Why Simulation Theory Is a Bust

By all appearances, Elon Musk is a smart man. And when he says that the odds are literally billions to one that we’re living in a Matrix-like simulation, people listen. Fortunately, he’s wrong.

On June 1st of this year, during an appearance at the Code Conference in Rancho Palos Verdes, CA, Musk said:

The strongest argument for us probably being in a simulation is … that 40 years ago we had Pong — two rectangles and a dot. That’s what games were. Now, 40 years later, we have photo-realistic, 3-D simulations with millions of people playing simultaneously. And it’s getting better every year …. If you assume any rate of improvement at all, then the games will become indistinguishable from reality …. Given that we’re clearly on a trajectory to have games that are indistinguishable from reality, and those games could be played on any set-top box … and there would probably be billions of such computers and set-top boxes, it would seem to follow that the odds that we’re in base reality is one in billions.

Since Musk made his remarks, other luminaries have stepped forward to echo his beliefs; Pulitzer-winning New York Times science writer John Markoff, a Simulation Theory sceptic, is quoted in Vanity Fair saying “it’s basically a religious belief system in the Valley.” Given that the idea holds such sway with thought leaders, it deserves careful consideration.

Musk, after laying out the theory in June, defied his audience to find the flaw in his argument. And no one did — despite the fact that it’s actually very easy.

¶ The strength of Musk’s argument lies in its appeal to reason, not in an appeal to our perceptions of reality. We know, intuitively, that it’s ridiculous to believe we’re trapped in a video game and don’t know it — but our intuition has lied to us before; the earth looks flat to us as we walk its surface, but we’ve adopted a non-intuitive way of thinking about the earth’s shape based on received evidence that contradicts our experience. Musk’s argument is weak when it comes to our perception, so if we’re going to adopt it, the reasoning behind it has to be unassailable.

In Musk’s version of the theory, we can only be certain of our own reality — after all, he says our reality is “one in billions,” with “one” referring to the reality we perceive and “billions” referring to the possible number of simulated realities in the universe. There are billions of people on earth, each with their own perception of reality; Musk treats these like a single shared reality. But this reality, assumed by the theory to be a simulation, is also the pseudo-reality that supposedly gives us evidence that the universe is full of simulations — that is, any evidence we have to support Simulation Theory comes from the simulation, not “base reality.” How can we know the universe is full of simulations when, if you follow the logic of the theory, we have no direct knowledge of the universe? Musk’s reasoning is sloppy, but it’s not enough to fatally undermine the theory.

Musk’s version is not the only one; Oxford philosopher Nick Bostrom first laid out the tenets of Simulation Theory much more rigorously in 2003. According to Bostrom’s deductive reasoning, one of the following statements must be true:

1. We’re literally living in a computer simulation
2. There is a strong aversion in advanced civilizations to making “ancestor simulations” — that is, highly evolved civilizations have no interest in making what are essentially video games of their past
3. Something destroys all civilizations before they’re able to advance to the point where they are technologically capable of simulating consciousness

If none of these can be disproved, it follows that the theory is true.

Bostrom’s three scenarios, like Musk’s assertion that the universe must contain billions of computers capable of creating simulations indistinguishable from reality, are all based on a big assumption: that advanced civilizations have emerged elsewhere in the universe. Here’s where the theory breaks down.

There is zero direct evidence that life has ever existed anywhere but on earth. What “evidence” we have rests on deductive reasoning, which stretches into ever more precarious territory.

¶ If we want to estimate how many advanced civilizations there are in the universe, the formula is simple enough. The Drake equation lists all the variables that would affect this number, so it’s just a question of plugging in the numbers: N = R* • fp • ne • fl • fi • fc • L, where N is the number of advanced civilizations, and the remaining variables stand in, respectively, for the rate at which stars capable of hosting life form, the fraction of those stars with planets, the number of planets capable of hosting life, the fraction of suitable planets where life actually forms, the fractions of life-bearing planets where civilizations form (fc and L refer to civilizations we might communicate with directly by radio, which is not really relevant to this discussion).

In the years since Frank Drake conceived his equation, astronomers have made great progress in determining how many planets — even planets like ours — there might be scattered throughout the universe (the estimate is about a billion earth-like planets in our galaxy alone). If you confidently embrace a teleological view of life, you might even believe that civilizations have evolved from most or all of the life that has originated on other planets.

