1: % PACS number: 0460P, 0470D

2: % Keywords: Quantum Geometry, Loop Quantum Cosmology

3: %\documentstyle[psfrag,preprint,graphicx,tighten,eqsecnum,floats,aps,amssymb]{revtex}

4: %\documentstyle[psfrag,preprint,graphicx,tighten,floats,aps]{revtex}

5: \documentclass[preprint,showpacs,preprintnumbers,amsmath,amssymb,nofootinbib]{revtex4}

6: \usepackage{amssymb}

7: \usepackage{graphicx}

8: \usepackage[mathscr]{eucal}

9: 

10: \def\be{\begin{equation}}

11: \def\ee{\end{equation}}

12: \def\ba{\begin{eqnarray}}

13: \def\ea{\end{eqnarray}}

14: \def\P{{\cal P}}

15: \def\A{{\cal A}}

16: \def\O{{\cal O}}

17: \def\T{{\cal T}}

18: \def\H{{\cal H}}

19: \def\Hb{\bar{\cal H}}

20: \def\G{{\cal G}}

21: 

22: \def\const{{\rm const}}

23: \def\Cyl{{\rm Cyl}}

24: \def\cyl{\Cyl}

25: \def\su{{\rm su}}

26: \def\SU{{\rm SU}}

27: 

28: \def\phys{{\rm phys}}

29: 

30: \def\dual{{\Cyl}^\star}

31: \def\a{\alpha}

32: \def\g{\gamma}

33: \def\e{\epsilon}

34: \def\k{\kappa}

35: \def\t{\triangle}

36: \def\lp{{\ell}_{\rm Pl}}

37: \def\Real{{\mathbb R}}

38: \def\Comp{{\mathbb C}}

39: \def\Nat{\mathbb N}

40: 

41: \def\q{{}^o\!q}

42: \def\e{{}^o\!e}

43: \def\w{{}^o\!\omega}

44: \newcommand{\ip}[2]{{\langle#1\,|\,#2\rangle}}

45: 

46: \def\b{$\bullet$}

47: \def\ni{\noindent}

48: \def\f{\frac}

49: 

50: \newcommand\scri{\mathscr{I}}

51: \newcommand{\md}{{\mathrm d}}

52: \newcommand{\vt}{\vartheta}

53: \newcommand{\vp}{\varphi}

54: \newcommand{\R}{{\mathbb R}}

55: \newcommand{\diag}{{\mathrm{diag}}}

56: \newcommand{\sgn}{{\mathrm{sgn}}}

57: 

58: \begin{document}

59: 

60: \preprint{IGPG04/8-4, AEI--2004--072}

61: \title{Black hole evaporation: A paradigm}

62: \author{Abhay\ Ashtekar${}^{1}$ and Martin Bojowald${}^{2,1}$}

63: \affiliation{1. Institute for Gravitational Physics and Geometry,\\

64: Physics Department, Penn State, University Park, PA 16802, USA\\

65: 2. Max-Planck-Institut f\"ur Gravitationsphysik,

66: Albert-Einstein-Institut, Am M\"uhlenberg 1, D-14476 Potsdam,

67: Germany}

68: 

69: \begin{abstract}

70: 

71: A paradigm describing black hole evaporation in non-perturbative

72: quantum gravity is developed by combining two sets of detailed

73: results: i) resolution of the Schwarzschild singularity using

74: quantum geometry methods \cite{m,ab1}; and ii) time-evolution of

75: black holes in the trapping and dynamical horizon frameworks

76: \cite{sh1,sh2,ak1,ak2}. Quantum geometry effects introduce a major

77: modification in the traditional space-time diagram of black hole

78: evaporation, providing a possible mechanism for recovery of

79: information that is classically lost in the process of black hole

80: formation. The paradigm is developed directly in the Lorentzian

81: regime and necessary conditions for its viability are discussed.

82: If these conditions are met, much of the tension between

83: expectations based on space-time geometry and structure of quantum

84: theory would be resolved.

85: \end{abstract}

86: 

87: \pacs{0460P, 0470D}

88: 

89: \maketitle

90: 

91: 

92: 

93: \section{Introduction}

94: \label{s1}

95: 

96: In classical general relativity, a rich variety of initial data on

97: past null infinity, ${\scri^-}$, can lead to the formation of

98: a black hole.%

99: %

100: \footnote{For simplicity of discussion, in this article we will

101: consider only zero rest mass matter fields and assume that past

102: null infinity is a good initial value surface. To include massive

103: fields, one can suitably modify our discussion by adjoining past

104: (future) time-like infinity to past (future) null infinity.}

105: %

106: Once it is formed, space-time develops a new, future boundary at

107: the singularity, whence one can not reconstruct the geometry and

108: matter fields by evolving the data \emph{backward} from future

109: null infinity, ${\scri^+}$. Thus, whereas an appropriately chosen

110: family of observers near ${\scri}^-$ has full information needed

111: to construct the entire space-time, no family of observers near

112: ${\scri^+}$ has such complete information. In this sense, the

113: classical theory of black hole formation leads to information

114: loss. Note that, contrary to the heuristics often invoked (see,

115: e.g. \cite{swh2}), this phenomenon is not directly related to

116: black hole uniqueness results: it occurs even when uniqueness

117: theorems fail, as with `hairy' black holes \cite{hairy} or in

118: presence of matter rings non-trivially distorting the horizon

119: \cite{gh}. The essential ingredient is the future singularity,

120: hidden from ${\scri^+}$, which can act as the sink of information

121: (see, in particular, Penrose's remarks in \cite{rp1}.)

122: 

123: A natural question then is: what happens in quantum gravity? Is

124: there again a similar information loss? Hawking's celebrated work

125: of 1974 \cite{swh1} analyzed this issue in the framework of

126: quantum field theory in curved space-times. In this approximation,

127: three main assumptions are made: i) the gravitational field can be

128: treated classically;  ii) one can neglect the back-reaction of the

129: spontaneously created matter on the space-time geometry; and iii)

130: the matter quantum field under investigation is distinct from the

131: collapsing matter, so one can focus just on spontaneous

132: emission.%

133: %

134: \footnote{Generally, only the first two assumptions are

135: emphasized. However, we will see that the third also has a bearing

136: on the validity of semi-classical considerations.}

137: %

138: Under these assumptions, Hawking found that there is a steady

139: emission of particles to ${\scri^+}$ and the spectrum is thermal

140: at a temperature dictated by the surface gravity of the final

141: black hole. In particular, pure states on ${\scri^-}$ evolve to

142: mixed states on ${\scri^+}$. In a next step, one can include

143: back-reaction. To our knowledge, a detailed, systematic

144: calculation is still not available. In essence one argues that, as

145: long as the black hole is large compared to the Planck scale, the

146: quasi-stationary approximation should be valid. Then, by appealing

147: to energy conservation and the known relation between the mass and

148: the horizon area of \emph{stationary} black holes, one concludes

149: that the area of the event horizon should steadily decrease. This

150: then leads to black hole evaporation depicted in figure

151: \ref{Traditional} \cite{swh1}.

