Frank Wilczek [1.14.09]

The most exciting thing that can happen is when theoretical dreams that started as fantasies, as desires, become projects that people work hard to build. There is nothing like it; it is the ultimate tribute. At one moment you have just a glimmer of a thought and at another moment squiggles on paper. Then one day you walk into a laboratory and there are all these pipes, and liquid helium is flowing, and currents are coming in and out with complicated wiring, and somehow all this activity is supposedly corresponds to those little thoughts that you had. When this happens, it's magic.

FRANK WILCZEK, a theoretical physicist at MIT and recipient of the Nobel Prize in Physics (2004), is known, among other things, for the discovery of asymptotic freedom, the development of quantum chromodynamics, the invention of axions, and the discovery and exploitation of new forms of quantum statistics (anyons). He is the author of Lightness of Being: Mass, Ether, and the Unification of Forces.

Frank Wilczek's Edge Bio Page


[FRANK WILCZEK:] In retrospect, I realize now that having the Nobel Prize hovering out there but never quite arriving was a heavy psychological weight; it bore me down. It was a tremendous relief to get it. Fortunately, it turns out I didn't anticipate that getting it is fantastic fun—the whole bit: there are marvelous ceremonies in Sweden, it's a grand party, and it continues, and is still continuing. I've been going to big events several times a month.

The most profound aspect of it, though, is that I've really felt from my colleagues something I didn't anticipate: a outpouring of genuine affection. It's not too strong to call it love. Not for me personally—but because our field, theoretical fundamental physics, gets recognition and attention. People appreciate what's been accomplished, and it comes across as recognition for an entire community and an attitude towards life that produced success. So I've been in a happy mood.

But that was a while ago, and the ceremonial business gets old after a while, and takes time. Such an abrupt change of life encourages thinking about the next stage. I was pleased when I developed a kind of three-point plan that gives me direction. Now I ask myself, when I'm doing something in my work: Is it relating to point one? Is it relating to point two? Is it relating to point three? If it's not relating to any of those, then I'm wasting my time.

Point one is in a sense the most straightforward. An undignified way to put it would be to say it's defending turf, or pissing on trees, but I won't say that: I'll say it's following up ideas that I've had physics in the past that are reaching fruition. There are several that I'm very excited about now. The great machine at CERN, the LHC, is going to start operating in about a year. Ideas—about unification and supersymmetry and producing Higgs particles—that I had a big hand in developing 20-30 years ago, are finally going to be tested. Of course, if they're correct that'll be a major advance in our understanding of the world, and very gratifying to me personally.

Then there's the area of exotic behavior of electrons at low temperature, so-called anyons, which is a little more tech It was thought for a long time that all particles were either bosons or fermions. In the early 80s, I realized there were other possibilities, and it turns out that there are materials in which these other possibilities can be realized, where the electrons organize themselves into collective states that have different properties from individual electrons and actually do obey the peculiar new rules, and are anyons. This is leading to qualitatively new possibilities for electronics. I call it anyonics. Recently advanced anyonics has been notionally bootstrapped into strategy for building quantum computers that might even turn out to be successful.

In any case, whether it's successful or not, the vision of anyonics—this new form of electronics—has inspired a lot of funding and experimentalists are getting into the game. Here similarly, there are kinds of experiments that have been in my head for 20 years but are very difficult, and people needed motivation and money to do them, that are now going to be done. It's a lot of fun to be involved in something that might actually have practical consequences and might even change the world. This stuff also, in a way, brings me back to my childhood because when I was growing up, my father was an electrical engineer and was taking home circuit diagrams, and I really admired these things. Now I get to think about making fundamentally new kinds of circuits, and it's very cool. I really like the mixture of abstract and concrete.

