Reality+ : Virtual Worlds and the Problems of Philosophy (9780393635812), page 20
More generally, given that a simulation can replicate the structure of standard physics, with digital entities playing the core mathematical roles and producing the right sort of observations in us, the simulation will make standard physics real. That’s enough for the it-from-bit hypothesis to be true.
I’m not saying that necessarily, any simulation of physics will make physics real. If we’re not in a simulation, then our physics is unsimulated. If so, photons—that is, the entities that play the structural role of photons in our world—are probably nondigital entities. As a result (as I’ll discuss in chapter 20), a simulated photon won’t be a genuine photon. However, if we are in a simulation, then our physics has been digital all along. Digital entities have always played the structural role of photons in our world. As a result, simulated photons have been genuine photons all along.
Of course, you can always resist the argument by denying structuralism. You could say that to have photons in our world, it’s not enough to have entities that play the structural role associated with photons. More generally, for standard physics to be true in our world, more than the right mathematical structure is needed. The right substrate is needed. A simulation has the right structure but not the right substrate.
For example, some people think that objects in a simulation are not solid enough to yield a genuine reality. Nonsimulated objects are solid and substantial, but simulated objects are not. According to this view, the only way to get a genuine it-from-bit physical reality from digital physics is if the bits are realized by something more substantial. If so, the simulation thesis does not imply the it-from-bit hypothesis.
We already know that this is the wrong way to think about solidity. In physics, the basic level of reality consists of evanescent quantities, such as quantum wave functions, with no special properties of solidity. Science also tells us that solid objects are mostly empty space. What makes objects count as solid is the way they interact with one another. A solid object is (roughly) one that other objects cannot easily penetrate. Solidity is really defined in terms of a certain pattern of interaction. This pattern can be present just as easily in a simulated reality.
Some people worry about space—that if we’re in a simulation, objects aren’t spread out in space the way they seem to be. There’s not really a desk three feet in front of me, and so on. In order for digital physics to yield genuine objects in a genuine space, the bits will have to be arranged in the right spatial relations themselves. If so, the simulation hypothesis doesn’t imply the it-from-bit hypothesis.
This view of space as a sort of primitive container of matter is intuitive, but physical theories increasingly suggest that it’s incorrect. Relativity theory suggests that space is not absolute. Many physicists entertain theories in which space isn’t present at the fundamental level but emerges only at a higher level. If that’s right, then the presence of space is not a constraint on the fundamental level. Instead, like solidity, space is grounded in the patterns by which things interact with one another. Those patterns can be equally present in a simulation.
You could worry about the stuff that photons and quarks are made of. Quarks have a special “quark-ish” intrinsic nature, and a simulation of a quark won’t have genuine quarkishness.
There’s nothing about this sort of quarkishness in modern physics. Quarks are characterized in mathematical terms, and that’s that. Some philosophers and physicists do speculate that quarks and other fundamental entities could have some sort of underlying nature. If the it-from-bit view is right, their underlying nature may involve bits. If the simulation view is right, their underlying nature could involve processes in the next world up. However, physics is neutral on what this intrinsic nature is. If it turns out that quarks are made of bits, so be it. They’re still quarks.
I’ll discuss these issues about structuralism further in chapters 22 and 23, where I’ll develop the argument in more detail. For now, I’ve made a case for the second premise of my argument: If the simulation hypothesis is true, the it-from-bit creation hypothesis is true. I’ve also made a case for the first premise: If the it-from-bit creation hypothesis is true, most of our ordinary beliefs are true. Together these make an argument for my conclusion.
Simulation realism
The conclusion of my argument is simulation realism: If we’re in a simulation, most of our ordinary beliefs are true. There really are cats and chairs around us. That really is a tree outside my window.
Suppose God tells us tomorrow that the simulation hypothesis is true. We should react much as we would if God told us the it-from-bit creation hypothesis is true. We should say that ordinary objects like cats and chairs were brought into existence by a creator, and that they’re made of bits. That will be surprising and interesting, but most of our ordinary beliefs—say, that a cat is sitting on a chair over there—remain unthreatened.
Perhaps we’ll have to revise some of our more theoretical beliefs. If we thought our universe wasn’t created, we were wrong. If we thought the level of quarks in our universe was the bottom level of reality, we were wrong. If we thought our spacetime was the whole cosmos, we were wrong. But most ordinary beliefs, such as my belief that there are two chairs in my office, will still be true.
Even if we’re not in a simulation, beings who are in simulations can make a case that their own world is real. Suppose we create a Matrix-style simulation containing pure sims with many beliefs about their world. Then the simulation hypothesis will be true for them: Their world is a simulation. The it-from-bit creation hypothesis will also be true for them: Their world was created and is made of bits. Most of their ordinary beliefs will be true. The objects they interact with are perfectly real. They’re just made of bits.
You may find these conclusions counterintuitive on first encounter. I’ll briefly address some common objections, with pointers to later chapters in which I discuss the issues in more depth.
Objection: What about the simulator in the next world up? If the simulator can stop the simulation at any time, doesn’t that threaten our reality? If the simulator based cats and chairs in our world on cats and chairs in its nonsimulated world, doesn’t that mean that the simulator’s cats and chairs are real and ours aren’t?
Response: These issues also arise given a standard creator. God might have the power to stop the world at any time, but that doesn’t mean that the world around us isn’t real. God might have based our cats and chairs on cats and chairs in Heaven, but this doesn’t mean that our cats and chairs aren’t real.
Objection: There are no cats in a computer simulation. Computers don’t contain cats and chairs. The brain in a vat believes it sees a cat and a chair, but there are no such objects in the simulation.
Response: The computer simulation contains a virtual cat and a virtual chair. These are real digital objects, each made of bits. I discuss virtual objects as digital objects in chapter 10.
Objection: A virtual cat isn’t a real cat. A simulated hurricane doesn’t make you wet. So how can a virtual cat or a simulated hurricane make my beliefs about cats and hurricanes true?
Response: If we’re in a simulation, cats have been virtual cats all along. Our word “cat” has referred to virtual cats all along. Our beliefs about cats have been about virtual cats all along. I discuss these issues about language and thought in chapter 20.
Objection: A simulation doesn’t contain genuine minds, brains, and bodies. If someone has a brain and a body outside the simulation, like Neo in the Matrix, doesn’t this mean that the brains and bodies in the simulation aren’t real? If someone has no brain and body outside the simulation, like the Agents in the Matrix, aren’t they akin to mindless nonplayer characters in a video game?
Response: Neo has a physical body outside the Matrix, and a virtual body inside the Matrix. Both are perfectly real. As an impure sim, he has a brain outside the Matrix supporting a mind of its own. In a pure simulation, people have virtual brains, but these can still support minds. I discuss these issues in chapters 14 and 15, which focus on the relationship between minds and bodies in simulations.
Objection: Isn’t anything a computer if we interpret it the right way? Won’t that mean that it is trivial to generate realities? How can there be genuine causal processes in a computer-based reality?
Response: Computers involve genuine causal processes connecting different elements of the computer. To run a simulation, you need a system in which these causal processes are set up the right way, which is highly nontrivial. I’ll talk about these issues in chapter 21 on computers and computation.
Objection: Is simulated physics really physics? Aren’t bits too insubstantial to make up a world? Does simulation really yield a genuine space?
Response: As we’ve seen in the previous section, we can address these objections by developing a structuralist view of physics and making analogies with quantum mechanics, digital physics, and other views in which physical reality seems evanescent. I’ll discuss some of these issues further in chapters 22 and 23.
Objection: What about other skeptical scenarios? Even if perfect lifelong simulations are not illusory, what about others? What if I’ve only recently entered a simulation? What if it’s just a local simulation? What if I’m dreaming? What about Descartes’s evil demon?
Response: In chapter 24, I’ll argue that my reasoning generalizes to any global Cartesian scenario threatening global skepticism: the evil-demon scenario, the lifelong dream, and so on. I’ll also discuss the prospects for local skepticism, based on local skeptical scenarios such as the hypothesis that I entered a simulation yesterday. I don’t claim to have overcome all forms of skepticism, but if I’m right, we’ve at least put a dent in one of the most severe forms of skepticism about the external world.
Taking stock
This case for simulation realism caps off my initial response to Descartes’s problem of the external world. By arguing for simulation realism, I’ve blocked the central Cartesian argument that uses simulations to argue for global skepticism.
The Cartesian argument combines two main premises: We can’t know we’re not in a simulation and In a simulation, nothing is real. It concludes: We can’t know that anything is real. If my argument for simulation realism is correct, it shows that the second premise is false. Even if we’re in a simulation, things are still real, and most of what we believe is still true. As a result, the simulation hypothesis does nothing to undercut our knowledge of the world.
If you’ve followed the book straight through so far, there are many directions you can take now. None of the next four parts of the book presuppose each other, so you can read them in any order you like. If you’re interested in how all this applies to real virtual reality technology, try part 4. If you’re interested in connections to issues about the mind and consciousness, try part 5. If you’re interested in issues about ethics and value, try part 6. If you’d like to pursue the argument for simulation realism in more depth and understand its philosophical foundations, try part 7.
Part 4
REAL VIRTUAL REALITY
Chapter 10
Do virtual reality headsets create reality?
IN HIS 1992 NOVEL SNOW CRASH, NEAL STEPHENSON DESCRIBED a definitive virtual reality world: the Metaverse. The Metaverse is a shared, computer-generated world in which people socialize, work, and play. The novel’s central character, Hiro Protagonist, accesses the Metaverse through goggles and an internet connection. The main drag of the Metaverse is a giant boulevard known as “the Street.”
The Metaverse may sound a lot like the Matrix, but there’s an important difference. The Matrix is a simulated universe in which most people have spent their whole lives. The Metaverse is a virtual world that no one spends an entire lifetime in and that people can enter and exit as they choose. Everyone in the Metaverse was born in ordinary physical reality and is still based there. When they choose to, they can don a headset and perhaps a bodysuit and enter the virtual world of the Metaverse. The Matrix is still science fiction (unless it has been reality all along!), but the Metaverse is gradually becoming reality.