152: 

153: \begin{figure}

154: \begin{center}

155: \includegraphics[height=2in]{Traditional.eps}

156: \caption{The standard space-time diagram depicting black hole

157: formation and evaporation.} \label{Traditional}

158: \end{center}

159: \end{figure}

160: 

161: If one does not examine space-time geometry but uses instead

162: intuition derived from Minkowskian physics, one may be surprised

163: that although there is no black hole at the end, the initial pure

164: state has evolved in to a mixed state. Note however that while

165: space-time is now dynamical even after the collapse, \emph{there

166: is still a final singularity, i.e., a final boundary in addition

167: to} $\scri^+$. Therefore, it is not at all surprising that, in

168: this approximation, information is lost ---it is still swallowed

169: by the final singularity \cite{rp1}. Thus, provided figure

170: \ref{Traditional} is a reasonable approximation of black hole

171: evaporation and one does not add new input `by hand', then pure

172: states must evolve in to mixed states.

173: 

174: The question then is to what extent this diagram is a good

175: representation of the physical situation. The general argument in

176: the relativity community has been the following (see e.g.

177: \cite{rmw1}). Figure \ref{Traditional} should be an excellent

178: representation of the actual physical situation as long as the

179: black hole is much larger than the Planck scale. Therefore,

180: problems, if any, are associated \emph{only} with the end point of

181: the evaporation process. It is only here that the semi-classical

182: approximation fails and one needs full quantum gravity. Whatever

183: these `end effects' are, they deal only with the Planck scale

184: objects and would be too small to recover the correlations that

185: have been steadily lost as the large black hole evaporated down to

186: the Planck scale. Hence pure states must evolve to mixed states

187: and information is lost.

188: 

189: Tight as this argument seems, it overlooks two important

190: considerations. First, one would hope that quantum theory is free

191: of infinities whence figure \ref{Traditional} can not be a good

192: depiction of physics near the \emph{entire singularity} ---not

193: just near the end point of the evaporation process. Second, the

194: event horizon is a highly global and teleological construct. (For

195: a recent discussion of limitations of this notion, see

196: \cite{ak3}). Since the structure of the \emph{quantum} space-time

197: could be very different from that of figure \ref{Traditional} near

198: (and `beyond') the singularity, the causal relations implied by

199: the presence of the event horizon of figure \ref{Traditional} is

200: likely to be quite misleading. Indeed, Hajicek \cite{ph} has

201: provided explicit examples to demonstrate that the Vaidya

202: solutions which are often used to model the evaporating black hole

203: of figure \ref{Traditional} can be altered just in a Planck scale

204: neighborhood of the singularity to change the structure of the

205: event horizon dramatically and even make it disappear.

206: 

207: The purpose of this article is to point out that these

208: considerations are important and conclusions drawn from figure

209: \ref{Traditional} are therefore incomplete. More precisely, we

210: will argue that \emph{the loss of information is not inevitable

211: even in space-time descriptions favored by relativists.} As in

212: other discussions of the black hole evaporation process, we will

213: not be able to present rigorous derivations. Rather, we will

214: present a paradigm%

215: %

216: \footnote{In this article, the term `paradigm' is used in the

217: modest, dictionary sense, `One that serves as a pattern or model'.

218: The paradigm presented here was briefly sketched in section 8 of

219: \cite{ak3}.}

220: %

221: by drawing on two frameworks where detailed and systematic

222: calculations have been performed: i) analysis of the fate of the

223: Schwarzschild singularity in loop quantum gravity; and ii) the

224: dynamical horizon formalism which describes evolving black holes

225: in classical general relativity. Even without these details,

226: certain general conclusions could be drawn simply by assuming that

227: the space-time geometry is somehow modified near the singularity

228: and analyzing the Hawking process on this new space-time. But then

229: there is a multitude of possibilities. As we will see below, loop

230: quantum gravity and dynamical horizon considerations serve to

231: focus the discussion and suggest concrete directions for future

232: work. The manner in which black hole (and cosmological)

233: singularities are resolved in loop quantum gravity provides a

234: specific type of quantum extension of space-time and the fact that

235: thermodynamical considerations apply also to dynamical horizons

236: makes it plausible to think of the Hawking process as evaporation

237: of these quasi-local horizons. The final result of these

238: considerations is a space-time description of black hole

239: evaporation in the physical, Lorentzian setting in which one

240: allows for a \emph{quantum} extension of the space-time geometry

241: \emph{beyond} singularity. Since the space-time no longer has a

242: future boundary at the singularity, pure quantum states on

243: ${\scri}^-$ can evolve to pure quantum states on ${\scri}^+$.

244: 

245: The plausibility of this scenario is supported by the fact that

246: its 2-dimensional version is realized \cite{av} in the CGHS black

247: hole \cite{cghs}. (For earlier work along these lines, see

248: especially \cite{am}.) There, it is possible to isolate the true

249: degree of freedom and carry out an exact quantization using, e.g.,

250: Hamiltonian methods. On the resulting Hilbert space, one can in

251: particular define the quantum (inverse) metric operator. The

252: classical black hole metric arises as the expectation value in a

253: suitable quantum state, i.e., in the \emph{mean field

254: approximation.} Hawking effect emerges through the study of small

255: fluctuations on this mean field. One can explicitly check that

256: this mean field approximation is good in a significant portion of

257: the quantum space-time. However, the quantum fluctuations are very

258: large near the entire singularity, whence the approximation fails

259: there. The quantum (inverse) metric operator itself is

260: well-defined everywhere; only its expectation value vanishes at

261: the classical singularity. Thus, quantum geometry is defined on a

262: manifold which is \emph{larger} than the black hole space-time of

263: the mean field approximation. The mean field metric is

264: well-defined again in the asymptotic region `beyond' the

265: singularity.%

266: %

267: \footnote{There is a qualitative similarity with the theory of

268: ferromagnetism. The (inverse) metric is analogous to the

269: magnetization vector. If you have a large ferromagnet (such as the

270: earth) a small, central portion of which is heated beyond the

271: Curie temperature, the mean field approximation will hold far away

272: from this central region and the magnetization operator will have

273: a well-defined mean value there. That region is analogous to the

274: part of the full, quantum space-time where there is a well-defined

275: classical metric. The analysis of the Hawking effect is analogous

276: to that of spin-waves on this part of the ferromagent, where the

277: mean field approximation holds. While the mean field approximation

278: fails in the central region where the expectation value of

279: magnetization vanishes,  quantum theory provides a good

280: description of the entire magnet, including the central region, in

281: terms of microscopic spins.}

282: %

283: Thus, there is a single asymptotic region in the distant past

284: \emph{and} distant future and pure states on ${\scri}^-$ evolve to

285: pure states on $\scri^+$ of the full quantum space-time.

