Saints & Simulators 2: The #SimulationArgument

Second in a series on the nexus between religion and technology. The previous essay is here.

In the year 1999, just on the cusp of a new millennium, the then Wachowski Brothers released what would become one of the most influential, imitated, and widely discussed movies of its times. The Matrix was a stylishly paranoid thriller about a future world that looked just like America in 1999, but that was actually just an elaborate fake created by a vast, parasitic computer. With its “red pill, blue pill” existential choice to either wake up to reality as it is, or continue to dream a computerized dream, the movie launched thousands of late-night dorm room conversations about whether we really live in the real world. As ground-breaking as the movie may have been, however, the ideas it contained were not new. And around that same time, people were beginning to take them seriously.

It was only four years later, in 2003, that philosopher Nick Bostrom of Oxford University introduced what would become known as the Simulation hypothesis. This advanced the argument that it is not only possible we are living inside a computer simulation, it is actually significantly likely. Although it may have sounded like a high-concept science-fiction thriller, the argument drew upon well-established lines of logic from earlier arguments that had already entered the realm of legitimate philosophical discourse, and established its conclusion by drawing on a series of assumptions that (while not held universally, or without controversy) are widely believed.

The suspicion that the world is somehow less than wholly real is an ancient one. People inclined to a mystical frame of mind have long speculated that the superficial appearances of life are dependent on some deeper, more fundamental level of reality. This line of thought traces as least as far back as the ancient Greek philosopher Plato, who, in his influential book The Republic (c. 380 BCE) compared what most people think of as reality to shadow images projected on a cave wall by the flickering light of a fire. It may well go back much further. The beliefs of the indigenous Australian people, for example, which are believed to stretch back tens of thousands of years, revolve around a mythological level of reality, often described as "The Dreaming," that is considered more fundamental and foundational than our ordinary waking lives.

The idea that our reality might be computer-generated is of course much newer. Yet it cannot be called unfamiliar, having been explored extensively in books, by bestselling authors such as William Gibson and Neal Stephenson, and in hit movies like The Matrix and The Thirteenth Floor. It is only recently, however, that this concept has been taken seriously by a small but growing cadre of scientists, academics, and other intellectuals (notably including billionaire investor and inventor Elon Musk). This group believes that computer simulation in some higher-order universe is a legitimately possible truth about the existence we experience.

The shift from speculation to belief in simulationism is most directly traceable to Bostrom’s argument. Although its original formulation is relatively simple, we can create our own, even more clear version, reduced to just two basic propositions:

1. Someday, our descendants may be able to create computer simulations of fully conscious human beings; people like ourselves.

2. To whatever extent it is likely that descendants of ours will simulate people like ourselves, then it is proportionally likely that we ourselves are simulated by entities such as we hypothesize our descendants to be.

The first proposition is one that seems perfectly reasonable to some people, and yet utterly absurd to others. If you believe artificial intelligence is possible; if you believe consciousness is a natural result of the pure complexity of the human brain’s structure, and not a spiritual or otherwise extra-physical property; if you believe technological advance is both unstoppable and accelerating; then the first proposition is difficult to deny. If you do deny it, it is likely either that you reject one of the clauses of the previous sentence, or that you have not fully considered all their implications.

To expand on this line of reasoning: At the dawn of history, human beings lived (or so we thought) in a spiritual universe. We believed the animals and even the plants around us had human-like traits and motivations. We perceived a mysterious quality called “soul” even within those elements of the world more modern humans assume are composed only of inert, senseless matter: the sun, the moon, the mountains, the sea.

Later, this sense of soul became reduced from that all-encompassing vision, and removed from the realm of everyday experience. It was reconstituted as the exclusive reserve of religion, which commanded belief in a spiritual realm, ruled by a spiritual being, but largely separated from the ordinary world of reliable matter. The accompanying concept of two separate but inseparable realms, joined uneasily at the nexus of the human individual, a material body somehow inhabited and animated by an immortal soul, became known as “substance dualism” and was popularized by the great mathematician and philosopher, René Descartes.

Almost before the ink on his manuscript was dry, however, Descartes’s view was already being opposed by those who felt that having two fundamental substances in the universe was excessive, if one could do the trick by itself. Even those who were united in their opposition to dualism, however, sometimes found themselves at odds when it came to naming their monism. If only one out of matter and spirit was necessary, then which one was it? Which substance was fundamental, and which was secondary?

There were those who continued to argue that spirit, not matter, was the all-encompassing foundational substance of the universe. Slowly but surely, however, the materialist viewpoint began to dominate. In the 1700s, the materialist Samuel Johnson took issue with the idealist Bishop Berkeley over the question of whether matter really existed, or whether all things that seemed to be real and physical were actually immaterial ideas within the mind of God. According to his biographer, James Boswell, when challenged to refute Berkeley’s argument, Johnson kicked a large stone with his foot, and shouted “I refute it thus.” This was seen by his many admirers as a robust and practical response to the fuzzy thinking and strange ideas of idealist philosophers, who were so distanced from physical reality they were capable of disbelieving in its very existence.

Spiritualism made a vivid and largely ill-advised comeback in the nineteenth century, when all manner of superstitions seized the popular imagination; and fairies, fortune tellers, magicians, and mystics of every stripe and level of plausibility were celebrated with full credulity by even those who ought to have known better. Aside from this, however, the general trend has been a march toward a realist, physicalist monism; the belief in a wholly and solely physical real universe, wholly composed of matter and energy, objectively behaving in rational and predictable ways, obedient to the dictates of knowable and unbreakable physical laws. This, scientific consensus seems to agree, is the universe we live in. Our understanding of the world is no longer based in strange dreams and visions, but rather is gifted to us by the inexorable advance of science and technology, which has demystified all the strange corners of the world, leaving us with a cosmos still wild and wonderful, but wild and wonderful for strictly physicalist reasons.