Perhaps the first genuine VR system was devised in 1968 by the computer scientist Ivan Sutherland. Sutherland combined computer simulation technology with stereoscopic vision technology—the technology once used in View-Master headsets for viewing stereoscopic color pictures—to yield an enormous headset system. The system was nicknamed the “Sword of Damocles” because it was attached to the ceiling and hung over the head of the user, like the threatening sword described by the Roman orator Cicero. Sutherland’s system was immersive and computer-generated, though interaction was limited to tracking the head movements of the user in order to change perspective on the image. Over ensuing decades, the headsets grew smaller and cheaper, and the interfaces and the computer simulations grew more sophisticated. These days, a number of consumer-grade virtual reality headsets are in widespread use.
Many people have tried to create a Metaverse—a common virtual universe for everyone to spend time in, living out day-to-day lives with many forms of social interaction. The most successful attempt to date has been the virtual world of Second Life, which at its peak around 2008 had over a million users. But Second Life is a world displayed on a two-dimensional screen. It has proved infeasible to port it to genuine VR because of the much higher number of frames per second that VR requires. There have been a few attempts at a Metaverse in virtual reality, but none has yet come close to establishing itself as definitive. Where VR headsets are concerned, the most common use remains game playing. The use of VR for social interaction—also known as social VR—is advancing, though, and it would not be surprising to see a flourishing ecosystem of Metaverses (or one giant Metaverse, depending on how one carves up virtual space) before long.
You needn’t invoke the Metaverse to raise philosophical questions about temporary VR. They arise even for simpler VR environments, such as those used for gaming. The Knowledge Question (How do we know we’re not in a virtual world?) may not arise; most users of VR headsets know they’re using VR. But the Value Question (Can one live a good life in a virtual world?) and the Reality Question (Are virtual worlds real or illusory?) are as pertinent as ever. Some of what we’ve said about the simulation hypothesis carries over to ordinary VR, but there are important differences as well.
By far the most common view is that virtual objects aren’t real. Stephenson himself tells us that the Street in the Metaverse is unreal: “This boulevard does not really exist; it is a computer-rendered view of an imaginary place.”
As you might expect, I disagree. If the boulevard is a virtual boulevard as described, it really exists. It is a real place in a virtual world. It’s grounded in computer processes, but no less real for that.
Even regular users of VR commonly distinguish between the “real world” and the unreal domain of VR. If I’m right, that’s the wrong way to talk. Instead of talking about the “real world,” we should talk about the “physical world” or the “nonvirtual world.” Instead of talking about “imaginary” objects, we should talk about “virtual” objects. Virtual objects are real, too!
What is virtual reality?
How can we best define “virtual reality”? Philosophers have learned how troublesome definitions are. Define “chair.” I’ll bet that for any definition you come up with, there will be counterexamples. Is a chair something you can sit on? Rocks, floors, and beds meet this definition, but they aren’t chairs. How about a flat surface with a back, designed for sitting? Loungers and sedan chairs don’t meet this definition. You can refine definitions, but generally you can never dispatch every last counterexample.
In his 1953 Philosophical Investigations, Wittgenstein noted that there seems to be no characteristic common to everything we call a “game.” Instead there is at best a family resemblance involving a few common threads. The Berkeley cognitive psychologist Eleanor Rosch has used behavioral experiments to argue that in the human mind, most concepts are represented using prototypes instead of definitions. Chairs might be represented using a few prototypical chairs, for example. In fact, most philosophers doubt that perfect definitions of ordinary words are possible in a natural language, such as English. Still, we can try to define “virtual reality” and see where the effort gets us.
Let’s start by trying to define “virtual.” The word comes from the Latin word “virtus,” which originally meant manliness but which came to mean strength or power. “Virtus” is also the root of the word “virtue,” which we now use for strengths or powers of a person in a general sense. In the medieval era, a virtual X was something which had the strengths or powers of X—and most importantly, the effects of X. In a 1902 dictionary of philosophy, the American philosopher Charles Sanders Peirce enshrined this definition: “A virtual X (where X is a common noun) is something, not an X, which has the efficiency (virtus) of an X.”
When defined this way, “virtual” means something like as if. A virtual duck is an as-if duck—something that looks like a duck and has some of the effects of a duck but isn’t a genuine duck. A virtual object in optics arises when one has the appearance of an object without an object really being there. According to Peirce’s definition, a “virtual” reality would be an as-if reality—something that has some of the effects of reality, but which isn’t real. If we approach VR this way, it is illusory more or less by definition.
When the French polymath Antonin Artaud introduced the expression “la réalité virtuelle” as a description of theater in his 1932 essay “The Alchemical Theater,” he seemed to have this conception of the virtual in mind. He likens theater to the “fictitious and illusory” world of alchemy. Both of them are “virtual arts,” and both involve a “mirage”:
