HUO: Your book, The Fabric of Reality , had a real influence on the art and architecture fields. Philippe Parreno wants to ask you this question: “Can reality be produced?”
DD: I think the deepest answer is that we don’t know yet. But I believe the best answer available is “no.” The whole of reality including the multiverse and all the “production”—all the creation—that has happened and will ever happen within the multiverse is in some sense already there. The trouble is that that answer doesn’t explain the fact that there is a vital distinction between knowledge that was already there but was merely transformed, and new knowledge being created. For fundamental reasons we know that there must be a difference between them. One of the symptoms of the fact that there is something missing in our philosophical understanding of knowledge creation is that we have so far been unable to produce artificial evolution. I know there are a lot of people in various fields of computer science who would plaintively reply that they are doing artificial evolution every day and that it’s already produced lots of useful results. But I don’t think so. I think that what is called artificial evolution today is nothing more than exploring a given landscape for the best its best features, rather than creating a new landscape. Real creativity is creating a new landscape. I went to a lecture a couple of years ago about making robots that learn to walk. When these robots start out, they’ve got no program for walking. All they do is wave their legs. You vary their program randomly and you measure how far each of them walked. From there you take the one that walked the furthest, which initially might be just one inch, and vary it. And eventually you find that, remarkably, just by these random variations and selections, the robot walks along quite well. At the end, they think that’s artificial evolution.
Learning, perhaps, in that any useful change can be considered learning. But it is not artificial evolution. I think that all the learning and creation of new knowledge in the robot happened in the mind of the graduate student that created the evolution program. And the token that that’s the case is that once it has walked as well as I described, it never walks any better. That particular program will never, for instance, learn to walk up stairs. It’s not because the hardware can’t do it; it could be programmed to do so. It’s just that the evolutionary system can’t reach that particular knowledge. And that’s a token of the fact that our present conception of computation, knowledge, evolution and physics, is missing something. Thus, I think there must be such a thing as genuine creation. Whether you call that “creating reality” doesn’t matter much. The important thing is that there is such a thing as the genuine creation of knowledge, ex nihilo.
It leads to the question of what is missing in the “theory of everything.” You wrote in the Fabric of Reality that it summarizes a theory of “everything that is known” and that it does not mean “all the knowledge in terms of data.” That’s a very important distinction. We often have this idea that we can’t actually grasp the complexity of so much knowledge being around. You don’t necessarily agree with that.
I don’t agree with that because what has increased is merely the complexity of data, information. But since the Enlightenment, the complexity of knowledge, of understanding, has actually gone down over the centuries, just as its content has gone up. That is thanks to the great unifications that have happened—physics with chemistry, biochemistry with biology, or more subtle unifications, like the unification of the theory of evolution with evolutionary theories of knowledge, or the unification of computation and physics. As a result, understanding all that is understood requires less complexity than it used to. I think that the process will continue, so that when we start to understand new things, like the real nature of knowledge and consciousness for instance, it won’t be an extra thing that you have to learn. It won’t mean that instead of doing a three-year university course you have to do a fouryear course or anything like that. The course will be the same length, but the amount of truth, the amount of understanding you gain per day, will be higher.
That is what we mean when we say that “depth” is growing?
In the theory of everything you have basically four main strands of knowledge: quantum physics, epistemology, theory of computation and theory of evolution. Could you explain those four elements a little bit? Also to what extent is your theory of everything different from those espoused by Edward Witten or Stephen Wolfram?
Let me answer the second part of your question first because it’s easier: the theory of everything in the sense I use the term in my book is not only the theory of physics. Out of those four strands, only one is reductionist: that’s the quantum physics strand. It tells us how large things are made up of small things, what the laws of nature are according to which the small things behave, and what kinds of small things exist in the first place. I say “small things,” but of course it includes also universes— parallel universes and so on. Now, theories like superstring theory, and theories of everything as they are known within elementary particle physics, are not actually theories of everything. They are simply attempts to give a complete theory of subatomic particles and forces, everything that is reductionistically fundamental. But in my view, that’s not the only way that a piece of knowledge can be fundamental. I think, for instance, that epistemology is also a completely fundamental field, even though from the point of view of physics it is the study of the behavior of very, very complex objects like the brain. Nevertheless, the fact that we can have a theory of epistemology shows that complexity resolves itself into a simplicity at the higher level, about which we can have knowledge that we cannot express with the form of knowledge about the lower level. Even if we could express it, it wouldn’t be explanatory. In my book I give the example: how do you explain the presence of a particular copper atom at the tip of the nose of the statue of Churchill in Parliament Square? Even if you could write down the entire history of all the forces acting on that atom starting from when it was created in a supernova and how it was then put into a seam below the earth and then mined and so on, that still wouldn’t explain why it is there, because the explanation should include concepts such as “this was put up to honor a human being because of certain things that he did.” And that is, in one sense, already in the description about the atoms, and in another sense isn’t there at all. So to understand all that is understood, we need higherlevel concepts.
