John Archibald Wheeler, The Physicist Who Made Information A Foundation

The physicist who named the black hole and then went further

John Archibald Wheeler was the American theoretical physicist who worked out the mechanism of nuclear fission with Niels Bohr and trained many of the people who built modern gravitational physics. He was born in Jacksonville, Florida on 9 July 1911 and he died at Hightstown, New Jersey on 13 April 2008, aged ninety-six. His working life ran the length of the century’s physics. It began at the atomic nucleus, moved to general relativity at a time when almost nobody serious was working on it, and ended with a claim about the nature of existence itself.

Two things set him apart from other physicists of his rank. The first is that he was a gifted namer, and the words he fixed to things have outlived most of the equations his generation produced. The second is that he taught, and taught relentlessly. Forty-six students took a doctorate under him at Princeton. The list runs from Richard Feynman to Hugh Everett, Charles Misner, Kip Thorne, Jacob Bekenstein and Bill Unruh.

The late Wheeler is the one most often quoted and least often read properly. From the late 1970s he argued that the physical world is not made of matter with information written on top of it, but the other way round. He put the claim in three words in 1989. Those three words, “it from bit”, sit under a good deal of what quantum information science now says about itself. They are also the part of his work that gets mangled most reliably in the retelling.

From Johns Hopkins to Copenhagen, and then to fission

Wheeler left Baltimore City College high school in 1926 and went to Johns Hopkins University on a Maryland scholarship. He took his doctorate there in 1933, at the age of twenty-one, with a thesis on the dispersion and absorption of helium supervised by Karl Herzfeld. A National Research Council fellowship followed. He spent 1933 to 1934 with Gregory Breit at New York University and then 1934 to 1935 with Niels Bohr in Copenhagen, which shaped much of what he later thought about measurement.

Otto Hahn and Fritz Strassmann showed on 19 December 1938 that a uranium nucleus had split, and Lise Meitner and Otto Frisch worked out what that meant the following month. Bohr carried their interpretation across the Atlantic on his way to lecture at Princeton, and he and Wheeler set to work on what was actually happening inside the nucleus when it broke. Their paper, The Mechanism of Nuclear Fission, appeared in Physical Review on 1 September 1939, the day Germany invaded Poland. It treated the nucleus as a drop of charged liquid, wobbling under the competition between surface tension holding it together and electrical repulsion pulling it apart.

The paper’s most consequential conclusion was about which uranium does the splitting. It deduced that uranium-235, the rare isotope, was the major contributor to slow-neutron fission, while the far more abundant uranium-238 fissions mainly at high neutron energies. That distinction is not a detail. Separating uranium-235 from uranium-238 is hard, slow and enormously expensive, and the Manhattan Project had to do it at industrial scale because of what this paper said. Pencil-and-paper nuclear theory set the shape of the plants.

Hanford, Matterhorn and the weapons work

Wheeler joined the Metallurgical Laboratory in Chicago in January 1942 and moved to Hanford in Washington State in July 1944 to work on the reactors that would breed plutonium. The B Reactor there was completed on 13 September 1944 and went critical on 26 September 1944. It then did something nobody had planned for. It shut itself down, and about fifteen hours later it started up again on its own.

Wheeler worked out what had happened. A fission product, iodine-135, was decaying into xenon-135, and xenon-135 turned out to soak up neutrons with extraordinary greed. In plain terms the reactor was poisoning itself with its own waste, taking in the neutrons it needed to keep going. He was not alone in the diagnosis, since a paper by Chien-Shiung Wu on xenon-135 had reached Enrico Fermi and fed the same conclusion. The cure was to charge the pile with more uranium, and the B Reactor produced its first plutonium on 6 November 1944.

He did not stop when the war did. In 1951 Wheeler established Project Matterhorn at Princeton, a laboratory run in two halves, and he took charge of Matterhorn B, the half doing thermonuclear weapons work. That work fed directly into the Ivy Mike test at Enewetak Atoll on 1 November 1952, the first full-scale hydrogen device, with a yield of about 10.4 megatons. Wheeler campaigned for the hydrogen bomb and recruited for it. He regarded it as a duty, not a regret. Any account of him that skips this is not an account of him.

The names that stuck, and the one he did not coin

The black hole story is told wrongly almost every time. Wheeler did not coin the phrase. The science writer Marcia Bartusiak credits Robert Dicke in the early 1960s with comparing a collapsed star to the Black Hole of Calcutta. The term reached print in Life and Science News during 1963. The first recorded use in writing is normally given to the journalist Ann Ewing. Her article Black Holes in Space is dated 18 January 1964 and reports a meeting of the American Association for the Advancement of Science in Cleveland, Ohio.

What Wheeler did was adopt it. In December 1967, at a lecture he was giving to the same association, somebody in the audience offered the phrase, and Wheeler took it up for its brevity and its advertising value. His standing in the field did the rest, and within a few years it had pushed aside the longer descriptive phrases physicists had been using. He was a populariser here rather than an inventor. The distinction matters, because it is the pattern of his whole naming career.

Wormhole is closer to his own. Hermann Weyl had described something similar in 1928 but called them one-dimensional tubes. The term entered physics properly in a 1957 paper by Charles Misner and Wheeler, who wrote of what physicists might perhaps be excused for more vividly terming a wormhole. Quantum foam is the third. The idea that spacetime is violently uneven at the smallest scales came from Geons, a Physical Review paper of 1955. The scale he had in mind is the Planck length, about ten to the minus thirty-five metres.

None of this has been seen. Quantum foam remains a picture rather than a measurement, and attempts to catch it have come back empty. A 2005 result from the MAGIC gamma-ray telescope hinted that high-energy photons might travel very slightly slower through a grainy vacuum. Later work could not confirm it, and searches for violations of Lorentz invariance have found nothing.

The equation with no time in it

For about fifteen years Wheeler pursued a programme he called geometrodynamics, an attempt to describe spacetime and everything in it using geometry alone. The slogans were mass without mass, charge without charge and field without field. If you could bend empty space hard enough, he thought, you would not need to add particles to it, because the bends would behave like particles all by themselves. His geons were the test case, bundles of gravitational wave energy held together by their own pull.

It did not work out. Spin one-half particles, which is to say electrons and everything else that makes ordinary matter, proved very hard to build out of geometry alone, and Wheeler himself punctured several of the early hopes. The programme is remembered less for its results than for its by-products. Wormholes came out of it, quantum foam came out of it, and so did the hardest surviving problem in the subject.

That problem sits inside the Wheeler-DeWitt equation. Bryce DeWitt first published it in 1967, in a paper called Quantum Theory of Gravity. It is an attempt to write down quantum mechanics for the whole universe, geometry included, and it has a startling feature. There is no time variable in it. The ordinary Schrödinger equation tells you how a state changes from one moment to the next. Here the energy operator becomes a constraint rather than a generator of change. The equation no longer describes evolution at all.

This is the problem of time, and it is easy to state and very hard to answer. Quantum mechanics treats time as a background parameter that ticks along outside the system, while general relativity treats time as part of the system, bendable and observer-dependent, with no outside to tick in. Put the two together honestly and the wave function of the universe comes out frozen, a single unchanging object, even though everything inside the universe plainly changes. Nearly sixty years on there is no agreed way out.

The students, and the one whose conclusion he rejected

Feynman came first. As a graduate student at Princeton he and Wheeler built a theory of radiation. On their account light is not something a charge does by itself. It is a transaction between an emitter and everything that eventually absorbs it. Their absorber theory appeared in Reviews of Modern Physics in April 1945, with a follow-up in the same journal in July 1949. It treated waves going forwards and backwards in time on equal footing, and it removed the infinities that came from a charge acting on itself. The theory is not part of standard physics today, but the habit of starting from an action over whole histories carried straight into Feynman’s path integral.

Hugh Everett is the harder case. Everett took his Princeton doctorate under Wheeler in 1957 with a dissertation on the foundations of quantum mechanics, having first written a much longer manuscript called The Theory of the Universal Wave Function. Wheeler pushed him to cut it and soften it. The published version ran in Reviews of Modern Physics in 1957 as a compromise between the two men, with a favourable note from Wheeler attached. Everett was not happy with the final form, and Bohr’s circle in Copenhagen rejected it outright. Our own account of what happened to Hugh Everett covers what happened to him afterwards.

Wheeler never accepted the interpretation he had supervised into existence. By 1980 he was saying that the many-worlds interpretation creates too great a load of metaphysical baggage to carry along, and he meant it. He had defended the work, placed it, and got it into print, and he still thought it was wrong, which is a combination rarer in physics supervision than it ought to be.

The gravitational students are a roll call. Charles Misner finished in 1957, Kip Thorne in 1965, Bill Unruh in 1971 and Jacob Bekenstein in 1972. Thorne went on to co-found the LIGO project in 1984 and shared the 2017 Nobel Prize in Physics with Rainer Weiss and Barry Barish for the detection of gravitational waves. Unruh showed in 1976 that an accelerating observer sees a warm bath of radiation where an observer moving steadily sees nothing at all. Misner, Thorne and Wheeler wrote the textbook that taught the subject to two generations, Gravitation, published by W. H. Freeman in September 1973 and known to everyone in the field as MTW.

Bekenstein matters most for what comes next. In 1972 he was the first to suggest that black holes have a well-defined entropy proportional to the area of the event horizon. Stephen Hawking objected at first, on the reasonable ground that a thing which cannot radiate cannot have a temperature. Then in 1974 he found that black holes do radiate after all, and he fixed the constant at one quarter. A black hole’s information content scales with its surface, not its volume, as our explainer on the the black hole information paradox sets out. That single fact is the seed of everything later called holographic.

Delayed choice, and what it does not show

In 1978 Wheeler published The Past and the Delayed-Choice Double-Slit Experiment, and it is the sharpest thing anyone has written about quantum measurement. Send a single photon into an apparatus with two possible paths. If you recombine the paths at the end, you see interference, which is the behaviour we call wave-like. If you leave them apart, you find out which path the photon took, which is the behaviour we call particle-like. Wheeler’s move was to delay the decision, and choose after the photon is already inside.

He then scaled it up until it could not be dismissed. Take light from a quasar billions of light years away, split into two paths by the gravity of an intervening galaxy, and make the choice tonight, at the telescope. In Wheeler’s own words, we have a strange inversion of the normal order of time. By moving the mirror in or out we have an unavoidable effect on what we have a right to say about the already past history of that photon. Note the careful phrasing. He wrote about what we have a right to say, not about what happened.

The experiment has been done. Thomas Hellmuth, Herbert Walther, Arthur Zajonc and Wolfgang Schleich ran a delayed-choice interferometer with a fast optical shutter in 1987. The clean version came from Vincent Jacques and colleagues in Science in 2007. They used genuine single photons from a defect in diamond, and a quantum random number generator made the choice, far enough away that no signal could have reached the photon. Closed, the interferometer gave interference with a visibility of ninety-four per cent, and open, it identified the path taken with an error probability below one per cent.

Now the part readers get wrong. This is not evidence that the future changes the past. Xiao-song Ma, Johannes Kofler and Anton Zeilinger put it flatly in their 2016 review. They write that quantum effects can mimic an influence of future actions on past events. They also write that no physical interactions or signals, let alone into the past, are necessary to explain the experimental results. The paradox dissolves once you stop treating the quantum state as a physical object sitting out there in the apparatus. Treat it instead as a bookkeeping device for what you can predict, and the temporal order of your measurements stops mattering. Our explainer on how reading a qubit works covers the same ground for a working qubit.

It from bit, in his own words

Wheeler presented Information, Physics, Quantum: The Search for Links at the ISQM conference in Tokyo held from 28 to 31 August 1989, and it was printed in the proceedings at pages 354 to 368. The abstract poses one question, and it is not a modest one. How come existence? His answer is that nothing in the description of physics is closer to the bottom than the act of putting a yes-or-no question to nature through an instrument and getting an answer.

Here is the passage itself, quoted exactly as he wrote it. The long dashes are his own.

It from bit. Otherwise put, every it — every particle, every field of force, even the spacetime continuum itself — derives its function, its meaning, its very existence entirely — even if in some contexts indirectly — from the apparatus-elicited answers to yes or no questions, binary choices, bits.

It from bit symbolizes the idea that every item of the physical world has at bottom — at a very deep bottom, in most instances — an immaterial source and explanation; that what we call reality arises in the last analysis from the posing of yes-no questions and the registering of equipment-evoked responses; in short, that all things physical are information-theoretic in origin and this is a participatory universe.

