls into a singularity in finite time, there is the

231: instability problem of a classical hydrogen atom, where the electron

232: would fall into the nucleus after a brief time. From quantum mechanics

233: we know how the instability problem is solved: There is a finite

234: ground state energy $E_0=-\frac{1}{2} me^4/\hbar^2$, implying that the

235: electron cannot radiate away all its energy and not fall further once

236: it reaches the ground state. From the expression for $E_0$ one can see

237: that quantum theory with its non-zero $\hbar$ is essential for this to

238: happen: When $\hbar\to0$ in a classical limit, $E_0\to-\infty$ which

239: brings us back to the classical instability. One expects a similar

240: role to be played by the Planck length $\lP=\sqrt{8\pi G\hbar/c^3}\approx

241: 10^{-35}{\rm m}$ which is tiny but non-zero in quantum theory. If,

242: just for dimensional reasons, densities are bounded by $\lP^{-3}$,

243: this would be finite in quantum gravity but diverge in the classical

244: limit.

245: 

246: Secondly, a classical treatment of black body radiation suggests the

247: Rayleigh--Jeans law according to which the spectral density behaves as

248: $\rho(\lambda)\propto\lambda^{-4}$ as a function of the wave

249: length. This is unacceptable since the divergence at small wave

250: lengths leads to an infinite total energy. Here, quantum mechanics

251: solves the problem by cutting off the divergence with Planck's formula

252: which has a maximum at a wave length $\lambda_{\rm max}\sim h/kT$ and

253: approaches zero at smaller scales. Again, in the classical limit

254: $\lambda_{\max}$ becomes zero and the expression diverges.

255: 

256: In cosmology the situation is similar for matter in the whole universe

257: rather than a cavity. Energy densities as a function of the scale

258: factor behave as, e.g., $a^{-3}$ if matter is just diluted or $a^{-4}$

259: if there is an additional redshift factor. In all cases, the energy

260: density diverges at small scales, comparable to the Rayleigh--Jeans

261: law. Inflation already provides an indication that the behavior must

262: be different at small scales. Indeed, inflation can only be achieved

263: with negative pressure, while all matter whose energy falls off as

264: $a^{-k}$ with non-negative $k$ has positive pressure. This can easily

265: be seen from the thermodynamical definition of pressure as the

266: negative change of energy with volume. Negative pressure then requires

267: the energy to increase with the scale factor at least at small scales

268: where inflation is required (e.g., an energy $\Lambda a^3$ for

269: exponential inflation). This could be reconciled with standard forms

270: of matter if there is an analog to Planck's formula, which

271: interpolates between decreasing behavior at large scales and a

272: behavior increasing from zero at small scales, with a maximum in

273: between.

274: 

275: \subsection{Early Quantum Cosmology}

276: 

277: Since the isotropic reduction of general relativity leads to a system

278: with finitely many degrees of freedom, one can in a first attempt try

279: quantum mechanics to quantize it. Starting with the Friedmann equation

280: (\ref{Friedmann}) and replacing $\dot{a}$ by its momentum

281: $p_a=3a\dot{a}/8\pi G$ gives a Hamiltonian which is quadratic in

282: the momentum and can be quantized easily to an operator acting on a

283: wave function depending on the gravitational variable $a$ and possibly

284: matter fields $\phi$. The usual Schr\"odinger representation yields

285: the Wheeler--DeWitt equation \cite{DeWitt,QCReview}

286: \begin{equation}\label{WdW}

287:  \frac{3}{2}\left(-\frac{1}{9}\lP^4

288:   a^{-1}\frac{\partial}{\partial a}a^{-1}

289:   \frac{\partial}{\partial a}\right) a

290: \psi(a,\phi)= 8\pi G\hat{H}_{\phi}(a) \psi(a,\phi)

291: \end{equation}

292: with the matter Hamiltonian $\hat{H}_{\phi}(a)$. This system is

293: different from usual quantum mechanics in that there are factor

294: ordering ambiguities in the kinetic term, and that there is no

295: derivative with respect to coordinate time $t$. The latter fact is a

296: consequence of general covariance: the Hamiltonian is a constraint

297: equation restricting allowed states $\psi(a,\phi)$, rather than a

298: Hamiltonian generating evolution in coordinate time.  Nevertheless,

299: one can interpret equation (\ref{WdW}) as an evolution equation in the

300: scale factor $a$, which is then called internal time.  The left hand

301: side thus becomes a second order time derivative, and it means that

302: the evolution of matter is measured relationally with respect to the

303: expansion or contraction of the universe, rather than absolutely in

304: coordinate time.

305: 

306: Straightforward quantization thus gives us a quantum evolution

307: equation, and we can now check what this implies for the singularity.

308: If we look at the equation for $a=0$, we notice first that the matter

309: Hamiltonian still leads to diverging energy densities. If we quantize

310: (\ref{Hscalar}), we replace $p_{\phi}$ by a derivative, but the

311: singular dependence on $a$ does not change; $a^{-3}$ would simply

312: become a multiplication operator acting on the wave function. Moreover,

313: $a=0$ remains a singular point of the quantum evolution equation in

314: internal time. There is nothing from the theory which tells us what

315: physically happens at the singularity or beyond (baring intuitive

316: pictures which have been developed from this

317: perspective \cite{tunneling,nobound}).

318: 

319: So one has to ask what went wrong with our expectations that

320: quantizing gravity should help. The answer is that quantum theory

321: itself did not necessarily fail, but only our simple implementation.

322: Indeed, what we used was just quantum mechanics, while quantum gravity

323: has many consistency conditions to be fulfilled which makes

324: constructing it so complicated. At the time when this formalism was

325: first applied there was in fact no corresponding full quantum theory

326: of gravity which could have guided developments. In such a simple case

327: as isotropic cosmology, most of these consistency conditions

328: trivialize and one can easily overlook important issues. There are

329: many choices in quantizing an unknown system, and tacitly making one

330: choice can easily lead in a wrong direction.

331: 

332: Fortunately, the situation has changed with the development of

333: strong candidates for quantum gravity. This then allows us to

334: reconsider the singularity and other problems from the point of view

335: of the full theory, making sure that also in a simpler cosmological

336: context only those steps are undertaken that have an analog in the full

337: theory.

338: 

339: \subsection{Loop quantum gravity}

340: 

341: Singularities are physically extreme and require special properties of

342: any theory aimed at tackling them. First, there are always strong

343: fields (classically diverging) which requires a non-perturbative

344: treatment. Moreover, classically we expect space to degenerate at the

345: singularity, for instance a single point in an isotropic model. This

346: means that we cannot take the presence of a classical geometry to

347: measure distances for granted, which is technically expressed as

348: background independence. A non-perturbative and background independent

349: quantization of gravity is available in the form of loop quantum

350: gravity \cite{Rov:Loops,ThomasRev,ALRev}, which by now is understood

351: well enough in order to be applicable in physically interesting

352: situations.

353: 

354: Here, we only mention salient features of the theory which will turn

355: out to be important for cosmology. The first one is the kind of basic

356: variables used, which are the Ashtekar

357: connection \cite{AshVar,AshVarReell} describing the curvature of space

358: and a densitized triad describing the metric by a collection of three

359: orthonormal vectors in each point. These variables are important since

360: they allow a background independent representation of the theory,

361: where the connection $A_a^i$ is integrated to holonomies

362: \begin{equation} \label{Hol}

363:  h_e(A)={\cal P}\exp\int_e A_a^i\tau_i \dot{e}^a\md t

364: \end{equation}

365: along curves $e$ in space and the densitized triad $E^a_i$ to fluxes

366: \begin{equation} \label{Flux}

367:  F_S(E)=\int_S E^a_i\tau^in_a\md^2y

368: \end{equation}

369: along surfaces $S$. (In these expressions, $\dot{e}^a$ denotes the

370: tangent vector to a curve and $n_a$ the conormal to a surface, both of

371: which are defined without reference to a background metric. Moreover,

372: $\tau_j=-\frac{1}{2}i\sigma_j$ in terms of Pauli matrices). While

373: usual quantum field theory techniques rest on the presence of a

374: background metric, for instance in order to decompose a field in its

375: Fourier modes and define a vacuum state and particles, this is no

376: longer available in quantum gravity where the metric itself must be

377: turned into an operator. On the other hand, some integration is

378: necessary since the fields themselves are distributional in quantum

379: field theory and do not allow a well-defined representation. This

380: ``smearing'' with respect to a background metric has to be replaced by

381: some other integration sufficient for resulting in honest

382: operators \cite{LoopRep,ALMMT}. This is achieved by the integrations in

383: (\ref{Hol}) and (\ref{Flux}), which similarly lead to a well-defined

384: quantum representation. Usual Fock spaces in perturbative quantum

385: field theory are thereby replaced by the loop representation, where an

386: orthonormal basis is given by spin network states \cite{RS:Spinnet}.

387: 

388: This shows that choosing basic variables for a theory to quantize has

389: implications for the resulting representation. Connections and

390: densitized triads can naturally be smeared along curves and surfaces

391: without using a background metric and then represented on a Hilbert

392: space. Requiring diffeomorphism invariance, which means that a

393: background independent theory must not change under deformations of

394: space (which can be interpreted as changes of coordinates), even

395: selects a unique

396: representation \cite{FluxAlg,Meas,HolFluxRep,SuperSel,WeylRep}. These are

397: basic properties of loop quantum gravity, recognized as important

398: requirements for a background independent quantization. Already here

399: we can see differences to the Wheeler--DeWitt quantization, where the

400: metric is used as a basic variable and then quantized as in quantum

401: mechanics. This is possible in the model but not in a full theory, and

402: in fact we will see later that a loop quantization will give a

403: representation inequivalent to the Wheeler--DeWitt quantization.

404: 

405: The basic properties of the representation have further consequences.

406: Holonomies and fluxes act as well-defined operators, and fluxes have

407: discrete spectra. Since spatial geometry is determined by the

408: densitized triad, spatial geometry is discrete, too, with discrete

409: spectra for, e.g., the area and volume

410: operator \cite{AreaVol,Area,Vol2}. The geometry of space-time is more

411: complicated to understand since this is a dynamical situation which

412: requires solving the Hamiltonian constraint. This is the analog of the

413: Wheeler--DeWitt equation in the full theory and is the quantization of

414: Einstein's dynamical equations. There are candidates for such

415: operators \cite{QSDI}, well-defined even in the presence of

416: matter \cite{QSDV} which in usual quantum field theory would contribute

417: divergent matter Hamiltonians. Not surprisingly, the full situation is

418: hard to analyze, which is already the case classically, without assuming

419: simplifications from symmetries. We will thus return to symmetric, in

420: particular isotropic models, but with the new perspective provided by

421: the full theory of loop quantum gravity.

