1:
2:
3: \input harvmac
4: %\draftmode
5: \noblackbox
6: %-------------------------
7: % This paper uses harvmac
8: %-------------------------
9:
10: \font\ticp=cmcsc10
11: \font\cmss=cmss10 \font\cmsss=cmss10 at 7pt
12: \def\Title#1#2{\rightline{#1}\ifx\answ\bigans\nopagenumbers\pageno0\vskip1in
13: \else\pageno1\vskip.8in\fi \centerline{\titlefont #2}\vskip .5in}
14: \font\titlerm=cmr10 scaled\magstep3 \font\titlerms=cmr7 scaled\magstep3
15: \font\titlermss=cmr5 scaled\magstep3 \font\titlei=cmmi10 scaled\magstep3
16: \font\titleis=cmmi7 scaled\magstep3 \font\titleiss=cmmi5 scaled\magstep3
17: \font\titlesy=cmsy10 scaled\magstep3 \font\titlesys=cmsy7 scaled\magstep3
18: \font\titlesyss=cmsy5 scaled\magstep3 \font\titleit=cmti10 scaled\magstep3
19: \font\ticp=cmcsc10
20: \font\ttsmall=cmtt10 at 8pt
21:
22: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
23: %The following lines are needed to insert the accompanying figures
24: %in the paper. If you do not have epsf, then comment out the line
25: % ``\input epsf'', and print the figures separately.
26: \input epsf
27: %
28: \ifx\epsfbox\UnDeFiNeD\message{(NO epsf.tex, FIGURES WILL BE
29: IGNORED)}
30: \def\figin#1{\vskip2in}% blank space instead
31: \else\message{(FIGURES WILL BE INCLUDED)}\def\figin#1{#1}\fi
32: %
33: \def\ifig#1#2#3{\xdef#1{fig.~\the\figno}
34: \goodbreak\topinsert\figin{\centerline{#3}}%
35: \smallskip\centerline{\vbox{\baselineskip12pt
36: \advance\hsize by -1truein\noindent{\bf Fig.~\the\figno:} #2}}
37: \bigskip\endinsert\global\advance\figno by1}
38: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
39:
40: %references
41:
42: \lref\fgl{B.~Freivogel, S.~B.~Giddings and M.~Lippert,
43: {\it Toward a theory of precursors},
44: hep-th/0207083.}
45: \lref\dkk{U. Danielsson, E. Keski-Vakkuri and M. Kruczenski, {\it
46: Spherically Collapsing Matter in AdS, Holography, and Shellons},
47: Nucl. Phys. B563 (1999) 279, hep-th/9905227.}
48: \lref\dkkk{U. Danielsson, E. Keski-Vakkuri and M. Kruczenski, {\it
49: Vacua, Propagators, and Holographic Probes in AdS/CFT},
50: JHEP 9901 (1999) 002, hep-th/9812007.}
51: \lref\dkkbh{
52: U.~H.~Danielsson, E.~Keski-Vakkuri and M.~Kruczenski,
53: {\it Black hole formation in AdS and thermalization on the boundary},
54: JHEP 0002, 039 (2000)
55: hep-th/9912209.}
56: \lref\simon{V. Balasubramanian and S. F. Ross,
57: {\it Holographic Particle Detection},
58: Phys. Rev. D61 (2000) 044007,
59: hep-th/9906226}
60: \lref\juan{J. Maldacena,
61: {\it The Large N Limit of Superconformal Field Theories and Supergravity},
62: Adv. Theor. Math. Phys. 2 (1998) 231, hep-th/9711200.}
63: \lref\magoo{O. Aharony, S.S. Gubser, J. Maldacena, H. Ooguri, Y. Oz,
64: {\it Large N Field Theories, String Theory and Gravity},
65: Phys. Rept. 323 (2000) 183, hep-th/9905111.}
66: \lref\witten{ E. Witten,
67: {\it Anti De Sitter Space And Holography},
68: Adv. Theor. Math. Phys. 2 (1998) 253, hep-th/9802150.}
69: \lref\witt{ E. Witten,
70: {\it Anti-de Sitter Space, Thermal Phase Transition,
71: And Confinement In Gauge Theories},
72: Adv. Theor. Math. Phys. 2 (1998) 505, hep-th/9803131.}
73: \lref\gkp{S. Gubser, I. Klebanov, and A. Polyakov,
74: {\it Gauge Theory Correlators from Non-Critical String Theory},
75: Phys. Lett. B428 (1998) 105,
76: hep-th/9802109.}
77: \lref\hoit{G. Horowitz and N. Itzhaki,
78: {\it Black Holes, Shock Waves, and Causality in the AdS/CFT Correspondence},
79: JHEP 9902 (1999) 010, hep-th/9901012}