At a deeper level, what excites me about quantum computing and this whole subject of quantum information processing is that it touches such fundamental questions that potentially it could lead to qualitatively new kinds of intelligences. It's notorious that human beings have a hard time understanding quantum mechanics; it's hard for humans to relate to its basic notions of superpositions of states—that you can have Schrödinger's cat that's both dead and alive—that are not in our experience. But an intelligence based on quantum computers—mechanical quantum thinking—from the start would have that in its bones, so to speak, or in its circuits. That would be its primary way of thinking. It's quite challenging but fascinating to try to put yourself in the other guy's shoes, when that guy has a fundamentally different kind of mind, a quantum mind.

It's almost an embarrassment of riches, but some of the ideas I had about axions turn out to go together very very well with inflationary cosmology, and to get new pictures for what the dark matter might be. It ties into questions about anthropic reasoning, because with axions you get really different amounts of dark matter in different parts of the multiverse. The amounts of dark matter would be different elsewhere and the only way to argue about how much dark matter there should be turns out to be, if you have too much dark matter, life as we know it couldn't arise.

There's a lot of stuff in physics that I really feel I have to keep track of, and do justice to. That's point one.

The second point is another way of having fun: looking for outlets, cultivating a public, not just thinking about science all the time. I'm in the midst of writing a mystery novel that combines physics with music, philosophy, sex, the rule that only three people at most can share a Nobel Prize—and murder (or was it suicide?). When a four-person MIT-Harvard collaboration makes a great discovery in physics (they figure out what the dark matter is) somebody's got to go. That project and I hope other subsequent projects will be outlets in reaching out to the public and bringing in all of life and just having fun.

The third point is what I like to call the Odysseus project. I'm a great fan of Odysseus, the wanderer who had adventures and was very clever. I really want to do more great work—not following up what I did before, but doing essentially different things. I got into theoretical physics almost by accident; when I was an undergraduate, I had intended to study how minds work and neurobiology. But it became clear to me rather quickly, at Chicago, that that subject at that time wasn't ripe for the kind of mathematical analytical approach that I really like and get excited about, and am good at. I switched and majored in mathematics and eventually wound up in physics.

But I've always maintained that interest and in the meantime the tools available for addressing those questions have improved exponentially. Both in terms of studying the brain itself—imaging techniques and genetic techniques and a variety of others—but also the inspiring model of computation. The explosion of computational ability and understanding of computer science and networks is a rich source of metaphors and possible ways of thinking about the nature of intelligence and how the brain works. That's a direction I really want to explore more deeply. I've been reading a lot; I don't know exactly what I want to do, but I have been nosing out what's possible and what's available. I think it's a capital mistake, as Sherlock Holmes said, to start theorizing before you have the data. So I'm gathering the data.

Quantum Computers and Anyons

Quantum computing is an inspiring vision, but at present it's not clear what the technical means to carry it off are. There is a variety of proposals. It's not clear which is the best, or if any of them is practical.

Let me backtrack a little bit, though, because even before you get to a full-scale quantum computer, there are information processing tasks for which quantum mechanics could be useful with much less than a full-scale quantum computer. A full-scale quantum computer is extremely demanding: you have to build various kinds of gates, you have to connect them in complicated ways, you have to do error correction—it's very complicated. That's sort of like envisioning a supersonic aircraft when you're at the stage of the Wright brothers. However, there are applications that I think are almost in-hand.

The most mature is for a kind of cryptography: you can exploit the fact that quantum mechanics has this phenomenon that's roughly called 'collapse of the wave' function—I don't like it—I don't think that's a really good way to talk about it—but for better or worse, that's the standard terminology. Which in this case means that if you send a message that's essentially quantum mechanical—in terms of the direction of spins of photons, for instance—then you can send photons one by one with different spins and encode information that way. If someone eavesdrops on this, you can tell because the act of observation necessarily disturbs the information you're sending. So that's very useful. If you want to transmit messages and make sure that they haven't been eavesdropped, you can have that guaranteed by the laws of physics. If somebody eavesdrops, you'll be able to tell.

You can't prevent it, necessarily, but you can tell. If you do things right, the probability of anyone being able to eavesdrop successfully can be made negligibly small. So that's a valuable application that's almost tangible. People are beginning to try to commercialize that kind of idea.