422: 

423: \section{Quantum cosmology}

424: 

425: Symmetries can be introduced in loop quantum gravity at the level of

426: states and basic operators \cite{PhD,SymmRed,SphSymm}, such that it is not

427: necessary to reduce the classical theory first and then quantize as in

428: the Wheeler--DeWitt quantization. Instead, one can view the procedure

429: as quantizing first and then introducing symmetries which ensures that

430: consistency conditions of quantum gravity are observed in the first

431: step before one considers treatable situations. In particular, the

432: quantum representation derives from symmetric states and basic

433: operators, while the Hamiltonian constraint can be obtained with

434: constructions analogous to those in the full theory. This allows us to

435: reconsider the singularity problem, now with methods from full quantum

436: gravity. In fact, symmetric models present a class of systems which

437: can often be treated explicitly while still being representative for

438: general phenomena. For instance, the prime examples of singular

439: situations in gravity, and some of the most widely studied physical

440: applications, are already obtained in isotropic or spherically

441: symmetric systems, which allow access to cosmology and black holes.

442: 

443: \subsection{Representation}

444: 

445: Before discussing the quantum level we reformulate isotropic cosmology

446: in connection and triad variables instead of $a$. The role of the

447: scale factor is now played by the densitized triad component $p$ with

448: $|p|=a^2$ whose canonical momentum is the isotropic connection

449: component $c=-\frac{1}{2}\dot{a}$ with $\{c,p\}=8\pi G/3$. The main

450: difference to metric variables is the fact that $p$, unlike $a$, can

451: take both signs with $\sgn p$ being the orientation of space. This is

452: a consequence of having to use triad variables which not only know

453: about the size of space but also its orientation (depending on whether

454: the set of orthonormal vectors is left or right handed).

455: 

456: States in the full theory are usually written in the connection

457: representation as functions of holonomies. Following the reduction

458: procedure for an isotropic symmetry group leads to orthonormal states which are

459: functions of the isotropic connection component $c$ and given by \cite{Bohr}

460: \begin{equation} \label{basis}

461:  \langle c|\mu\rangle= e^{i\mu c/2} \qquad \mu\in{\mathbb R}\,.

462: \end{equation}

463: On these states the basic variables $p$ and $c$ are represented by

464: \begin{eqnarray}

465: \hat{p}|\mu\rangle &=& {\textstyle\frac{1}{6}}\lP^2\mu|\mu\rangle\\

466: \widehat{e^{i\mu'c/2}}|\mu\rangle &=& |\mu+\mu'\rangle 

467:  \end{eqnarray}

468: with the properties:

469: \begin{enumerate}

470:  \item $[\widehat{e^{i\mu'

471:        c/2}},\hat{p}]=-\frac{1}{6}\lP^2\mu'\widehat{e^{-i\mu' c/2}}=

472:    i\hbar(\{e^{i\mu' c/2},p\})^{\wedge}$,

473:  \item $\hat{p}$ has a discrete spectrum and

474:  \item only exponentials $e^{i\mu'c/2}$ of $c$ are represented, not

475:    $c$ directly.

476: \end{enumerate}

477: These statements deserve further explanation: First, the classical

478: Poisson relations between the basic variables are indeed represented

479: correctly, turning the Poisson brackets into commutators divided by

480: $i\hbar$.  On this representation, the set of eigenvalues of $\hat{p}$

481: is the full real line since $\mu$ can take arbitrary real values.

482: Nevertheless, the spectrum of $\hat{p}$ is discrete in the technical

483: sense that eigenstates of $\hat{p}$ are normalizable. This is indeed

484: the case in this non-separable Hilbert space where (\ref{basis})

485: defines an orthonormal basis. The last property follows since the

486: exponentials are not continuous in the label $\mu'$, for otherwise one

487: could simply take the derivative with respect to $\mu'$ at $\mu'=0$

488: and obtain an operator for $c$. The discontinuity can be seen, e.g., from

489: \[

490:  \langle\mu|\widehat{e^{i\mu'c/2}}|\mu\rangle= \delta_{0,\mu'}

491: \]

492: which is not continuous.

493: 

494: These properties are quite unfamiliar from quantum mechanics, and

495: indeed the representation is inequivalent to the Schr\"odinger

496: representation (the discontinuity of the $c$-exponential evading the

497: Stone--von Neumann theorem which usually implies uniqueness of the

498: representation). In fact, the loop representation is inequivalent to

499: the Wheeler-DeWitt quantization which just assumed a Schr\"odinger

500: like quantization. In view of the fact that the phase space of our

501: system is spanned by $c$ and $p$ with $\{c,p\}\propto 1$ just as in

502: classical mechanics, the question arises how such a difference in the

503: quantum formulation arises.

504: 

505: As a mathematical problem the basic step of quantization occurs as

506: follows: given the classical Poisson algebra of observables $Q$ and

507: $P$ with $\{Q,P\}=1$, how can we define a representation of the

508: observables on a Hilbert space such that the Poisson relations become

509: commutator relations and complex conjugation, meaning that $Q$ and $P$

510: are real, becomes adjointness? The problem is mathematically much

511: better defined if one uses the bounded expressions $e^{isQ}$ and

512: $e^{it\hbar^{-1}P}$ instead of the unbounded $Q$ and $P$, which still

513: allows us to distinguish any two points in the whole phase space. The

514: basic objects $e^{isQ}$ and $e^{it\hbar^{-1}P}$ upon quantization will

515: then not commute but fulfill the commutation relation (Weyl algebra)

516: \begin{equation}\label{QP}

517:  e^{isQ}e^{it\hbar^{-1}P} = e^{ist} e^{it\hbar^{-1}P} e^{isQ}

518: \end{equation}

519: as unitary operators on a Hilbert space.

520: 

521: In the Schr\"odinger representation this is done by using a Hilbert

522: space $L^2({\mathbb R},\md q)$ of square integrable functions

523: $\psi(q)$ with $\int_{\mathbb R}\md q|\psi(q)|^2$ finite. The

524: representation of basic operators is

525: \begin{eqnarray*}

526:  e^{isQ}\psi(q) &=& e^{isq}\psi(q)\\

527:  e^{it\hbar^{-1}P}\psi(q) &=& \psi(q+t)

528: \end{eqnarray*}

529: which indeed are unitary and fulfill the required commutation

530: relation. Moreover, the operator families as functions of $s$ and $t$

531: are continuous and we can take the derivatives in $s=0$ and $t=0$,

532: respectively:

533: \begin{eqnarray*}

534:  -i\left.\frac{\md}{\md s}\right|_{s=0} e^{isQ} &=& q\\

535:  -i\hbar\left.\frac{\md}{\md t}\right|_{t=0} e^{it\hbar^{-1}P} &=&

536:  \hat{p}=-i\hbar\frac{\md}{\md q}\,.

537: \end{eqnarray*}

538: This is the familiar representation of quantum mechanics which,

539: according to the Stone--von Neumann theorem is unique under the

540: condition that $e^{isQ}$ and $e^{it\hbar^{-1}P}$ are indeed continuous

541: in both $s$ and $t$.

542: 

543: The latter condition is commonly taken for granted in quantum

544: mechanics, but in general there is no underlying physical or

545: mathematical reason. It is easy to define representations violating

546: continuity in $s$ or $t$, for instance if we use a Hilbert space

547: $\ell^2({\mathbb R})$ where states are again maps $\psi_q$ from the

548: real line to complex numbers but with norm $\sum_q|\psi_q|^2$ which

549: implies that normalizable $\psi_q$ can be non-zero for at most

550: countably many $q$. We obtain a representation with basic operators

551: \begin{eqnarray*}

552:  e^{isQ}\psi_q &=& e^{isq}\psi_q\\

553:  e^{it\hbar^{-1}P}\psi_q &=& \psi_{q+t}

554: \end{eqnarray*}

555: which is of the same form as before. However, due to the different

556: Hilbert space the second operator $e^{it\hbar^{-1}P}$ is no longer

557: continuous in $t$ which can be checked as in the case of $e^{i\mu

558:   c/2}$. In fact, the representation for $Q$ and $P$ is isomorphic to

559: that of $p$ and $c$ used before, where a general state

560: $|\psi\rangle=\sum_{\mu}\psi_{\mu}|\mu\rangle$ has coefficients

561: $\psi_{\mu}$ in $\ell^2({\mathbb R})$. 

562: 

563: This explains mathematically why different, inequivalent

564: representations are possible, but what are the physical reasons for

565: using different representations in quantum mechanics and quantum

566: cosmology? In quantum mechanics it turns out that the choice of

567: representation is not that important and is mostly being done for

568: reasons of familiarity with the standard choice. Physical differences

569: between inequivalent representations can only occur at very high

570: energies \cite{PolymerParticle} which are not probed by available

571: experiments and do not affect characteristic quantum effects related

572: to the ground state or excited states. Thus, quantum mechanics as we

573: know it can well be formulated in an inequivalent representation, and

574: also in quantum field theory this can be done and even be

575: useful \cite{QEDBohr}.

576: 

577: In quantum cosmology we have a different situation where it is the high

578: energies which are essential. We do not have direct observations of

579: this regime, but from conceptual considerations such as the

580: singularity issue we have learned which problems we have to face. The

581: classical singularity leads to the highest energies one can imagine,

582: and it is here where the question of which representation to choose

583: becomes essential. As shown by the failure of the Wheeler--DeWitt

584: quantization in trying to remove the singularity, the Schr\"odinger

585: representation is inappropriate for quantum cosmology. The

586: representation underlying loop quantum cosmology, on the other hand,

587: implies very different properties which become important at high

588: energies and can shed new light on the singularity problem.

589: 

590: Moreover, by design of the symmetric models as derived from the full

591: theory, we have the same basic properties of a loop representation in

592: cosmological models and the full situation where they were recognized

593: as being important for a background independent quantization: discrete

594: fluxes $\hat{F}_S(E)$ or $\hat{p}$ and a representation only of

595: holonomies $h_e(A)$ or $e^{i\mu c/2}$ but not of connection components

596: $A_a^i$ or $c$. These basic properties have far-reaching consequences

597: as discussed in what follows \cite{LoopCosRev}:

598: \begin{center}

599: \begin{picture}(160,80)(0,0)

600: \put(80,70){\makebox(0,0){discrete triad \hspace{2cm} only holonomies}}

601: \put(30,60){\vector(0,-1){10}} \put(130,60){\vector(0,-1){10}}

602: \put(80,45){\makebox(0,0){finite inverse volume \hspace{2cm} discrete evolution}}

603: \put(40,40){\vector(1,-1){10}} \put(125,40){\vector(-1,-1){10}}

604: \put(80,25){\makebox(0,0){ non-singular}}

605: \put(30,35){\vector(0,-1){20}} \put(130,35){\vector(0,-1){20}}

606: \put(80,5){\makebox(0,0){\hspace{-1cm}non-perturbative modifications \hspace{1cm}

607: higher order terms}}

608: \end{picture}

609: \end{center}

610: By this reliable quantization of representative and physically

611: interesting models with a known relation to full quantum gravity we

612: are finally able to resolve long-standing issues such as the

613: singularity problem.