80: \lref\gary{G. Horowitz,
81: {\it Comments on Black Holes in String Theory},
82: Class. Quant. Grav. 17 (2000) 1107, hep-th/9910082.}
83: \lref\peet{A. Peet and J. Polchinski,
84: {\it UV/IR Relations in AdS Dynamics},
85: Phys. Rev. D59 (1999) 065011,
86: hep-th/9809022.}
87: \lref\suwi{L. Susskind and E. Witten,
88: {\it The Holographic Bound in Anti-de Sitter Space},
89: hep-th/9805114.}
90: \lref\suto{L. Susskind and N. Toumbas,
91: {\it Wilson Loops as Precursors},
92: Phys.Rev. D61 (2000) 044001, hep-th/9909013 .}
93: \lref\pst{J. Polchinski, L. Susskind, and N. Toumbas,
94: {\it Negative Energy, Superluminosity and Holography},
95: Phys.Rev. D60 (1999) 084006, hep-th/990322.}
96: \lref\lmr{J. Louko, D. Marolf, and S. F. Ross,
97: {\it On geodesic propagators and black hole holography},
98: Phys.Rev. D62 (2000) 044041, hep-th/0002111.}
99: \lref\giro{S. Giddings and S. Ross,
100: {\it D3-brane shells to black branes on the Coulomb branch},
101: Phys.Rev. D61 (2000) 024036,
102: hep-th/9907204.}
103: \lref\gili{S.~B.~Giddings and M.~Lippert,
104: {\it Precursors, black holes, and a locality bound},
105: Phys.\ Rev.\ D 65, 024006 (2002),
106: hep-th/0103231.}
107: \lref\wl{D.~Beckman, D.~Gottesman, A.~Kitaev and J.~Preskill,
108: {\it Measurability of Wilson loop operators},
109: Phys.\ Rev.\ D 65, 065022 (2002),
110: hep-th/0110205.}
111: \lref\so{G.~T.~Horowitz and V.~E.~Hubeny,
112: {\it CFT description of small objects in AdS},
113: JHEP 0010, 027 (2000),
114: hep-th/0009051.}
115: \lref\matschull{
116: H.~J.~Matschull,
117: {\it Black hole creation in 2+1-dimensions},
118: Class.\ Quant.\ Grav.\ 16, 1069 (1999),
119: gr-qc/9809087;
120: S.~Holst and H.~J.~Matschull,
121: {\it The anti-de Sitter Gott universe: A rotating BTZ wormhole},
122: Class.\ Quant.\ Grav.\ 16, 3095 (1999)
123: gr-qc/9905030.}
124: \lref\loma{
125: J.~Louko and D.~Marolf,
126: {\it Single-exterior black holes and the AdS-CFT conjecture},
127: Phys.\ Rev.\ D 59, 066002 (1999),
128: hep-th/9808081.}
129: \lref\maldacena{J.~M.~Maldacena,
130: {\it Eternal black holes in Anti-de-Sitter},
131: hep-th/0106112.}
132: \lref\bklt{
133: V.~Balasubramanian, P.~Kraus, A.~E.~Lawrence and S.~P.~Trivedi,
134: {\it Holographic probes of anti-de Sitter space-times},
135: Phys.\ Rev.\ D 59, 104021 (1999),
136: hep-th/9808017.}
137: \lref\gregross{
138: J.~P.~Gregory and S.~F.~Ross,
139: {\it Looking for event horizons using UV/IR relations},
140: Phys.\ Rev.\ D 63, 104023 (2001),
141: hep-th/0012135.}
142: \lref\kali{
143: D.~Kabat and G.~Lifschytz,
144: {\it Gauge theory origins of supergravity causal structure},
145: JHEP 9905, 005 (1999), hep-th/9902073.}
146: \lref\bdhm{
147: T.~Banks, M.~R.~Douglas, G.~T.~Horowitz and E.~J.~Martinec,
148: {\it AdS dynamics from conformal field theory},
149: hep-th/9808016.}
150:
151:
152: %-------------------
153: % title page
154: %-------------------
155: %
156: \baselineskip 16pt
157: \Title{\vbox{\baselineskip12pt
158: \line{\hfil SU-ITP-2/32}
159: \line{\hfil \tt hep-th/0208047} }}
160: {\vbox{
161: {\centerline{Precursors see inside black holes}
162: }}}
163: \centerline{\ticp Veronika E. Hubeny\footnote{}{\ttsmall
164: veronika@itp.stanford.edu}}
165: \bigskip
166: \centerline{\it Department of Physics, Stanford University,
167: Stanford, CA 94305, USA}
168: \bigskip
169: \centerline{\bf Abstract}
170: \bigskip
171: We argue that, given the nonlocal nature of precursors in AdS/CFT
