Jason Shankel

The Fabric of Reality


In his book The Fabric of Reality, physicist David Deutsch presents four key concepts which define the universe: quantum theory, computation, evolution by natural selection and Popperian epistemology. Deutsch calls these four concepts the strands of the fabric of reality.


The Four Strands

It would be impossible to do justice to the totality of Deutsch's argument here, but in summary it goes a little something like this: we live in a multiverse where every single possibility is played out according to the many-worlds interpretation of quantum physics. In this multiverse, there are two kinds of events: classical events, which are limited to one universe (or an infinite set of identical universes) and quantum mechanical events, which are distributed across large numbers of nearly identical universes. The universes described by these events are inherently computable and renderable using the theory of computation established by Alan Turing. From the theory of computation, we can see that there are a large number of universes where systems emerge which concentrate information over successive generations of entities. Darwinian evolution is the most familiar form of this phenomenon. And finally all of these laws can be discerned by the epistemology recommended by Karl Popper, which holds that knowledge is obtained not through repetition of experience but via the selection of simpler and more explanatory theories over more complex and opaque theories.

Deutsch attacks each of these strands in turn, eagerly devouring such philosophical mainstays as induction, solipsism, instrumentalism and logical positivism. While his views on epistemology, computation and evolution are interesting and well thought out, his most interesting and compelling argument is the defense of the many-worlds interpretation of quantum physics.

Throughout the book, Deutsch emphasizes the value of explanation in our theories of knowledge. It's not enough for us to simply derive a series of equations to account for blips we see on a screen so we can predict the future positions of future blips. What's important is for us to know WHY those blips move, what those blips are, how our equations model the underlying reality of blippiness and what would constitute a better, simpler, broader and more explanatory model of blip behavior than the one we have today. Deutsch believes that science does its best work when scientists take the conclusions of their best theories seriously.

Comparing oneself to Galileo is always a dicey proposition at best, but Deutsch makes the case that modern physicists who deny the many-worlds interpretation are akin to the Church of Galilleo's day clinging to geocentrism. To understand what the many worlds interpretation is and how it fits into Deutsch's overall fabric, it's important to understand how some of the spookier elements of quantum mechanics (indeterminacy, non-locality and quantum interference) are accounted for in single-universe and multi-universe models of quantum physics.


Indeterminacy is the property of quantum physics that holds that we can never be quite sure when or how a quantum state is going to change or what state a quantum system is in after it is measured. The best you can do is describe a distribution of probabilities defined by a wave equation. This concept was famously modeled as the thought experiment commonly known as Schrödinger's cat. For those unfamiliar, a cat is placed in a sealed box with a poison capsule set to open if a certain quantum event occurs, for instance the decay of a radioactive isotope. Classically, of course, we say the cat either lives or dies depending on what happens with the isotope. In the quantum world, however, we don't speak of the cat being alive or dead but rather in a wave superposition state, both dead and alive simultaneously. Only when we look in the box does the cat's status resolve into a defined reality.

Under the single-universe Copenhagen interpretation of quantum physics, we say that a quantum event either occurs or doesn't occur at the point it is "observed." Some have interpreted this to mean that something like human consciousness is required to actualize reality, which leaves us with the question of where consciousness came from in the first place, but most consider any interaction with the macroscopic world to count as a suitable observation, thus eliminating the need for quantum Cheshire cats. Other single universe interpretations such the pilot wave and related de Broglie-Bohm interpretation eliminate the need for observation by suggesting that the wave equation is a real, physical object and the quantum state it describes always has an actual value which is classically determined by underlying hidden variables.

Under the many-worlds interpretation of quantum indeterminacy, when you put the cat in the box, the universe splits into two, one in which the cat lives and another in which it dies. When you open the box, you find out which universe you're in and your counterparts in the other universes find out which one they're in. In this interpretation, the wave equation is neither physically real nor does it collapse but rather the universe divides to accommodate all possible values of the wave equation. The wave equation describes the topology of the multiverse in the region the observation is made.

Be Here, And There, Now

The second phenomenon, nonlocality, is a consequence of the first. Given that quantum events are indeterminate, when you make a measurement of a quantum system, you potentially change it. Say, for example, you measure the spin on an electron and find it to be +x. The next time around, it could be +y or -z or maybe still +x. No way to know. The best you can hope for is to define the probability function that governs the values of spin.

The tricky bit comes when you have not one, but two particles being measured, particles which start off in a single quantum state (say, as part of an atomic nucleus) and remain bound together when separated. This phenomenon is called "quantum entanglement." Particles bound by quantum entanglement remained synchronized so their spins will be opposite. If you measure both particles at the same time, you will get opposite values. If you read +x on particle A, you'll read -x on particle B, so long as you read them simultaneously. Thing is, this remains true even if you read particle A a bunch of times on its own before reading particle B. If you read A three times and get +x, -y and +z and then read A and B simultaneously, they'll still be synchronized, no matter how far apart they are at the time of measurement. While it appears on first glance that this phenomenon violates General Relativity, it does not. No information can be transmitted using this phenomenon because there is no way to control the values that will be read in the future or to determine if the other particle has been read or not read recently. But philosophically, it seems that there's some faster than light causality going on here, even if we can't transmit a signal with it. This is what Einstein called "spooky action at a distance" and by "spooky" he meant "I don't cotton to it."

The Copenhagen interpretation of non-locality is very technical, but it boils down to "Particles. What're you gonna do?" Under Copenhagen, all that matters is the math. There is no underlying physical explanation. And for many, the math is good enough, but not for Deutsch. For Deutsch, math is a tool you use to describe things you understand. It's not a substitute for understanding.

The many worlds interpretation of nonlocality leaves plenty of questions as well, but it does at least make sense in the light of General Relativity. This is because the many worlds interpretation does not require causality to explain the particle synchronization. There simply do not exist universes where the particles are desynchronized. When you measure a particle and change its state, you and your counterparts move into each one of the possible universes that exist after the measurement, and in each one the particles remain synchronized. Every time you measure, the universe splits and resplits so that every possible combination of measurements is experienced by one of your counterparts. And while that's a lot of universes, not one of them requires the particles to communicate faster than the speed of light or to possess hidden variables that pre-determine the outcomes of your measurement.

Quantum Interference

The final phenomenon is quantum interference. Quantum interference is most famously illustrated by the two slit experiment. Light behaves like a wave, so when a light beam is split and sent through two narrow slits, the two scattered waves interfere with each other and create a pattern of light and dark bands on the screen. If there is only one open slit, there is no interference and the light creates a smooth pattern. The strange thing is that light is also a particle, so these patterns persist even if we only send one particle at a time through the mechanism. The question is, how does a single particle traveling through an open slit "know" whether the other slit is open and hence "know" which regions of the screen it can land on?

The single-universe interpretations say that the particle isn't just a particle but a particle and an extended object, a pilot wave, describing other potential paths of the particle. It's basically like a roller coaster track that gets built in front of the photon as it travels. The shape of the track is determined by the configuration of the apparatus.

Deutsch's argument is that every formulation of this "pilot wave" approach boils down to "a probability field that acts exactly like a bunch of other particles in a bunch of other universes interfering with the particle we see in our universe." The problem with this is not just that the single universe interpretations are overly complex. It's that they discourage us from embracing the explanatory power of our most useful quantum theory and thereby discourage us from seeking deeper understanding of the world.

Quantum Computers

Okay, but isn't this all so much dorm room hoo-hah? Does it really matter if we believe in literal alternate universes or just pilot waves that act like alternate universes? The math is all the same and no one is ever going to be able to contact any of these other universes with more than a few particles. And those universes will be so nearly identical to ours that it's not likely we'd even see them as being different. No one is ever going to shine a quantum flashlight in a mirror and see a version of themselves with strikingly different facial hair.

Fundamentally, the argument over which position is ontologically correct boils down to Occam's Razor, which tells us to prefer the simpler explanation. And as is usually the case with Occam's Razor, the problem comes with the effort to define "simple." For single-universists, it's simpler to say there is but one universe. For many worlders, it's simpler to say that particles only interact with real physical things, not unrealized possibilities. Is it simpler to suggest that everything in the solar system, including the Earth, moves in smooth ellipses? Or that the apparent motionlessness of the Earth is true and everything else moves in a complex dance of epicycles?

Deutsch argues that the many worlds interpretation is superior because it explains how our perceptions of quantum reality actually work. It's not that the universe has some hidden array of mechanisms...some of which that aren't even "real" in any meaningful sense...that choose which possible reality to make actual using some mechanism that depends on unseen variables and can only be described probabilistically. Rather, it's that every potential is equally realized, none preferred, and we and all our counterparts are all equally baffled as to why our respective universes were "chosen" simply because we can't pick up the phone and call each other.

Philosophy aside, the practical reason it's important to embrace the explanatory power of the many worlds interpretation is that it will give us the insight we need to produce useful quantum computers. Deutsch holds that quantum computing represents the next great breakthrough in computation, but if we are unwilling to take seriously the idea that quantum computers work their magic by distributing parallel work across many different universes, then we would be like NASA scientists trying to launch space probes using geocentric models of the solar system. It may be possible, but it will be complicated and opaque to understanding. Every epicycle we add makes the model of the solar system less comprehensible and ultimately less useful.

Was Blinded By Science, But Now I See

It's important to note that while he makes a compelling argument, not everyone is on the Deutsch Train. The late physicist David Bohm and Deutsch took at times amusing pot shots at each other, including this one from Deutsch: "pilot-wave theories are parallel-universe theories in a state of chronic denial." In the posthumously-published TheUndividedUniverse, Bohm and his co-author B.J. Hiley make a strong defense of the pilot-wave approach, both mathematically and philosophically. TheUndividedUniverse is a much more turgid text than the more accessible TheFabricofReality. It's more of a traditional textbook, with a good deal of math. But the chapters on the ontology of pilot waves are fairly accessible and make a fascinating critique of Deutsch.

Ultimately, the question isn't whether we endorse Deutsch's multitude of universes or Bohm's realistic unrealized realities. What we must embrace is the spirit of competitive philosophy and the search for explanations that Deutsch, Bohm and their colleagues have brought to their writing and their research. Science is not valuable unless it is meaningful. In an age when physics seems to be losing its way, we could use more physicists like David Deutsch and David Bohm to light candles in the dark.