Monday, July 25, 2011

Mailbag Monday: The Simulation Argument

Mailbag Monday: A weekly segment that covers readers' questions and concerns about all things Philosophy, Bro, and Philosophy Bro that don't quite fit anywhere else. Send your questions to with 'Mailbag Monday' in the subject line.


zepfan5150 writes,
Hey Philosophy Bro,I've been hearing a lot recently on the simulation argument..any chance you wanna try and tackle it? You've explained everything else it seems like.
Oh man, the simulation argument. Alright, look. The simulation argument falls into the category of Cartesian Mindfucks, the sort of arguments that ask, "Is the universe really, truly the way we perceive it to be or are we somehow being deceived?" 

This particular argument proceeds by presenting three different propositions, at least one of which is probably true. Those three propositions are, in a nutshell: (1). We'll probably go extinct before we evolve past humanity, to what the original author calls "posthumanity", (2) If we do achieve posthumanity, we won't really care enough to run tons of simulations of our ancestors, or (3) We are living in a simulation.

At face value, it's easy to see how one of those has to be true. Let's say there's this race of superbros who are technologically way the fuck ahead of us. That's not hard to imagine - you're likely reading this on a computer that is thousands of times more powerful than the very first supercomputers, built only decades ago, and they stored those fuckers in entire rooms. Imagine the shit we'll have in a couple centuries - crazy fast computing. Ridiculous. So let's say these bros have that level of technology. They could, if they wanted to, run simulations of human evolution, right down to the very thoughts of each simulated person. In fact, they could run millions of these things. And if the did run these things, then the chances that we're a simulation, instead of a race that will eventually run simulations, is literally millions-to-one. Those are not good odds, kids. Either we're almost definitely a simulation, or these simulations just don't happen.

What exactly are these simulations I keep talking about? This is the Cartesian Mindfuck part. What if our brains are actually just giant processors in a computer somewhere? After all, as far as we know, the brain is just a bunch of individual cells linked together, but we only know that from looking at brains, feeling them, examining them; this could all be an illusion. Given sufficiently advanced technology, we could definitely simulate a neural network of connections large enough to mimic a brain; sure, it would be crazy now, but these guys are way advanced, remember. What is the difference between a quadrillion connections in the brain and a quadrillion connections on a computer chip if they can process all the same stuff? Throw in some software that mimics our desires to have lots of sex and to belong to groups, and baby, you've got yourself a stew.

So, we could be a computer simulation? I mean, maybe. For some reason, people think that if a machine gains sentience, it realizes that it's a machine and starts mowing down humans as fast as it possibly fucking can. But really all we mean by 'sentience' is 'having consciousness', or 'able to have subjective experiences', or something like that - philosophers don't exactly agree on what 'sentience' is, but we mostly agree that it's something special that sets us apart from our laptops. Self-awareness, maybe. But self-awareness doesn't mean that you know everything about the conditions of self. Why couldn't we just as easily be a program in a computer as a brain in a vat? What's so special about biological tissue? Why can't a computer chip be the basis for consciousness?

Look at it this way - if, in The Sims 47 or whatever iteration we're on, one of your characters became sentient, would he know he was in a computer, or would he just be making more complex decisions? All he's ever seen is the world he inhabits, the house you've built for him. He would flirt with a girl down the street and wonder why she's so boring; he'd have no reason to suspect she was a computer program - I know I've met girls who couldn't pass the Turing Test, but I never check them for batteries.  He'd have even less reason to suspect he was a computer program. What if we're just like the Sims, and it's taken us thousands of years to realize it?

So now that I've sufficiently freaked you the fuck out, let's take a deep breath and see if there's a way out. Is there any way these simulations would not get run? Well, sure, and those are the first two propositions. The cool thing about this particular paper is that it gives us the way out of skepticism, ready-formed. We should very seriously consider whether anyone makes it to this level of technology, and if they do, whether they would run the simulations at all.

First, the simulations won't get run if civilizations tend to go extinct before they have powerful enough technology. Have you met us? I'm not saying we're all absolute dickheads, but given enough dickheads with too much power, it's not hard at all to imagine the myriad ways we might kill each other off before we can build the level of supercomputer we'd need. If this is an evolutionary pattern, if advanced civilizations keep fucking themselves up, then the same thing would probably happen to any civilization that approached this level of technology, in which case no one would ever get advanced enough to run simulations, and you can sleep soundly.

Hell, we wouldn't even necessarily need to kill each other off; we would just need to never reach the point where we can run these super-complex programs. Maybe we'll be around for another million years, but we never have the raw materials to make a computer big and complex enough. Maybe we're at the edge of computing power - we are talking about ridiculous amounts of processing here.

Even if some civilbrozation made it to the point where they can run these simulations, they just don't give enough fucks to run them. Maybe they figured out evolution some other way; maybe they don't want to waste the computing power. Who the fuck knows why? There are lots of reasons they might not run simulations; maybe something about getting past the dickhead stage of human civilization made them decide not to create sentient simulations and expose them to pain - maybe this is an incredibly ethical bunch of geniuses.

Those are the things that might prevent us from being a simulation, but if it's possible to develop the technology, and bros with the technology decide they want to simulate their ancestors for whatever reason, maybe it's possible that we really are just a part of that simulation. Like all the other Cartesian Mindfucks, there's no real way to prove otherwise. I mean, if one of your Sims became sentient and started fucking shit up, you could just delete him. Just because he's sentient doesn't mean he knows how to get out of the game and into the rest of the computer: he is, as we are, still constrained by the environment around him. On the other hand, this is unlike those other Mindfucks in that it gives us clear alternatives to the skepticism. The argument isn't just that we could be computer simulations; the argument is that either we're a simulation, or simulations don't get run. Perhaps there's something about the direction of our advancement that isn't obvious, but that we should look into - unlike the other skeptical scenarios, this one kindly suggests really interesting avenues of inquiry into human nature. 


The popularity of this idea led the paper's original author to set up a website that you can find here. That site has other resources as well as a link to the original paper, which contains some mathematics of probability but is otherwise fairly accessible. The FAQ is also particularly clear.


  1. "I know I've met girls who couldn't pass the Turing Test, but I never check them for batteries."

    Hands down the winningest sentence ever written on this site.

  2. I am pretty sure this is just dwarf fortress

  3. I think I missed the point that shows why it's most likely we are in a simulation (if such technology were possible). Isn't the general idea of science to make observations and derive facts from them, not derive facts from nothing and look for things to disprove them?

  4. @Wes: Dito.

    And some arty-farty post-humans surely would put up a "project" where they simulate the old human race just for the sake of art or to state some deep, meaningful somethingness. You know, the stuff artists do for a living.

  5. Wes:

    I think the more convincing 'we could be in a simulation' argument is simply the fact that we can never prove either way. Not that our brains may be plugged into a Matrix, but that we may only exist as software, not hardware at all.

    It is, after all, entirely feasible that we can understand all of the 'laws' of our Universe, but be completely wrong about its origin, since by definition as a similation it will be an internally consistent closed system. It may even have a simulated God! We simply cannot answer this question.

    The big bang may just be the switch being flicked on and our Universe booting up. No matter how much scientific knowledge we obtain, this will still be a possibility and beyond our ability to ascertain. Even if our programmer gods reveal themselves, they are still straddled with the same uncertainty, the same limit to their own knowledge. Are we a simulation within a simulation??

    This is also why atheists are wrong when they say their view is not a belief system, but a lack of belief. They're calculating a probability they cannot possibly calculate using tools completely inappropriate to the task.

  6. @Wes:
    1. Philosophy is by no means science. There is no testing or observation involved. That doesn't, however, mean that it is "deriving facts from nothing.
    2. The argument is that if any species discovers how to create simulations with sentient beings, or, "The Sims 1000," and run some simulations, for each simulation that exists, the ratio of sentient beings that are simulated to the sentient beings that are "real" becomes near infinite. This is assuming they run a shit ton of simulations. There's no particular reason why it would have to be future humans or anything. In fact, by this argument, even if future humans figured it out, it would be like a simulation inside a simulation.
    3. @Grayzie this has little to do with atheism (which is not a belief system) because atheism says nothing about whether or not we are living inside a simulation. Atheism's standpoint on the simulation would be that we have no evidence to support that we are, so we have no reason to believe it (whether or not it is true).

  7. (1) We'll probably go extinct before we grow in population and artistic ability to the point where there are lots of standup comedians and they are very good at impressions

    (2) If we do grow in population and artistic ability, we won't really care enough to do lots of impressions of our ancestors, or

    (3) I am not a person but actually an impression being done by a standup comedian

  8. OK, so.... If it's possible that our brains are nothing but pieces in a computer.... And Searle argues that computers cannot be thinking, sentient things... Does that mean that there's a possibility that we are not sentient, thinking beings either?

    I think I need to sit down...

  9. if dwarf fortress ever gets finished

  10. Spot on, ML. I had already copied that sentence to paste into my comment to write essentially the same comment.

    Well done, PB.

  11. @Anonymous re "atheism says nothing about whether or not we are living inside a simulation"

    Well exactly, because it can't.

    Of course there is evidence for a God, it's called existence. There are only two possibilities regarding the nature of our Universe, either it is created or it simply just is. Either option is largely unsatisfactory since they both require that something "be" from "nothing". This is why you get theists asking atheists "Where did everything came from?" while atheists are asking theists "Where did God come from?" They both think they have a valid point, and they both do.

    Since we cannot make any meaningful statements about what my exist or occur outside of our closed system we also can't make any judgement regarding the origin of that system. Since there are only two options, to disregard one is tantamount to choosing the other. It's a positive claim that atheists make using spurious reasoning and a claim to 'likelihood' and 'probability' that they are not justified in making. Since we can't say anything meaningful about the topic, the probability ratio between God:NoGod is exactly 50/50. We simply cannot know and probably never will. To claim one is more likely than the other is then a belief.

    Don't buy into this 'new atheism' rubbish so quickly. It's turning into a religion itself unfortunately.

  12. I agree with everything that you said, as long as by "God" you mean "something that created the universe" and of course not "biblical christian god" or "Zeus." Also, existence is at best circumstantial evidence for the universe being created.

    I think that our dispute was simply over what you meant by the term "atheist." In my opinion, as an atheist, I believe that while we have no idea how our universe came to be, we do know that the religions created by man are all pretty shitty. That doesn't mean that it's impossible that "bible god" came down to earth and did all the incredible stuff that is claimed in the bible, but it does mean that we shouldn't worry about it due to the lack of any evidence.

    Same with creation. I don't think that it's impossible (or even really unlikely) that the universe was created, but as of now, we don't have any evidence for who or what created it, or whether or not it was created. And as we have no evidence, it doesn't really make sense to believe one way or the other. And that brings us back to (weak) atheism.

  13. I think we're in agreement.

    However, agnosticism, strong or weak, is not synonymous with weak atheism, and I find the attempts by the New Atheist movement to paint it as such a blatant political move to try and recruit agnostics in their militant campaign against organised religion.

    Both weak and strong agnosticism are compatible with a suspension of judgment, not necessarily a lack of belief. On first glance it may look like there's no difference, if one suspends judgment one is not a theist, therefore one is an atheist. But like I said, if there are only two possibilities, to reject one is making a positive affirmation of the other.

    To elucidate what I'm getting at a little further let me give you an analogy: Say I have two balls, one blue and one red, and I ask you to choose one. If you say, "Not the red one," then you have effectively chosen the blue ball. It is a positive affirmation as determined by a negative declaration.

    Since there are only two possibilities for the origin of our universe (if you can come up with others let me know), this is what the atheist is doing. They discount the evidence for God, so fall by default to the only other option. All god so far.

    The agnostic, however, is doing something different. The agnostic isn't saying, "Not the red one," or, "Not the blue one." They are refusing to make a choice at all.

    It is true that agnostic/gnostic deals with knowledge and theism/atheism deals with belief, and in this sense one can be an agnostic atheist or a gnostic theist, or any range of combinations. But that does not mean an agnostic MUST be an atheist. There needs to be a distinction, because of the ball analogy I listed earlier, between an atheist's position and an agnostic's, and there absolutely IS in the public mind. This attempt by New Atheists to paint all agnostics as atheists is pure political semantics, and certainly NOT what Thomas Huxley had in mind when he coined the word.

  14. @Grayzie

    Thank god (or not, hehe) for your comments! As an agnostic myself, I've always thought that agnosticism is the most humble intellectual stance. I loathe how atheism is becoming a religion itself, while a lot of their adherents are blatantly unaware of this similarity. To agree on anything out of habit is to forsake our capabilities as sentient beings, simulations or not.

    Of course, knowing that we can't make meaninful statements about the God:No God issue shouldn't stop us from engaging in good conversation. I think what a lot of people should realize, particularly those who take themselves too seriously, be they theists or atheists, is that, since neither side is correct, there's no need to get angry about it. We should value discussion itself as a platform for intellectual excercise, not as a sort of mountaintop from which to look down on those who disagree with our viewpoints.

  15. @Grayzie

    "There needs to be a distinction, because of the ball analogy I listed earlier, between an atheist's position and an agnostic's, and there absolutely IS in the public mind. This attempt by New Atheists to paint all agnostics as atheists is pure political semantics, and certainly NOT what Thomas Huxley had in mind when he coined the word."

    1. While a distinction does exist in the public mind between "agnostic" and "atheist," that doesn't necessarily mean one should because...

    2. This history of differentiation between "non-theists" is pure verbal semantics. I suspect it has been perpetuated by the ability of the cultural majority - namely, theists - to define the framework of the argument over these many years. And that framework is continually structured in ideological terms: theism versus atheism. In this ideological battle the so-called agnostic is free and alone in claiming objectivity. You reflect this sentiment when you say that the "new atheism" is "turning into a religion itself." This is all sophistry at its most pervasive. Why?

    Because an atheist is nothing more than someone who has examined the world's religions and found their doctrines to be equally specious. No one need embrace a particular dogma to discount those dead gods of mythology, who were once as real to their adherents as the sun - likewise for those gods' variants, which persist to this day. The only difference between the "New Atheists" and their forerunners is their stridency in highlighting the absurdities of theism. And even that is not unique to them; Bertrand Russell and Robert Ingersoll spring to mind as less recent examples of the great skeptical tradition.

    And make no mistake, it is quite a tradition. Because as long as humans have pondered their origins and the possibility of some prime mover, there have been those who arrogate knowledge of such a being, those who claim to be privy to its intentions. And for just as long - though not in equal numbers - there have been those who call them frauds and charlatans for claiming to know things they cannot possibly know.

    Frankly, all these metaphysics and cosmological arguments, like the one you touch upon, make my head hurt. How did the universe arise? I don't know (<--agnosticism). But I do know that invoking God is not - and never, ever has been - a real answer (<--atheism). It simply moves the question of creation up one step along an infinite regress. It amounts to special pleading: "Let's stop here - at this suspiciously anthropomorphic point - along our metaphysical path of speculation!" I also know that the above question is one for physicists and cosmologists to consider - not theologians. And if you're interested in seeing just how far their considerations have progressed, you should check out Lawrence Krauss's lecture on modern cosmology, which you can find on Youtube. Of course, all of this is perfectly irrelevant to theism. Even if you could establish that some intelligence created the universe (which you can't), you still must traverse nearly 14 billion light-years of space - and as many years in time - before you can relate that intelligence and its designs to our present condition. Good luck.

    You are free to call yourself an agnostic if indeed you rate the probability of (G/g)od(s) at 50%. But in that case I must respectfully disagree with your estimation and call myself an anthropogenic creationist: religion is man-made, as is the concept of a god.

    Whew… sorry for the rant. What were we talking about? Oh yes, ancestor simulations. I sometimes like to think that we do exist in such a simulation, one in which our more reasonable progeny tweaked the superstition knob and watched how things played out. I’m sure they’d be amused.

  16. Ok, wow, that was really ranty. Let me just clarify my main points in the order I expound on them.

    1. You paint atheists in one corner of an ideological fight about religion. This is misleading, as atheism is not an ideology. It should not even exist as a word; it is substantively similar to non-words such as "a-Santa Claus-ist" or "a-astrolger."

    2. The "New Atheists" are nothing new; their sentiments have been voiced before, and by many more than we recognize in history.

    3. The metaphysical claims of modern theism - even those as abstract as the origin of the universe - should not be taken seriously.

    4. We are superstitious.

  17. Bostrom's formula is incorrect, he confuses the average number of simulated individuals with the average number of non-simulated individuals in real civilizations. The number 1 in the denominator of his equation (*) should be replaced with the average number of real individuals in civilizations (whether or not they are post-human) divided by the average number of simulated individuals in a simulation.

    His first proposition is stated incorrectly. It should be, 'Civilizations are extremely unlikely to develop technology whereby ancestor simulations are possible'. I don't know why he limits that statement to extinctions.

    His third proposition, 'It is extremely likely we are in a simulation', is also incorrect. His formula counts/estimates the fraction of simulated individuals from the perspective of an observer in the current universe, it has no bearing on whether that observer is simulated (because he includes himself as a real person in the formula), only on the probability that future individuals are simulated. To estimate the probability that the observer is simulated, the observer would have to be able to count/estimate the fraction of simulated individuals in the entire multiverse, but this is of course impossible.

  18. Personally, I do check for batteries.

  19. A question I have, is there any real difference between us being a simulation or created by a deity. And taken further, say we are a simulation, what would really be the difference between the person running the simulation and a deity?