614: 

615: \subsection{Quantum evolution}

616: 

617: We will first look at the quantum evolution equation which we obtain as

618: the quantized Friedmann equation. This is modeled on the Hamiltonian

619: constraint of the full theory such that we can also draw some

620: conclusions for the viability of the full constraint.

621: 

622: \subsubsection{Difference equation}

623: 

624: The constraint equation will be imposed on states of the form

625: $|\psi\rangle=\sum_{\mu}\psi_{\mu}|\mu\rangle$ with summation over

626: countably many values of $\mu$. Since the states $|\mu\rangle$ are

627: eigenstates of the triad operator, the coefficients $\psi_{\mu}$ which

628: can also depend on matter fields such as a scalar $\phi$ represent the

629: state in the triad representation, analogous to $\psi(a,\phi)$ before.

630: For the constraint operator we again need operators for the conjugate

631: of $p$, related to $\dot{a}$ in the Friedmann equation. Since this is

632: now the exponential of $c$, which on basis states acts by shifting the

633: label, it translates to a finite shift in the labels of coefficients

634: $\psi_{\mu}(\phi)$. Plugging together all ingredients for a

635: quantization of (\ref{Friedmann}) along the lines of the constraint in

636: the full theory leads to the difference equation \cite{IsoCosmo,Bohr}

637: \begin{eqnarray} \label{DiffEq}

638: &&    (V_{\mu+5}-V_{\mu+3})\psi_{\mu+4}(\phi)- 2

639: (V_{\mu+1}-V_{\mu-1})\psi_{\mu}(\phi)\\\nonumber

640: &&+    (V_{\mu-3}-V_{\mu-5})\psi_{\mu-4}(\phi)

641:   = -{\textstyle\frac{4}{3}}\pi

642: G\ell_{\rm P}^2\hat{H}_{\rm matter}(\mu)\psi_{\mu}(\phi)

643: \end{eqnarray}

644: with volume eigenvalues $V_{\mu}=(\ell_{\rm P}^2|\mu|/6)^{3/2}$

645: obtained from the volume operator $\hat{V}=|\hat{p}|^{3/2}$, and the

646: matter Hamiltonian $\hat{H}_{\rm matter}(\mu)$.

647: 

648: We again have a constraint equation which does not generate evolution

649: in coordinate time but can be seen as evolution in internal time.

650: Instead of the continuous variable $a$ we now have the label $\mu$

651: which only jumps in discrete steps. As for the singularity issue,

652: there is a further difference to the Wheeler--DeWitt equation since

653: now the classical singularity is located at $p=0$ which is in the

654: interior rather than at the boundary of the configuration space.

655: Nevertheless, the classical evolution in the variable $p$ breaks down

656: at $p=0$ and there is still a singularity.  In quantum theory,

657: however, the situation is very different: while the Wheeler--DeWitt

658: equation does not solve the singularity problem, the difference

659: equation (\ref{DiffEq}) uniquely evolves a wave function from some

660: initial values at positive $\mu$, say, to negative $\mu$. Thus, the

661: evolution does not break down at the classical singularity and can

662: rather be continued beyond it. Quantum gravity is thus a theory which

663: is more complete than classical general relativity and is free of

664: limitations set by classical singularities.

665: 

666: An intuitive picture of what replaces the classical singularity can be

667: obtained from considering evolution in $\mu$ as before. For negative

668: $\mu$, the volume $V_{\mu}$ decreases with increasing $\mu$ while

669: $V_{\mu}$ increases for positive $\mu$. This leads to the picture of a

670: collapsing universe before it reaches the classical big bang

671: singularity and re-expands. While at large scales the classical

672: description is good \cite{SemiClass}, when the universe is small close

673: to the classical singularity it starts to break down and has to be

674: replaced by discrete quantum geometry. The resulting quantum

675: evolution does not break down, in contrast to the classical space-time

676: picture which dissolves. Using the fact that the sign of $\mu$, which

677: defines the orientation of space, changes during the transition

678: through the classical singularity one can conclude that the universe

679: turns its inside out during the process. This can have consequences

680: for realistic matter Hamiltonians which violate parity symmetry.

681: 

682: \subsubsection{Meaning of the wave function}

683: \label{sec:wavefct}

684: 

685: An important issue in quantum gravity which is still outstanding even

686: in isotropic models is the interpretation of the wave function and its

687: relation to the problem of time. In the usual interpretation of

688: quantum mechanics the wave function determines probabilities for

689: measurements made by an observer outside the quantum system. Quantum

690: gravity and cosmology, however, are thought of as theories for the

691: quantum behavior of a whole universe such that, by definition, there

692: cannot be an observer outside the quantum system. Accordingly, the

693: question of how to interpret the wave function in quantum cosmology is

694: more complicated. One can avoid the separation into a classical and

695: quantum part of the problem in quantum mechanics by the theory of

696: decoherence which can explain how a world perceived as classical

697: emerges from the fundamental quantum description \cite{Decoherence}.

698: The degree of ``classicality'' is related to the number of degrees of

699: freedom which do not contribute significantly to the evolution but

700: interact with the system nonetheless. Averaging over those degrees of

701: freedom, provided there are enough of them, then leads to a classical

702: picture. This demonstrates why macroscopic bodies under usual

703: circumstances are perceived as classical while in the microscopic

704: world, where a small number of degrees of freedom is sufficient to

705: capture crucial properties of a system, quantum mechanics prevails.

706: This idea has been adapted to cosmology, where a large universe comes

707: with many degrees of freedom such as small inhomogeneities which are

708: not of much relevance for the overall evolution. This is different,

709: however, in a small universe where quantum behavior becomes dominant.

710: 

711: Thus, one can avoid the presence of an observer outside the quantum

712: system. The quantum system is described by its wave function, and in

713: some circumstances one can approximate the evolution by a quantum part

714: being looked at by classical observers within the same

715: system. Properties are then encoded in a relational way: the wave

716: function of the whole system contains information about everything

717: including possible observers. Now, the question has shifted from a

718: conceptual one --- how to describe the system if no outside observers

719: can be available --- to a technical one. One needs to understand how

720: information can be extracted from the wave function and used to

721: develop schemes for intuitive pictures or potentially observable

722: effects. This is particularly pressing in the very early universe

723: where everything including what we usually know as space and time are

724: quantum and no familiar background to lean on is left.

725: 

726: One lesson is that evolution should be thought of as relational by

727: determining probabilities for one degree of freedom under the

728: condition that another degree of freedom has a certain value. If the

729: reference degree of freedom (such as the direction of the hand of a

730: clock) plays a distinguished role for determining the evolution of

731: others, it is called internal time: it is not an absolute time outside

732: the quantum system as in quantum mechanics, and not a coordinate time

733: as in general relativity which could be changed by coordinate

734: transformations. Rather, it is one of the physical degrees of freedom

735: whose evolution is determined by the dynamical laws and which shows

736: how other degrees of freedom change by interacting with them. From

737: this picture it is clear that no external observer is necessary to

738: read off the clock or other measurement devices, such that it is

739: ideally suited to cosmology. What is also clear is that now internal

740: time depends on what we choose it to be, and different questions

741: require different choices. For a lab experiment the hand of a clock

742: would be a good internal time and, when the clock is sufficiently

743: isolated from the physical fields used in the experiment and other

744: outside influence, will not be different from an absolute time except

745: that it is mathematically more complicated to describe. The same

746: clock, on the other hand, will not be good for describing the universe

747: when we imagine to approach a classical singularity. It will simply

748: not withstand the extreme physical conditions, dissolve, and its parts

749: will behave in a complicated irregular manner ill-suited for the

750: description of evolution. Instead, one has to use more global objects

751: which depend on what is going on in the whole universe.

752: 

753: Close to a classical singularity, where one expects monotonic

754: expansion or contraction, the spatial volume of the universe is just

755: the right quantity as internal time. A wave function then determines

756: relationally how matter fields or other gravitational degrees of

757: freedom change with respect to the expansion or contraction of the

758: universe. In our case, this is encoded in the wave function

759: $\psi_{\mu}(\phi)$ depending on internal time $\mu$ (which through the

760: volume defines the size of the universe but also spatial orientation)

761: and matter fields $\phi$. By showing that it is subject to a

762: difference equation in $\mu$ which does not stop at the classical

763: singularity $\mu=0$ we have seen that relational probabilities are

764: defined for all internal times without breaking down anywhere. This

765: shows the absence of singularities and allows developing intuitive

766: pictures, but does not make detailed predictions before relational

767: probabilities are indeed computed and shown how to be observable at

768: least in principle. 

769: 

770: Here, we encounter the main issue in the role of the wave function: we

771: have a relational scheme to understand what the wave function should

772: mean but the probability measure to be used, called the physical inner

773: product, is not known so far. We already used a Hilbert space which we

774: needed to define the basic operators and the quantized Hamiltonian

775: constraint, where wave functions $\psi_{\mu}$, which by definition are

776: non-zero for at most countably many values $\mu\in{\mathbb R}$, have

777: the inner product $\langle\psi|\psi'\rangle=

778: \sum_{\mu}\bar{\psi}_{\mu}\psi'_{\mu}$. This is called the kinematical

779: inner product which is used for setting up the quantum theory. But

780: unlike in quantum mechanics where the kinematical inner product is

781: also used as physical inner product for the probability interpretation

782: of the wave function, in quantum gravity the physical inner product

783: must be expected to be different. This occurs because the quantum

784: evolution equation (\ref{DiffEq}) in internal time is a constraint

785: equation rather than an evolution equation in an external absolute

786: time parameter. Solutions to this constraint in general are not

787: normalizable in the kinematical inner product such that a new physical

788: inner product on the solution space has to be found. There are

789: detailed schemes for a derivation, but despite some

790: progress \cite{Golam,IsoSpinFoam} they are difficult to apply even in

791: isotropic cosmological models and research is still ongoing. An

792: alternative route to extract physical statements will be discussed in

793: Sec.~\ref{sec:pheno} together with the main results.

794: 

795: A related issue, which is also of relevance for the classical limit of

796: the theory is that of oscillations on small scales of the wave

797: function. Being subject to a difference equation means that

798: $\psi_{\mu}$ is not necessarily smooth but can change rapidly when

799: $\mu$ changes by a small amount even when the volume is large. In such

800: a regime one expects classical behavior, but small scale oscillations

801: imply that the wave function is sensitive to the Planck scale.  There

802: are also other issues related to the fact that now a difference rather

803: than differential equation provides the fundamental

804: law \cite{FundamentalDisc}. Before the physical inner product is known

805: one cannot say if these oscillations would imply any effect observable

806: today, but one can still study the mathematical problem of if and when

807: solutions with suppressed oscillations exist. This is easy to answer

808: in the affirmative for isotropic models subject to (\ref{DiffEq})

809: where in some cases one even obtains a unique wave

810: function \cite{DynIn,Essay}. However, already in other homogeneous but

811: anisotropic models the issue is much more complicated to

812: analyze \cite{GenFunc,GenFuncBI}.