286: 

287: In this paper, we will focus on 4 dimensions where the qualitative

288: picture is similar but the arguments are based on a number of

289: assumptions. We will spell these out at various steps in the

290: discussion. As we will see, specific calculations need to be

291: performed to test if the assumptions are valid and the scenario is

292: viable also in 4 dimensions. Our hope is that the proposed

293: paradigm will provide direction and impetus for the necessary

294: detailed analysis which will deepen our understanding of the

295: evaporation process, irrespective of whether or not the paradigm

296: is realized.

297: 

298: The paper is organized as follows. In section 2, we summarize the

299: resolution of the Schwarzschild singularity by effects associated

300: with the quantum nature of geometry. The new paradigm for black

301: hole evaporation is presented in section 3. Section 4 contains

302: some concluding remarks.

303: 

304: \section{Quantum geometry and the Schwarzschild interior}

305: \label{s2}

306: 

307: Since the key issues involve the final black hole singularity and

308: since this singularity is expected to be generically space-like

309: (see, e.g. \cite{md}), the situation is similar to cosmology. In

310: fact, the interior of the Schwarzschild horizon is naturally

311: foliated by 3-manifolds which are spatially homogeneous with the

312: Kantowski-Sachs isometry group. Accordingly, the result of absence

313: of singularities in homogeneous loop quantum cosmology \cite{b3}

314: can be applied to this situation of the Kantowski-Sachs

315: `mini-superspace' of vacuum, spatially homogeneous space-times.

316: That this is possible has been shown explicitly using

317: Arnowitt-Deser-Misner (ADM) variables \cite{m}: unlike classical

318: evolution, the dynamical quantum equation does not break down at

319: the location of the classical singularity. However, since the ADM

320: variables allow only non-degenerate metrics, the geometrical

321: meaning of the resulting space-time extension has remained obscure

322: in this framework. The connection dynamics phase space, by

323: contrast, is an extension of the ADM phase space where the

324: (density weighted) triad, which captures the Riemannian geometry,

325: is allowed to vanish. Thanks to this larger phase space, the

326: extended space-time has a clearer interpretation: the `other side'

327: of the singularity corresponds to the new domain of the enlarged

328: phase space where the triad reverses its orientation.%

329: %

330: \footnote{In addition, the elementary variables that feature in

331: the quantization used in \cite{m} ---the exponentials of $i$ times

332: extrinsic curvature components--- do not have natural analogs in

333: full geometrodynamics based on the ADM variables. In the

334: connection mini-superspace, by contrast, the elementary variables

335: are just holonomies of homogeneous connections, i.e., restrictions

336: to the basic variables used in the full theory to the symmetry

337: reduction under consideration.}

338: %

339: Therefore, in this section we will summarize the results that have

340: been obtained in the connection-dynamics framework \cite{ab1}.

341: 

342: The first result is that, although the co-triad and curvature

343: diverge at the singularity in the classical theory, the

344: corresponding quantum operators are in fact bounded on the full

345: kinematic Hilbert space. This analysis is analogous to that which

346: established the boundedness of the quantum operator representing

347: the inverse scale factor in the spatially homogeneous, isotropic

348: quantum cosmology \cite{b1,abl}. As in that analysis, the co-triad

349: operator has various nice properties one expects of it and

350: departures from the classical behavior appear only in the deep

351: Planck regime (i.e. very near what was classical singularity).

352: This finiteness results from the fact that the `polymer

353: representation' of the Weyl relations underlying our quantum

354: description is inequivalent to the `standard representation' used

355: in quantum geometrodynamics (for details, see, e.g.,

356: \cite{alrev}). It is analogous to the fact that matter

357: Hamiltonians in the full theory are densely defined \cite{tV}

358: operators. This result suggests that quantum dynamics may well be

359: singularity-free. But a definitive conclusion can only be drawn

360: through a detailed analysis.

361: 

362: Using quantum geometry, one can write down a well-defined

363: Hamiltonian constraint. In the mini-superspace under

364: consideration, there are only two degrees of freedom. One can be

365: interpreted as the radius of any (round) 2-sphere in the slice and

366: the other (the norm of the translational Killing field) is a

367: measure of the anisotropy. It is natural to use the first as an

368: intrinsic `clock' and analyze how anisotropy `evolves' with

369: passage of this `time'. In quantum theory, one can expand out the

370: state $|\Psi\rangle$ as $|\Psi\rangle = \sum_{\phi,\tau}

371: \psi(\phi,\tau)|\phi, \tau\rangle$ where $\phi$ are eigenvalues of

372: the anisotropy operator and $\tau$ of the radius operator. The

373: Hamiltonian constraint is of the form:

374: %

375: \be \label{1} f_+(\tau)\, {\hat{O}_+}\, \psi(\phi, {\tau+2\delta})

376: + f_o(\tau)\, {\hat{O}_o}\,

377:  \psi(\phi, {\tau}) + f_-(\tau)\,

378: {\hat{O}_-}\, \psi(\phi, {\tau-2\delta})  =0 \ee

379: %

380: where $f_{\pm}, f_o$ are rather simple functions of $\tau$,

381: $\hat{O}_\pm, \hat{O}_o$ are rather simple operators on functions

382: of $\phi$ alone and $\delta$ is a number whose value is determined

383: by the smallest area eigenvalue in Planck units. Being a

384: constraint, it simply restricts the physically allowed states.

385: However, one can also regard it as providing `time-evolution' of

386: the quantum state through discrete time steps of magnitude

387: $2\delta$ (in Planck units). The functions $f$ and the operators

388: $\hat{O}$ are such that this evolution does not break down at

389: $\tau =0$ (which corresponds to the classical singularity). Thus,

390: as in quantum cosmology \cite{b2,abl} one finds that the

391: \emph{quantum} evolution does not stop at the singularity; one can

392: evolve right through it \cite{ab1}. The state remains pure.

393: However one expects that, in the \emph{deep} Planck regime around

394: the singularity, the notion of a classical space-time geometry

395: would fail to make even an approximate sense in general.

396: Nonetheless, there is no longer a final boundary in the interior,

397: whence the full quantum evolution is quite different from the

398: classical one.

399: 

400: This calculation was done \cite{ab1} in the Kantowski-Sachs

401: mini-superspace and $|\Psi\rangle$ represents the state of the

402: Schwarzschild black hole interior in loop quantum gravity. This

403: black hole can not evaporate: there is no matter and, because of

404: the restriction to spherical symmetry, there can not be Hawking

405: radiation of gravitons either. However, since the generic

406: singularity is expected to be space-like (see, e.g., \cite{md}),

407: one may hope that the general intuition about the resolution of

408: the Schwarzschild singularity provided by this calculation can be

409: taken over to models in which gravity is coupled to scalar fields,

410: where the evaporation does occur. Indeed, there is already some

411: work on the spherical model without restriction to the interior

412: \cite{bs,hw,b4} and its extension is now in progress. The initial

413: results support expectations from the homogeneous models. Here, we

414: will assume that the \emph{overall, qualitative} features of our

415: singularity resolution will continue to be valid in these models.