172: correspondence,
173: the boundary field theory contains information about events
174: inside a black hole horizon.
175: The essence of our proposal is sketched in figure 1, and relies
176: on the global nature of event horizons.
177:
178: \Date{August, 2002}
179: %
180: \newsec{Introduction}
181:
182: With the pressing problems related to string theory in cosmological
183: settings, there has been a resurgence of interest in physics associated
184: with interiors of black holes.
185: So far, the best handle we have appears in the context of the AdS/CFT
186: correspondence \refs{\juan,\witten,\gkp,\magoo},
187: where string theory on AdS (which is a gravitational
188: theory, and can contain black holes) has a dual description in terms of
189: a field theory ``living on the boundary''.
190: The details of the correspondence remain yet to be unraveled, and
191: its causal properties to be better understood; nevertheless, there are
192: certain expectations on how the bulk information is to be holographically
193: encoded on the boundary.
194:
195: One system in which we can try to ask these questions is the large
196: Schwarzschild black hole in $AdS_5 \times S^5$.
197: This is well known to correspond to an approximately thermal state in the
198: gauge theory.
199: The spacetime has an event horizon, i.e.\ the black hole interior is causally
200: disconnected from the boundary.
201: One may then ask: {\it Is the physics inside the event horizon encoded by the
202: boundary theory?}
203: Affirmative answer immediately leads to the
204: more ambitious question\foot{
205: Such questions have been considered previously
206: \refs{\simon,\lmr,\gili,\maldacena}, but
207: despite extensive work, these issues still remain somewhat obscure.
208: The holographic representation of horizons in terms of the
209: dual gauge theory has also been discussed
210: \refs{\bdhm,\bklt,\kali,\dkkbh,\gregross};
211: we believe that the lack of complete success in this venue
212: stems largely from the global nature of the horizon.}: {\it How?}
213: Once this is understood, one may hope
214: to come closer to the resolution of the black hole information paradox,
215: as well as to get a better handle on the singularity inside the horizon.
216: While these are certainly worthy---and rather lofty---goals by themselves,
217: a better
218: understanding and ``resolution'' of spacelike singularities would provide
219: a new window into studying cosmological singularities in string theory.
220:
221: In this brief note, we point out that under the commonly-accepted assumption
222: of precursors, at least part of the spacetime inside the horizon will be
223: accessible to the field theory.
224: In the following we will explain this idea in more detail, using a
225: simple gedanken experiment which exemplifies it. No explicit calculations
226: are presented, since it is the matter-of-principle point that we wish to
227: stress.
228:
229: \newsec{Gedanken experiment}
230:
231: Let us, for the moment, consider pure AdS, without any black hole present.
232: Even certain local boundary operators contain a
233: remarkable amount of information about static objects deep inside
234: the bulk of AdS \so;
235: but since such operators are determined only from the asymptotic values of
236: the bulk fields,
237: any dynamical information is propagated out in a manner
238: consistent with bulk causality.