813: 

814: In a more general situation than homogeneous cosmology there is an

815: additional complication even if the physical inner product would be

816: known. In general, it is very difficult to find an internal time to

817: capture the evolution of a complicated quantum system, which is called

818: the problem of time in general relativity. In cosmology, the volume is

819: a good internal time to understand the singularity, but it would not

820: be good for the whole history if the universe experiences a recollapse

821: where the volume would not be monotonic. This is even more complicated

822: in inhomogeneous situations such as the collapse of matter into a

823: black hole. Since we used internal time $\mu$ to show how quantum

824: geometry evolves through the classical singularity, it seems that the

825: singularity problem in general cannot be solved before the problem of

826: time is understood. Fortunately, while the availability of an internal

827: time simplifies the analysis, requirements on a good choice can be

828: relaxed for the singularity problem. An internal time provides us with

829: an interpretation of the constraint equation as an evolution equation,

830: but the singularity problem can be phrased independently of this as

831: the problem to extend the wave function on the space of metrics or

832: triads. This implies weaker requirements and also situations can be

833: analyzed where no internal time is known. The task then is to find

834: conditions which characterize a classical singularity, analogous to

835: $p=0$ in isotropic cosmology, and find an evolution parameter which at

836: least in individual parts of an inhomogeneous singularity allows to

837: see how the system can move through it. Inhomogeneous cases are now

838: under study but only partially understood so far, such that in the

839: next section we return to isotropic cosmology.

840: 

841: \subsection{Densities}

842: 

843: In the previous discussion we have not yet mentioned the matter

844: Hamiltonian on the right hand side, which diverges classically and in

845: the Wheeler--DeWitt quantization when we reach the singularity. If

846: this were the case here, the discrete quantum evolution would break

847: down, too. However, as we will see now the matter Hamiltonian does not

848: diverge, which is again a consequence of the loop representation.

849: 

850: \subsubsection{Quantization}

851: 

852: For the matter Hamiltonian we need to quantize the matter field and in

853: quantum gravity also coefficients such as $a^{-3}$ in the kinetic term

854: which now become operators, too. In the Wheeler--DeWitt quantization

855: where $a$ is a multiplication operator, $a^{-3}$ is unbounded and

856: diverges at the classical singularity. In loop quantum cosmology we

857: have the basic operator $\hat{p}$ which one can use to construct a

858: quantization of $a^{-3}$. However, a straightforward quantization

859: fails since, as one of the basic properties, $\hat{p}$ has a discrete

860: spectrum containing zero. In this case, there is no densely defined

861: inverse operator which one could use. This seems to indicate that the

862: situation is even worse: an operator for the kinetic term would not

863: only be unbounded but not even be well-defined. The situation is much

864: better, however, when one tries other quantizations which are more

865: indirect. For non-basic operators such as $a^{-3}$ there are usually

866: many ways to quantize, all starting from the same classical

867: expression. What we can do here, suggested by constructions in the

868: full theory \cite{QSDV}, is to rewrite $a^{-3}$ in a classically

869: equivalent way as

870: \[

871:  a^{-3}=(\pi^{-1} G^{-1}{\rm

872:    tr}\tau_3e^{c\tau_3}\{e^{-c\tau_3},\sqrt{V}\})^6 

873: \]

874: where we only need a readily available positive power of $\hat{p}$.

875: Moreover, exponentials of $c$ are basic operators, where we just used

876: su(2) notation $e^{c\tau_3}=\cos\frac{1}{2}c+2\tau_3\sin\frac{1}{2}c$

877: in order to bring the expression closer to what one would have in the

878: full theory, and the Poisson bracket will become a commutator in

879: quantum theory.

880: 

881: This procedure, after taking the trace, leads to a densely defined

882: operator for $a^{-3}$ despite the nonexistence of an inverse of

883: $\hat{p}$ \cite{InvScale}:

884: \begin{equation}

885:  \widehat{a^{-3}} = \left(8i\lP^{-2} (\sin{\textstyle\frac{1}{2}}c

886:    \sqrt{\hat{V}}

887:  \cos{\textstyle\frac{1}{2}}c - \cos{\textstyle\frac{1}{2}}c

888:  \sqrt{\hat{V}} \sin{\textstyle\frac{1}{2}}c)\right)^6\,.

889: \end{equation}

890: That this operator is indeed finite can be seen from its action on

891: states $|\mu\rangle$ which follows from that of the basic operators:

892: \begin{equation}

893:  \widehat{a^{-3}}|\mu\rangle = \left(4\lP^{-2}

894:  (\sqrt{V_{\mu+1}}-\sqrt{V_{\mu-1}}\,)\right)^6 |\mu\rangle

895: \end{equation}

896: immediately showing the eigenvalues which are all finite. In

897: particular, at $\mu=0$ where we would have the classical singularity

898: the density operator does not diverge but is zero.

899: 

900: This finiteness of densities finally confirms the non-singular

901: evolution since the matter Hamiltonian

902: \begin{equation} \label{HmatterQuant}

903:  \hat{H}_{\rm

904:    matter}={\textstyle\frac{1}{2}}\widehat{a^{-3}}\hat{p}_{\phi}^2+ 

905:   \hat{V} V(\phi)

906: \end{equation}

907: in the example of a scalar is well-defined even on the classically

908: singular state $|0\rangle$. The same argument applies for other

909: matter Hamiltonians since only the general structure of kinetic and

910: potential terms is used.

911: 

912: \subsubsection{Confirmation of indications}

913: 

914: The finiteness of the operator is a consequence of the loop

915: representation which forced us to take a detour in quantizing inverse

916: powers of the scale factor. A more physical understanding can be

917: obtained by exploiting the fact that there are quantization

918: ambiguities in this non-basic operator. This comes from the rewriting

919: procedure which is possible in many classically equivalent ways,

920: which all lead to different operators. Important properties such as

921: the finiteness and the approach to the classical limit at large volume

922: are robust under the ambiguities, but finer details can change. The

923: most important choices one can make are selecting the representation

924: $j$ of SU(2) holonomies before taking the trace \cite{Gaul,Ambig} and

925: the power $l$ of $|p|$ in the Poisson bracket \cite{ICGC}. These

926: values are restricted by the requirement that $j$ is a half-integer

927: ($j=1/2$ in the above choice) and $0<l<1$ to obtain a well-defined

928: inverse power of $a$ ($l=3/4$ above). The resulting eigenvalues can be

929: computed explicitly and be approximated by the formula \cite{Ambig,ICGC}

930: \begin{equation} \label{effdens}

931:  (a^{-3})_{\rm eff}= a^{-3} p_l(a^2/a_{\rm

932:   max}^2)^{3/(2-2l)}

933: \end{equation}

934: where $a_{\rm max}=\sqrt{j/3}\,\ell_{\rm P}$ depends on the first

935: ambiguity parameter and the function

936: \begin{eqnarray}

937:  p_l(q) &=&

938: \frac{3}{2l}q^{1-l}\left((l+2)^{-1}

939: \left((q+1)^{l+2}-|q-1|^{l+2}\right)\right. \label{plq}\\

940:  &&\qquad- \left.(l+1)^{-1}q

941: \left((q+1)^{l+1}-{\rm sgn}(q-1)|q-1|^{l+1}\right)\right) \,. \nonumber

942: \end{eqnarray}

943: on the second.

944: 

945: \begin{figure}[th]              

946: \centerline{\includegraphics[width=14cm]{pl.eps}}

947: \vspace*{8pt}

948: \caption{The function $p_l(q)$ in (\ref{plq}) for some values of $l$,

949:   including the limiting cases $l=0$ and $l=1$.

950: \label{pl}}

951: \end{figure}

952: 

953: The function $p_l(q)$, shown in Fig.~\ref{pl}, approaches one for

954: $q\gg1$, has a maximum close to $q=1$ and drops off as $q^{2-l}$ for

955: $q\ll1$. This shows that $(a^{-3})_{\rm eff}$ approaches the classical

956: behavior $a^{-3}$ at large scales $a\gg a_{\rm max}$, has a maximum

957: around $a_{\rm max}$ and falls off like $(a^{-3})_{\rm eff}\sim

958: a^{3/(1-l)}$ for $a\ll a_{\rm max}$. The peak value can be

959: approximated, e.g.\ for $j=1/2$, by $(a^{-3})_{\rm eff}(a_{\rm

960:   max})\sim 3l^{-1}2^{-l}(1-3^{-l})^{3/(2-2l)} \lP^{-3}$ which indeed

961: shows that densities are bounded by inverse powers of the Planck

962: length such that they are finite in quantum gravity but diverge in the

963: classical limit.  This confirms our qualitative expectations from the

964: hydrogen atom, while details of the coefficients depend on the

965: quantization.

966: 

967: Similarly, densities are seen to have a peak at $a_{\rm max}$ whose

968: position is given by the Planck length (and an ambiguity

969: parameter). Above the peak we have the classical behavior of an

970: inverse power, while below the peak the density increases from

971: zero. As suggested by the behavior of radiation in a cavity whose

972: spectral energy density

973: \[

974:  \rho_T(\lambda)=8\pi h\lambda^{-5}(e^{h/kT\lambda}-1)^{-1} = h

975:  \lambda^{-5} f(\lambda/\lambda_{\rm max})

976: \]

977: can, analogously to (\ref{effdens}), be expressed as the diverging

978: behavior $\lambda^{-5}$ multiplied with a cut-off function

979: $f(y)=8\pi/((5/(5-x))^{1/y}-1)$ with $x=5+W(-5e^{-5})$ (in terms of

980: the Lambert function $W(x)$, the inverse function of $xe^x$) and

981: $\lambda_{\rm max}=h/xkT$, we obtain an interpolation between

982: increasing behavior at small scales and decreasing behavior at large

983: scales in such a way that the classical divergence is cut off.

984: 

985: \begin{figure}[th]              

986: \centerline{\includegraphics[width=15cm]{Density.eps}}

987: %\vspace*{8pt}

988: \caption{Comparison between the spectral energy density of black body

989:   radiation (wide curve) and an effective geometrical density with

990:   their large scale approximations (dashed).

991: \label{Density}}

992: \end{figure}

993: 

994: We thus have an interpolation between increasing behavior necessary

995: for negative pressure and inflation and the classical decreasing

996: behavior (Fig.~\ref{Density}). Any matter density turns to increasing

997: behavior at sufficiently small scales without the need to introduce an

998: inflaton field with tailor-made properties. In the following section

999: we will see the implications for cosmological evolution by studying

1000: effective classical equations incorporating this characteristic loop

1001: effect of modified densities at small scales.

1002: 

1003: \subsection{Phenomenology}

1004: \label{sec:pheno}

1005: 

1006: The quantum difference equation (\ref{DiffEq}) is rather complicated

1007: to study in particular in the presence of matter fields and, as

1008: discussed in Sec.~\ref{sec:wavefct}, difficult to interpret in a fully

1009: quantum regime. It is thus helpful to work with effective equations,

1010: comparable conceptually to effective actions in field theories, which

1011: are easier to handle and more familiar to interpret but still show

1012: important quantum effects. This can be done

1013: systematically \cite{Bohr,Perturb,Josh}, starting with the Hamiltonian

1014: constraint operator, resulting in different types of correction terms

1015: whose significance in given regimes can be estimated or studied

1016: numerically \cite{Time}. There are perturbative corrections to the

1017: Friedmann equation of higher order form in $\dot{a}$, or of higher

1018: derivative, in the gravitational part on the left hand side, but also

1019: modifications in the matter Hamiltonian since the density in its

1020: kinetic term behaves differently at small scales. The latter

1021: corrections are mainly non-perturbative since the full expression for

1022: $(a^{-3})_{\rm eff}$ depends on the inverse Planck length, and their

1023: range can be extended if the parameter $j$ is rather large. For these

1024: reasons, those corrections are most important and we focus on them

1025: from now on.

1026: 

1027: The effective Friedmann equation then takes the form

1028: \begin{equation} \label{effFried}

1029:   a\dot{a}^2={\textstyle\frac{8\pi}{3}}G

1030: \left({\textstyle\frac{1}{2}}(a^{-3})_{\rm eff}\, p_{\phi}^2+a^3

1031: V(\phi)\right)

1032: \end{equation}

1033: with $(a^{-3})_{\rm eff}$ as in (\ref{effdens}) with a choice of

1034: ambiguity parameters. Since the matter Hamiltonian does not just act

1035: as a source for the gravitational field on the right hand side of the

1036: Friedmann equation, but also generates Hamiltonian equations of

1037: motion, the modification entails further changes in the matter

1038: equations of motion. The Klein--Gordon equation (\ref{KG}) then takes

1039: the effective form

1040: \begin{equation} \label{effKG}

1041:  \ddot{\phi}=\dot{\phi}\,\dot{a}\frac{\md\log(a^{-3})_{\rm

1042: eff}}{\md a}-a^3(a^{-3})_{\rm eff}V'(\phi)

1043: \end{equation}

1044: and finally there is the Raychaudhuri equation

1045: \begin{equation} \label{effRay}

1046:  \frac{\ddot{a}}{a}= -\frac{8\pi G}{3}\left( a^{-3}d(a)_{\rm

1047: eff}^{-1}\dot{\phi}^2 

1048: \left(1-{\textstyle\frac{1}{4}}a\frac{\md

1049: \log(a^3d(a)_{\rm eff})}{\md a}\right) -V(\phi)\right)

1050: \end{equation}

1051: which follows from the above equation and the continuity equation of

1052: matter.

1053: 

1054: \subsubsection{Bounces}

1055: 

1056: The resulting equations can be studied numerically or with qualitative

1057: analytic techniques. We first note that the right hand side of

1058: (\ref{effFried}) behaves differently at small scales since it

1059: increases with $a$ at fixed $\phi$ and $p_{\phi}$. Viewing this

1060: equation as analogous to a constant energy equation in classical

1061: mechanics with kinetic term $\dot{a}^2$ and potential term ${\cal

1062:   V}(a):=-{\textstyle\frac{8\pi}{3}}G a^{-1}

1063: \left({\textstyle\frac{1}{2}}(a^{-3})_{\rm eff}\, p_{\phi}^2+a^3

1064:   V(\phi)\right)$ illustrates the classically attractive nature of

1065: gravity: The dominant part of this potential behaves like $-a^{-4}$

1066: which is increasing. Treating the scale factor analogously to the

1067: position of a classical particle shows that $a$ will be driven toward

1068: smaller values, implying attraction of matter and energy in the

1069: universe. This changes when we approach smaller scales and take into

1070: account the quantum modification. Below the peak of the effective

1071: density the classical potential ${\cal V}(a)$ will now decrease,

1072: $-{\cal V}(a)$ behaving like a positive power of $a$. This implies

1073: that the scale factor will be repelled away from $a=0$ such that there

1074: is now a small-scale repulsive component to the gravitational force if

1075: we allow for quantum effects. The collapse of matter can then be

1076: prevented if repulsion is taken into account, which indeed can be

1077: observed in some models where the effective classical equations alone

1078: are sufficient to demonstrate singularity-free evolution.

1079: 

1080: This happens by the occurrence of bounces where $a$ turns around from

1081: contracting to expanding behavior. Thus, $\dot{a}=0$ and $\ddot{a}>0$.

1082: The first condition is not always realizable, as follows from the

1083: Friedmann equation (\ref{Friedmann}). In particular, when the scalar

1084: potential is non-negative there is no bounce, which is not changed by

1085: the effective density. There are then two possibilities for bounces in

1086: isotropic models, the first one if space has positive curvature rather

1087: than being flat as assumed here \cite{BounceClosed,BounceQualitative},

1088: the second one with a scalar potential which can become

1089: negative \cite{Oscill,Cyclic}. Both cases allow $\dot{a}=0$

1090: even in the classical case, but this always corresponds to a maximum

1091: rather than minimum. This can easily be seen for the case of negative

1092: potential from the Raychaudhuri equation (\ref{effRay}) which in the

1093: classical case implies negative $\ddot{a}$. With the modification,

1094: however, the additional term in the equation provides a positive

1095: contribution which can become large enough for $\ddot{a}$ to become

1096: positive at a value of $\dot{a}=0$ such that there is a bounce.

1097: 

1098: This provides intuitive explanations for the absence of singularities

1099: from quantum gravity, but not a general one. The generic presence of

1100: bounces depends on details of the model such as its matter content or

1101: which correction terms are being used \cite{EffHam,GenericBounce}, and

1102: even with the effective modifications there are always models which

1103: classically remain singular. Thus, the only general argument for

1104: absence of singularities remains the quantum one based on the

1105: difference equation (\ref{DiffEq}), where the conclusion is model

1106: independent.

1107: 

1108: \subsubsection{Inflation}

1109: 

1110: A repulsive contribution to the gravitational force can not only

1111: explain the absence of singularities, but also enhances the expansion

1112: of the universe on scales close to the classical singularity. Thus, as

1113: seen also in Fig.~\ref{InflAeff} the universe accelerates just from

1114: quantum effects, providing a mechanism for inflation without choosing

1115: special matter.

1116: 

1117: \begin{figure}[th]              

1118: \centerline{\includegraphics[width=16cm]{InflAeff.eps}}

1119: \vspace*{8pt}

1120: \caption{Numerical solution to the effective Friedmann equation

1121:   (\ref{effFried}) with a vanishing scalar potential. While the

1122:   modification in the density on the left is active the expansion is

1123:   accelerated, which stops automatically once the universe expands to

1124:   a size above the peak in the effective density.

1125: \label{InflAeff}}

1126: \end{figure}

1127: 

1128: Via the generation of structure, inflationary phases of the universe

1129: can have an imprint on the observable cosmic microwave background.

1130: Observations imply that the predicted power spectrum of anisotropies

1131: must be nearly independent of the scale on which the anisotropies are

1132: probed, which implies that the inflationary phase responsible for

1133: structure formation must be close to exponential acceleration. This is

1134: true for slow-roll inflation, but also for the inflationary phase

1135: obtained from the effective density once a non-zero scalar potential

1136: is taken into account \cite{GenericInfl}. For more detailed comparisons

1137: between theory and observations one needs to consider how

1138: inhomogeneous fields evolve, which already requires us to relax the

1139: strong symmetry assumption of homogeneity. The necessary methods are

1140: not well-developed at the current stage (see

1141: \cite{SphSymm,Horizon,SphSymmHam} for the basic

1142: formulation), but preliminary calculations of implications on the

1143: power spectrum have been undertaken nonetheless.

1144: Ref.~\cite{PowerLoop} indicates that loop inflation can be

1145: distinguished from simple inflaton models because the power depends

1146: differently on scales. 

1147: 

1148: It turns out that this loop phase alone can provide a sufficient

1149: amount of inflation only for unnatural choices of parameters (such as

1150: extremely large $j$), and those cases are even ruled out by

1151: observations already. At this point, the modified matter dynamics of

1152: (\ref{effKG}) and its $\dot{\phi}$-term becomes important.

1153: Classically, it is a friction term which is used for slow-roll

1154: inflation. But in the modified regime at small scales the sign of the

1155: term changes once $(a^{-3})_{\rm eff}$ is increasing. Thus, at those

1156: small scales friction turns into antifriction and matter is driven up

1157: its potential if it has a non-zero initial momentum (even a tiny one,

1158: e.g., from quantum fluctuations). After the loop phase matter fields

1159: slow down and roll back toward their minima, driving additional

1160: inflation. The potentials need not be very special since structure

1161: formation in the first phase and providing a large universe happen by

1162: different mechanisms. When matter fields reach their minima they start

1163: to oscillate and usual re-heating to obtain hot matter can commence.

1164: 

1165: Loop quantum cosmology thus presents a viable alternative to usual

1166: inflaton models which is not in conflict with current observations

1167: but can be verified or ruled out with future experiments such as the

1168: Planck satellite. Its attractive feature is that it does not require

1169: the introduction of an inflaton field with its special properties, but

1170: provides a fundamental explanation of acceleration from quantum

1171: gravity. This scenario is thus encouraging, but so far has not been

1172: developed to the same extent as inflaton models.

1173: 

1174: Even if we assume the presence of an inflaton field are its properties

1175: less special than in the purely classical treatment. We still need to

1176: assume a potential which is sufficiently flat, but there is now an

1177: explanation of initial values far away from the minimum. For this we

1178: again use the effective Klein--Gordon equation and the fact that

1179: $\phi$ is driven up its potential. One can then check that for usual

1180: inflaton potentials the value of typical initial conditions, as a

1181: function of chosen ambiguity parameters and initial fluctuations of

1182: the scalar, is just what one needs for sufficient inflation in a wide

1183: range \cite{InflationWMAP,Robust}. After the modifications in the

1184: density subside, the inflaton keeps moving up the potential from its

1185: initial push, but is now slowed down by the friction term. Eventually,

1186: it will stop and turn around, entering a slow roll phase in its

1187: approach to the potential minimum.  Thus, the whole history of the

1188: expansion is described by a consistent model as illustrated in

1189: Fig.~\ref{Push}, not just the slow roll phase after the inflaton has

1190: already obtained its large initial values.

1191: 

1192: \begin{figure}

1193: \centerline{\includegraphics[width=17cm]{Push.eps}}

1194: \caption{History of the scale factor (top) and inflaton (bottom) with

1195:   the left hand side in slow motion. (Tics on the right horizontal

1196:   axis mark increments in $t$ by 100.) The upper right data are

1197:    rescaled so as to fit on the same plot. Units for $a$ and $\phi$

1198:    are Planck units, and parameters are not necessarily realistic but

1199:    chosen for plotting purposes. Dashed lines mark

1200:   the time and scale factor where classical behavior of $(a^{-3})_{\rm

1201:     eff}$ starts.