416: 

417: 

418: \section{Evaporation Process}

419: \label{s3}

420: 

421: The physical situation we wish to analyze is the following: some

422: radiation field on ${\scri^-}$ collapses and forms a large,

423: macroscopic black hole which then evaporates. For simplicity, we

424: will restrict ourselves to the \emph{spherically symmetric sector

425: of Einstein gravity coupled to a massless Klein-Gordon field}. The

426: incoming state on ${\scri}^-$ will be assumed to be a coherent

427: state peaked at a classical scalar field representing a large

428: `pulse', i.e., a field which is large over a compact region of

429: ${\scri}^-$ and vanishes (or become negligible) outside this

430: region. Note that there is a single scalar field, coupled to

431: gravity, whose collapse from ${\scri}^-$ leads to the formation of

432: the black hole and whose quanta are radiated to ${\scri}^+$ during

433: the evaporation process. There are no test fields; the system is

434: `closed'.

435: 

436: In this setting, conclusions drawn from classical general

437: relativity should be valid to an excellent approximation until we

438: are in the Planck regime near the singularity. Thus, marginally

439: trapped surfaces would emerge and their area would first grow. In

440: this phase the world tube of marginally trapped surfaces  would be

441: a \emph{trapping horizon} \cite{sh1}. For the massless scalar

442: field under consideration, during and for a long time after the

443: collapse, it would be space-like \cite{gv,acg} and thus constitute

444: a \emph{dynamical horizon} \cite{ak1,ak2}. When Hawking radiation

445: starts to dominate the in-falling scalar field, the trapping

446: horizon would be time-like and thus constitute a \emph{time-like

447: membrane} \cite{ak3}. In the spherical symmetric case now under

448: consideration, this scenario was discussed already in the eighties

449: (see, in particular \cite{jy,ph}). However, constructions were

450: tailored just to spherical symmetry and made use of some heuristic

451: considerations involving an `ergo-region of an approximate Killing

452: field.' Therefore, although well-motivated, the discussions

453: remained heuristic. Laws governing the growth of the area of

454: dynamical horizons and shrinkage of area of time-like membranes

455: are now available in a general and mathematically precise setting

456: \cite{ak2,sh2}. Furthermore, laws of black hole mechanics have

457: been extended to these dynamical situations. These results

458: strengthen the older arguments considerably and reenforce the idea

459: that what evaporates is the trapping horizon.

460: 

461: Let us now combine this semi-classical picture with the discussion

462: of section \ref{s2} on the resolution of the singularity to draw

463: qualitative conclusions on what the black hole evaporation process

464: would look like in full loop quantum gravity. Once the Planck

465: regime is reached, a priori there are two possibilities:\\

466: $\bullet\,\,$ a) States which start out semi-classical on

467: ${\scri}^-$ never become semi-classical on the `other side' of the

468: singularity (say, in the sense discussed in \cite{bd,aps}).  Then

469: only a part of the process can be described in space-time terms.

470: However, one \emph{can} look at the problem quantum mechanically

471: and conclude that pure states remain pure. If we restricted them

472: only to the classical part of the space-time and measure

473: observables which refer only to this part, we would get a density

474: matrix but this is not surprising; it happens even in laboratory

475: physics when one ignores a part of the system. \\

476: $\bullet\,\,$ b) As in spatially homogeneous, isotropic

477: cosmologies coupled to a massless scalar field \cite{aps}, after

478: evolving through the deep Planck regime, the state becomes

479: semi-classical again on the `other side' so we can use a classical

480: space-time description also in `distant future'.

481: 

482: In the CGHS model, possibility b) holds. Furthermore, using the

483: underlying conformal structure, one can show that the classical

484: region in the distant future remains causally connected to that in

485: the distant past in the full quantum theory; there is no baby

486: universe. Such a calculation is yet to be undertaken in four

487: space-time dimensions. If it turns out that the possibility a)

488: holds, it would be impossible to speak of a scattering matrix

489: since there would not be an adequate ${\scri}^+$ or a space-like

490: surface in the distant future for the `final' states to live on.

491: Hence, it would be quite difficult to say anything beyond the

492: statement that pure states remain pure. If b) holds, one can

493: compare various scenarios. Therefore, in the rest of the article,

494: we will focus on b).

495: 

496: 

497: \begin{figure}

498: \begin{center}

499: \includegraphics[height=3in]{NonSingH.eps}

500: \caption{Space-time diagram of black hole evaporation where the

501: classical singularity is resolved by quantum geometry effects. The

502: shaded region lies in the `deep Planck regime' where geometry is

503: genuinely quantum mechanical. $H$ is the trapping horizon which

504: is first space-like (i.e., a dynamical horizon) and grows because

505: of infalling matter and then becomes time-like (i.e., a time-like

506: membrane) and shrinks because of Hawking evaporation. In region I,

507: there is a well-defined semi-classical geometry.} \label{NonSingH}

508: \end{center}

509: \end{figure}

510: 

511: A space-time diagram that could result in scenario b) is depicted

512: in figure \ref{NonSingH}. Here, the extended, `quantum space-time'

513: has a single asymptotic region in the future, i.e., there are no

514: `baby universes'. In four dimensions, this is an

515: \emph{assumption}. It is motivated by two considerations: i) the

516: situation in the CGHS model where detailed calculations are

517: possible and show that the quantum space-time has this property;

518: and ii) experience with the action of the Hamiltonian constraint

519: in the spherically symmetric midi-superspace in four dimensions.

520: However, only detailed calculations can decide whether this

521: assumption is borne out. Since our goal in this paper is only to

522: point out the existence of a possible space-time description in

523: which information can be recovered at future null infinity, for

524: our purposes it suffices to note only that none of the existing

525: arguments rule out this mechanism.

526: 

527: We will refer to figure \ref{NonSingH} as a `Penrose diagram'

528: where the inverted commas will serve as a reminder that we are not

529: dealing with a purely classical space-time. Throughout the quantum

530: evolution, the pure state remains pure and so we again have a pure

531: state on ${\scri^+}$. In this sense there is no information loss.

532: Noteworthy features of this `Penrose diagram' are the following.

533: 

534: i) \textsl{Effect of the resolution of the classical singularity:}

535: Region marked I is well-approximated by a classical geometry.

536: Modulo small quantum fluctuations, this geometry is determined via

537: Einstein's equations by the classical data on ${\scri^-}$ at which

538: the incoming quantum state is peaked. The key difference between

539: figures \ref{Traditional} and \ref{NonSingH} is that while

540: space-time `ends' at the singularity in figure \ref{Traditional}

541: it does not end in figure \ref{NonSingH}. But there may not be

542: even an approximate classical space-time in the shaded region

543: representing the `deep Planck regime'.