Read literally, that is a claim about what a physical quantity is. It is not a claim that the universe is a computer. Wheeler’s example is a photon detector. We ask whether the counter clicked during a given second, we get a yes or a no, and if it is yes we say a photon did it. He is blunt about the status of that photon. He writes that we know perfectly well it existed neither before the emission nor after the detection. Any talk of it existing in between is only a blown-up version of the raw fact, a count.

The participatory universe is the consequence he drew, and it comes with a closed loop rather than a foundation. He rejected what he called a tower of turtles, meaning an infinite regress of deeper levels, and offered a circle instead. In his words, physics gives rise to observer-participancy, observer-participancy gives rise to information, and information gives rise to physics. Observer here means any apparatus that makes an irreversible record. It does not mean a conscious mind, whatever the popular retellings suggest.

What working physicists make of it now

The respectable descendants are easy to point at. Black hole entropy came out of Wheeler’s own student, and it behaves exactly as an information count should. Wheeler said so in the 1989 paper, writing that the bits swallowed by a black hole show up as horizon area. Gerard ‘t Hooft generalised the area law in 1993, and Leonard Susskind gave it a string-theoretic form in 1995. The result is the holographic principle, which says the physical information inside any region is bounded by the area of its boundary rather than its volume.

The quantum information line is more direct than most people realise. The acknowledgements of the 1989 paper thank Charles Bennett, David Deutsch, Rolf Landauer, Benjamin Schumacher, William Wootters and Wojciech Zurek, which is close to a founding roster for the field. Anton Zeilinger and Caslav Brukner took the programme furthest. Their 2002 paper argues that if you allow one definite proposition per elementary constituent of nature, quantum physics becomes an elementary theory of information. Zeilinger shared the 2022 Nobel Prize in Physics with Alain Aspect and John Clauser, in part for pioneering quantum information science.

Now the objections, because this is not settled and should not be presented as though it were, and the most pointed of them comes from Rolf Landauer, whom Wheeler thanked. Landauer showed in 1961 that erasing a single bit must dissipate at least kT ln 2 of heat, a prediction confirmed by Bérut and colleagues in Nature in 2012. That result runs the opposite way to Wheeler, because it says information has a physical price. That makes the substrate look primary rather than derived.

The broader family of information-first ideas has not fared well empirically. Konrad Zuse proposed in 1967 that the universe computes its own behaviour and published Rechnender Raum in 1969, and the tradition that followed has produced no confirmed novel prediction. Discrete models of space tend to break the continuous symmetries that experiment confirms to high precision. Feynman argued in 1981 that a local classical system cannot efficiently reproduce quantum amplitudes, and that objection still stands. There is also the plainest complaint of all, that computation without a substrate has no causal properties to speak of. The modern form of the argument reaches the public through the simulation hypothesis, which our profile of Rizwan Virk examines on its own terms.

Two caveats are owed to Wheeler. His position is not the cellular-automaton one, and the it from bit slogan gets quoted approvingly by physicists who reject discreteness entirely. He was also unusually candid about what was missing, closing the 1989 paper by refusing to define his central term on the grounds that theory and measurement are born together. He named the one prediction by which his idea most clearly exposed itself to destruction in Popper’s sense, which was whether the bits available could ever balance the bits required.

What he left

Wheeler left Princeton in 1976 at sixty-five and directed the Center for Theoretical Physics at the University of Texas at Austin until 1986. The honours are a fair measure of how he was regarded. He took the Enrico Fermi Award in 1968, the Franklin Medal in 1969, the National Medal of Science in 1970 and the Wolf Prize in Physics in 1997. None of them were for it from bit.

What survives is a change in what counts as a respectable question. For most of the twentieth century, making a career out of asking what a measurement is marked you as a philosopher rather than a physicist. He asked it with the authority of a man who had built weapons and written the standard textbook on gravity, and he asked it in the language of bits. Quantum information science did not need his metaphysics to get going, and most of its practitioners do not hold it. It did need the questions to be askable, and that is the part he supplied. The line runs on through what quantum supremacy actually means and every argument about what a quantum machine really demonstrates.

Frequently asked questions