Micah Redding — faith in humanity's future

In this series:

An Introduction to the Simulation Argument

In 2003, Nick Bostrom published his Simulation Argument, and shook the landscape of philosophy.

The basic idea is easy to describe.

For the past few decades, we’ve been building simulations of the world around us. We created Simcity, and then the Sims, and then one after another, a barrage of increasingly realistic video games.

We can only imagine this trend will continue. Our games and simulated worlds will become more and more complex, more and more concrete, more and more capable of deceiving us into believing they are reality. Until one day, they do.

Sometime in the future, we will manage to create a simulation that will convince us, in every way, that it is real. It may not happen until 2050, or 2150, or til the year 3000. But it is almost certain that we will eventually do so, and that once we’ve done it, we will do it over and over and over again.

One of the things that we will simulate is our own past. After all, we currently play Oregon Trail, and make slideshows out of our old photo albums. Why wouldn’t we try to experience them as in-depth as possible? Why stop with photos and video? Why not re-live the past as it really was?

In the next million years or so, we will have created trillions upon trillions of such simulations, including an unimaginable number that simulate this moment right now.

But if the number of people experiencing this moment as a simulation is in the millions or trillions, and the number of people experiencing this moment as reality is just one, then we have to ask ourselves whether or not we are, in fact, in the simulation right now.

In fact, it’s obvious that if we accept the above sequence of events, then the probability that we are in a simulation is overwhelming.

We can and should generalize this argument. The basic premise is that in all realities (whether past, future, in other galaxies or in other universes) where sufficiently capable intelligent beings exist, they will create an almost infinite number of simulated worlds.

In any reality which allows simulated worlds to exist, the simulated worlds will outnumber the real worlds, trillions to one.

Thus, the vast majority of beings will exist in simulated worlds.

And every intelligent being should make the assumption that their world is a simulation.

There are two basic objections to this idea, which leads to what I describe as the trilemma: Suicide, Censorship, or Simulation.

The first one (suicide), argues that these virtual worlds never come into existence because every intelligent civilization ultimately destroys itself before becoming capable of such simulations.

If this is true, then we should expect that humanity will die out in the very near future, probably before the end of the century.

The second one (censorship), argues that all intelligent beings will see these simulated worlds as being unethical, and will permanently ban all such technology.

Note that for this to impact the argument, it’s not sufficient that one country or species ban the practice. It actually requires that every country, every race, every species in every world, ban the practice for all eternity.

This brings us to the third scenario. If intelligent species don’t all die out, and if they don’t all ban the practice, then we’re left with the inevitable conclusion:

Virtual worlds outnumber real worlds, and we’re living in somebody’s simulation.

What are we to make of this?

I have more to say on the subject, but I think the basic idea holds firm. This isn’t the traditional “maybe the world is a dream” musing, which starts with doubt about what we know and experience. This argument starts with the assumption that the world as we know it makes sense, and ends up concluding that if so, then the world as we know it isn’t real.

But this brings us to a question about what reality really means. And that’s what I want to talk about next.

Next: The Simulation Argument, remixed


Lincoln Cannon:

Enjoyed the article, Micah. It's worth noting, too, that the Simulation Argument can be generalized beyond computational creative mechanisms to any feasible creative mechanism. The generalized Creation Argument is part of the New God Argument: http://www.new-god-argument.com/p/creation-argument.html


Suicide, Censorship, or Simulation? What about apathy, technological barriers, practical costs in terms of engineering, energy requirements, development issues? Any of these could negate all three options. Your trilemma is no such thing.


Hi Hugh, thanks for commenting. I agree that all those are possibilities. I was just trying to give the options presented by the original argument, expressed in an easy-to-remember way. However, that said, I think it could easily be argued that all the things you've mentioned fall into those categories. It's really just the question of whether we will be _able_ and _willing_ to create these simulations. If we're not able to gain that ability EVER, it's probably because we've committed suicide. If we don't want to do it EVER, that's what I'm calling censorship. Again, I'm just using some catchwords to get the gist of the alternatives.


"If we're not able to gain that ability EVER, it's probably because we've committed suicide." The question is will we ever attain a practical realization of this technology, not will we be able to. There are many ways in which this technology might never be realized that do not connote suicide or extinction, I listed some above and they cannot fit into categories of suicide or censorship. I'm pointing this out because this is precisely what I dislike most in Bostrom's own presentation of the argument. He is needlessly polemical in calling the first proposition of his trilemma DOOM and focusing solely on extinction, when even he will admit elsewhere that the truth of the first proposition of the trilemma does not necessarily imply extinction. Presenting, as he does, the options of the trilemma as either, (1) extinction, (2) no interest in ancestor simulations or (3) we are most likely simulated, is intellectually dishonest.


This is a common science-philosophy thing, where the argument starts with "on a long enough timeline." So, on a long enough time line, technological barriers and practical costs are assumed to be covered by the apparent exponential growth of human capabilities, and we're left with suicide, censorship, or simulation. That's the devil's advocate version anyway. I'm not a Buddhist, but my guess is Buddhists would have no problem with these existential questions, whereas for other it might be a moral issue.


Wouldn't each subsequent simulation and simulation in simulation become further and further removed from the original reality as it becomes less and less complex or maybe its a perfect program?


Rev, each level of simulation would definitely have to be less complex than the one containing it, however, the inhabitants wouldn't necessarily know it was less complex. For one, they wouldn't know anything different. For another, you could be in a very small simulation that _felt_ like a very large simulation.


Less complex, or perhaps just running more slowly, relative to the host world. Computer virtual machines today are just as complex as a physical machine, but have some (negligible to large) overhead. Or some combination of less complex and slower.

Lincoln Cannon:

A hybrid creation mechanism (computation leading to instantiation outside the original computational substrate) would not necessitate decreasing complexity. To use a biological example, my wife and I planned to have a child (including pregnancy diet and exercise plans, and financial plans for baby gear and medical costs), then we conceived that child, and my wife gestated and birthed that child, and finally we nurtured and educated the child. The planning, conception and gestation were the computation in my wife's body which led to and shaped aspects of the instantiation of the child outside her substrate. I assure you that child is now at least as complex as my wife!