239: In particular, no local operators on the boundary
240: can see events inside a causally disconnected region, such as a black hole.
241: However, as has been argued \refs{\pst,\suto,\fgl},
242: an event in the bulk is actually encoded by
243: {\it nonlocal}
244: operators on the boundary, with the associated scale dictated by the UV/IR
245: correspondence \refs{\suwi,\peet}.
246: A concrete proposal has been put forth that such operators are decorated
247: Wilson loops.
248: While this has not yet been proved\foot{
249: The identification of precursors with Wilson loops has been suggested in
250: \pst\ and expanded in \suto; while the latter's calculation was challenged
251: by \gili, further analysis \fgl\ suggested that the precursor
252: information is coded in high frequency components of Wilson loops, or
253: the so-called decorated Wilson loops.},
254: the essential point is the existence of {\it some} set
255: of nonlocal operators encoding the bulk, apparently ``instantaneously'',
256: i.e.\ before any message sent from the event could reach the boundary.
257:
258: If we accept this assumption, the fact that the boundary field theory
259: must also contain some information about physics inside a horizon
260: follows almost immediately.
261: (A similar observation has been made recently in \gili, but their
262: attempts at exemplifying the process through ``flossing the black hole''
263: cast some doubt on the possibility of extracting information from inside
264: the horizon.
265: Here we suggest a much simpler process, which does not rely on
266: inconclusive tree level computations.)
267: The event horizon is a global object, defined as the boundary of the
268: past of the future infinity,
269: which means that we cannot determine the presence or position of the
270: event horizon until we know the entire future evolution of the spacetime.
271: In particular, an innocent-looking, empty, locally AdS part
272: of spacetime can in fact be an interior of a black hole, if later
273: there forms a large enough black hole for its horizon to stretch far enough
274: into the past to encompass the event in question.
275:
276: \ifig\figa{Sketch of the proposed process of abstracting information
277: from inside of the horizon:
278: a) An event $p$, in locally pure AdS space is measured by $W$.
279: b) After this measurement is performed, a shell
280: $s$ collapses, forming a black hole with a horizon $H$ which encompasses
281: $p$.}
282: {\epsfxsize=10cm \epsfysize=8cm \epsfbox{figa.eps}}
283:
284: Taking advantage of this fact, we can consider the following gedanken
285: experiment:
286: Start with ``empty'' AdS space,
287: and consider some event, labeled $p$ in Fig.1a.
288: As a boundary observer, one can obtain instantaneous information
289: about the event $p$, for instance by measuring\foot{
290: While it was argued \wl\ that a Wilson loop cannot be measured in
291: the conventional non-demolition sense, it can be measured in a
292: suitable demolition experiment, such as sketched out in \fgl, which
293: is all we need for the present argument.}
294: appropriately decorated Wilson loop $W$.
295: {\it After} this measurement has been made, i.e.\ entirely to the
296: future of $W$, one can send in a shell $s$ of radiation with sufficient
297: energy such that when it implodes in the center of AdS,
298: it forms a large black hole, as sketched in Fig.1b.
299: If the shell is spherically symmetric, the spacetime everywhere outside
300: this shell will be that of Schwarzschild-$AdS_5 \times S^5$, whereas
301: inside the shell the spacetime is described by pure $AdS_5 \times S^5$.
302: The global event horizon for this spacetime will be obtained by tracing back
303: the radially outgoing null rays, denoted by $H$ in Fig.1b.
304: As apparent from Fig.1, these can originate at the center of AdS prior to $p$.
305: In particular, provided the shell is sent in
306: within time of order the AdS radius,
307: one can always make the resulting black hole large
308: enough for its horizon to encompass the event $p$.
309:
310: \ifig\figb{Penrose diagram of the proposed process of abstracting information
311: from inside of the horizon corresponding to the sketch in Fig.1b.}
312: {\epsfxsize=4cm \epsfysize=6.4cm \epsfbox{figb.eps}}
313:
314: This implies that one has in fact succeeded in ``measuring'' an event,
315: $p$, which is inside a black hole horizon.
316: This is seen more clearly from the corresponding Penrose diagram, Fig.2.
317:
318: \newsec{Discussion}
319:
320: Several comments about this process are in order.