I think in the long run the killer application of quantum computers will be doing quantum mechanics. Doing chemistry by numbers, designing molecules, designing materials by calculation. A capable quantum computer would let chemists and materials scientists work at a another level, because instead of having to mix up the stuff and watch what happens, you can just compute. We know exactly what the equations are that govern the behavior of nuclei and electrons and the things that make up atoms and molecules.

So in principle, it's a solved problem to figure out chemistry: just compute. We don't know all the laws of physics, but it's essentially certain that we know the adequate laws of physics with sufficient accuracy to design molecules and to predict their properties with confidence. But our practical ability to solve the equations is limited. The equations live in big multi-dimensional spaces, and they have a complicated structure and, to make a long story short, we can't solve any but very simple problems. With a quantum computer we'll be able to do much better.

As I sort of alluded to earlier, it's not decided yet what the best long-term strategy is for achieving powerful quantum computers. People are doing simulations and building little prototypes. There are different strategies being pursued based on nuclear spins, electron spins, trapped atoms, anyons.

I am very fond of anyons because I worked at the beginning on the fundamental physics involved. It was thought, until the late 70s and early 80s, that all fundamental particles, or all quantum mechanical objects that you could regard as discrete entities fell into two classes: so-called bosons after the Indian physicist Bose, and fermions, after Enrico Fermi. Bosons are particles such that if you take one around another, the quantum mechanical wave function doesn't change. Fermions are particles such that if you take one around another the quantum mechanical wave function is multiplied by a minus sign. It was thought for a long time that those were the only consistent possibilities for behavior of quantum mechanical entities. In the late 70s and early 80s, we realized that in two plus one dimensions, not in our everyday three dimensional space (plus one dimension for time), but in planar systems, there are other possibilities. In such systems, if you take one particle around another, you might get not a factor of one or minus one, but multiplication by a complex number—there are more general possibilities.

More recently, the idea that when you move one particle around another, it's possible not only that the wave function gets multiplied by a number, but that it actually gets distorted and moves around in bigger space, has generated a lot of excitement. Then you have this fantastic mapping from motion in real space as you wind things around each other, to motion of the wave function in Hibert space—in quantum mechanical space. It's that ability navigate your way through Hilbert space—that connects to quantum computing and gives you access to a gigantic space with potentially huge bandwidth that you can play around with in highly parallel ways, if you're clever about the things you do in real space.

But in anyons we're really at the primitive stage. There's very little doubt that the theory is correct, but the experiments are at a fairly primitive stage—they're just breaking now.

Quantum Logic and Quantum Minds

Quantum mechanics is so profound that it genuinely changes the laws of logic. In classical logic a statement is either true or false, there's no real sense of in-between. But in quantum mechanics you can have statements or propositions encoded in wave functions that have different components, some of which are true, some of which are false. When you measure the result is indeterminate. You don't know what you are going to get. You have states, meaningful states of computation, what you can think of as states of consciousness, that simultaneously contain contradictory ideas and can work with them simultaneously. I find that concept tremendously liberating and mind expanding. The classic structures of logic are really far from adequate to do justice to what we find in the physical world.

To do justice to the possible states, the possible conditions that just a few objects can be in, say, five spins, classically you would think you would have to say for each one whether it's up or down. At any one time they are in some particular configuration. In quantum mechanics, every single configuration—there are 32 of them, up or down for each spin—has some probability of existing. So simultaneously to do justice to the physical situation, instead of just saying that there is some configuration these objects are in, you have to specify roughly that there is a certain probability for each one and those probabilities evolve. But that verbal description is too rough because what's involved is not probabilities, it's something called amplitudes. The difference is profound.

Whereas probabilities have a kind of independence, with amplitudes the different configurations can interact with one other. There are different states which are in the physical reality and they are interacting with each other. Classically they would be different things that couldn't happen simultaneously. In quantum theory they coexist and interact with one another. That also goes to this issue of logic that I mentioned before. One way of representing true or false that is famously used in computers is, you have true as one and false as zero, spin up is true, spin down in false. In quantum theory the true statement and the false statement can interact with each other and you can do useful computations by having simultaneous propositions that contradict each other, sort of interacting with each other, working in creative tension. I just love that idea. I love the idea of opposites coexisting and working with one another. Come to think of it, it's kind of sexy.

More on Quantum Computers

Realizing this vision will be a vast enterprise. It's hard to know how long it's going to take to get something useful, let alone something that is competitive with the kind of computing we already have developed, which is already very powerful and keeps improving, let alone create new minds that are different from and more powerful than the kind of minds we're familiar with. We'll need to progress on several fronts.

You can set aside the question of engineering, if you like, and ask: Suppose I had a big quantum computer, what would I do with it, how would I program it, what kind of tasks could it accomplish? That is a mathematical investigation. You abstract the physical realization away. Then it becomes a question for mathematicians, and even philosophers have got involved in it.

Then there is the other big question: how do I build it? How do I build it in practice? That's a question very much for physicists. In fact, there is no winning design yet. People have struggled to make even very small prototypes. My intuition, though, is that when there is a really good idea, progress could be very rapid. That is what I am hoping for and going after. I have glimmers of how it might be done, based on anyons.

I've been thinking about this sort of thing on and off for a long time. I pioneered some of the physics, but other theorists including Alexei Kitaev and my former student Chetan Nayak have taken things to another level. There's now a whole field called "topological quantum computing" with its own literature, and conferences, and it's moving fast. What has changed is that now a lot of people, and in particular experimentalists, have taken it up.

The most exciting thing that can happen in the career of a theoretical physicist is when theoretical dreams that started as fantasies, as desires, become projects that people work hard to build. There is nothing like it; it is the ultimate tribute. At one moment you have just a glimmer of a thought and at another moment squiggles on paper. Then one day you walk into a laboratory and there are all these pipes, and liquid helium is flowing, and currents are coming in and out with complicated wiring, and somehow all this activity is supposedly corresponds to those little thoughts that you had. When this happens, it's magic.

Methods and Styles in Physics

The great art of theoretical physics is the revelation of surprising things about reality. Historically there have been many approaches to that art, which have succeeded in different ways. In the early days of physics, people like Galileo and Newton were very close to the data and stressed that they were trying to put observed behavior into mathematical terms. They developed some powerful abstract concepts, but by today's standards those concepts were down-to-earth; they were always in terms of things that you could touch and feel, or at least see through telescopes. That approach very much dominated physics, at least through the 19th century. Maxwell's great synthesis of electricity and magnetism and optics, and leading to the understanding that light was a form of electricity and magnetism, predicting new kinds of light that we call radio and microwaves and so forth—that came from a very systematic review of all that was known about electricity and magnetism experimentally and trying to put it into equations, noticing an inconsistency and fixing it up. That's the kind of classic approach.

In the 20th century, some of the most successful enterprises have looked rather different. Without going into the details it's hard to do justice to all the subtleties, but it's clear that theories like special relativity—especially general relativity—were based on much larger conceptual leaps. In constructing special relativity, Einstein abstracted just two very broad regularities about the physical world: that is that the laws of physics should look the same if you're moving at a constant velocity, and that the speed of light should be a universal constant. This wasn't based on a broad survey of a lot of detailed experimental facts and putting them together; it was selecting a few very key facts and exploiting them conceptually for all they're worth. General relativity even more so: it was trying to make the theory of gravity consistent with the insights of special relativity. This was a very theoretical enterprise, not driven by any specific experimental facts*, but led to a theory that changed our notions of space and time, did lead to experimental predictions, and to many surprises. (*Actually, there was a big "coincidence" that Newtonian gravity left unexplained, the equality of inertial and gravitational mass, which was an important guiding clue.)

The Dirac equation is a more complicated case. Dirac was moved by broad theoretical imperatives; he wanted to make the existing equation for quantum mechanical behavior of electrons—that's the Schrödinger equation—consistent with special relativity. To do that, he invented a new equation—the Dirac equation—that seemed very strange and problematic, yet undeniably beautiful, when he first found it. That strange equation turned out to require vastly new interpretations of all the symbols in it, that weren't anticipated. It led to the prediction of antimatter and the beginnings of quantum field theory.

This was another revolution that was, in a sense, conceptually driven. On the other hand, what gave Dirac and others confidence that his equation was on the right track, was that it predicted corrections to the behavior of electrons in hydrogen atoms that were very specific, and that agreed with precision measurements. This support forced them to stick with it, and find an interpretation to let it be true! So there was important empirical guidance, and encouragement, from the start.

Our foundational work on QCD falls in the same pattern. We were led to specific equations by theoretical considerations, but the equations seemed problematic. They were full of particles that aren't observed (quarks and—especially—gluons), and didn't contain any of the particles that are observed! We persisted with them nevertheless, because they explained a few precision measurements, and that persistence eventually paid off.

In general, as physics has matured in the 20th century, we've realized more and more the power of mathematical considerations of consistency and symmetry to dictate the form of physical laws. We can do a lot with less experimental input. (Nevertheless the ultimate standard must be getting experimental output: illuminating reality.) How far can esthetics take you? Should you let that be your main guide, or should you try to assemble and do justice to a lot of specific facts? Different people have different styles; some people try to use a lot of facts and extrapolate a little bit; other people try not to use any facts at all and construct a theory that's so beautiful that it has to be right and then fill in the facts later. I try to consider both possibilities, and see which one is fruitful. What's been fruitful for me is to take salient experimental facts that are somehow striking, or that seem anomalous—don't really fit into our understanding of physics—and try to improve the equations to include just those facts.

My reading of history is that even the greatest advances in physics, when you pick them apart, were always based on a firm empirical foundation and straightening out some anomalies between the existing theoretical framework and some known facts about the world. Certainly QCD was that way; when we developed asymptotic freedom to explain some behaviors of quarks, that they seem to not interact when they're close together seemed inconsistent with quantum field theory, but we were able to push and find very specific quantum field theories in which that behavior was consistent which essentially solved the problem of the strong interaction, and has had many fruitful consequences. Axions also—similar thing—a little anomaly— there's a quantity that happens to be very small in the world, but our theories don't explain why it's small; you can change the theories to make them a little more symmetrical—then we do get zero—but that has other consequences: the existence of these new particles rocks cosmology, and they might be the dark matter—I love that kind of thing.

String theory is sort of the extreme of non-empirical physics. In fact, its historical origins were based on empirical observations, but wrong ones. String theory was originally based on trying to explain the nature of the strong interactions, the fact that hadrons come in big families, and the idea was that they could be modeled as different states of strings that are spinning around or vibrating in different ways. That idea was highly developed in the late 60s and early 70s, but we put it out of business with QCD, which is a very different theory that turns out to be the correct theory of the strong interaction.

But the mathematics that was developed around that wrong idea, amazingly, turned out to contain, if you do things just right, and tune it up, to contain a description of general relativity and at the same time obeys quantum mechanics. This had been one of the great conceptual challenges of 20th century physics: to combine the two very different seeming kinds of theories—quantum mechanics, our crowning achievement in understanding the micro-world, and general relativity, which was abstracted from the behavior of space and time in the macro-world. Those theories are of a very different nature and, when you try to combine them, you find that it's very difficult to make an entirely consistent union of the two. But these evolved string theories seem to do that.

The problems that arise in making a quantum theory of gravity, unfortunately for theoretical physicists who want to focus on them, really only arise in thought experiments of a very high order—thought experiments involving particles of enormous energies, or the deep interior of black holes, or perhaps the earliest moments of the Big Bang that we don't understand very well. All very remote from any practical, do-able experiments. It's very hard to check the fundamental hypotheses of this kind of idea. The initial hope, when the so-called first string revolution occurred in the mid-1980s, was that when you actually solved the equations of string theory, you'd find a more or less unique solution, or maybe a handful of solutions, and it would be clear that one of them described the real world.

From these highly conceptual considerations of what it takes to make a theory of quantum gravity, you would be led "by the way" to things that we can access and experiment, and it would describe reality. But as time went on, people found more and more solutions with all kinds of different properties, and that hope—that indirectly by addressing conceptual questions you would be able to work your way down to description of concrete things about reality—has gotten more and more tenuous. That's where it stands today.

My personal style in fundamental physics continues to be opportunistic: To look at the phenomena as they emerge and think about possibilities to beautify the equations that the equations themselves suggest. As I mentioned earlier, I certainly intend to push harder on ideas that I had a long time ago but that still seem promising and still haven't been exhausted in supersymmetry and axions and even in additional applications of QCD. I'm also always trying to think of new things. For example, I've been thinking about the new possibilities for phenomena that might be associated with this Higgs particle that probably will be discovered at the LHC. I realized something I'd been well aware of at some low level for a long time, but I think now I've realized has profound implications, which is that the Higgs particle uniquely opens a window into phenomena that no other particle within the standard model would be sensitive to. If you look at the mathematics of the standard model, you discover that there are possibilities for hidden sectors—things that would interact very weakly with the kind of particles we've had access to so far, but would interact powerfully with the Higgs particles. We'll be opening that window. Very recently I've been trying to see if we can get inflation out of the standard model, by having the Higgs particle interact in a slightly nonstandard way with gravity. That seems promising too.

Most of my bright ideas will turn out to be wrong, but that's OK. I have fun, and my ego is secure.

On National Greatness

In 1983 the Congress of the United States canceled the SSC project, the Superconducting SuperCollider, that was under construction near Waxahachie, Texas. Many years of planning, many careers had been invested in that project, also $2 billion had already been put into the construction. All that came out of it was a tunnel from nowhere to nothing.

Now it's 2009 and a roughly equivalent machine, the Large Hadron Collider LHC, will coming into operation at CERN near Geneva. The United States has some part in that. It has invested half a billion dollars out of the $15 billion total. But it's a machine that is in Europe, really built by the Europeans; there's no doubt that they have contributed much more.

Of course, the information that comes out will be shared by the entire scientific community. So the end result, in terms of tangible knowledge, is the same. We avoided spending the extra money. Was that a clever thing to do?

I don't think so. Even in the narrowest economic perspective, I think it wasn't a clever thing to do. Most of the work that went into this $15 billion was local, locally subcontracted within Europe. It went directly into the economies involved and furthermore into dynamic sectors of the economy for high-tech industries involved in superconducting magnets, fancy cryogenic engineering and civil engineering of great sophistication and of course computer technology. All that know-how is going to pay off much more than the investment in the long run. But even if it weren't the case that purely economically it was a good thing to do, The United States missed an opportunity for national greatness. A 100 years or 200 years from now, people will largely have forgotten about the various spats we got into, the so-called national greatness of imposing our will on foreigners, and they will remember the glorious expansion of human knowledge that is going to happen at the LHC and the gigantic effort that went into getting it. As a nation we don't get many opportunities to show history our national greatness, and I think we really missed one there.

Maybe we can recoup. The time is right for an assault on the process of aging. A lot of the basic biology is in place. We know what has to be done. The aging process itself is really the profound aspect of public health, eliminating major diseases, even big ones like cancer or heart disease, would only increase life expectancy by a few years. We really have to get to the root of the process.

Another project on a grand scale would be to search systematically for life in the galaxy. We have tools in astronomy, we can design tools, to find distant planets that might be earth like, study their atmospheres, and see if there is evidence for life. It would be feasible, given a national investment of will and money, to survey the galaxy and see if there are additional earth-like planets that are supporting life. We should think hard about doing things we will be proud to be remembered for, and think big.