On Hawking’s Concession

John Preskill
24 July 2004

Last Wednesday (21 July 2004) at the GR17 meeting in Dublin, Ireland, Stephen Hawking conceded this bet he had made with me about whether black holes destroy information. (Kip Thorne, Stephen’s partner in the bet, has not yet decided whether to concede.) This event has received quite a bit of attention in the media. Of course, the main reason for all the fuss is that the public is fascinated by Stephen’s personal story, which is indeed inspiring. Though it is fun for my family and friends to see my name in the paper, the media coverage has been embarrassing for me, because some of the stories give the impression that Stephen’s concession has something to do with my own scientific contributions, which isn’t the case at all. (An exception is Dennis Overbye’s story in the New York Times, which does a pretty good job of putting the scientific issues in context.) Anyway, since so many people have asked for my reaction to this development, I am putting these comments on my website.

The black hole information loss puzzle

The bet concerns what happens to information that is hidden behind the event horizon of a black hole: Is the information destroyed and lost forever, or might it in principle be recovered from the radiation that is emitted as the black hole evaporates? This question was first raised by Hawking in this very important paper (written in 1975 and published over a year later). Hawking pointed out that the process of black hole evaporation (which he had discovered earlier) could not be reconciled with the principles of quantum physics and gravitational physics that were then generally accepted. This was a brilliant insight, and for much of the theoretical physics community it took many years to appreciate fully the depth of the problem that Hawking had posed.

Hawking had precipitated a genuine crisis in fundamental physics, and it seemed that we would have to give up at least one of our cherished beliefs. Hawking's radical suggestion was that the foundations of quantum theory needed to be revised. According to quantum mechanics, although physical processes can transform the information that is encoded in a physical system into a form that is inaccessible in practice, in principle the information can always be recovered. If I burn volume A of the encyclopedia or volume B of the encyclopedia, the flames and the ashes look pretty much alike in both cases, but actually there are subtle differences --- with sufficient cleverness and sufficiently advanced technology, I ought to be able to decipher the content of the volume by observing the flames and the ashes.
If black holes respect the usual rules of quantum physics, then if we toss volume A into a black hole, eventually we ought to be able (in principle) to read the encyclopedia by closely observing the light emitted as the black hole evaporates. But Hawking argued that black hole evaporation is fundamentally different that ordinary physical processes, that information that falls behind the event horizon of a black hole will be lost forever, remaining concealed even after the black hole has evaporated completely and disappeared.

The controversy

Many physicists, especially those whose background (like Hawking's) is in gravitational physics, accepted Hawking conclusion (and still do). But there is another possible point of view that has been favored by many physicists whose background (like mine) is in elementary particle physics: Perhaps information really can escape from a black hole, but we will not be able to understand just how the information escapes without deep new insights into the physics of processes in which both quantum effects and gravitational effects are important.

Two physicists who argued persuasively for this latter point of view were Gerard 't Hooft (Utrecht) and Leonard Susskind (Stanford) --- Susskind in particular had a big influence on my own thinking. Gradually, work by string theorists, especially Joseph Polchinski (Kavli Institute for Theoretical Physics), Andrew Strominger (Harvard), Cumrun Vafa (Harvard) and Juan Maldacena (Institute for Advanced Study) built a strong but rather circumstantial case that black holes really preserve information. Despite this impressive progress, we still lack a compelling picture of exactly how information escapes from an evaporating black hole.

Hawking’s reversal

Hawking has followed this work by the string theorists with great interest; I think he has been especially impressed by Maldacena's contributions, which suggest (as ’t Hooft, Susskind, and others had anticipated) that information is encoded in black hole spacetimes in a very subtle way. This past year he has been thinking a lot about how his earlier conclusions about information loss might be evaded, and in his talk in Dublin last Wednesday he outlined a new argument supporting the conclusion that information loss does not occur after all. Unfortunately, I don’t understand this argument well enough to attempt to summarize it here. (I hope that Hawking's forthcoming paper, which is expected to appear within a month of so, will provide details that were missing from the talk.) I do have the strong impression that Hawking's new approach, even if it becomes widely accepted, will still leave lingering questions about the explicit mechanism by which information escapes. Some of the physicists at the GR17 meeting who still agree with Stephen's earlier position were distressed by his reversal. And Kip Thorne, Stephen's partner in the bet, wants to get a chance to study Stephen's paper before he decides whether to concede.

Stephen’s change of heart surprised me; it isn't yet clear to me why he finds his new argument so compelling that he is willing to reverse a position that he has held adamantly for nearly 30 years. And it saddens me a little in a way --- Stephen and I have had many discussions about the black hole information puzzle over the years, and I have always enjoyed those discussions and have learned from them, even though we disagreed. If we are really on the same side now, our future discussions won't be as much fun. We'll find other things to disagree about, I suppose, but probably nothing else as deep and engrossing.

Settling the bet

There were quite a few print, television, and radio reporters present at Stephen’s GR17 talk, and at a press conference that immediately followed. Stephen had requested that Kip and I participate in the press conference, so we were actually on the stage during the talk. (Kip and I were both in Dublin to give talks at the conference on other topics.) At the conclusion of the talk, Stephen offered me a baseball encyclopedia to settle the bet, and I accepted it. (It would not have made sense to refuse, even though I’m not sure whether I agree with Stephen’s reason for conceding.)

A lot of people ask me why I requested a baseball encyclopedia. Well, the timing was good --- the 8th edition of Total Baseball: the Ultimate Baseball Encyclopedia was published this month. This edition is the first to compile Win Shares for every major league season in baseball history. That's a new statistic invented by my hero, Bill James, so I'm glad to have the book. (It was a schlep getting it home, though; it’s 2,688 pages.) My favorite article about the bet is Joe Friesen’s story for the Globe and Mail, describing the reactions of the author and publisher to the news. Incidentally, I wondered whether I should ask for a CD-ROM encyclopedia instead of a print encyclopedia. But print is the right choice --- it’s heavy like a black hole. And it takes much longer to get the information out --- that's like a black hole, too.

Stephen, Kip, and I make bets for fun, but the scientific question at issue in this case is an important one that we (and many other physicists) deeply care about. I don’t think any of us could have imagined the attention this bet would receive (I know I didn’t). It’s a bit humbling to think that no matter what else I do in life, my claim to fame will be that I won a couple of bets, but I figure I should just enjoy it.


Hits since 26 July 2004:
Free Hit Counter
Free Hit Counter