The sticking point, though, is the variable fl, “the fraction of suitable planets on which life actually appears.” Given ideal conditions, how often will life spontaneously emerge? We can look at the evidence at our disposal: we know of one earth-like planet that we can examine for life (our own), and indeed life exists here. That is a rate of 100% — which, to some, suggests that life must occur pretty much anywhere it’s possible for it to occur.

Unfortunately, a sample size of one is of no statistical use. Not even limited use — of no use. Because the condition we’re looking for — life — is a necessary condition for the looking to occur in the first place, so it will necessarily occur on any planet where people are looking. That’s the anthropic principle in a nutshell, and some version of it has existed for over a century.

That being said, while the earth itself offers only a single test-case, looking at the events of earth’s history might reveal more about how likely life is to arise, given the right conditions. After all, if life could arise on earth once, it might be able to do it any number of times — anywhere the conditions are right. In fact, as life spread across the planet, it would presumably become more likely that life would spontaneously emerge a second or third time: the oceans are now awash with organic molecules, there are random strands of complex proteins floating throughout the environment, and the composition of the atmosphere today is demonstrably capable of supporting not just one form of life but many diverse varieties. Given all this action happening across the globe, and given the assumption that life will emerge wherever it can, it follows that life should have sprung from nothing many times throughout earth’s history. Or at least a few times.

That’s not what we see. Life, to our knowledge, has only arisen once in the billions of years that earth has offered prime conditions for it to do so. For that matter, scientists trying to jump-start life have been working for decades to create ideal conditions for molecules to self-replicate, thus creating new life in a lab. The fact that even experts on the origins of life can’t recreate it when they’re trying to suggests that, in all likelihood, life does not spontaneously emerge on any planet capable of supporting it. It suggests the opposite: that the emergence of life is so unlikely that, apart from a single exception, it never happened at all.

¶ Remove the assumption that life is everywhere throughout the universe, and the whole house of cards collapses. Gone are the alien civilizations, and with them the “billions of computers and set-top boxes” that might convincingly replicate reality. Two of Bostrom’s three scenarios deflate; advanced civilizations have no “strong aversion” to simulating primitive worlds, nor are they destroyed before they achieve the capability of doing so — they just simply never were. We might develop the ability to create convincing simulations ourselves, some day, but for us to assume that we already have, and we’re looking at it, is just another tautology, and believing it relies on faith, not evidence.

Why should we assume there is life anywhere but earth anyway? All we know of life first-hand is that it emerged only once, evolved and diversified, slowly at first, then more rapidly We know life surmounted seemingly insurmountable obstacles on its journey from a strand of self-replicating protein to complex, sentient life — the shift from single-celled prokaryotes to slightly more complex but still single-celled eukaryotic organisms took more than 3 billion years. That it happened only once during an inestimably huge number of generations suggests that it’s improbable to the point of virtual impossibility.

Can we know whether what we are experiencing is a simulation or “base reality”? Is Simulation Theory falsifiable, or is it ever-retreating in the face of science? The philosopher G. E. Moore, speaking in the middle of the last century, would only laugh at the question; asked to prove that reality was really real and not just a dream, he replied,

Of course empirical propositions are not certain the way that a priori propositions are. Of course they can be denied without self-contradiction. If this were not so, they would not be empirical propositions. But it does not follow from this that they cannot properly be said to be certain in any sense at all. It does not follow that they cannot be known for certain to be true. Do you mean to tell me that you do not know that you are awake and reading this? Do you mean to tell me that I do not know that I have a pen in my hand? How improper that would be, what a misuse of English, to say that it was not certain that I had a sheet of paper in front of me, but only highly probable. How absurd it would be to say, ‘Perhaps this is not a pen. I believe that it is, but I do not know it.’

Moore’s point was that we don’t have to prove something beyond any question in order to know it’s true, and to put such a high burden of proof on the reality very clearly before us calls into question literally everything. At some point, these lines of questioning cease to be meaningful. Simulation Theory is an interesting thought experiment, but one that can ultimately tell us nothing about the reality we live in.