1202: \label{Push}} 

1203: \end{figure}

1204: 

1205: One may think that such a second phase of slow-roll inflation washes

1206: away potential quantum gravity effects from the early expansion. That

1207: this is not necessarily the case has been shown in

1208: \cite{InflationWMAP}, based on the fact that around the turning point

1209: of the inflaton the slow-roll conditions are violated.  In this

1210: scenario, structure we see today on the largest scales was created at

1211: the earliest stages of the second inflationary phase since it was

1212: enlarged by the full inflationary phase. If the second inflationary

1213: regime did not last too long, these scales are just observable today

1214: where in fact the observed loss of power can be taken as an indication

1215: of early violations of slow-roll expansion. Thus, loop quantum

1216: cosmology can provide an explanation, among others, for the

1217: suppression of power on large scales.

1218: 

1219: There are diverse scenarios since different phases of inflation can be

1220: combined, and eventually the right one has to be determined from

1221: observations. One can also combine bounces and inflationary regimes in

1222: order to obtain cyclic universes which eventually reach a long phase

1223: of accelerated expansion \cite{InflOsc}. In particular, this allows the

1224: conctruction of models which start close to a simple, static initial

1225: state and, after a series of cycles, automatically reach values of the

1226: scalar to start inflation. In this way, a semiclassical non-singular

1227: inflationary model \cite{Emergent,Emergent2,EmergentLoop} is formulated

1228: which evades the singularity theorem of \cite{InflSing}.

1229: 

1230: Current observations are already beginning to rule out

1231: certain, very large values of the ambiguity parameter $j$ such that

1232: from future data one can expect much tighter limits. In all

1233: these scenarios the non-perturbative modification of the density is

1234: important, which is a characteristic feature of loop quantum

1235: cosmology. At larger scales above the peak there are also perturbative

1236: corrections which imply small changes in the cosmological expansion

1237: and the evolution of field modes. This has recently been

1238: investigated \cite{PowerPert} with the conclusion that potential

1239: effects on the power spectrum would be too small to be noticed by the

1240: next generation satellites. The best candidates for observable effects

1241: from quantum gravity thus remain the non-perturbative modifications in

1242: effective densities.

1243: 

1244: \section{Conclusions and Outlook}

1245: 

1246: What we have described is a consistent picture of the universe which

1247: is not only observationally viable but also mathematically

1248: well-defined and non-singular. There are instances where quantum

1249: gravity is essential, and others where it is helpful in achieving

1250: important effects. The background independent quantization employed

1251: here is very efficient: There are a few basic properties, such as the

1252: discreteness of spatial geometry and the representation only of

1253: exponentials of curvature, which are behind a variety of applications.

1254: Throughout all the developments, those properties have been known to

1255: be essential for mathematical consistency before they were recognized

1256: as being responsible for physical phenomena.

1257: 

1258: For instance, for the singularity issue the basic properties are all

1259: needed in the way they turn out to be realized. First, the theory had

1260: to be based on densitized triad variables which now not only provides

1261: us with the sign of orientation, and thus two sides of the classical

1262: singularity, but also in more complicated models positions the

1263: classical singularity in phase space such that it becomes accessible

1264: by quantum evolution.  Then, the discreteness of spatial geometry

1265: encoded in triad operators and the representation of exponentials of

1266: curvature play together in the right way to remove divergences in

1267: densities and extend the quantum evolution through the classical

1268: singularity. These features allow general results about the absence of

1269: singularities without any new or artificial ingredients, and lead to a

1270: natural solution of a long-standing problem which has eluded previous

1271: attempts for decades.  Symmetry assumptions are still important in

1272: order to be able to perform the calculations, but they can now be

1273: weakened considerably and are not responsible for physical

1274: implications. The essential step is to base the symmetry reduction on

1275: a candidate for full quantum gravity which is background independent

1276: so as to allow studying quantum geometry purely.

1277: 

1278: Absence of singularities in this context is a rather general

1279: statement about the possibility to extend a quantum wave function

1280: through a regime which classically would appear as a singularity. More

1281: explicit questions, such as what kind of new region one is evolving to

1282: and whether it again becomes classical or retains traces of the

1283: evolution through a quantum regime, depend on details of the relevant

1284: constraint operators. This includes, for instance, quantization

1285: ambiguities and the question whether a symmetric operator has to

1286: be used. The latter aspect is also important for technical concepts

1287: such as a physical inner product.

1288: 

1289: Here we discussed only isotropic models which are classically

1290: described solely by the scale factor determining the size of space.

1291: But a more realistic situation has to take into account also the shape

1292: of space, and changes of the distribution of geometry and matter

1293: between different points of space. The methods we used have been

1294: extended to homogeneous models, allowing for anisotropic spaces, and

1295: recently to some inhomogeneous ones, defined by spherical symmetry and

1296: some forms of cylindrical symmetry. In all cases, essential aspects of

1297: the general mechanism for removing classical singularities which has

1298: first been seen only in the simple isotropic models are known to be

1299: realized.\footnote{This does not refer to the boundedness of densities

1300:   or curvature components for {\em all geometries}, which is known not

1301:   to be present in anisotropic models \cite{HomCosmo,Spin} or even on

1302:   some degenerate configurations in the full theory \cite{BoundFull}.

1303:   What is relevant is the behavior on configurations seen along the

1304:   dynamical evolution.}  Moreover, in the more complicated systems it

1305: is acting much more non-trivially, again with the right ingredients

1306: provided by the background independent quantization. Nevertheless,

1307: since the inhomogeneous constraints are much more complicated to

1308: analyze, absence of singularities for them has not yet been proven

1309: completely.  The inhomogeneous systems now also allow access to black

1310: hole and gravitational wave models such that their quantum geometry

1311: can be studied, too.

1312: 

1313: Effective equations are a useful tool to study quantum effects in a

1314: more familiar setting given by classical equations of motion. They

1315: show diverse effects whose usefulness in cosmological phenomenology is

1316: often surprising. Also here, the effects were known to occur from the

1317: quantization and the transfer into effective classical equations,

1318: before they turned out to be helpful. In addition to inflationary

1319: scenarios and bounces which one can see in isotropic cosmologies,

1320: modified densities have more implications in less symmetric models.

1321: The anisotropic Bianchi IX model, for instance, is classically chaotic

1322: which is assumed to play a role in the complicated approach to a

1323: classical singularity \cite{BKL}. With the effective modifications the

1324: dynamics changes and simplifies, removing the classical

1325: chaos \cite{NonChaos}. This has implications for the effective approach

1326: to a classical singularity and can provide a more consistent picture

1327: of general singularities \cite{ChaosLQC}. Effective classical equations

1328: can also be used to study the collapse of matter to a black hole, with

1329: modifications in the development of classical singularities and

1330: horizons \cite{Collapse}. This can now also be studied with

1331: inhomogeneous quantum models which allow new applications for black

1332: holes and cosmological phenomenology where the evolution of

1333: inhomogeneities is of interest in the context of structure formation.

1334: 

1335: With these models there will be new effects not just in cosmology but

1336: also for black holes and other systems which further check the overall

1337: consistency of the theory. Moreover, a better understanding of

1338: inhomogeneities evolving in a cosmological background will give us a

1339: much better computational handle on signatures in the cosmic microwave

1340: or even gravitational wave background, which may soon be testable with

1341: a new generation of observations. One may wonder how it can be

1342: possible to observe quantum gravity effects, given that the Planck

1343: scale is so many orders of magnitude away from scales accessible by

1344: today's experiments. The difference in scales, however, does not

1345: preclude the observation of indirect effects even though direct

1346: measurements on the discreteness scale are impossible, as illustrated

1347: by a well-known example: Brownian motion allows to draw conclusions

1348: about the atomic structure of matter and its size by observations on

1349: much larger scales \cite{Brownian}. Similarly, cosmological

1350: observations can carry information on quantum gravity effects which

1351: otherwise would manifest themselves only at the Planck scale.

1352: 

1353: %\bibliographystyle{../preprint}

1354: %\bibliography{../Bib/QuantGra}

1355: 

1356: \begin{thebibliography}{10}

1357: 

1358: \bibitem{Guth}

1359: A.~H.\ Guth,

1360: \newblock The inflationary universe: A possible solution to the horizon and

1361:   flatness problems,

1362: \newblock {\em Phys.\ Rev.\ D} 23 (1981) 347--356

1363: 

1364: \bibitem{NewInfl}

1365: A.~D.\ Linde,

1366: \newblock A new inflationary universe scenario: A possible solution of the

1367:   horizon, flatness, homogeneity, isotropy and primordial monopole problems,

1368: \newblock {\em Phys.\ Lett.\ B} 108 (1982) 389--393

1369: 

1370: \bibitem{InflAS}

1371: A.\ Albrecht and P.~J.\ Steinhardt,

1372: \newblock Cosmology for grand unified theories with radiatively induced

1373:   symmetry breaking,

1374: \newblock {\em Phys.\ Rev.\ Lett.} 48 (1982) 1220--1223

1375: 

1376: \bibitem{WMAPParam}

1377: D.~N.\ Spergel et~al.,

1378: \newblock First year Wilkinson Microwave Anisotropy Probe (WMAP) observations:

1379:   Determination of cosmological parameters,

1380: \newblock {\em Astrophys.\ J.\ Suppl.} 148 (2003) 175, [astro-ph/0302209]

1381: 

1382: \bibitem{SingTheo}

1383: S.~W.\ Hawking and R.\ Penrose,

1384: \newblock The singularities of gravitational collapse and cosmology,

1385: \newblock {\em Proc.\ Roy.\ Soc.\ Lond.\ A} 314 (1970) 529--548

1386: 

1387: \bibitem{Friedmann}

1388: A.~Friedmann,

1389: \newblock \"Uber die Kr\"ummung des Raumes,

1390: \newblock {\em Z.\ Phys.} 10 (1922) 377--386

1391: 

1392: \bibitem{ChaoticInfl}

1393: A.~D.\ Linde,

1394: \newblock Chaotic Inflation,

1395: \newblock {\em Phys.\ Lett.\ B} 129 (1983) 177--181

1396: 

1397: \bibitem{InflSing}

1398: A.\ Borde, A.~H.\ Guth, and A.\ Vilenkin,

1399: \newblock Inflationary spacetimes are not past-complete,

1400: \newblock {\em Phys.\ Rev.\ Lett.} 90 (2003) 151301, [gr-qc/0110012]

1401: 

1402: \bibitem{DeWitt}

1403: B.~S.\ DeWitt,

1404: \newblock Quantum Theory of Gravity. I. The Canonical Theory,

1405: \newblock {\em Phys.\ Rev.} 160 (1967) 1113--1148

1406: 

1407: \bibitem{QCReview}

1408: D.~L.\ Wiltshire,

1409: \newblock An introduction to quantum cosmology,

1410: \newblock In B.\ Robson, N.\ Visvanathan, and W.~S.\ Woolcock, editors, {\em

1411:   Cosmology: The Physics of the Universe}, pages 473--531. World Scientific,

1412:   Singapore, 1996, [gr-qc/0101003]

1413: 

1414: \bibitem{tunneling}

1415: A.\ Vilenkin,

1416: \newblock Quantum creation of universes,

1417: \newblock {\em Phys.\ Rev.\ D} 30 (1984) 509--511

1418: 

1419: \bibitem{nobound}

1420: J.~B.\ Hartle and S.~W.\ Hawking,

1421: \newblock Wave function of the Universe,

1422: \newblock {\em Phys.\ Rev.\ D} 28 (1983) 2960--2975

1423: 

1424: \bibitem{Rov:Loops}

1425: C.\ Rovelli,

1426: \newblock Loop Quantum Gravity,

1427: \newblock {\em Living Reviews in Relativity} 1 (1998) 1, [gr-qc/9710008],

1428: \newblock http://www.livingreviews.org/Articles/Volume1/1998-1rovelli

1429: 

1430: \bibitem{ThomasRev}

1431: T.\ Thiemann,

1432: \newblock Introduction to Modern Canonical Quantum General Relativity,

1433:   [gr-qc/0110034]

1434: 

1435: \bibitem{ALRev}

1436: A.\ Ashtekar and J.\ Lewandowski,

1437: \newblock Background independent quantum gravity: A status report,

1438: \newblock {\em Class.\ Quantum Grav.} 21 (2004) R53--R152, [gr-qc/0404018]

1439: 

1440: \bibitem{AshVar}

1441: A.\ Ashtekar,

1442: \newblock New Hamiltonian Formulation of General Relativity,

1443: \newblock {\em Phys.\ Rev.\ D} 36 (1987) 1587--1602

1444: 

1445: \bibitem{AshVarReell}

1446: J.~F.\ Barbero~G.,

1447: \newblock Real Ashtekar Variables for Lorentzian Signature Space-Times,

1448: \newblock {\em Phys.\ Rev.\ D} 51 (1995) 5507--5510, [gr-qc/9410014]

1449: 

1450: \bibitem{LoopRep}

1451: C.\ Rovelli and L.\ Smolin,

1452: \newblock Loop Space Representation of Quantum General Relativity,

1453: \newblock {\em Nucl.\ Phys.\ B} 331 (1990) 80--152

1454: 

1455: \bibitem{ALMMT}

1456: A.\ Ashtekar, J.\ Lewandowski, D.\ Marolf, J.\ Mour\~ao, and T.\ Thiemann,

1457: \newblock Quantization of Diffeomorphism Invariant Theories of Connections with

1458:   Local Degrees of Freedom,

1459: \newblock {\em J.\ Math.\ Phys.} 36 (1995) 6456--6493, [gr-qc/9504018]

1460: 

1461: \bibitem{RS:Spinnet}

1462: C.\ Rovelli and L.\ Smolin,

1463: \newblock Spin Networks and Quantum Gravity,

1464: \newblock {\em Phys.\ Rev.\ D} 52 (1995) 5743--5759

1465: 

1466: \bibitem{FluxAlg}

1467: H.\ Sahlmann,

1468: \newblock Some Comments on the Representation Theory of the Algebra Underlying

1469:   Loop Quantum Gravity, [gr-qc/0207111]

1470: 

1471: \bibitem{Meas}

1472: H.\ Sahlmann,

1473: \newblock When Do Measures on the Space of Connections Support the Triad

1474:   Operators of Loop Quantum Gravity?, [gr-qc/0207112]

1475: 

1476: \bibitem{HolFluxRep}

1477: A.\ Okolow and J.\ Lewandowski,

1478: \newblock Diffeomorphism covariant representations of the holonomy-flux

1479:   star-algebra,

1480: \newblock {\em Class.\ Quantum Grav.} 20 (2003) 3543--3568, [gr-qc/0302059]

1481: 

1482: \bibitem{SuperSel}

1483: H.\ Sahlmann and T.\ Thiemann,

1484: \newblock On the superselection theory of the Weyl algebra for diffeomorphism

1485:   invariant quantum gauge theories, [gr-qc/0302090]

1486: 

1487: \bibitem{WeylRep}

1488: C.\ Fleischhack,

1489: \newblock Representations of the Weyl Algebra in Quantum Geometry,

1490:   [math-ph/0407006]

1491: 

1492: \bibitem{AreaVol}

1493: C.\ Rovelli and L.\ Smolin,

1494: \newblock Discreteness of Area and Volume in Quantum Gravity,

1495: \newblock {\em Nucl.\ Phys.\ B} 442 (1995) 593--619, [gr-qc/9411005],

1496: \newblock Erratum: {\em Nucl.\ Phys.\ B} 456 (1995) 753

1497: 

1498: \bibitem{Area}

1499: A.\ Ashtekar and J.\ Lewandowski,

1500: \newblock Quantum Theory of Geometry I: Area Operators,

1501: \newblock {\em Class.\ Quantum Grav.} 14 (1997) A55--A82, [gr-qc/9602046]

1502: 

1503: \bibitem{Vol2}

1504: A.\ Ashtekar and J.\ Lewandowski,

1505: \newblock Quantum Theory of Geometry II: Volume Operators,

1506: \newblock {\em Adv.\ Theor.\ Math.\ Phys.} 1 (1997) 388--429, [gr-qc/9711031]

1507: 

1508: \bibitem{QSDI}

1509: T.\ Thiemann,

1510: \newblock Quantum Spin Dynamics {(QSD)},

1511: \newblock {\em Class.\ Quantum Grav.} 15 (1998) 839--873, [gr-qc/9606089]

1512: 

1513: \bibitem{QSDV}

1514: T.\ Thiemann,

1515: \newblock {QSD V}: Quantum Gravity as the Natural Regulator of Matter Quantum

1516:   Field Theories,

1517: \newblock {\em Class.\ Quantum Grav.} 15 (1998) 1281--1314, [gr-qc/9705019]

1518: 

1519: \bibitem{PhD}

1520: M.\ Bojowald,

1521: \newblock {\em Quantum Geometry and Symmetry},

1522: \newblock PhD thesis, RWTH Aachen, 2000,

1523: \newblock published by Shaker-Verlag, Aachen

1524: 

1525: \bibitem{SymmRed}

1526: M.\ Bojowald and H.~A.\ Kastrup,

1527: \newblock Symmetry Reduction for Quantized Diffeomorphism Invariant Theories of

1528:   Connections,

1529: \newblock {\em Class.\ Quantum Grav.} 17 (2000) 3009--3043, [hep-th/9907042]

1530: 

1531: \bibitem{SphSymm}

1532: M.\ Bojowald,

1533: \newblock Spherically Symmetric Quantum Geometry: States and Basic Operators,

1534: \newblock {\em Class.\ Quantum Grav.} 21 (2004) 3733--3753, [gr-qc/0407017]

1535: 

1536: \bibitem{Bohr}

1537: A.\ Ashtekar, M.\ Bojowald, and J.\ Lewandowski,

1538: \newblock Mathematical structure of loop quantum cosmology,

1539: \newblock {\em Adv.\ Theor.\ Math.\ Phys.} 7 (2003) 233--268, [gr-qc/0304074]

1540: 

1541: \bibitem{PolymerParticle}

1542: A.\ Ashtekar, S.\ Fairhurst, and J.\ Willis,

1543: \newblock Quantum Gravity, Shadow States, and Quantum Mechanics,

1544: \newblock {\em Class.\ Quant.\ Grav.} 20 (2003) 1031--1062, [gr-qc/0207106]

1545: 

1546: \bibitem{QEDBohr}

1547: W.\ Thirring and H.\ Narnhofer,

1548: \newblock Covariant QED without Infinite Metrics,

1549: \newblock {\em Rev.\ Math.\ Phys.} SI1 (1992) 197--211

1550: 

1551: \bibitem{LoopCosRev} 

1552: M.\ Bojowald and H.~A.\ Morales-T\'ecotl,

1553:   \newblock Cosmological applications of loop quantum gravity,

1554:   \newblock In {\em Proceedings of the Fifth Mexican School (DGFM):

1555:     The Early Universe and Observational Cosmology},

1556:   {\em Lect.\ Notes Phys.} {\bf 646} (2004), 421--462

1557:   (Springer-Verlag), [gr-qc/0306008]

1558: 

1559: \bibitem{IsoCosmo}

1560: M.\ Bojowald,

1561: \newblock Isotropic Loop Quantum Cosmology,

1562: \newblock {\em Class.\ Quantum Grav.} 19 (2002) 2717--2741, [gr-qc/0202077]

1563: 

1564: \bibitem{SemiClass}

1565: M.\ Bojowald,

1566: \newblock The Semiclassical Limit of Loop Quantum Cosmology,

1567: \newblock {\em Class.\ Quantum Grav.} 18 (2001) L109--L116, [gr-qc/0105113]

1568: 

1569: \bibitem{Decoherence}

1570: D.\ Giulini, C.\ Kiefer, E.\ Joos, J.\ Kupsch, I.~O.\ Stamatescu, and H.~D.\

1571:   Zeh,

1572: \newblock {\em Decoherence and the Appearance of a Classical World in Quantum

1573:   Theory},

1574: \newblock Springer, Berlin, Germany, 1996

1575: 

1576: \bibitem{Golam}

1577: G.~M.\ Hossain,

1578: \newblock Hubble operator in isotropic loop quantum cosmology,

1579: \newblock {\em Class.\ Quantum Grav.} 21 (2004) 179--196, [gr-qc/0308014]

1580: 

1581: \bibitem{IsoSpinFoam}

1582: K.\ Noui, A.\ Perez, and K.\ Vandersloot,

1583: \newblock On the Physical Hilbert Space of Loop Quantum Cosmology,

1584: \newblock {\em Phys.\ Rev.\ D} 71 (2005) 044025, [gr-qc/0411039]

1585: 

1586: \bibitem{FundamentalDisc}

1587: M.\ Bojowald and G.\ Date,

1588: \newblock Consistency conditions for fundamentally discrete theories,

1589: \newblock {\em Class.\ Quantum Grav.} 21 (2004) 121--143, [gr-qc/0307083]

1590: 

1591: \bibitem{DynIn}

1592: M.\ Bojowald,

1593: \newblock Dynamical Initial Conditions in Quantum Cosmology,

1594: \newblock {\em Phys.\ Rev.\ Lett.} 87 (2001) 121301, [gr-qc/0104072]

1595: 

1596: \bibitem{Essay}

1597: M.\ Bojowald,

1598: \newblock Initial Conditions for a Universe,

1599: \newblock {\em Gen.\ Rel.\ Grav.} 35 (2003) 1877--1883, [gr-qc/0305069]

1600: 

1601: \bibitem{GenFunc}

1602: D.\ Cartin, G.\ Khanna, and M.\ Bojowald,

1603: \newblock Generating function techniques for loop quantum cosmology,

1604: \newblock {\em Class.\ Quantum Grav.} 21 (2004) 4495--4509, [gr-qc/0405126]

1605: 

1606: \bibitem{GenFuncBI}

1607: D.\ Cartin and G.\ Khanna,

1608: \newblock Absence of pre-classical solutions in Bianchi I loop quantum

1609:   cosmology,

1610: \newblock {\em Phys.\ Rev.\ Lett.} 94 (2005) 111302, [gr-qc/0501016]

1611: 

1612: \bibitem{InvScale}

1613: M.\ Bojowald,

1614: \newblock Inverse Scale Factor in Isotropic Quantum Geometry,

1615: \newblock {\em Phys.\ Rev.\ D} 64 (2001) 084018, [gr-qc/0105067]

1616: 

1617: \bibitem{Gaul}

1618: M.\ Gaul and C.\ Rovelli,

1619: \newblock A generalized Hamiltonian Constraint Operator in Loop Quantum Gravity

1620:   and its simplest Euclidean Matrix Elements,

1621: \newblock {\em Class.\ Quantum Grav.} 18 (2001) 1593--1624, [gr-qc/0011106]

1622: 

1623: \bibitem{Ambig}

1624: M.\ Bojowald,

1625: \newblock Quantization ambiguities in isotropic quantum geometry,

1626: \newblock {\em Class.\ Quantum Grav.} 19 (2002) 5113--5130, [gr-qc/0206053]

1627: 

1628: \bibitem{ICGC}

1629: M.\ Bojowald,

1630: \newblock Loop Quantum Cosmology: Recent Progress,

1631: \newblock In {\em Proceedings of the International Conference on Gravitation

1632:   and Cosmology (ICGC 2004), Cochin, India},

1633:   {\em Pramana} {\bf 63} (2004) 765--776, [gr-qc/0402053]

1634: 

1635: \bibitem{Perturb}

1636: A.\ Ashtekar, M.\ Bojowald, and J.\ Willis, in preparation

1637: 

1638: \bibitem{Josh}

1639: J.\ Willis,

1640: \newblock {\em On the Low-Energy Ramifications and a Mathematical Extension of

1641:   Loop Quantum Gravity},

1642: \newblock PhD thesis, The Pennsylvania State University, 2004

1643: 

1644: \bibitem{Time}

1645: M.\ Bojowald, P.\ Singh, and A.\ Skirzewski,

1646: \newblock Time dependence in quantum gravity,

1647: \newblock {\em Phys.\ Rev.\ D} 70 (2004) 124022, [gr-qc/0408094]

1648: 

1649: \bibitem{BounceClosed}

1650: P.\ Singh and A.\ Toporensky,

1651: \newblock Big Crunch Avoidance in ${\rm k}=1$ Semi-Classical Loop Quantum

1652:   Cosmology,

1653: \newblock {\em Phys.\ Rev.\ D} 69 (2004) 104008, [gr-qc/0312110]

1654: 

1655: \bibitem{BounceQualitative}

1656: G.~V.\ Vereshchagin,

1657: \newblock Qualitative Approach to Semi-Classical Loop Quantum Cosmology,

1658: \newblock {\em JCAP} 07 (2004) 013, [gr-qc/0406108]

1659: 

1660: \bibitem{Oscill}

1661: J.~E.\ Lidsey, D.~J.\ Mulryne, N.~J.\ Nunes, and R.\ Tavakol,

1662: \newblock Oscillatory Universes in Loop Quantum Cosmology and Initial

1663:   Conditions for Inflation,

1664: \newblock {\em Phys.\ Rev.\ D} 70 (2004) 063521, [gr-qc/0406042]

1665: 

1666: \bibitem{Cyclic}

1667: M.\ Bojowald, R.\ Maartens, and P.\ Singh,

1668: \newblock Loop Quantum Gravity and the Cyclic Universe,

1669: \newblock {\em Phys.\ Rev.\ D} 70 (2004) 083517, [hep-th/0407115]

1670: 

1671: \bibitem{EffHam}

1672: G.\ Date and G.~M.\ Hossain,

1673: \newblock Effective Hamiltonian for Isotropic Loop Quantum Cosmology,

1674: \newblock {\em Class.\ Quantum Grav.} 21 (2004) 4941--4953, [gr-qc/0407073]

1675: 

1676: \bibitem{GenericBounce}

1677: G.\ Date and G.~M.\ Hossain,

1678: \newblock Genericity of Big Bounce in isotropic loop quantum cosmology,

1679: \newblock {\em Phys.\ Rev.\ Lett.} 94 (2005) 011302, [gr-qc/0407074]

1680: 

1681: \bibitem{GenericInfl}

1682: G.\ Date and G.~M.\ Hossain,

1683: \newblock Genericity of inflation in isotropic loop quantum cosmology,

1684: \newblock {\em Phys.\ Rev.\ Lett.} 94 (2005) 011301, [gr-qc/0407069]

1685: 

1686: \bibitem{Horizon}

1687: M.\ Bojowald and R.\ Swiderski,

1688: \newblock Spherically Symmetric Quantum Horizons,

1689: \newblock {\em Phys.\ Rev.\ D} 71 (2005) 081501(R), [gr-qc/0410147]

1690: 

1691: \bibitem{SphSymmHam}

1692: M.\ Bojowald and R.\ Swiderski,

1693: \newblock Spherically Symmetric Quantum Geometry: Hamiltonian Constraint,

1694: \newblock to appear

1695: 

1696: \bibitem{PowerLoop}

1697: G.~M.\ Hossain,

1698: \newblock Primordial Density Perturbation in Effective Loop Quantum Cosmology,

1699: \newblock {\em Class.\ Quantum Grav.} 22 (2005) (to appear), [gr-qc/0411012]

1700: 

1701: \bibitem{InflationWMAP}

1702: S.\ Tsujikawa, P.\ Singh, and R.\ Maartens,

1703: \newblock Loop quantum gravity effects on inflation and the CMB,

1704: \newblock {\em Class.\ Quantum Grav.} 21 (2004) 5767--5775, [astro-ph/0311015]

1705: 

1706: \bibitem{Robust}

1707: M.\ Bojowald, J.~E.\ Lidsey, D.~J.\ Mulryne, P.\ Singh, and R.\ Tavakol,

1708: \newblock Inflationary Cosmology and Quantization Ambiguities in Semi-Classical

1709:   Loop Quantum Gravity,

1710: \newblock {\em Phys.\ Rev.\ D} 70 (2004) 043530, [gr-qc/0403106]

1711: 

1712: \bibitem{InflOsc}

1713: D.~J.\ Mulryne, N.~J.\ Nunes, R.\ Tavakol, and J.\ Lidsey,

1714: \newblock Inflationary Cosmology and Oscillating Universes in Loop Quantum

1715:   Cosmology,

1716: \newblock {\em Int.\ J.\ Mod.\ Phys.\ A} (2004) to appear, [gr-qc/0411125]

1717: 

1718: \bibitem{Emergent}

1719: G.~F.~R.\ Ellis and R.\ Maartens,

1720: \newblock The Emergent Universe: inflationary cosmology with no singularity,

1721: \newblock {\em Class.\ Quant.\ Grav.} 21 (2004) 223--232, [gr-qc/0211082]

1722: 

1723: \bibitem{Emergent2}

1724: G.~F.~R.\ Ellis, J.\ Murugan, and C.~G.\ Tsagas,

1725: \newblock The Emergent Universe: An Explicit Construction,

1726: \newblock {\em Class.\ Quant.\ Grav.} 21 (2004) 233--250, [gr-qc/0307112]

1727: 

1728: \bibitem{EmergentLoop}

1729: D.~J.\ Mulryne, R.\ Tavakol, J.~E.\ Lidsey, and G.~F.~R.\ Ellis,

1730: \newblock An emergent universe from a loop, [astro-ph/0502589]

1731: 

1732: \bibitem{PowerPert}

1733: S.\ Hofmann and O.\ Winkler,

1734: \newblock The Spectrum of Fluctuations in Inflationary Quantum Cosmology,

1735:   [astro-ph/0411124]

1736: 

1737: \bibitem{HomCosmo}

1738: M.\ Bojowald,

1739: \newblock Homogeneous loop quantum cosmology,

1740: \newblock {\em Class.\ Quantum Grav.} 20 (2003) 2595--2615, [gr-qc/0303073]

1741: 

1742: \bibitem{Spin}

1743: M.\ Bojowald, G.\ Date, and K.\ Vandersloot,

1744: \newblock Homogeneous loop quantum cosmology: The role of the spin connection,

1745: \newblock {\em Class.\ Quantum Grav.} 21 (2004) 1253--1278, [gr-qc/0311004]

1746: 

1747: \bibitem{BoundFull}

1748: J.\ Brunnemann and T.\ Thiemann,

1749: \newblock Unboundedness of Triad-Like Operators in Loop Quantum Gravity,

1750:   [gr-qc/0505033]

1751: 

1752: \bibitem{BKL}

1753: V.~A.\ Belinskii, I.~M.\ Khalatnikov, and E.~M.\ Lifschitz,

1754: \newblock A general solution of the Einstein equations with a time singularity,

1755: \newblock {\em Adv.\ Phys.} 13 (1982) 639--667

1756: 

1757: \bibitem{NonChaos}

1758: M.\ Bojowald and G.\ Date,

1759: \newblock Quantum suppression of the generic chaotic behavior close to

1760:   cosmological singularities,

1761: \newblock {\em Phys.\ Rev.\ Lett.} 92 (2004) 071302, [gr-qc/0311003]

1762: 

1763: \bibitem{ChaosLQC}

1764: M.\ Bojowald, G.\ Date, and G.~M.\ Hossain,

1765: \newblock The Bianchi IX model in loop quantum cosmology,

1766: \newblock {\em Class.\ Quantum Grav.} 21 (2004) 3541--3569, [gr-qc/0404039]

1767: 

1768: \bibitem{Collapse}

1769: M.\ Bojowald, R.\ Goswami, R.\ Maartens, and P.\ Singh,

1770: \newblock A black hole mass threshold from non-singular quantum gravitational

1771:   collapse, [gr-qc/0503041]

1772: 

1773: \bibitem{Brownian}

1774: A.\ Einstein,

1775: \newblock \"Uber die von der molekularkinetischen Theorie der W\"arme

1776:   geforderte Bewegung von in ruhenden Fl\"ussigkeiten suspendierten Teilchen,

1777: \newblock {\em Annalen Phys.} 17 (1905) 549--560

1778: 

1779: \end{thebibliography}

1780: 

1781: 

1782: \end{document}