So it is the opposite of inductive reasoning.
Inductive reasoning is invalid, but more importantly, it can’t take place at all. It’s an illusion. It’s just a description that people gave to the scientific process when it wasn’t understood very well. We now know that science is conjectural and creative, as creative as any human activity. The same is true with evolution. Before the time of Darwin, there was a mystery: how did the animals get their adaptive complexity? If you look at the eye, for example, it looks very much like a camera. We know the camera was designed, what about the eye? There was a religious way of answering that question, namely that the eye was designed too. But once reason became more established it was realized that this answer was no answer because it just says that a piece of complexity arose from a different piece of complexity, namely God, which by definition does not have an explanation. But if you’re going to accept that, you might as well say that animals don’t have an explanation in the first place and be done with it. On the other hand, a no-nonsense realist could have said, say in 1800, that an animal is composed of atoms, that atoms obey laws of nature and that’s the explanation. But he would have been wrong too, for although that would be a true statement, it is not an explanation. The explanation was Darwin’s theory of evolution, and you can’t express that literally as a statement about atoms. Of course, it underlies the whole theory that animals are made of atoms and don’t have immortal souls (or at least that they are not relevant to the evolution). But that doesn’t explain it. This difference between description or prediction on the one hand, and explanation on the other is central to my view of the world. And it’s why you need all four strands to understand the world, because to understand you need to explain what is actually there and how things relate to each other. You simply can’t do that entirely at the level of fundamental physics.
You need the four components: quantum physics, epistemology, theory of computation and theory of evolution?
Yes. And three of those four are emergent properties. But as I also say in the book, even that is a kind of biased way of putting it. You could say that any one of the four is emergent with respect to the other three. You could regard any one of them as being the basis of everything. Indeed, in all four cases there are people who insist that this particular strand is the real underlying truth and that the others are built on top of it. For instance Stephen Wolfram is one of those who believe that the computational strand alone is the real underlying truth.
There is a huge polemic against superstring theory at the moment. Peter Woit wrote a book called Not Even Wrong(1) which makes an incredible full frontal attack. The tide seems to be turning against the string theory and its speculative attempts to produce a theory of everything. How do you place yourself in this polemic?
I never joined in the enthusiasm about superstring theory. My personal thought was that it was unlikely to work because it was not motivated by trying to solve a problem within theoretical physics. It’s very rare for foundational problems in physics to be solved by first wishing you had a theory with a given property. Instead it’s nearly always the case that there is some problem within physics. For instance, Einstein’s general theory of relativity was trying to reconcile things like the constancy of the speed of light on the one hand with the existence of gravity on the other. And the answer was first conceptual and only later mathematical. Einstein was trying to get relativity to work for several years after he had the basic explanatory idea of what the solution was. He once gave a lecture that was attended by the great mathematician David Hilbert. The story is that Hilbert went home after Einstein’s lecture and wrote down Einstein’s field equations. Just like that, because to him, as a mathematician, it wasn’t that hard. To him, Einstein had already done the hard part, the physics part.
Einstein was unaware of what Hilbert had done and it took him, I think,
another two years.
It could have been a shortcut.
It would have been a shortcut if they had got together at that point. In fact some people afterwards even tried to say, “Hilbert discovered it first!” but Hilbert said himself that he didn’t discover what the equations he had obtained meant. He was a pure mathematician, one of the greatest of the twentieth century. Einstein was by common consent the greatest physicist, and the two disciplines needed each other, but, it almost never happens that an advance in physics is made by just looking for a piece of mathematics and seeing what application it has. So that’s why superstring theory is a worthwhile thing to explore but unlikely to work. And if people are beginning to see it won’t work, I’m not let down. Through it, we’ve understood a whole slew of things about how fundamental physics can be put together, how it can’t be put together, what kind of ideas can or can’t work. It is progress even if it’s not the
What about Wolfram?
You might think since I’m so keen on unifying computation with physics that I would be keen on Wolfram’s idea but I’m not. There are two reasons. One of them is simply that his type of computation is classical not quantum, and I think it’s a great mistake to think that there is anything fundamental about classical computation. It’s just an accident of history from the way we discovered things, and the accident of our own physical constitution that we can’t easily detect quantum interference. So we are tempted to think that the particular subset of possible computations that we call classical computations and which our present day computers can do, has a fundamental significance, that it sort of stands outside physics as a preferred set of operations in terms of which you can then try to build physics. I think the relationship is entirely the other way around: which operations we can and cannot perform, such as computations, depends on what the laws of physics are.
One of the elements of The Fabric of Reality that the art world has particularly found resonance with is the idea of multiverse.
Well the first thing to say is that the multiverse theory is only a minority opinion amongst my physics colleagues. Perhaps no more than ten percent of physicists would endorse what I am about to say. In my opinion the evidence is overwhelming that the reality we see around us—the stars, the galaxy, each other—is only one slice of a bigger reality which contains many entities of that type: the type that we call universe. So the whole collection of those entities, we call the multiverse. But as I said, many colleagues would disagree. Now, if you say artists are interested, there are two points I would like to make. One is that it is in my opinion hopeless to try to look outside the laboratory, for phenomena directly caused by the multiverse. People often ask me if dreams might give us a window into other universes, or if artistic inspiration comes from a collaboration of different instances of oneself in different universes. The answer to that is no, because the laws of quantum mechanics itself, the very theory from which we know that these things exist in the first place, says that this kind of collaboration between universes is not possible. So that’s the bad news. The good news is that the kind of reality in which we are, is one in which alternatives to the events we experience really do happen. Now, why is that important if they can’t physically affect us?
My best metaphor or parable to explain why is to go back again 150 years to the time before Darwin’s theory of evolution was discovered. At that time, there was a problem: Why do giraffes have long necks? So that they can eat the leaves. But is that providential? Is it divine providence that lengthened their necks or is there some other explanation? The thing is that for most of human history, we didn’t have the conceptual framework that Darwin had to answer that question, because all theories, including theories of astronomy and so on, were about the end result. They were teleological theories, saying that the giraffe had a long neck so that it could eat the leaves or so that God would be pleased or so that it would stay alive. Similarly, even in astronomy and the other hard sciences where they had real scientific theories, they were still in the form of predicting outcome and not predicting the process. For instance, [Johannes] Kepler’s most famous law was that the planets move in ellipses with the sun at one focus. He was saying that ellipses are the answer—not circles, not parabolas, but ellipses, with the sun not at the center but at the focus. So that’s how the world is. Now starting with Galileo or Newton, the character of scientific theories changed. Instead of wanting merely descriptive theories about what the end result is, we wanted explanatory theories about what produces that result. Newton’s theory, which likewise predicts ellipses, doesn’t mention ellipses. The word doesn’t appear in Newton’s laws at all, nor does the word “planets.” Nor does the word “sun.” None of that appears. It’s a general theory, not about outcomes but about forces, momentum and so on. It’s about the thing that causes the outcome. Now we can go back to Darwin. His key philosophical innovation, the thing that distinguishes him from all previous attempts, including all previous theories of evolution like [Jean-Baptiste] Lamarck’s, is that Darwin is all about the laws under which giraffes get their long neck. Again, Darwin’s theory doesn’t mention “giraffes,” “necks,” “trees,” anything like that. It’s a general theory stating what kind of a world makes a long-necked giraffe explicable, without a divine being having to create it by fiat.
That’s a toolbox.
Yes. The worldview of Newton’s physics was a toolbox that opened the door to Darwin understanding the theory of evolution. Now in the same way, I believe that the unsolved problems of the present day—such as the nature of consciousness and creativity—can only be solved within the worldview of the parallel universes. It can only be solved by somebody who knows that there are parallel copies of them, that they’re all doing different things, that the world is all connected together, just as Darwin needed the Newtonian worldview in order to understand giraffes. Note that Darwin never needed to use Newton’s actual laws of motion. He never needed to write down the formulae, the equations of Newtonian mechanics. Similarly, I don’t believe that philosophers —whatever philosophers or scientists solve the problem of consciousness — will need to use the equation of quantum mechanics. What they will need is the worldview that quantum mechanics tells us.
Yes. The explanation. If you think in those terms you can be on the path to the answer just as Darwin had to think in terms of a Newtonian universe. If he had thought in terms of a Keplerian universe he could never have understood evolution. And if we think in terms of a classical single universe cosmology like that before quantum theory, we will not be able to understand things like consciousness and knowledge. There are some examples in The Fabric of Reality of what different things such as genes, knowledge or brains look like when you look at them from a multiverse point of view. For instance, although a gene is a microscopic object in any one universe, it is a gigantic object in the multiverse. Because a gene is one of those very rare types of objects that get errorcorrected back to their original form if they change or mutate. So if you have a population with a certain genome and a mutation happens, a few of them have a variant. Then if you go on through the generations, the original form will be restored because the ones with the wrong form won’t reproduce as well. That means it will get set back to the way it was in most universes.
In the multiverse there can be change of scale: something small can suddenly be very big.
Absolutely. And the things which are big in the multiverse are the ones that contain a lot of knowledge or a lot of evolution.
Another key idea for the art and architecture fields is your conception of time. You said that the common-sense view on time as a flow is nonsense. Could you explain your notion of time played out in The Fabric of Reality?
In science we often find that common sense is wrong, but we very rarely find that common sense is nonsense. But in the case of the flow of time, common sense is nonsense. It contains an internal contradiction: time can’tpossibly flow because the concept of flow presupposes an externally existing time against which something is flowing. So the quantum concept of time is that other times like yesterday and tomorrow are just special cases of other universes. If you think about it, there are parallel universes in which you and I are sitting at a different place in the room and other universes in which we even fail to meet today and so on. There are also universes where we did meet, but on different dates. If you think about universes that just resemble ours, there are universes where we are sitting just a millimeter displacement from where we are in this universe. There are ones that are very like ours, there are some that are genuinely like ours but have significant differences. Now if you think about it, that’s exactly the same as in the flow of time, because a minute ago this room was in many ways just the same way as it is now, but we were sitting in a slightly different position and the contexts of our
brain were slightly different.
It wasn’t dark behind you but blue.
Exactly. So different times have a striking resemblance to other universes. And philosophically, just as with other universes, people often ask, when they’re trying to get their heads around these concepts: “Ok, if there are all these different instances of me in other universes, which one is the real me?” This problem has arisen in regard to time since time immemorial. People have wondered if the version of them who did reckless things in their youth was really them. For instance, if we’re thinking of crime and punishment: If a criminal has truly repented at a later time, is it worth punishing him? Is it still the same person? And even more, suppose the criminal has had brain damage since he committed the crime and can’t even remember doing it or why he did it, is he still the same person? So the issue of where, along this continuum of different versions of this entity that are continuously related to each other, you stop being the same person, arises in regard to time and parallel universes in the same way. And it turns out that this isn’t just an analogy. The physicists Don Page and Bill Wootters analyzed time in the quantum framework.
Is this recent?
No, this is quite an old paper published in the 1980’s.(2) It turns out from their work that the different universes and the different times just are special cases of the same thing. Think of the multiverse as a collection of snapshots. For simplicity, just consider the multiverse in regard to this room, all the configurations of this room in all universes at all times, and let’s forget about the ones where this room doesn’t exist at all. Well, some of those are other universes at this time, some of them are other times in this universe, and some of them just can’t be described like that—they are just elsewhere in the multiverse.
But they’re all snapshots.
Yes, and they’re all treated uniformly in quantum theory.
In your book you say it is a universe at a particular time.
I call that a snapshot. Although this metaphor connotes just a visual thing, I’m talking about everything: where all the atoms are and so on. To a good approximation the multiverse consists of a collection of such snapshots which are related to each other in different ways. Some of them are related because different objects in the different parts of the multiverse are part of a same bigger object, like a gene. The gene really extends throughout many parallel universes. Another kind of relation is that the laws of physics sometimes determine what is in one snapshot if you know what is in another, and those snapshots are called the past and future of each other. Those are only two kinds of the relationships that can exist between snapshots; the multiverse allows other more complex relationships where again it’s not really meaningful to say whether it’s past or future or another universe or a part of the same object and so on. We don’t really have the language to describe it.
What about time travel?
I’ve worked on time travel almost entirely from the point of view of quantum theory, and quantum theory doesn’t tell you whether time travel is actually possible or not. The issue of time travel in physics resolves itself into two almost completely separate questions. One of them: “Can you build a time machine?” The other one: “If you could, what would happen?”
It’s a “what if” scenario?
That’s right. Those are two different questions. And remarkably, the first question is dealt with almost entirely by the theory of relativity whereas the second one is dealt with almost entirely by quantum physics. And it’s the latter one that I’ve been concerned with. Most people are more interested in the first question: Could you build one and if so, how? But I haven’t worked in that direction. Basically the answer is it’s simply too complicated. For a while the experts were saying, “Yes, it is possible,” and then they slowly started changing their minds and saying that maybe it isn’t possible. We don’t know. The other question, the one that I worked on is pretty well worked out: If you had a time machine, what would happen? The answer is if you used it to go into the past, you would go into the past of another universe. Which other universe? Well, obviously the one in which you appeared in the past at that moment. That’s the answer.
It’s that simple. But I just would like to point out that it’s not a matter of saying, “how can we resolve the paradox? Oh look, it’s resolved if there are parallel universes.” It’s not like that. What happens is you take the equations of quantum mechanics and you work out what they say will happen and it says that the time travel will happen. That’s equivalent to saying that the initial premise (that a time machine could exist at all) is consistent. But it could have said otherwise. We don’t have the freedom to tell the parallel universes what to do. The laws of physics already tell them what to do. And they either permit or don’t permit time travel. It turns out that they do permit it. But we don’t know whether general relativity provides a mechanism for actually doing it.
The possibility of an impossibility.
Yes, if something is forbidden by the laws of physics, then it definitely is impossible.
That’s not the case.
We don’t know if that’s the case yet. But if it’s not forbidden by the laws of physics then it’s just a matter of knowing how.