It is arguably the successes of science and technology that have tipped the scales towards physicalism. Science has explained much of our world, and the effectiveness of technology as the application of scientific principles is an advertisement for the validity of the worldview that produced it. In other words, technology gives us visual proof that science works, a proof we need not take on faith; and science offers us a universe free of the explanatory need for any independent realities of mind, soul, spirit, or consciousness—except as epiphenomena, side effects of the interactions of matter in motion.

It may seem that these assumptions give us a very solid, rational, scientific, reliably real universe. But they also give us the conviction that anything we experience in the world is something we can potentially re-create. If everything is just matter in motion, then by sufficiently sophisticated manipulations of matter, we are theoretically capable of building anything that exists. This extends even to something as seemingly miraculous as the human body, and even beyond that, to the human mind, or to what we once understood as “the soul.” This is the essential core of our first Bostrom-inspired proposition. Rejecting it requires rejecting or modifying at least some part of the standard belief in a real, purely physical universe.

The second proposition is actually more secure, although it is trickier to understand. From a technical standpoint, it relies on some of the more esoteric implications of statistics, but in practice it comes down to this idea: It is roughly just as likely that the simulation of people like ourselves will happen in the future as it is that it has already happened in the past. In other words, if we think it is likely that people like ourselves are capable of being simulated, then there is no compelling reason to be sure that we ourselves are not simulated.

This can be a bit hard to work your mind around. We can understand it a bit better, perhaps, by removing the time element. Suppose it were possible, right now, to clone, or otherwise create a perfect duplicate of yourself, as an adult, complete with all your thoughts and memories intact. If you woke up one morning and found two of you, how would you know you were the original? Your chances of being the original person would only be one in two (less if there were more than one clone). Similarly, to the extent Mr. A is likely to have a great-great-great-grandson who creates utterly realistic computer simulations, then it is equally likely that Mr. A is actually Mr. A-Prime, the perfectly accurate computer generated re-creation of the original Mr. A created by that future grandson. And in the case that we know the grandson creates multiple copies of his re-creation, then any person who thinks he is Mr. A is more than equally likely to be a duplicate as the original.

But what would it mean to be re-created, not in the real world, but on a computer? One way of describing this is as an extension of the concept of virtual reality, the illusionary simulation of a real universe through the use of advanced computers and sensory interfaces; toward the goal of creating an experience that feels convincingly real to the user, yet without actual correspondence with objective physical reality. Given the role science has played in building and supporting our current collective beliefs, it is deeply ironic that this newest and perhaps ultimately the most damning assault on our confidence at living in a physical, objective universe that really exists emerges from within our science-based technology itself.

Virtual reality is far from perfect, but even as of the year 2017, it had advanced a considerable distance toward the achievement of perfection. Were Samuel Johnson alive today, he could put on virtual reality goggles, and see something that looked like a rock, and he could put on a virtual reality boot and kick something that felt like a rock, and put on virtual reality headphones, and hear something that sounded like a rock being kicked, all without there actually being a rock in front of him. After taking off the gear, would he still be so confident in his refutation of Berkeley? Could any of us be?

What lends the Bostrom argument strength is that it starts from the idea that either human beings will at some point advance technologically to the point where they are able to create realistic computer simulations of life as we now know it, or human beings will die out before they ever reach that point technologically. At first glance, it may seem like Bostrom has chosen two relatively unlikely scenarios, and that it is possible for neither to occur; but a closer examination shows that this is essentially a binary choice that covers all options. Either our descendants will one day reach the point where they are able to simulate the lives of us, their ancestors, or they will not, one or the other. They may not reach it anytime soon, they may not progress there in a linear fashion (in other words, without reversals and cycles of decline and recovery), they may not reach it for millions of years, but they will get there eventually or they will not. It might be that it is impossible to ever reach the level of technology needed to simulate realistic human beings, but this still fails to escape Bostrom’s dilemma, it merely deposits us permanently on the “will not” side.

Having taken us this far, Bostrom offers another fork in the road. If our descendants do reach that level of technology, they will either simulate the lives of we, their ancestors (either specific people, if their knowledge of us is good enough, or else generic “best-guess” people of our times), or they will collectively choose not to. Everyone may not agree that ancestor-simulation is an obvious thing to do, but Bostrom is not relying on a claim that everyone will be doing it. Either at least one person (or perhaps one technologically augmented “transhuman”) will decide to simulate his or her ancestors, or no one will.

These are all the choices Bostrom needs. His argument to this point can be summarized as follows. Either humanity dies out before it gains the technological ability to simulate its own ancestors, or it decides unilaterally not to simulate its own ancestors, or, it can and does simulate its ancestors. If you reject both of the first two options, you are committed, according to the argument, to the third. But if our descendants will be able to simulate us, then it is possible that we could actually be the simulated ancestors, rather than the originals. And—here is Bostrom’s key move—it is not only possible, it is likely, since there is little reason to expect our descendants will stop with just one simulation, but rather that they are likely to run many of them, so many that far more of the “people” who ever exist in the universe will be simulated than original.

Bostrom’s argument is not without its flaws, but it's harder to debunk than it might seem at first glance, and deserves to be taken seriously. Significantly, it has consequential implications for our core beliefs about the world, regardless of the strength or weakness of the argument itself. The very act of trying to explain why and how the argument goes wrong inevitably forces us to reevaluate some of our most cherished assumptions about what is, and is not, real. We are still a distance from virtual reality that convincingly duplicates real life to the point of indistinguishability, but we get closer every day. Once we reach that point, how will we know if we are in the simulated reality or outside of it? And once we reach that point, how could we be sure if we were ever outside the simulated reality in the first place?

by