321: First, while we claim that, through precursors,
322: the boundary field theory encodes some information about events in
323: the interior of black holes, our claim does not imply violations of causality.
324: This is simply because the ``measurement'', $W$, is not local. No message
325: is being sent between any two spacelike-separated points.
326: Correspondingly, one might then object that 1) a local field theory observer
327: can't set up the collapse of the shell fast enough if he waits till after
328: $W$ to decide whether to collapse the shell or not, and 2) a single local
329: observer cannot get any information out of the nonlocal operator $W$ before
330: the shell is to be sent in.
331: Both of these statements are true, but they do not contradict the claim that
332: there is some quantity, $W$, which encodes some information about $p$, and
333: yet is unaffected (due to the causality of the local quantum field theory
334: on the boundary) by the eventual creation of the horizon.
335: In fact, this is suggestive of the generic expectation that extracting
336: information out of black holes must necessarily involve some nonlocal
337: physics.
338:
339: Another objection that one might raise concerns the sudden creation of the
340: shell's energy, i.e.\ that there is a source introduced
341: on the boundary.
342: This however does not refute the basic point that
343: $p$ can be measured despite being inside the horizon.
344: One may alternately set up the experiment with the shell's energy
345: being present for all time, as long as the shell implodes and creates
346: a black hole when required.
347: We have been talking about a ``shell'' without specifying how it
348: is to be constructed from string theory, because such
349: detailed information is also irrelevant to the basic point.
350: One can in fact consider more complicated nonspherical geometries, such as
351: several colliding gravitational shock waves,
352: which may nevertheless be easier to discuss in
353: the dual field theory \hoit.
354:
355: Clearly, for obtaining a specific demonstration of extracting
356: information from a black hole,
357: we would first need to specify exactly all the components
358: in our set-up before any explicit calculation could be carried out.
359: So far, the main obstacle to verifying our claim by an
360: explicit computation is the incomplete
361: understanding of the holographic encoding of the bulk, namely the
362: exact nature of $W$.
363: A possible simplification is to work in lower dimensions
364: where exact solutions are easier to find, such
365: as in $AdS_3 \times S^3 \times T^4$; however, even in this context,
366: sufficient understanding of the precursors has not yet been achieved.
367:
368: It is not clear from our construction
369: how much information one can actually extract out of
370: the black hole, and what part of the spacetime inside the black hole
371: can one extract information about.
372: In the above example, we claimed that we can learn something about
373: the part of spacetime to the past of the shell, since there the
374: picture is the same as for pure AdS.
375: In terms of Fig.2, our process can access the left wedge,
376: but it is less clear whether we can similarly extract any information
377: from the top wedge.
378: Since the left wedge is still far form the singularity, one might
379: worry that our proposal cannot hope address the motivation
380: of resolving the singularity.
381: This does not appear to be a fundamental obstacle, since one may hope
382: to get closer to the singularity
383: by considering different types of collapse.
384:
385: In summary,
386: we are still faced with the outstanding question of whether or not the gauge
387: theory can probe physics inside the event horizon.
388: Since this region is causally disconnected from the boundary,
389: one may fear that the existence of the horizon will prevent
390: the gauge theory from ``seeing'' what goes on behind it,
391: as is
392: true for the local operators.
393: In this note
394: we have shown that, given the existence of precursors,
395: which are nonlocal operators, one can find situations
396: in which the boundary theory does know about events behind the horizon.
397: While our gedanken experiment does not address extracting
398: any information from an existing black hole,
399: it demonstrates the important point that
400: the causal properties
401: of the horizon do not pose a fundamental obstacle to learning something
402: about the physics behind it.
403: Once this crucial causal obstacle is removed, extracting {\it all}
404: information out of black holes may seem more hopeful.
405:
406:
407: \vskip 1cm
408:
409:
410: \centerline{\bf Acknowledgements}
411: I would like to thank
412: Lenny Susskind and Steve Shenker for encouraging me to write-up the idea,
413: and Gary Horowitz for initial discussions.
414: This work was supported by NSF Grant PHY-9870115.
415:
416:
417: %
418: % ====================================================
419: %
420:
421: \listrefs
422: \end
423: