Lee Smolin Does Not Believe in the Scientific Method

Lee Smolin is a theoretical physicist, a co-inventor of loop quantum gravity, and the author of *The Trouble with Physics* and *Einstein’s Unfinished Revolution*. Lex Fridman opens with the easiest hard question: what is real, and how do we tell the real from a product of human perception and imagination? Smolin’s first answer is that we don’t know. What he believes, warming up past science into metaphysics, is that there is a world independent of his existence, his experience, and his knowledge of it. That world he calls the real world. That stance is realism.

Fridman asks whether it is crazy to think that everything that exists is a creation of the mind — that the constructs of the visual cortex are reality, and that whatever lies outside can never be grasped. Smolin says a version of that is not crazy at all. What we experience is constructed by our brains in an active mode. “We don’t see the real world,” he says. “We see a very processed world,” run through an organ he calls incredible.

He still believes that behind that experience, that mirror or veil, there is a real world, and he is curious about it. How we get a sense of it is not, for him, a choice between physics and some ape-ancestor intuition. We are in the same position as people in the caves, or before the caves: curious, needing to explain fire and food, and also looking up at the sun and moon because we cannot help it. We make up stories and explanations. That is the human starting point, not a laboratory protocol.

Two things, Smolin thinks, are just true of being human. We judge fast because we have to survive: is that a tiger or not? We act on incomplete information. And we are often wrong — or at least sometimes wrong, which is all he needs for the argument. We fool ourselves and we fool other people readily. Lots of stories get told; some of them produce a concrete benefit and some do not. To be “right,” if you are a realist, is not a binary. There is a relative scale of more right and less right, and you are more right if you come closer to an exact description of that independent world.

Does the scientific method tell us how close we are getting? No, on two counts, and the first is that Smolin does not believe in the scientific method. In graduate school he was marked by Paul Feyerabend, a student of Karl Popper, who argued by logic and by historical example that the thing does not exist. Name any rule that is supposed to be part of the practice of science — say, that your theories must always agree with all the data already taken — and Feyerabend will show you times when some scientist had to contradict that advice for science to progress overall. It is not a simple matter.

What Smolin puts in its place is sociological and ethical. Science is a community of people bound by precepts: report all your results, whether they agree with your hypothesis or not. Training consists mostly of methods of checking, because we are very error-prone, with tools on both the mathematical and the experimental side to check and double-check and triple-check. You cannot walk in off the street and declare yourself a scientist. The test that ends the training is whether you can form a convincing case that colleagues cannot shut down with “did you check this?” and “what about this contradiction?” When you can produce that defensive argument, they give you a degree. You will still be questioned and you may still publish mistakes, but the community will waste less time fixing them.

The gap between that community practice and the popular ideal of method is that there is no simple relationship between experiment and hypothesis. Galileo drops a ball from a tower; it falls at the base. An Aristotelian says of course it does: that shows the Earth is not moving while the ball falls. Galileo says no: the principle of inertia gives the ball the same motion as the Earth, so the tower, the ball, and the Earth move together, and the ball lands at the base. The same observation is claimed by two opposite theories. You have to assemble an interconnected bunch of cases and work hard to line them up. The transition from Aristotelian physics to the new physics took centuries. It was not finished until Newton in 1687.

Smolin is not a relativist about the long arc. Century by century, scientists have come up with ideas that engineers can turn into rockets and computers and things that improve life. He wants the dates right, though. There was a scientific revolution that partly succeeded from the late 1880s and 1890s into the 1930s and 1940s, with some history stretching into the 1970s: relativity, electromagnetism, the confirmation that matter is made of atoms, nuclei with electrons, and quantum mechanics from the early twentieth century to a basically final form in the late 1920s. That partial revolution is the basis of the technologies Fridman has in mind. Electrical technology grew in close relation with the electrification of cities in the United States and Europe.

Fundamental physics since the early 1970s does not have a story like that. There is not a series of triumphs and progresses, and there is not a practical application. Experimental validation, Smolin says when Fridman presses, was absolutely necessary for the earlier key predictions of quantum mechanics and of electricity and magnetism. The later stall is not an argument against experiment. It is a fact about what the community has and has not delivered.

A realist, for Smolin, believes in an external world independent of our existence, perception, belief, or knowledge. A realist as a physicist believes there should be a completely objective description of each and every process at the fundamental level, explaining exactly what happens and why. That sounds like determinism: no fuzzy magic you can never get to the bottom of. Some people, he notes, are not interested in determinism and could live with a fundamental world that had some chance in it. So could he. Asked whether God plays dice, he says it is probably much worse than that. Theories can change, and they can change without warning. The future is open. If the fundamental laws themselves can change, there is no static list of ultimates waiting to be found.

Does the human mind play a role in making things real? Smolin splits the question. He does not think our perceptions make things become real — though the answer depends on whether a naturalistic, realist view can yet make a place for consciousness, which he hopes it can and says it certainly does not yet. On the anti-realist side he declines to play scholar of the camps. The founders of quantum physics were extremely smart people living in a perilous time between the world wars, with cultural trends running that way. Niels Bohr’s line, as Smolin renders it, is that the purpose of science is not an objective description of nature as it would be in our absence. Science is an extension of our conversations with each other, used to describe our interactions with nature. We are free to invent terms like particle, wave, causality, time, and space if they are useful and carry some intuitive meaning. We should not believe they correspond to what nature would be like without us, about which we have nothing to say.

Does he find any of that compelling, given that he has already said humans are storytellers? Sure, he says. He would just like to believe the aspiration is the other thing: that the stories eventually lead us to discovering the real world as it is. Will we ever get there, and would we know if we had? That, he says, is a question for people five thousand years in the future. We are certainly nowhere near there yet.

Is there a limit to what ape-descended biological minds can grasp, even with the tools of physics? Again, Smolin says, five thousand years from now. He does not agree with David Deutsch about everything, but he admires how Deutsch put the case in his last book: exploration, the universality of certain languages, of mathematics, of computing, and an argument that because symbolic systems can refer to themselves, there is a universality of possibility for our understanding of whatever is out there. Smolin admires the argument. We are doing okay so far. Whether there is a limit, we will have to see. For now there is plenty to play with.

Some of that plenty is sitting in front of us and we miss it. He quotes his friend Eric Weinstein: Einstein carried his luggage, Freud carried his luggage, Marx carried his luggage, Martha Graham carried hers. All those geniuses carried luggage, and not until relatively recently did it occur to anybody to put a wheel on it and pull it. The wheel was right there waiting to be invented for centuries. Every morning Smolin wakes up and thinks, why can’t he be that person walking through the airport?

What does it take? He started a long way from physics and read the biographies, beginning with Einstein and Galileo. Some of it is being in the right place at the right time. Some of it is stubbornness and arrogance, which can easily go wrong: slightly the wrong speed or the wrong angle and they are ways to fail. Get the right measure of confidence and you can hold a question others drop. Einstein, he thinks, went after the nature of nature with a ferocity and commitment that exceeded other people of his time, so he asked more stubborn, deeper questions. There is also ability, partly born and partly developed, like musical talent. On his own confidence he has no verdict. He has contributed what he has contributed. Whether more confidence would have taken him further, he does not know. He knows people whose self-confidence vastly exceeds his, and sometimes he thinks it is justified and sometimes he does not.

Finishing Einstein’s unfinished revolution is something Smolin has been trying to do his whole life. The unfinished work is the twin revolutions of relativity — special and especially general — and quantum theory. Einstein was, Smolin reminds Fridman, among the founders of quantum mechanics, even if people do not always associate him with it. In 1935 he already saw that a radically different theory would be needed to resolve the wave-particle duality of the photon. At first he did not believe the quantum mechanics of the late 1920s at all. Then he was convinced it was consistent but incomplete. That is Smolin’s view as well. For various reasons it needs additional degrees of freedom, particles, forces, something, to reach the stage where it gives a complete description of each phenomenon. Realism, he says, demands that.

What bothers him most is the measurement problem. He will not speak for Einstein, but the basic formulation of quantum mechanics gives you two ways to evolve a situation in time. One is when no observer is observing and no measurement is taking place. The other is when measurements and observations take place. They basically contradict each other. That is the popular version of the measurement problem, and it is enough to leave the theory incomplete.

There is a second reason the revolution is unfinished: nobody understands the relationship between general relativity, our best theory of space, time, gravitation, and cosmology, and quantum theory. For the most part general relativity describes big things and quantum theory describes little things. Both toolkits are extremely powerful and they are two totally separate systems. If you believe quantum mechanics as now understood is correct, you bring general relativity or some extension of it into the quantum domain: you quantize gravity. If you believe with Einstein that quantum mechanics needs to be completed — Smolin’s view — then part of finding the correct completion is finding one that incorporates spacetime and gravity.

Asked what spacetime is, Smolin does not give the textbook answer. He believes spacetime is an intellectual construction we make of the events in the universe. The events are real, and the causal relationships among them are real. The idea that there is a four-dimensional smooth geometry with a metric and a connection, satisfying the equations Einstein wrote, is a good description to some scale, a good approximation that captures some of what is going on in nature. He does not believe for a minute that it is fundamental. Fundamental means part of the description as far down as you go, as real as real could be. Time, he thinks, is fundamental and goes all the way down. Space does not. The combination we call spacetime in general relativity also does not.

The activity of time is the continual creation of events from existing events. If there are no events, there is no time — and then there is not only no time, there is no nothing. The history of the universe has a history that goes to the past. The future does not exist. There is a notion of a present and a notion of the past, and the past is a story about events that took place to our past. An event, at this level, is a primitive concept: it does not have a definition in terms of other things. In general relativity it looks like two particles intersecting in their paths, or something changing along a path. At the fundamental level it is something elementary happening, and it is involved with energy, with the exchange of energy.

The version of a theory of events he has been developing is joint work, and he wants the collaborators named. Marina Cortes has been central to the work on causality since about 2012 or 2013. In the period before that, Roberto Mangabeira Unger, a philosopher and professor of law, had a huge influence on his thinking. In this picture, events create other events, but they do not live in space. They have an order in which they cause each other, and that order is part of the nature of time. Space is an emergent property: a property that arises at a level of complexity larger than the fundamental, not directly derivable from the properties of the fundamental things. In a sufficiently complex universe, three-dimensional position emerges. In some models, he says, they have seen how this happens in detail, both computationally and analytically.

Locality is the comforting idea that you can affect things close to you and not things far away — the same discomfort Newton had about action at a distance. It seems crazy, Fridman says, for locality not to be fundamental. Smolin’s first move is to split definitions. In quantum field theory, the mix of special relativity and quantum mechanics, there is a precise definition: field operators corresponding to spacelike-separated events commute. That is a property of the theory, it matches the intuition that things in a field affect nearby things more than distant ones, and it is well tested.

There is another definition, expressed by Einstein and made more precise by John Bell, which has been tested experimentally and found to fail. Two particles can interact and then share a property that is not a property of either particle alone: entanglement. Separate them. An observer measures particle A; someone else measures particle B. Bell locality assumes that whatever is real about B is unaffected by the *choice* of what to measure on A — not the outcome, just the choice — because the systems are spacelike-separated and it will take a while for any information about the choice to affect reality at B. From that assumption you derive an inequality that certain correlation functions have to satisfy.

You can test it pretty directly with pairs of photons or other particles. It is wrong by many sigma. That means the definition of locality Einstein was playing with is false. It is not true that whatever is real about particle B is unaffected by the choice of what to measure on particle A, no matter how long the particles have been propagating at almost the speed of light, or the speed of light, away from each other. The experiments have been done over distances on the order of tens to a couple of hundred kilometers; Smolin does not remember whether it is 25 or 170. Causality as fundamental, and a break with this kind of locality, is in his view a crucial step toward finishing the revolution.

How do we finish the unfinished revolution? Smolin can only say what he is trying and what he has abandoned. He is one ant in an ant colony, and maybe that is enough. Nice theories have been invented, related to quantum mechanics and to quantum gravity, and there is a lot to admire in many of them. To his understanding none of them completely solve the problems he cares about. The situation is either terrifying for a student or full of opportunity: more than a dozen attempts, none of which, he thinks anybody anticipated, work out all the way. Each works partly and then hits an issue nobody can figure out how to go around.

His reaction is twofold, and the first is sociological. Communities have formed around some of these approaches. Borrowing a metaphor from Eric Weinstein, they sit on hills in the landscape of theories and throw rocks at each other. What is needed, Weinstein says, is for people to get off the hills, come down into the valleys, talk, become friendly, and learn to say not no but yes: your idea goes this far, but maybe if we put it together with mine we could go further. In that spirit Fridman brings up Sean Carroll, a friend of Smolin’s and a public champion of the many-worlds interpretation. Smolin loves Carroll, calls him an articulate ambassador of science who takes philosophy seriously and does the homework. The mysterious thing is that they so often end up on opposite sides. It is fun.

Many worlds does not appeal to him. It does not answer the questions he wants answered, so strongly that when he began looking for principles, the first was: there is just one world that happens once. He has still thought a lot about it and struggled with it. Everett’s interpretation cuts off the measurement half of quantum evolution and keeps only the smooth Schrödinger evolution — but probability lives in the half you threw away, and experimentalists use probabilities to check the theory. You might try to count branches and recover frequencies; Everett had an argument in that direction, but it gets very subtle when there are infinitely many possibilities, as in most quantum systems. Smolin’s understanding is that Everett’s own proposal did not work.

Decoherence may be relevant, and a number of people after Everett have tried to make versions of many-worlds quantum mechanics. It is subtle and not the kind of thing Smolin does well, so he consults; that is why the book has a chapter on Everett and another on the problem. At Oxford, Simon Saunders, David Wallace, Harvey Brown, and David Deutsch have put years into a sophisticated attempt. The flavor is that maybe there are really no probabilities, but if the Everett story were true and you lived in that multiverse, how would you bet? They use decision theory from probability and gambling to shape that story. When Smolin checked in while finishing the book, Saunders and Wallace told him they were not sure any of the proposals was yet correct. Carroll has his own approach involving self-locating observers; Smolin tried to read it, did not study it hard, did not talk to Carroll, and so had nothing to say about it in the book.

The other half of Smolin’s reaction to the stall is methodological. When he feels very frustrated about quantum gravity, he goes back and reads history. Einstein’s achievements are a huge lesson and something like a role model. It is very clear, Smolin says, that Einstein thought the first job when you enter theoretical physics is to discover and invent principles, then make models of how those principles might be applied in some experimental situation, which is where the mathematics comes in. Before Einstein there was no unified spacetime. He invented the idea. For him it was a model of his principles, of his postulates.

Smolin’s view now is that we do not know the principles of quantum gravity. He can think of candidates and has papers discussing them. His belief is that the partly successful approaches are all models which might describe some quantum-gravity physics in some domain, in some aspect, but would ultimately matter because they model the principles. The first job is to get those principles down. That is the approach he is taking.

*The Trouble with Physics*, published in 2006, made Smolin a controversial figure. He cannot, as an author, claim after all that work that he was misread, but many reviewers not personally involved — and even some working on string theory — told him they thought the book was fair and balanced. His original purpose got diverted by an argument that was already hard. He had wanted to write about how science works as a community bound by ethics. He drafted a rather academic book whose first third was methodology, with case studies later, one of them the search for quantum gravity and string theory. He could not get that book published. Someone suggested flipping it and starting with the story of string theory, already controversial in 2004 and 2005. He was careful, he says, to criticize papers, not people.

The picture that worried him was concrete. Originally there were five string theories, perhaps unifiable into one, from which you might break symmetries, recover the Standard Model, and predict masses, charges, and coupling constants. Then a bunch of solutions were found, each with a different particle physics — Calabi-Yau manifolds. Then there were, nobody was sure, hundreds of thousands of versions. His friend Andy Strominger found a way to put a kind of mathematical curvature called torsion into the solutions, wrote a paper on string theory with torsion, and discovered he could invent no way to count or classify them. Strominger wrote that this was worrying: doing phenomenology the old-fashioned way, by solving the theory, was not going to work, because there would be loads of solutions for anything. They spoke about it. Smolin got seriously worried.

He tried to solve the problem himself. He was reading a lot of biology and evolutionary theory, and he borrowed the idea of a fitness landscape of DNA, proteins, or species. He made a scenario, cosmological natural selection, about how the universe as a whole could evolve to discover the parameters of the Standard Model: laws changing in a way that echoes natural selection. He wrote the paper in 1988 or 1989, published it in 1992, and made it the subject of his 1997 book *The Life of the Cosmos*. Because you would have an ensemble of universes related by descent, almost every universe would share the property that its fitness was maximized, or close to it, and from that you could deduce testable predictions. He contrasted it with the anthropic principle, which could not make tests or falsifications.

That is also why he worked on both string theory and loop quantum gravity. A strong belief in relational, background-independent principles led him to try to make string theory background-independent; the tools he developed could be applied directly to general relativity, and that became loop quantum gravity. In his mind the two have always been closely related. The idea of two communities, one devoted to strings and one to loops, was never his division. What he regrets about the 2006 book is not, as far as he knows, the diagnosis of string theory, which has largely stood up aside from underestimating some applications. It is that the same critique applies to many other communities in science, including his own — people working on quantum gravity outside string theory. He considered saying so explicitly. It was a small, intimate community, and saying it would have meant telling stories, naming names, and writing a history he had no right to write. He stayed away, and was misunderstood.

Fridman asks for a hopeful message for theoretical physics in the twenty-first century, looking at the community and not only at Smolin’s own work on causality and nonlocality. Smolin’s first hope is blunt: solve the unfinished problem of science. That is the thing he cares about most. Among the young people he works with, though, he often hears and senses a total disinterest in the arguments that older scientists have, and an interest in what one another is doing.

The change is starting to appear in conferences. Young people interested in quantum gravity make a meeting and invite loops and strings, causal dynamical triangulations, and the other tribes. The week after the taping there was a small workshop at Perimeter along those lines, and in the summer a big conference billed simply as quantum gravity — not strings, not loops — with organizers and speakers from all the different communities. To Smolin that is very helpful. The different ideas are coming together, or at least people are expressing interest in that. After they sign off, Fridman leaves listeners with a line of Smolin’s: one possibility is that God is nothing but the power of the universe to organize itself.