544: 

545: ii) \textsl{Event horizon:} Since the shaded region does not have

546: a classical metric, it is not meaningful to ask questions about

547: causal relations between this region and the rest. Therefore,

548: although it {is} meaningful to analyze the causal structure (to an

549: excellent approximation) within each local semi-classical region,

550: due care must be exercised to address \textit{global} issues which

551: require knowledge of the metric on the entire space-time. This is

552: in particular the case for the notion of the event horizon, the

553: future boundary of the causal past of ${\scri}^+$. Because there

554: is no classical metric in the shaded region, while one can

555: unambiguously find some space-time regions which are in the past

556: of ${\scri}^+$, we can not determine what the \emph{entire} past of

557: ${\scri}^+$ is. If we simply cut out this region and look at the

558: remaining classical space-time, we will find that the past is not

559: all of this space-time. But this procedure can not be justified

560: especially for purposes of quantum dynamics.  Thus, because the

561: geometry in the deep Planck regime is genuinely quantum

562: mechanical, the global notion of an event horizon ceases to be

563: useful. It may well be that there is a well-defined, new notion of

564: quantum causality and using it one may be able to reanalyze this

565: issue. However, the standard classical notion of the event horizon

566: is `transcended' because of absence of a useful

567: classical metric in the deep Planck region.%

568: %

569: \footnote{Some authors \cite{s'thw,sh3} have suggested that there

570: may be a classical metric on entire space-time but Einstein's

571: classical equations would be violated in the deep Planck region,

572: resulting in a metric which is continuous (or better behaved)

573: everywhere. Should this turn out to be the case the event horizon

574: would not just be `transcended' but simply disappear. Trapping

575: horizon would still be well-defined.}

576: %

577: 

578: iii) \textsl{Dynamical horizon:} Nonetheless, we can trust

579: classical theory in region I and this region will admit marginally

580: trapped surfaces. It is reasonable to expect that a spherical

581: dynamical horizon will be formed. It will be space-like and its

582: area will grow during collapse. In the classical theory, the

583: dynamical horizon will eventually settle down to a null, isolated

584: horizon which will coincide with (the late portion of) the event

585: horizon. However, in quantum theory eventually the horizon will

586: shrink because of Hawking radiation. While the black hole is

587: large, the process will be very slow. Semi-classical calculations

588: indicate that there is a positive flux of energy out of the black

589: hole. The dynamical horizon $H$ will now `evolve' into a time-like

590: membrane and its area loss will be dictated by the balance law

591: %\cite{ak2}

592: %

593: \be \label{2}\frac{dR}{dt} = - 8\pi G\, R^2\, T_{ab} \ell^a

594: \hat{r}^b \ee

595: %

596: where $R$ is the area radius of cross-sections of marginally

597: trapped 2-spheres in $H$, $\ell^a$ the (future directed) null

598: normal with vanishing expansion, and $\hat{r}^a$ is the unit

599: (outward) radial normal to $H$. (See Appendix B of the second

600: paper in \cite{ak2}). This process is depicted in figure

601: \ref{NonSingH}. The union of the dynamical horizon, the isolated

602: horizon and the timelike membrane constitutes the trapping

603: horizon. Thus, although we no longer have a well-defined notion

604: of an event horizon, we can still meaningfully discuss formation

605: and evaporation of the black hole using trapping horizons

606: because most of this process occurs in the semi-classical region

607: and, more importantly, because \emph{the notion of a trapping

608: horizon is quasi-local}. When the black hole is large, the

609: evaporation process is extremely slow. Therefore, it seems

610: reasonable to assume that the intuition developed from the quantum

611: geometry of isolated horizons \cite{abck,abk} will continue to be

612: valid. If so, the quantum geometry of the trapping horizon will be

613: described by the $U(1)$ Chern-Simons theory on a punctured $S^2$,

614: where the punctures result because the polymer excitations of the

615: bulk geometry pierce the dynamical horizon, endowing it with

616: certain area quanta. During the evaporation process, the punctures

617: slowly disappear, the horizon shrinks and quanta of area are

618: converted into quanta of the scalar field, seen as Hawking

619: radiation at infinity.%

620: %

621: \footnote{Equation (\ref{2}) relates the change in the area of the

622: time-like membrane part of the trapping horizon with the flux of

623: the energy flowing out of it. However, because of the dynamical

624: nature of geometry, there is no simple relation between this

625: ingoing flux at the time-like membrane and the energy carried by

626: the outgoing quanta on ${\scri}^+$. Indeed, not only are they

627: evaluated at very different locations, the two fluxes refer to

628: \emph{distinct} components of the stress-energy tensor,

629: $T_{ab}\ell^a \hat{r}^b$ at the horizon and $T_{ab}n^an^b$ at

630: ${\scri}^+$.}

631: %

632: The existence, in the classical theory, of a meaningful

633: generalization of the first law of black hole mechanics to

634: dynamical horizons \cite{ak1,ak2} supports the view that the

635: process can be interpreted as evaporation of the dynamical

636: horizon.

637: 

638: \begin{figure}

639: \begin{center}

640: \includegraphics[height=3in,keepaspectratio]{NonSingAB.eps}

641: \caption{The solid line with an arrow represents the world-line of

642: an observer restricted to lie in region I. While these observers

643: must eventually accelerate to reach ${\scri}^+$, if they are

644: sufficiently far away, they can move along an asymptotic time

645: translation for a long time. The dotted continuation of the world

646: line represents an observer who is not restricted to lie in region

647: I. These observers can follow an asymptotic time translation all

648: the way to $i^+$.} \label{NonSingAB}

649: \end{center}

650: \end{figure}

651: 

652: iv) \textsl{Reconciliation with the semi-classical information

653: loss:} Consider observers restricted to lie in region I (see

654: figure \ref{NonSingAB}). For a macroscopic black hole this

655: semi-classical region is very large. These observers would see the

656: radiation resulting from the evaporation of the horizon. This

657: would be approximately thermal, only approximately because, among

658: other things,  the space-time geometry is not fixed as in

659: Hawking's original calculation \cite{swh1}, but evolves slowly.

660: Although the full quantum state is `pure', there is no

661: contradiction because these observers look at only part I of the

662: system and trace over the rest which includes a purely quantum

663: part. In effect, for them space-time has a future boundary where

664: information is lost. Since the black hole is assumed to be

665: initially large, the evaporation time is long (about $10^{70}$

666: years for a solar mass black hole). Suppose we were to work with

667: an approximation that the black hole takes \emph{infinite} time to

668: evaporate. Then, the space-time diagram will be figure

669: \ref{Infinite} because the horizon area would shrink to zero only

670: at $i^+$. In this case, there would be an event horizon and

671: information would be genuinely lost for any observer in the

672: initial space-time; it would go to a second asymptotic region

673: which is inaccessible to observers in the initial space-time. Of

674: course this does not happen because the black hole evaporates only

675: in a finite time.

676: 

677: \begin{figure}

678: \begin{center}

679: \includegraphics[height=3.5in]{Infinite.eps}

680: \caption{The `would be' space-time if the black hole were to take

681: an infinite time to evaporate.} \label{Infinite}

682: \end{center}

683: \end{figure}

684: 

685: v) \textsl{`Recovery' of the `apparently lost' information:} Since

686: the black hole evaporates only in a finite amount of time, the

687: point at which the black hole shrinks to zero (or Planck) size

688: is \emph{not} $i^+$ and the space-time diagram looks like figure

689: \ref{NonSingAB} rather than figure \ref{Infinite}. Now, $i^+$ lies

690: to the `future' of the `deep Planck' region and there are

691: observers lying entirely in the asymptotic region going from $i^-$

692: to $i^+$ (represented by the dotted continuation of the solid line

693: in figure \ref{NonSingAB}). This family of observers will recover

694: the apparently lost correlations. Note that these observers always

695: remain in the asymptotic region where there is a classical metric

696: to an excellent approximation; they {never} go near the deep

697: Planck region. The total quantum state on ${\scri}^+$ will be pure

698: and will have the complete information about the initial state on

699: ${\scri^-}$. It looked approximately thermal at early times, i.e.,

700: to observers represented by the solid line, only because they

701: ignore a part of space-time. The situation has some similarity

702: with the EPR experiment in which the two subsystems are first

703: widely separated and then brought together (see also \cite{fw}).

704: 

705: vi) \textsl{Entropy:} Since the true state is always pure, one

706: might wonder what happens to black hole entropy. It is only the

707: observers in region I that `sense' the presence of a black hole.

708: In the quantum geometry approach to black hole entropy, entropy is

709: not an absolute concept associated objectively with a space-time.

710: Rather, it is associated with a family of observers who have

711: access to only a part of space-time.  Indeed, the entropy of an

712: isolated horizon calculated in \cite{abk} referred to the family

713: of observers for whom the isolated horizon serves as the internal

714: boundary of accessible space-time. So, for observers restricted to

715: region I, that entropy calculation is still meaningful, at least

716: so long as the black hole is macroscopic (i.e., the area of

717: marginally trapped surfaces on $H$ is much larger than Planck

718: area). And it is these observers who see the (approximate) Hawking

719: radiation. More precisely, since these observers have access only

720: to observables of the type $A_{\rm I}\otimes 1$, they trace over

721: the part of the system not in I, getting a density matrix

722: $\rho_{\rm I}$ on the Hilbert space ${\cal H}_{\rm I}$. Entropy

723: for them is simply $Tr_{\rm I} \rho_{\rm I} \ln \rho_{\rm I}$. Had

724: there been a true singularity `ending' the space-time, this

725: entropy would have become objective in the sense that it would be

726: associated with \emph{all} observers who do not fall into the

727: singularity.

728: 

729: \section{Concluding Remarks}

730: \label{s4}

731: 

732: In the last two sections we used a quantum gravity perspective to

733: argue that information loss is not inevitable in the space-time

734: description of black hole evaporation. The qualitative difference

735: between figures \ref{Traditional} and \ref{NonSingH} arises

736: essentially from the fact that the singularity is resolved in

737: quantum geometry, as per a general expectation that a satisfactory

738: quantum theory of gravity should not have infinities. In this

739: sense the paradigm shift is well-motivated. Furthermore,

740: conclusions of the traditional paradigm drawn from the usual

741: space-time diagram \ref{Traditional} are not simply discarded. For

742: a large black hole, they continue to be approximately valid for a

743: very long time. Figure \ref{NonSingAB} clarifies the approximation

744: involved. However, from the \emph{conceptual perspective of

745: fundamental physics}, conclusions drawn from the \emph{complete}

746: space-time diagram \ref{NonSingH} are qualitatively different from

747: the standard ones. A pure state from ${\scri}^-$ evolves to a pure

748: state on ${\scri}^+$ and there is no obstruction in quantum theory

749: to evolving the final state on ${\scri}^+$ backwards to recover

750: full space-time. However it is likely that the resulting geometry

751: would fail to be globally classical. In the shaded region, it is

752: likely to be genuinely quantum mechanical, described only in terms

753: of the quantum geometry states (i.e., in terms of spin-networks).

754: In the region in which one can introduce classical geometry to

755: an excellent approximation, it is meaningful to speak of

756: marginally trapped surfaces, trapping horizons and null infinity

757: ${\scri}^\pm$.%

758: %

759: \footnote{Because of the presence of the purely quantum part, the

760: space-time is not asymptotically simple \cite{rp}; the classical

761: region admits null geodesics which do not end on ${\scri}^\pm$.

762: However, it is asymptotically flat and admits a global null

763: infinity in the sense of \cite{ax}.}

764: %

765: What `evaporates' is the area of the trapping horizon.

766: 

767: From the perspective of this paradigm, the conclusion that a pure

768: state must evolve to a mixed state results if one takes the

769: classical space-time diagram \ref{Traditional}, \emph{including

770: the singular boundary in the future}, too seriously.%

771: %

772: \footnote{Perhaps an analogy from atomic physics would be to base

773: the analysis of the ground state of the hydrogen atom on the zero

774: angular momentum, classical electron trajectories, all of which

775: pass through the `singularity' at the origin.}

776: %

777: In the cosmological context, a combination of detailed analytical

778: and numerical calculations \cite{aps} has recently shown that

779: quantum geometry is well-defined at the classical big-bang

780: singularity; backward quantum evolution enables one to pass

781: through it; and, on the other side, there is again a classical

782: space-time. Thus, quantum geometry in the deep Planck regime

783: serves to bridge two large classical regimes. Classical

784: singularity is only a reflection of the failure of the mean field

785: approximation and quantum geometry is defined on a larger

786: manifold. Our paradigm is based on the assumption that the

787: situation is qualitatively similar with black hole

788: singularities. If this assumption is borne out, pure states will

789: evolve to pure states, without any information loss provided the

790: analysis pays due respect to this space-time extension.

791: 

792: The two dimensional analog of our paradigm is realized quite well

793: by CGHS black holes. However, 2 dimensional models have special

794: features that are not shared by higher dimensional theories. To

795: carry out the analogous analysis in 4 dimensions, one would have to

796: complete several difficult steps: \\

797: i) Discussion of quantum dynamics in the spherically symmetric

798: midi-superspace \cite{bs}. To be directly useful, we would need to

799: introduce a satisfactory generalization of the notion of `time'

800: used in \cite{ab1,aps}; \\

801: ii) demonstration of the semi-classical behavior of the quantum

802: state in regions where the dynamical horizon grows and the

803: time-like membrane shrinks (in the regime where its area is

804: large); \\

805: iii) extension of the available theory \cite{abk} of quantum

806: geometry from isolated to slowly evolving dynamical horizons;

807: and\\

808: iv) establishing that the quantum state becomes semi-classical

809: again on the `other side' of what was a classical singularity,

810: with a single asymptotic region. \\

811: Note, however, that any approach to quantum gravity will have

812: to resolve similar issues if it is to provide a detailed

813: `space-time description' of the black hole evaporation in the

814: Lorentzian framework. In particular, all discussions beyond the

815: semi-classical approximation that we are aware of implicitly

816: assume that there is a classical space-time in the future.

