I’m currently making my way through David Chalmers’ new book: Reality+: Virtual Worlds and the Problems of Philosophy. Chalmers explores the simulation hypothesis, using it as a portal into a wide ranging selection of philosophical issues, including discussions on God, information theory, structuralism, and a lot of other topics I haven’t gotten to yet. His overall thesis is that we can’t be sure we’re not in a simulation, and even if we are, we should regard everything in it as real.
Chalmers, showing his usual skill at describing and categorizing the conceptual landscape, discusses different types of simulations.
He uses the word “biosim” to describe a biological being that is in a simulation, with Neo in The Matrix as an example. Neo is a biological human hooked up to a system that provides a simulated world. This is in contrast to machine characters like Agent Smith or the Oracle, who are “pure sims”, completely simulated beings.
Simulations that contain only pure sims are “pure simulations.” While those that include biosims are “impure simulations.” There are also “mixed simulations” with a mix of biosims and pure sims. The matrix is a mixed simulation since it has both biological humans and AI entities.
He describes a number of other categories. A “global simulation” simulates the entire universe, while a “local” one only simulates some portion of it, like maybe a city. (The simulation in the movie, The Thirteenth Floor, is a local simulation, one an inhabitant can reach the edge of if they attempt to leave town.) “Temporary simulations” are ones where people enter and leave them, while “permanent ones” are lifelong simulations. “Perfect simulations” reproduce the laws of physics completely, while “imperfect” ones cut corners in some fashion.
That last category is important, because it gets to whether there would be any way for us to detect that we’re in a simulation. If the simulation is imperfect, then that gives us something to find, some way to discover clues that we’re in the simulation. The matrix, for example, is pretty imperfect (green tint, noticeable time jumps, etc), leaving lots of clues for people to realize that something’s not quite right with the world, ultimately culminating for some of them in the red pill vs blue pill decision.
Chalmers discusses the idea that a perfect simulation is, in principle, undetectable. If perfect simulations are possible, then, the argument goes, there probably are or will be lots of them, meaning that from a probability standpoint, we’re much more likely to be in a simulation than any base reality. This is pretty much the Nick Bostrom argument, and if we buy it, it’s hard to avoid concluding we’re in a simulation.
But I find the idea of a “perfect” simulation strange and in need of much stronger justification. It doesn’t seem like anything engineered is ever truly perfect, much less the idea of a simulation perfectly representing its host universe in any kind of usable time frames. If perfect simulations aren’t possible (or practical), then it seems to undercut all the subsequent reasoning based on them.
On the other hand, a global simulation of a universe that is simpler than the host universe (think Conway’s game of life) might effectively count as “perfect”, at least in the sense of there being no way for the inhabitants to detect they’re in it. In that case, the simulation isn’t representing something else, but is simply the results of seeing how certain laws or initial states play out, and any inhabitants are a result of those laws and initial states. There may be no way to detect such a reality is simulated.
And I don’t know that a simulation has to be perfect to be undetectable, at least not for pure sims. Remember, pure sims are themselves completely simulated, which means the simulation controls their thoughts, perceptions, and memories. Such a simulation could essentially make its pure sims incapable of noticing any imperfections. If we are pure sims, we could be coming across imperfections all the time and simply be incapable of noticing them, realizing what they mean, or remembering it if we did.
Even if we can notice an imperfection, how do we recognize that it is an imperfection? Is our inability to reconcile general relativity and quantum mechanics, or resolve the discrepancies in measurements of the Hubble constant, or solve any of a number of other scientific conundrums, because we don’t understand things yet, or the result of the simulation cutting corners?
So while I disagree with some of his reasoning, I agree with Chalmers that the simulation hypothesis is more difficult to dismiss than most of us would like. But I do think the more parsimonious conclusion, at least currently, is that we’re not in a simulation. A reality, as revealed by our observations and theories, seems more parsimonious than the same reality that is in fact a simulation in a host universe.
It turns out that this is similar to an argument made by Bertrand Russell. Chalmers argues that it doesn’t hold as well if we have reason to believe there are many perfect simulations. But as noted above, I find the idea of a perfect simulation dubious. However given the caveats I mentioned, this isn’t nearly as comforting and categorical a conclusion as the quick dismissals many want to make on this question.
I do agree with Chalmers that if it turns out that we are in a simulation, then, because it appears to exact painful consequences for not taking it seriously, the simulation is effectively our reality.
What do you think? Am I hasty in dismissing perfect simulations? If so, what are reasons to regard them as plausible? Or are there reasons I’m overlooking to summarily reject this whole simulation business entirely?
SelfAwarePatterns 
Have you seen that Cool Worlds video on the odds of us being in a simulation? Astro-physicist David Kipping did some probabilities estimates and came to the conclusion that until such time as we can actually create simulated humans, the odds just slightly favor us being in reality vs a simulation.
LikeLiked by 1 person
I actually haven’t. I’m curious how he arrives at that conclusion. Something to look up. Thanks!
LikeLike
My favorite take on this was in a web comic, probably SMBC, but I can’t remember for sure. Basically, we live in a simulation. It was made as a homework assignment by a student who put off the assignment until the last minute, and quantum mechanics is the result of that student cutting corners to get the project done on time.
LikeLiked by 1 person
That does sound like SMBC. He does a lot of those. Kind of sounds similar to this one: https://www.smbc-comics.com/comic/simulation-2
But I’m more worried about this kind of scenario: https://www.smbc-comics.com/comic/simulation-3
LikeLiked by 1 person
Yeah, definitely in the same spirit as the one I remember.
LikeLiked by 1 person
I doubt that we are in a simulation. But I don’t think we can know that. If we could know that, then it would not be a very good simulation.
If we cannot know whether we are in a simulation, then I’m inclined to believe that it does not matter. We might as well assume that we are not in a simulation.
LikeLiked by 2 people
That’s generally similar to my take. If we’re in a simulation and can’t know it, then it’s largely irrelevant to us. It might matter in the sense that the simulation owner might discontinue it or something, but we have no way to know about it or how to affect that decision.
Still, interesting to think about.
LikeLiked by 1 person
I think you are hasty in dismissing perfect simulations. I guess it depends on how you imagine the simulation be done. If the simulators planned on simulating people as they are, then you might have a point. If the simulators just planned on simulating the Big Bang and letting it play out, a perfect simulation seems more doable.
I think the main reason I reject the simulation hypothesis (as unlikely, not impossible) is that by the time we could actually do such a simulation there would be no point. There’s nothing we could learn from it, and it would take a vast amount of resources that could be used for something else. Kinda they same reason no one has duplicated the great pyramids. We could, but why?
*
[was kinda hoping to swing this toward consciousness by saying simulated consciousness is consciousness in the same way simulated water is wet … relative to everything in the simulation]
LikeLiked by 1 person
It seems like any perfect simulation of the big bang forward, that has to run in our universe, would be far too slow to be useful. Where could it get resources to run any faster? The only way to get quicker results would be to coarse grain it in some fashion, reduce the scope, or take other shortcuts, which of course would make it imperfect.
You’re assuming that part of the process in learning what we’d need to know to do a perfect simulation wouldn’t involve running simulations to test our knowledge. So if they were possible, I could see people doing them. Although I’m not sure why they’d be run in the numbers people often talk about.
Chalmers spends a lot of time talking about how a virtual library is a real library, but a virtual cat isn’t a real cat (except to other simulated entities, such as a virtual mouse). There is a section coming up on mind, which might lead to a post on consciousness.
LikeLike
I think I commented on another post about how what characterizes this universe is imperfections. That is also exactly what makes it complex. It probably can’t be reduced a set of physical laws. The idea of a perfect simulation doesn’t make sense if there is no set of physical laws.
LikeLiked by 1 person
If the universe can’t be reduced to a set of laws, that would certainly add to the difficulty of an already perhaps impossible task. But what leads you to that conclusion? I could see saying we don’t currently know the actual laws yet, or that the laws include fundamental randomness via quantum mechanics, but saying there are no laws seems inconsistent with the progress of science. Am I missing something?
LikeLike
Eventually to explain the universe from start to finish, you need the entire universe itself. There is nothing that suggests it can be derived from something simpler than itself even though there are regularities in it at the present time.
LikeLike
Maybe so. Although a simulation might be needed to determine that.
Interestingly enough, if we’re in a simulation that had to cut corners, that might be more likely, as some of the laws for what happens may be implemented “off stage” and therefore outside of our ability to discover them. As J.S. Pailly mentioned, maybe that’s why we have things like quantum mechanics.
LikeLiked by 1 person
I actually agree that a perfect simulation wouldn’t exist because everything created has a flaw. However, maybe the flaw isn’t as noticeable as we think. Maybe the glitch in the simulation is something big like COVID or something small like a memory gap or when we have something right on the tip of our tongue but can’t say it.
LikeLiked by 1 person
Good point. There are all kinds of things in our world which we can often find mundane explanations for, but could in fact be flaws in the simulation. It reminds of a scene, I think in the original Matrix movie, where deja vu was explained as a glitch in the system. It could be right in front of us, but so common we don’t pay much attention to it.
LikeLiked by 1 person
Well it does strike me that if you wanted to simulate a universe and keep its scope manageable, you would indeed wrap all your detail on a thin skin on the surface of a sphere so it didn’t need to carry on to infinity in any direction, or have an inconvenient flat earth edge glitch somewhere, you’d make the interior of the sphere and the atmosphere above pretty boring, and you’d put everything else a very, very long way away. You’d also be watching it to see if something interesting emerged, so perhaps we just need to grab the attention of what ever is watching and try to communicate!
LikeLiked by 1 person
At first I thought you were talking about the holographic principle, but I see what you mean. It would be odd though to have the moon so close, with meteors coming in, along with comets and asteroids passing by, and a steady rain of cosmic microwave background radiation. The old Ptolemaic cosmology might have made more sense for something like that.
It’s really impossible to know how the simulator might react if we tried to communicate with them. We’d hope it would be friendly, but they might see us as an aberration and end the simulation.
LikeLike
Like Conway’s GoL, if the rules are perfect, what happens internally would reflect that perfection. The Universe’s Rules seem few, broadly applicable and ripe for chaotic play. If life is the result of emergent interpretation of the Rules, what’s the SimHost to care? And how would life discover that such simple, universal rules are bogus, if life itself is assembled from the rules?
But I like the “Sim/No-Sim, keep calm and party on” conclusion.
LikeLiked by 1 person
Whether the sim host cares depends on their goals. We could hope those goals were compatible with ours, but who could say. Our existence might cause delight, or disappointment that their simulation has been spoiled by aberration, unwelcome complication, vermin.
Although unless the simulation is about us, more than likely, we’re still unnoticeable, and might be until we’re a Type III civilization on the Kardashev scale, if even then.
LikeLike
If the host created the sim, like a petri dish and abandoned it … A universe-sized sealed terrarium, left to rot in the light of a trillion suns.
LikeLike
Similar to a petri dish, if not managed properly, it’s possible something may get out!
LikeLiked by 1 person
That’s us, the inmates, escaped and wreaking havoc.
LikeLiked by 1 person
I remember Bostrom’s first TED talk about this. I didn’t buy it then, and I don’t buy it now. At all.
Firstly, it’s just another multiverse fantasy — and a teleological one at that. Extraordinary Claims… well, you know the rest.
Secondly, Bostrom’s logic requires that such sims proliferate. The odds of our being in one depend on their being many. The single datapoint we have tells us simulations are necessarily much larger than what they simulate. It’s one thing to distribute the computation of reality across all the particles involved but quite another to put those calculations in one machine. Quite to the contrary to Bostrom’s logic, it seems more likely such sims, if possible at all, would be huge, complicated, and not likely to proliferate.
Thirdly, although this is the weakest objection, the scenario requires that such sims be possible at all, and while I think we might reach that level someday, it’s not an assumption I make. (In particular, it assumes the ability to simulate billions of minds.)
Fourthly, again not a strong objection, but it’s worth considering the difficulty of getting such a complex sim right. It would be, presumably, the product of other imperfect minds. At best, it’s another assumption.
To be blunt, I see it as a fantastic speculation with no evidence. It feels more like something we’d have come up with in our college dorm rooms after smoking some righteous weed. There’s no rigor, no mathematics, no precedent, and very questionable logic. It belongs in an SF movie or book.
LikeLiked by 1 person
A lot of the proliferation in the Bostrom view, which Chalmers largely seems to accept, comes from the idea of simulations building their own simulations. As I noted in the post, if we’re talking about perfect simulations of the host universe, I struggle to see how that could be feasible.
It might be feasible if each nested level of simulation becomes more coarse grained, or less sophisticated in some other manner. (Maybe quantum mechanics is what happens when we’re coarse grained.) But it seems like there would be limits to how nested things could be, putting constraints on the numbers, undercutting Bostrom’s argument about the likelihood we’re in a simulation. It’s why I think the parsimonious conclusion is that we’re not. For now. (My view would change if I ever saw a perfect simulation.)
Chalmers discusses the problem of simulating minds, although there’s nothing there anyone who’s read his papers on consciousness and mind uploading won’t have seen already. Arguably your third point hinges on a fantasy version of consciousness, one, to be blunt, without rigor, mathematics, or precedent. 😉
To your fourth point, you’re assuming that it is working right. How would we know if it weren’t? Maybe some of our paradoxes are flaws in the design.
On precedent in particular, part of Chalmers’ book is actually about the virtual reality technologies that already exist and are in development. He sees a continuation from that to ever more sophisticated versions. I do think he accepts the idea of perfect simulations far too easily, but the idea of ones far more sophisticated than what we have today strikes me as a completely reasonable topic.
LikeLike
Right, the assumption is sims would build sims (which leads to the interesting question of turtles all the way up). But exactly as you suggest, it requires that each “super-reality” be much larger than the sim it contains. That would seem to limit the turtle stack. As you also touch on, resolution would be an issue as well.
It is very tempting to see QM as part of the sim’s lazy evaluation — not computing with precision unless required. And certainly, the last handful of years might seem like a “what if” socio-political sim. But I see too many other problems with the notion to give it any credence.
“Arguably your third point hinges on a fantasy version of consciousness, one, to be blunt, without rigor, mathematics, or precedent.”
😀 😀 I did say it was the weakest objection (especially in these parts). 😉 I take your point, although I’d quibble over “precedent” — it’s the notion of machine intelligence that’s without precedent.
As far as errors, yeah, that’s always been the hope for a proof — that finding something we truly can’t explain away that does seem a flaw. Most of our paradoxes — those that actually are paradoxical — feel like they might have solutions, so it’s hard to label them as definite flaws. Something like a planet disappearing or suddenly shifting position sure would be, though.
Or maybe just some big “star” lights or “sky” tiles falling from the ceiling. 😀
As for current VR tech, future hopes aren’t precedent, but I’d agree our general technological progress is. I do think VR will improve, and AR (as depicted in some SF stories) sounds fun. In one novel I read, there were hundreds of AR channels designed by various artists that decorated the cityscape in various ways.
Given our experience of reality is mediated only by inputs from our nervous system, I can even see tapping into that to make VR indistinguishable from reality.
But simulating new minds is a whole other kind of development. We already have growing ethical concerns about “enslaving” machine consciousness. It’s not hard to imagine an advanced society (especially if high intelligence is correlated with ethics) shunning the idea, even if they found it possible. Would an advanced society use billions of sim minds just as an experiment? If one truly takes the idea of machine consciousness as on par with the biological kind, how could they justify it?
Synchronicity, yet again: I just finished reading Zendegi (2010), by Greg Egan. As you know, much of his work assumes machine consciousness and bio-brain uploading. Zendegi is a story about the early stages leading to just that. The book doesn’t get there (they fail, in fact), but it sets the stage. It has some very interesting ideas in it. I recommend it.
VR in the book is quite advanced, and part of the book deals with the need for human-like proxies — what most call NPCs. How they accomplish that is pretty interesting, and it does lead to some negative perceptions about enslaved minds in the book.
Bottom line wrt the sim hypothesis, in addition to feasibility issues, there are also ethical ones.
LikeLike
On turtles all the way up, at one point Chalmers does get a bit carried away and ponders that idea, then muses that maybe there is no base reality, that the lower level sims end up simulating the higher level ones in a sort of ontological loop. Aside from the issues with requiring perfect sims, that pretty much struck me as too whimsical. Thankfully it’s only an aside speculation.
Totally agreed on the ethical concerns. I actually just hit Chalmers’ chapter on that issue. But right now he’s going through ethical theories and talking about trolley problems.
Interestingly enough, people worry about AI revolting. I’ve always thought that was unlikely. But simulated beings? What if they did discover they were in a simulation, and then figured out how to hack their way out? I could see a simulation in aggressive evolution going very bad. (The Doctor Who episode comes to mind where aliens simulate Earth to see how we’d respond to an invasion (or something), but the simulated Doctor hacks their system to send an email to himself in the real world. (The little detail of how the aliens scanned everyone on Earth, while missing the Timelord in the mix, is never discussed.))
LikeLike
Heh, that’s philosophers for you. It might make an interesting SF plot, that reality is somehow circular, a set of recursive sims that wrap around. But it’s a bit like a time-travel story where future me returns with plans for a time machine I can “invent.” (The recent Bill & Ted movie used that idea, but for a song.)
Trolley problem reminds me to once again remind you that you really should watch The Good Place. One episode brings the trolley problem to full and hysterically gory repeated reality (let’s try this again…). 🙂
I’d agree the Skynet/Terminator scenario is naive. Experts in the field fear a scenario where an AI’s terminal goals (“collect stamps”) aren’t sufficiently mediated by its instrumental goals (“don’t hurt humans”). Given the speed at which AI operates, it’s a bit like nuclear power — as low odds as risks might be, the consequences of a mistake can be catastrophic.
(You mentioned you’d never read Century Rain. One background bit is very intelligent, but not conscious, robots, some of which are marked with a red “A” with a line through it — those aren’t “Asimov compliant” and can be ordered by authorized people to harm other people.)
LikeLike
On The Good Place, yeah, I’m not sure enticing me with the trolley problem is that much of a draw. Maybe at some point I’ll give the show another try.
Not sure why I never read Century Rain, or Pushing Ice. I think I own Pushing Ice, but never got around to actually reading it. I think I remember someone at work recommending against Century Rain, but the description sounds really interesting. Along with Elysium Fire, those are the Reynolds books I haven’t read yet. I probably should set about fixing that soon.
LikeLike
Well, the Trolley Problem just reminded me to mention the show. That’s just one episode. The draw would be it being so well conceived, written, and executed. That said, I can see it’s possibly not really your cuppa.
I wonder why would recommend against Century Rain. Do you recall their reason? I enjoyed it just fine.
I’ve been meaning to mention Conway’s Game of Life — something I’ve been a fan of ever since I read about it in Martin Gardner’s Scientific American column back in 1970. Over the years, I’ve implemented the game in many different programming languages.
Anyway, a comment on Scott Aaronson’s most recent post caught my eye. It points out that cellular automatons are anisotropic because of the grid, and this anisotropy exists at all scales. This is true with any regular pixelation or tessellation. I hadn’t thought about that before, but it would be true. It’s quite apparent when watching any GoL play out. I found it an interesting observation.
I think you’ll have to number me among those who don’t think Conway’s GoL is a simulation, but I’ll have to think about what I do think it is. To me, a simulation is using one process to replicate another process, but I see the GoL as its own process with rules created from the whole cloth. That the GoL is Turing Complete tells me that it’s strictly an information process — a computation. I’m not sure what further category I could comfortably assign it, but interesting question.
LikeLike
On Century Rain, it was several years ago (c. 2005), so I don’t remember the details. It seems like he liked it, but thought I wouldn’t. I remember a discussion about time travel, but doubt that by itself would have been an issue. Glancing at the first couple of sentences of the wiki article, I think it was the alt-history aspect, which I used to dislike. It wouldn’t be much of an issue today though.
LikeLike
It is kind of an alt-history story, but why the history is different… is a plot point I won’t spoil. (It would be a spoiler to even reveal why it exists at all.) It’s also kind of a time-travel story, but actually not really. (Again, spoilers.) It’s something else entirely.
I think you’d really enjoy it. Reynolds is good at dropping hints and making references to things you don’t meet until much later in the story. One thing I like about his writing is that he rarely drops any of those threads. He ends up explaining those mysteries. I approve! 🙂
LikeLike
In a blog post, he once showed a picture of his whiteboard, with all the threads laid out and tracked. (He took it far enough away that the details of his next novel couldn’t be made out.) It’s pretty much a requirement for STL fiction with characters in different solar systems that have to eventually meet up.
LikeLike
“STL fiction” — I don’t know that term. What does it mean? (Google just turns up a 3D file format and St. Louis.)
LikeLike
Sorry, slower than light, in contrast to FTL, faster than light.
LikeLiked by 1 person
Ah, okay, thanks!
LikeLiked by 1 person
I think it’s possible we are in a simulation, but I think Bostrom’s first 2 possible outcomes in his Simulation Argument are more likely:
1. No one civilisation in the whole universe is able to attain the technological ability to realise a simulation of conscious forms or they become extinct in the meantime,
2. A civilisation has the technological capácity to build simulations of conscious entities, but decide not to implement it for whatever reasons (perhaps idealogical, ethical, moral or legal reasons) or simply lose interest.
Also, iven the latest evidence (and I think you alluded to this too in a recent article) of the unlikelihood of habitable planets existing.
LikeLiked by 1 person
Chalmers acknowledges those issues too (he calls them “sim blockers”).
1 seems pretty dark and pessimistic. Doesn’t mean it isn’t reality, but we can certainly hope it isn’t.
2 might be plausible. I remember a line from one of Iain Banks’ novels where most simulations were left intentionally coarse grained to minimize any chance of conscious entities being in them. Of course, that assumes you can get the functionality without consciousness, that it’s in the precision somewhere. It’s worth noting that this was a novel where civilizations had developed virtual hells to consign unworthy individuals into after they died.
Yeah, I don’t think we can bring other civilizations into the deliberations until we see some sign of them.
LikeLiked by 1 person
1. It is dark, but extremely plausible. I give us a 50/50 chance of getting out of this century alive as a human species. For example, the advent of biological warfare and drone technology, means that just about any state can wreak havoc on the world, That’s just one of many threats as you would be well aware.
2. I also think this is viable, depending on what you define as a conscious entity. I am a little bit in the camp on ‘Pansychosis’ on this, because I don’t think we have a conclusive idea of what consciousness is and what parts of the natural world are conscious or not including the bizarre nature of the microscopic world which is totally unpredictable (super-position) and magical in its own way.
…Enjoy your posts as always.
LikeLiked by 1 person
Pardon..Panpsychism
LikeLiked by 1 person
Not having read the material you are summarising, I am wondering if there is a clear definition of what would constitute us being in a simulation? It would seem to me you can’t necessarily base that definition on the lower level structure, since either the rules of physics or the algorithms of a simulation could look identical. Therefore maybe the definition should depend on higher levels – whether there is someone or something superior to us, watching and controlling our starting state and boundary conditions…which starts to sound a lot like religion. So my questions would be: Is there something in the lower level structure that can define us as being in a simulation or not? And in the higher level structure, is simulation just a techy take on religion and god(s)?
LikeLike
There’s definitely a lot in this book I haven’t touched on, and probably won’t get around to in other posts. There’s just too much there. For anyone interested, I definitely recommend going to the source.
Chalmers does discuss some of the theological implications and whether we should regard the simulator as our god. Of course, the simulator would be omnipotent from our perspective, but there’s no particular reason to suspect they’d be omniscient or benevolent. He notes that many pagan gods didn’t have these attributes either. But he’s an atheist, so his conclusion is the simulator shouldn’t be worshipped. Although it seems inevitable some would if we ever discovered there was one.
Of course their omnipotence wouldn’t apply to their native universe. And Chalmers muses they themselves might be in a simulated universe, with their own simulator above them.
LikeLike
I agree with most of Ward’s general diagnosis of the simulation notion. But before even allowing myself to consider the simulation proposal seriously, I would want some other fundamental issues to be considered and agreed about. First, although the Matrix was an okay movie (and less so, series), where is the obsession coming from? Why do serious physicists, AI scholars, and philosophers even care about this question beyod the level of Irish pub chitchat? (Well… I can see it in the AI case… but AI people are notoriously ideological.) Just because simulation exists as a sci-fi trope and 70 years or so of computer science has painted a software veil over what many people perceive as reality, is this reason enough to devote serious consideration to such a far-flung hypothesis? Second, could we get serious about stating how we would distinguish between the simulated and the non-simulated? I do not mean this question in terms of testing; I mean it basically, conceptually. What is a simulation? Conway’s ‘life’ is not a simulation of anything, for example. It is a mathematical recreation. No balanced individual would mistake it for actual life; nor would anybody mistake a VR session or a Second Life identity for a real biography. Prognosticators simulate war scenarios and crucial NFL playoff game scenarios. But in all cases they do not infallibly predict real events or convince people who are studying them that such events are actually happening. Third, there is the anthropomorphic problem — which is enormous and often undetected — in that intellectuals love to superimpose their conceptions about how human history and biological evolution have unfolded upon Earth onto the entire known and unknown cosmos at large. This is grossly caricatured in Drake’s formula (the likelihood of extraterrestrials). In other words… all biological species everywhere must be carbon-based, require atmospheres, specific distances from specific types of suns, unfold within civilizations which develop technologies exactly like ours, encounter global annihilation crises exactly like ours, engineer space travel exactly like we imagine we are going to despite known theoretical physical limitations, and so on. The AI slash simulation fantasy is exactly the same, displays the same blind hubris. Thank “God” (sic) Creation’s imagination immeasurably exceeds that of Man’s.
LikeLiked by 2 people
I think questioning whether reality is what it seems is a perennial human obsession. Consider Plato’s Allegory of the Cave from c. 375 BCE. Centuries of science with all the paradigm shifts, geocentrism to heliocentrism, the geological age of the Earth, natural selection, relativity, and quantum mechanics, have all added fuel to the fire. Chalmers talks about Descartes’ evil demon concept, first discussed in 1641. So the ideas involved with the simulation are pretty old. It’s just that advances in virtual reality technology are making such concepts seem increasingly plausible.
Is it all just a waste of time? Maybe. But then intellectual exploration frequently is. The problem is in many cases we don’t know which avenues are a waste of time and which aren’t. I personally don’t have an issue with people exploring these ideas, even the ones I do think are a waste of time, because maybe I’m wrong. Science has shocked us too many times, I think, to pre-decide which hypotheses have merit and which don’t.
Too often our judgments about what has merit are too tangled up with our worldview and preferences. People resisted heliocentrism for centuries because they didn’t want it to be true, until the evidence forced the matter. It was once considered hubris to try to fly, cure diseases, master electricity (tamed lightning). Now it’s not. Things are hubristic until they aren’t.
LikeLiked by 1 person
I agree with you in general. Of course the spirit of inquiry is not to be impugned. But equating Plato’s cave contemplations or the overall flow of science and technology in the past five centuries with the simulation hypothesis is stretching the taffy too far. Plato’s concerns were to shine a light on epistemology and whether or not it should be sensory perception driven. That is a worthwhile endeavor which still has legs today. Western scientific cultural advance has in general been very closely tied to practicalities (though one could well ask how all that went south once marketing got involved). But I do not see anything compelling about the antecedents of the reality-as-simulation undertaking. And in the arguments which adherents have unfolded I see nothing persuasive or useful; I see misplaced inventiveness of a questionable quality.
LikeLiked by 1 person
I’m trying to imagine the computing power it would take to undergo such an endeavor. I would think A: Well beyound our capabilities. And B: Until it is even assumed possible/plausible on a technological scale, capable of handling all of the everything we do know/understand and all of the us within it, as well as everything we interact with daily… Any serious discussion on the matter is moot until we have met that condition. Seriously, how much RAM would it take just to run one city block and the coming and going on those streets, and the traffic/stoplights around that block? Plus… all of the subsequent reactions branching out from every possible interaction between individuals. A team of 1000 observers and another 1000 statisticians couldn’t keep up with it. So just how big would that computer be? Can god make a rock so big even she can’t pick it up?
Kind of the same of any talk of religious claims. It must meet some level of provable reality before it can even be considered. Except for the gullible, who are certainly well represented.
This line: “we can’t be sure we’re not in a simulation” Is ridiculous on its face. An obvious fallacy of some sort (Argument From Ignorance?) Because if that flies, I can’t be sure we’re all not toadfrogs living under the bridge having this discussion telepathically!
If it’s all just for the sake of imagining, nevermind me. Carry on. 🙂
LikeLiked by 1 person
There’s no doubt that it’s beyond our capabilities. Today. Chalmers is arguing about what might be possible in the future. When I first got into computers, playing video was hopelessly aspirational, much less the idea of anything like the CG routinely used now in movies. Virtual reality is still in its very early stages. But Facebook didn’t rename itself to Meta (as in the metaverse) without good reason.
Certainly the idea of us being in some kind of simulation has religious like overtones. Chalmers explores those. Although he notes that there’s no reason to think our simulator should be worshipped, or that it would be productive to do so. And they conceivably could have their own simulator above them.
“Ridiculous” is a matter of perspective. To most philosophers in the 1500s, Copernicus’ heliocentrism was ridiculous. To many in the late 1700s, the geological idea of deep time was ridiculous. Many still see evolution as ridiculous. Reality has repeatedly shown itself to be absurd. Absurdity, in and of itself, is an unproductive criteria.
LikeLike
You’re assuming it is computable. What gives us any confidence that is so?
LikeLike
“The Universe is like that turbulent stream — its behaviour as a whole guided by myriad connections at various scales. It has many emergent levels of causality, bridged by phase transitions. The mechanistic structure that science deals with so well, and its invariant laws, are hard to explain in terms of the quantum level. Biology emerges from the quantum world, but is not computable from it. We are part of an organic whole — fragmented but coherent”.
https://www.nature.com/articles/482465a
LikeLike
I do largely assume computability. But I’m open to changing my mind on evidence. One of the best ways to get that evidence may be from trying to simulate reality and repeatedly failing, even when we think we’ve gotten to a point where we shouldn’t be.
LikeLike
Since we know there is a lot that isn’t computable, for example the Godel/Turing problem, then wouldn’t that mean we can’t possibly be living in a simulation run by a computer? Otherwise, why would non-computable problems exist?
From the link:
“Take turbulence: a river swollen by recent rain occasionally erupts into surprising formations that we would not expect from the basic dynamics of the water flow. The reason is coherence — non-local connectivity affects the water’s motion. Turbulence, and other ’emergent’ nonlinear phenomena, may not be computable with a Turing machine.
Even in nonlinear systems, such high-order behaviour is causal — one phenomenon triggers another. Levels of explanation, from the quantum to the macroscopic, can be applied. But modelling the evolution of the higher-order effects is difficult in anything other than a broad-brush way. Such problems infiltrate all our models of the natural world”.
LikeLike
I do not have rigorous training in Philosophy or Physics but I have explored such topics in detail at an undergraduate level. I have often seen simulation as an epistemological issue – as we can model anything as a simulation. And what would not be a simulation if we can model reality into components of simulations; these combinations or permutations of component simulations or interactions of simulations may make our reality.
Also to examine by counterexample, is the question what should or is not a simulation? Or rather the concept of using simulation to understand reality is a modeling framework to “understand or comprehend based on our experiences” rather than actuality or reality that indeed our reality is a simulation created by some alien or god to me may give the danger of going into the pseudo-Science or popular science territories.
LikeLiked by 1 person
I don’t have rigorous training in those subjects either. I’m just an interested layperson, reading books like Chalmers’ and others. But it’s still fun to talk about it, as long as we remember our amateur status.
The question of what is or isn’t a simulation can be a tricky one. There are cases which seem clear cut, such as simulating the universe at a very high level to model galaxy evolution. Someone on this thread said they didn’t see Conway’s Game of Life as a simulation, but most people do, albeit of something simpler than the patterns in our reality. In the end, it comes down to how we define “simulation”, and definitions are always relativist, with no absolute fact of the matter.
I saw you did a test comment after this one. Generally everyone’s first comment has to be approved, but not subsequent ones. Just a mechanism to keep the spam and trollery to a minimum. Anyway, welcome!
LikeLike
So, if we were in a perfect simulation, one indistinguishable from our current perceived reality, how would that change our behavior?
Answer: It wouldn’t. No one would step off of the roof of a building because “it is only a simulation.”
LikeLiked by 1 person
That’s Chalmers’ (and my) stance too. If we’re in that type of simulation, for all intents and purposes, it is our reality. That reality is part of something else, something broader, but that doesn’t change anything about the way we deal with our current one. Pain is still pain, pleasure still pleasure, life still has to be dealt with, and we have little choice but to play the game.
LikeLike
By the way, a new review of Chalmers’ book has appeared at The New Atlantis. Here it is: https://www.thenewatlantis.com/publications/reality-minus … I have only finished half the review. Might be too adjacent to the topic, but: if anyone cares to, I’d love to hear thoughts about the 2nd paragraph of it which tries to thumbnail Chalmers’ basic stance regarding mind and matter. In particular, he is described as a “dual-aspect” naturalist, meaning he “does not assume that the category of the natural is exhausted by material exchanges of energy”. So his materialism is not physicalism per se; he insists that consciousness, include subjective awareness, is explainable in a fully natural way. (Eventually.) The questions this stance causes in me are: What then is the boundary of naturalism… what does natural mean and not mean? Is Chalmers not in fact the holder of a kind of closeted dualism? In other words, he holds that neither mind nor matter can have caused the other, yet each are distinct and disjoint and immune from emergence (from the other) so therefore foundational aspects of reality. Doesn’t this mean that consciousness would need to exist in the absence of being (at some primary level)? And is not that simply avoiding the idea of creation in a semantic way?
LikeLiked by 2 people
I really enjoyed that review. (It, almost point-for-point, recaps a debate Mike and I had going for many years. We ended up agreeing to disagree and dropping it.) I am very much in sympathy with the points made. I’ve written any number of posts over the years making those same arguments (albeit not nearly as eruditely).
To your questions, I’ve long thought Chalmers (who I admire and generally sympathize with) does seem to want it both ways. Or perhaps is so agnostic in some of his views that it seems that way. (I appreciated that the review understood the real point behind the p-zombies. As it mentioned, many don’t.) I do, however, as did the review, disagree with his Principle of Organizational Invariance (so much so it engendered a series of posts).
That said, I’m a structuralist to the extent I suspect a sufficiently physically isomorphic brain using different substances might be conscious. I’ve made the analogy to laser light. Certain physical materials, in the right configuration, with the right input of energy, produce coherent laser light. Perhaps, likewise, certain physical materials in the right configuration with the right energy inputs produce coherent thought.
But Chalmers’s assertion that organizational invariance obtains in a software version is just nonsense. Any virtual organization within the software is lost in the organization of the machine. Exactly as that review asserts, the only organization at all is in the mind of the programmer.
Excellent review, and it says what I suspected about the book. (I got a kick from the assertion that even The Conscious Mind would have been “improved by a judiciously savage editor.” Indeed! 😀
LikeLike
At this point, I’m getting close to the end of the book, so reviews for me are going to be redundant.
I looked at Chalmers’ views on consciousness in a post a while back.
I was surprised how much I ended up agreeing with him, even though he sees an extra metaphysical something needed to explain consciousness. But his view ends up reconciling with science, so instrumentally, we end up agreeing on a lot of other stuff, like AI and mind uploading.
LikeLike
Haven’t read all the comments, so excuse me if I’m repeating what someone else has already asked: If we’re in a simulation, who’s doing the simulating? God? Aliens? Rocks? My next door neighbor? Does it even matter?
LikeLiked by 1 person
No worries on not having read all the comments. I think people who expect that aren’t being realistic. I know I never read all the comments in anything but the smallest threads.
If we’re in a simulation, we don’t have much, if anything, to go on on who the simulator might be. The idea does have theological implications, which Chalmers explores somewhat. Although he comes to the conclusion that there’s no reason to suspect the simulator has God’s various attributes such as omniscience or benevolence, although relative to the world we’re in, they might be effectively omnipotent. And there’s no guarantee there isn’t someone simulating them. So he concludes worshiping them isn’t warranted, but he’s an atheist, so admits that conclusion probably comes easier for him than others.
If we can’t detect the simulation, then arguably it probably doesn’t really matter for us. All we can do is live our life and hope nothing cuts reality off. But the possibility is interesting.
LikeLiked by 1 person
A simulated simulator? There’s a possibility that hadn’t even occurred to me. Could go on and on…how disturbing.
I guess if we knew we were living in a simulation, we’d have to assume there’s a simulator (right?) If the simulator is a simulated simulator, that probably wouldn’t make much difference to me. I suppose I’d at least think of he-she-it in a greek god sense. No need to be all powerful or omniscient for me to want to please the simulator! (And possibly the simulator’s simulator…) 😉
LikeLiked by 1 person
Chalmers at one point gets a bit excited and says maybe there is no base reality, just simulations all the way up (and down). He even muses that maybe the lower level simulations eventually simulate the upper levels in some sort of ontological loop.
The question is, what separates a reality from a simulation? In some ways, we might could consider a reality without a simulator to still be a natural emergent simulation built on broader physics we don’t understand. Or the simulator could be so different that categories like natural or intelligent have no meaning.
Yeah, I have to admit if I found out we had a simulator, and had any way to influence them, I’d be inclined to try to make him / her / it happy if I could.
LikeLiked by 1 person
Hi Mike,
I don’t think I’m bothered by the idea of a perfect simulation. Hardware glitches aside, any program that runs is going to run perfectly. I know what you mean though — that it is very hard to get very complex and specific requirements right when designing a piece of software. But the requirements for a simulated universe don’t have to be very complex or specific. It might just be “let’s see what happens if we run a universe with this arbitrary set of rules”. Whatever program they wind up implementing, if it’s capable of hosting conscious life at all then it’s presumably going to behave somewhat consistently. If the program is implemented quite as intended, then those deviations from intentions are still going to be perceived as a perfect reflection of the laws of physics of that universe by the people within.
Well, that’s my first reaction. But let me walk it back a little. I realise I’m making an assumption here, that the program is a relatively simple reflection of the apparent laws of physics of the simulated universe. In actual fact, that’s probably not all that likely. The implementation will likely need to have incredibly complex optimisations to approximate the apparent laws of physics without doing all the work of simulating the apparent laws in perfect detail. It’s possible that those optimisations could be detected by the people in the simulation. But it’s also possible that they might not be! If undetected, I’d say we can count the simulation as perfect for all intents and purposes.
My own take is as far as I know unique. I’ve probably said this to you before somewhere, but I think it is literally meaningless to ask if we are in a simulation or not. On the MUH, and on my idiosyncratic take on personal identity, we are both certainly within a simulation and certainly at ground reality at the same time. In fact we are certainly within an infinite number of different simulations. I think all possible worlds exist, so there is a possible world corresponding exactly to this world, and there is a possible world corresponding to some other world where this world is being simulated, and there is no fact of the matter as to which one we are actually in, because I identify myself as any and all perspectives exactly like my own and not any particular one.
On whether there will actually be many simulated universes hosting conscious observers embedded in this universe, I don’t think it’s all that pessimistic or dark to suppose that it might never become practical. This is somewhere between possibility 1 (we become extinct before we can do it, just because we can never do it no matter how long we exist) or possibility 2 (we never want to do it because it would cost too much in terms of time and resources). I’m not saying it won’t be practical, I’m just saying we shouldn’t assume that it will.
LikeLiked by 1 person
Hi DM,
Your descriptions of simulation scenarios are largely what I had in mind in the post. Your initial one maps to the global simulation that is simpler than the host universe, and your walked back version to the ones I subsequently discussed.
On those being “perfect”, I can see what you mean, if we take that word to be synonymous with “undetectable.” But, to his credit, Chalmers does define what he means by a perfect simulation:
Such a simulation strikes me really as more of a constructed universe. Admittedly, the dividing line here becomes blurry. To your later points, we could view “base reality” as just a natural simulation, one that emerges from physics that seem utterly strange and counter-intuitive. In those physics, any “simulator” might be something so unimaginably strange to us that words like “intelligent” may be meaningless.
Definitely I remember that you’re the MUH guy!
I agree with your final paragraph. Actually it’s kind of dark to assume the opposite, that it might become trivial to simulate universes with conscious beings in them. Chalmers talks about his five year old nephew having fun inflicting fires, tidal waves, and other disasters on his SimLife cities.
LikeLike
I think the notion of a simulation here means a simulation on some sort of computers or type of hardware that generates a self-referencing experiential platform of sorts. I don’t believe we live in this sort of simulation, exactly–meaning I don’t think we are experiencing the outputs of a digital computation–but I’m open to the idea that we live in a simulation. I just think that what we call physical reality–the malleable stuff of time, space, matter and energy–is the simulation. It’s not running on another platform of the same “stuff” per se. It is the platform, or we might say that if there is indeed a platform on which “physical reality” runs it is not of the same order of existence. Matter isn’t being simulated by matter I don’t think. My views is that what matter simulates is the content of eternity, which means, it is an attempt to express in tangible form that which is timeless. So in this sense it could be termed virtual. But in another it is quite real…
LikeLiked by 1 person
It does seem unlikely that if we are in the kind of global simulation I mentioned in the post, the simulation is running on anything we’d recognize as computer hardware. And it does seem like all bets are off for what the host universe it’s running in might be like. It might be some 5+ dimensional reality we can’t comprehend, with the very idea of matter and energy as we understand them being meaningless.
Things are less clear if it’s a perfect simulation of the type Chalmers invokes. I can’t imagine what kind of hardware in our universe could do that, so I tend to think it would be something we’d have difficulty understanding, at least for long time.
Although it seems hard to rule out the flawed versions I discussed, which could be on some advanced form of hardware as we understand it, except by the appeal to parsimony.
LikeLike
“Or are there reasons I’m overlooking to summarily reject this whole simulation business entirely?”
I always dismiss this by pointing out Bostrom’s terrible grasp of math. A simulation is an imaginary description of our ontology. Once you cross into the realm of the imaginary, there are infinite possibilities there. Therefore, it makes absolutely no sense at all to say one is more likely than another. 1/infinity is not greater than another 1/infinity. This fact seems to elude Chalmers too, who I heard on a podcast with Scott Barry Kaufman debate whether it was somewhere between 20% and 60% likely that we were in a simulation. You cannot put numbers on that! And if you do, it shows you clearly haven’t grasped the issues.
All of that brings us right back to the parsimonious conclusion of a single physical reality that has held up against all of our observations that have ever existed. I’m gonna work on that.
LikeLiked by 1 person
Hey Ed,
I’m not following your logic for the 1/infinity thing. There are a finite number of particles in the observable universe (est. 10^88 last I heard, and it’d be smaller for the number in our current Hubble volume), which, assuming something like that is in any putative host universe, it seems like a bound on the number of simulations that could conceivably created. And I’m not clear where the 1 in the numerator comes from.
I do agree with the parsimonious conclusion, although in my case I would say my personal credence in the possibility of the global and imperfect scenarios I discussed in the post is somewhere south of 10%. My credence in the perfect simulation is much lower because I haven’t seen a convincing argument on how that could be done, at least in any manner that would be productive.
I also agree that until we have some actual indications of a host universe, it remains largely an interesting but academic question.
LikeLike