817: 

818: Finally, in this paradigm correlations are restored by part of the

819: state that passes through the singularity and emerges on

820: ${\scri}^+$ to the future of region I of figure \ref{NonSingH}.

821: Therefore, it is presumably necessary that this part should carry

822: a non-trivial fraction of the total ADM mass of space-time (see,

823: however, \cite{fw}). This seems physically plausible because one

824: expects non-trivial space-time curvature also on the `other side

825: of the singularity'. However, whether this is realized in detailed

826: calculations remains to be seen. Thus, the paradigm is based on

827: pieces of calculations and analogy to the CGHS model, rather than

828: a systematic detailed analysis. Recall, however, that the

829: traditional reasoning that led to figure \ref{Traditional} was

830: based on general considerations and plausibility arguments and a

831: systematic analysis of the viability of approximations is still

832: not available. Nonetheless, it led to a paradigm which proved to

833: be valuable in focussing discussions. Our hope is that that the

834: paradigm presented here will play a similar role.

835: 

836: \emph{Remark:} After this work was posted on the archives, we

837: became aware of two discussions of black hole evaporation which

838: feature space-time diagrams similar to figure \ref{NonSingH}. The

839: first is due to Stephens, 't Hooft and Whiting \cite{s'thw} which

840: appeared more than a decade ago and the second is due to Hayward

841: \cite{sh3} which appeared very recently. In the first, one draws a

842: distinction between hard matter which creates curvature and soft

843: matter whose effect on gravity is negligible. A detailed

844: calculation is carried out in a 2-dimensional model, where the

845: focus is on the soft matter. The main idea is to first assume that

846: quantum gravity effects would halt the collapse and cause a bounce

847: and then do a calculation analogous to that of Hawking's

848: \cite{swh1} on this modified but classical background geometry.

849: The result is that although pure states evolve to pure states, in

850: the appropriate portion of ${\scri}^+$, the state is approximately

851: thermal. This scenario is similar to ours in that the space-time

852: under consideration has no singularity; pure states evolve to pure

853: states; and expectations based on semi-classical considerations

854: are not just discarded but recovered in a precise sense. However,

855: there are also some important differences. If our paradigm is

856: realized by detailed calculations, all matter would be `hard';

857: singularity would be resolved by specific quantum gravity effects;

858: and a genuinely quantum mechanical geometry would bridge the

859: space-time of classical general relativity with a new classical

860: space-time. In contrast to \cite{s'thw}, the new portion in the

861: geometry of any one space-time will not correspond to a simple

862: time-reversal of the standard, collapsing portion. Hayward's

863: considerations \cite{sh3} are different from those of

864: \cite{s'thw}. As in the current paradigm, he emphasizes trapping

865: horizons and his space-time diagram is closer to ours, especially

866: for a massless Klein-Gordon source. In particular, his space-time

867: is a singularity-free extension of standard one and the collapsing

868: matter re-emerges on ${\scri}^+$, in addition to the Hawking

869: radiation. However, he assumes that space-time will have a $C^2$

870: metric everywhere (which, however, violates the classical field

871: equations near what was the singularity), and the collapsing

872: matter which re-emerges is treated classically. Apart from the

873: Hawking radiation, genuine quantum considerations do not appear to

874: play a significant role. Recent numerical evolutions in quantum

875: cosmology \cite{aps} indicate that there may well exist initial

876: states on ${\scri}^-$ for which the physics of our deep Planck

877: regime can be approximated by an effective continuum classical

878: geometry. If this does happen for black hole space-times, then our

879: paradigm would essentially reduce to Hayward's in those

880: situations.

881: 

882: \section*{Acknowledgements:}

883: 

884: We thank Alex Corichi, Gary Horowitz, Bei-Lok Hu, Ted Jacobson,

885: Daniel Sudarsky and especially Jim Hartle, Sean Hayward and Don

886: Marolf for stimulating discussions. This work was supported in

887: part by NSF grants PHY-0090091, and PHY-0354932, the Alexander von

888: Humboldt Foundation, the C.V. Raman Chair of the Indian Academy of

889: Sciences and the Eberly research funds of Penn State.

890: 

891: \begin{thebibliography}{99}

892: 

893: \bibitem{m}

894: 

895: Modesto L 2004 The Kantowski-Sachs Space-Time in Loop Quantum

896: Gravity \textit{Preprint} gr-qc/0411032

897: 

898: \bibitem{ab1} Ashtekar A and Bojowald M 2004 Non-singular quantum

899: geometry of the Schwarzschild black hole interior (in

900: preparation)

901: 

902: \bibitem{sh1} Hayward S 1994 General laws of black hole dynamics

903: \textit{Phys. Rev.} \textbf{49} 6467--6474

904: 

905: \bibitem{sh2} Hayward S  2004 Energy and entropy conservation for

906: dynamical black holes \textit{Phys. Rev.} \textbf{D70} 104027

907: 

908: \bibitem{ak1} Ashtekar A and  Krishnan B 2002 Dynamical Horizons: Energy,

909: Angular Momentum, Fluxes and Balance Laws, \textit{Phys. Rev.

910: Lett.} \textbf{89} 261101

911: 

912: \bibitem{ak2} Ashtekar A and Krishnan B 2003 Dynamical horizons and

913: their properties, \textit{Phys. Rev.} \textbf{D68} 104030

914: 

915: \bibitem{swh2} Hawking S W 2004, Lecture at the 17th

916: International conference on General Relativity and Gravitation, see

917: \verb|http://math.ucr.edu/home/baez/week207.html| for a transcript

918: 

919: \bibitem{hairy} Volkov M S and Gal'tsov D V 1999 Gravitating nonabelian

920: solitons and black holes with Yang--Mills fields \textit{Phys. Rept.}

921: \textbf{319} 1--83

922: 

923: \bibitem{gh} Geroch R and Hartle J 1982 Distorted Black Holes,

924: \textit{J. Math. Phys.} \textbf{23} 680-692\\

925: Fairhurst S and Krishnan B 2001 Distorted Black Holes with Charge,

926: \textit{Int. J. Mod. Phys.} \textbf{D10} 691--710

927: 

928: \bibitem{rp1} Hawking S W and Penrose R 1996 \textit{The nature of space

929: and time} (Princeton University Press, Princeton), pages 62--63.

930: 

931: \bibitem{swh1} Hawking S W 1975 in \textit{Quantum gravity: An

932: Oxford Symposium}, eds Isham C J, Penrose R P and Sciama D W

933: (Oxford University Press, Oxford)

934: 

935: \bibitem{rmw1} Wald R M 1994 \textit{Quantum field theory in curved

936: space-time and black hole thermodynamics} (University of Chicago

937: Press, Chicago)

938: 

939: \bibitem{ak3} Ashtekar A and Krishnan B 2004 Isolated and

940: Dynamical horizons and their properties \textit{Living Rev. Rel.} \textbf{10}

941: 1--78, \texttt{gr-qc/0407042}

942: 

943: \bibitem{ph} Hajicek P 1987 Origin of Hawking radiation \textit{Phys.

944: Rev} \textbf{D36} 1065--1079

945: 

946: \bibitem{av} Ashtekar A and Varadarajan M 2005 (in preparation)

947: 

948: \bibitem{cghs} Callen C G, Giddings S B, Harvey J A and Strominger A

949: 1992 Evanescent black holes \textit{Phys. Rev} \textbf{D45}

950: R1005--R1009

951: 

952: \bibitem{am} Mikovic A 1996 Unitary theory of

953: evaporating 2D black holes \textit{Class. Quant. Grav.}

954: \textbf{13}, 209--220\\

955: Mikovic A and Radovanovic V 1996 Two-loop Back-reaction in 2D

956: Dilaton Gravity \textit{Nucl Phys.} \textbf{B481}, 719--742

957: 

958: \bibitem{md} Dafermos M 2003 Black hole formation from a complete

959: regular past, \texttt{gr-qc/0310040}

960: 

961: \bibitem{b3} Bojowald M 2002 Isotropic Loop Quantum Cosmology

962: \textit{Class.\ Quantum Grav.} \textbf{19} 2717--2741\\ Bojowald M 2003

963: Homogeneous loop quantum cosmology \textit{Class.\ Quantum Grav.} \textbf{20}

964: 2595--2615\\ Bojowald M, Date G and Vandersloot K 2004 Homogeneous loop

965: quantum cosmology: The role of the spin connection

966: \textit{Class.\ Quantum Grav.} \textbf{21} 1253--1278

967: 

968: \bibitem{b1} Bojowald M 2001 Inverse Scale Factor in Isotropic Quantum

969: Geometry \textit{Phys.\ Rev.\ D} \textbf{64} 084018

970: 

971: \bibitem{abl} Ashtekar A, Bojowald M and Lewandowski L 2003

972: Mathematical structure of loop quantum cosmology \textit{Adv.

973: Theor. Math. Phys.} \textbf{7}  233--268

974: 

975: \bibitem{alrev} Ashtekar A and Lewandowski L 2004 Background

976: independent quantum gravity: A status report \textit{Class. Quant.

977: Grav.} \textbf{21} R53-R152 (2004).

978: 

979: \bibitem{tV} Thiemann T 1998 QSD V: Quantum Gravity as the Natural

980: Regulator of Matter Quantum Field Theories \textit{Class.\ Quantum

981: Grav.} \textbf{15} 1281--1314

982: 

983: \bibitem{b2} Bojowald M 2001 Absence of a Singularity in Loop Quantum

984: Cosmology \textit{Phys.\ Rev.\ Lett.} \textbf{86} 5227--5230

985: 

986: \bibitem{bs} Bojowald M 2004 Spherically Symmetric Quantum Geometry:

987: States and Basic Operators \textit{Class.\ Quantum Grav.} \textbf{21}

988: 3733--3753\\ Bojowald M and Swiderski R 2004 The Volume Operator in

989: Spherically Symmetric Quantum Geometry \textit{Class.\ Quantum Grav.}

990: \textbf{21} 4881--4900

991: 

992: \bibitem{hw}

993: Husain V and Winkler O 2004 Quantum resolution of black hole

994: singularities \textit{Preprint} gr-qc/0410125

995: 

996: \bibitem{b4}

997: Bojowald M 2005 Non-singular black holes and degrees of freedom in 

998: quantum gravity \textit{Phys.\ Rev.\ Lett.} to appear

999: 

1000: \bibitem{gv} Booth I, Britt L, Gonzalez J and Van Den Broeck C 2005

1001: (in preparation)

1002: 

1003: \bibitem{acg} Alcubierre M, Corichi A and Gonzalez-Samaniego A

1004: 2005 (in preparation)

1005: % Dynamical horizons for self-gravitating scalar fields

1006: 

1007: \bibitem{jy} York J 1984 What happens to the horizon when the

1008: black hole radiates? in \textit{Quantum theory of gravity. Essays

1009: in honor of the 60th birthday of Bryce S. DeWitt}, ed Christensen

1010: S M (Adam Higler, Brostol)

1011: 

1012: \bibitem{bd} Bojowald M 2001 Dynamical Initial Conditions in Quantum

1013: Cosmology \textit{Phys.\ Rev.\ Lett.} \textbf{87} 121301\\

1014: Bojowald M and Date G 2004 Consistency conditions for

1015: fundamentally discrete theories \textit{Class.\ Quantum Grav.}

1016: \textbf{21} 121--143

1017: 

1018: \bibitem{aps} Ashtekar A, Pawlowski T and Singh P 2005

1019: Homogeneous, isotropic loop quantum cosmology coupled to a

1020: massless scalar field: A complete mathematical and numerical

1021: analysis, (in preparation)

1022: 

1023: \bibitem{s'thw} Stevens C R, 't Hooft G and Whiting B F 1994 Black

1024: hole evaporation without information loss, \textit{Class. Quantum.

1025: Grav.} \textbf{11}, 621--648

1026: 

1027: \bibitem{sh3}Hayward S 2005 The disinformation problem for black

1028: holes (conference version), \texttt{gr-qc/0504038}

1029: 

1030: \bibitem{abck} Ashtekar  A, Baez  J C, Corichi  A,

1031: and Krasnov K 1998 Quantum geometry and black hole entropy

1032: \textit{Phys. Rev. Lett.} \textbf{80} 904--907

1033: 

1034: \bibitem{abk} Ashtekar  A, Baez  J C   and Krasnov  K 2000 Quantum

1035: geometry of isolated horizons and black hole entropy  \textit{Adv.

1036: Theo. Math. Phys.}  \textbf{4}  1--95\\

1037: Ashtekar A, Engle J and Van Den Broeck C 2005 Quantum horizons and

1038: black hole entropy: Inclusion of distortion and rotation

1039: \textit{Class. Quantum Grav.} \textbf{22} L27-L38

1040: 

1041: \bibitem{fw} Wilczek F 1993 Quantum purity at a small price:

1042: Erasing a black hole paradox \texttt{hep-th/9302096}

1043: 

1044: \bibitem{rp} Penrose R 1965 Zero rest mass fields including

1045: gravitation: asymptotic behavior \textit{Proc. R. Soc. (London)}

1046: \textbf{A284} 159-203

1047: 

1048: \bibitem{ax} Ashtekar A and Xanthopoulos B C 1978 {Isometries compatible

1049: with asymptotic flatnes at null infinity: A complete description}

1050: \textit{J. Math. Phys.} \textbf{19} 2216--2222

1051: 

1052: %\bibitem{ps} Pawlowski T and Singh P 2005 personal communication

1053: 

1054: \end{thebibliography}

1055: \end{document}

1056: