It has been well over a year since I first read Permutation City and relating writings on the internet on Greg Egan's dust theory. It still haunts me. The theory has been discussed tangentially in this community, but I haven't found an article that directly addresses the rationality of Egan's own dismissal of the theory.

In the FAQ, Egan says things like:

I wrote the ending as a way of dramatising[sic] a dissatisfaction I had with the “pure” Dust Theory that I never could (and still haven't) made precise (see Q5): the universe we live in is more coherent than the Dust Theory demands, so there must be something else going on.


I have yet to hear a convincing refutation of it on purely logical grounds...

However, I think the universe we live in provides strong empirical evidence against the “pure” Dust Theory, because it is far too orderly and obeys far simpler and more homogeneous physical laws than it would need to, merely in order to contain observers with an enduring sense of their own existence. If every arrangement of the dust that contained such observers was realised, then there would be billions of times more arrangements in which the observers were surrounded by chaotic events, than arrangements in which there were uniform physical laws.

Isn't this, along with so many other problems, a candidate for our sometime friend the anthropic principle? That is: only in a conscious configuration field which has memories of perceptions of an orderly universe is the dust theory controversial or doubted? In the vastly more numerous conscious configuration fields with memories of perceptions of a chaotic and disorderly universe lacking a rational way to support the observer the dust theory could be accepted a priori or at least be a favored theory.

It is fine to dismiss dust theory because it simply isn't very helpful and because it has no predictions, testable or otherwise. I suppose it is also fine never to question the nature of consciousness as the answers don't seem to lead anywhere helpful either; though the question of it will continue to vex some instances of these configuration states.


102 comments, sorted by Click to highlight new comments since: Today at 8:35 PM
New Comment
Some comments are truncated due to high volume. (⌘F to expand all)Change truncation settings

I'm haunted by Solipsism. We should start a support group.

No, seriously. If LW had a forum, this would be one of its sections.

There's something terribly ironic that I can't put my finger on, about a support group for people haunted by solipsism.

I once received a letter from an eminent logician, Mrs. Christine Ladd Franklin, saying that she was a solipsist, and was surprised that there were no others. Coming from a logician, this surprise surprised me.

-- Bertrand Russell

Tell me about it. But surely I cannot get rid of it all alone by myself, can I?
I was haunted for years, and then developed some kind of immunity. I sometimes think it was a physical immunity; a change in genes as a result of getting older and -- to some extent -- being female. This dust argument is indeed Solipistic, and I find my brain cannot take it seriously.
Females are immune to to solipsism? Explain, please?
That's not what she said. Still, I would also appreciate elaboration of the involvement of being female. I can sort-of see the plausibility in how women were more strongly selected for modeling others' feelings, but I'm not sure that's what byrnema meant.
I don't think that solipsism results from a lack of empathy, but from feeling disconnected from (or emotionally unattached to) physical reality. I'm not qualified to say anything in general about gender, but for me, feeling more connected to physical reality waxes and wanes with my hormones, so I tend to think of a cluster of characteristics (being highly intellectual, detached, etc) as being hormonally regulated in myself. I feel most physically grounded when particular "female" hormones are high, as measured during pregnancy and breast-feeding, etc., and suspect that my immunity to solipsism now is at least partly a result of a steady level of those hormones now. I think that age has even more effect than gender; I feel most grounded now in my thirties, and felt least grounded at the ages 9-12 and in my early twenties.
I have also noticed this correlation between feeling connected to reality, and my predicted estrogen levels. I feel like this feeling is orthogonal to my rejection of soiipsism, though.
A change in genes? Your cells' gene complements don't change*. *exceptions: germline cells [] and immune system cells [\]J_recombination).
No, no... I just meant my gene expression.

Having never read Permutation City, I would find a summary of Dust theory essential to understanding this post, which goal presently eludes me.


The basic idea is this: If the physical world (including consciousness) is just a succession of states, then why would it matter that these states occur sequentially in time and at the same place? The Dust Theory is the idea that it doesn't matter -- that the same collection of bits that describes the universe we experience also describes some ridiculously large number of other universes.

There's more on Egan's site, although it might be hard to follow if you haven't read the book: (edit: linked in TFA, I see now.)

The book is excellent, by the way.

Hey, I actually understood this paragraph. It makes more sense to me than some of the longer, more complicated answers elsewhere. So, thanks.
Zack has written it here for you, but if you'd clicked the links in my first sentence you'd have a fuller explanation than anything I could provide myself. I also didn't want to risk getting a discussion hung on some discrepancy in my account of it so felt it better to refer to source material.
It's basically modal realism for compsci geeks. In a universe without ontologically fundamental mental things, we can't perceive reality directly; our thoughts are implemented in brains. But then, once you concede that thoughts are implemented somehow, it seems to follow that there's no way to tell that you're not actually being implemented in a computer situation. And if you buy some strong Machian [] principles, you start thinking, well, if I can't possibly tell whether or not I'm in a simulation or the "real world," then maybe there is no fact of the matter. Maybe the universe is just pure information, all possible programs, all possible mathematical structures, and I'm an observer that just happens to find itself embedded in the "dust" of random bits. Cf. Tegmark level IV []. ADDENDUM--- Yeah, I don't buy it, either.
I haven't read Permutation City, but that's pretty much my view of reality. I don't think there's a meaningful difference between "the real world" and a perfect simulation of it (at least seen "from the inside") - the same way a chess configuration is the same whether it's played on a computer or on a real chessboard today or on a real chessboard a couple centuries ago. Does the rook agonize about whether he's a real rook or a simulation? Does it mean something to him?
I agree, and say we are in a simulation. I'm not sure what the precise definition of 'simulation' is, but it should be a broad enough concept to include the universe, whatever the universe is. The universe may not be a directed simulation, it may not be a simulation that has a beginning and an end, and even the continuity of it may be a complete illusion. But I cannot imagine how anything at a sufficient level of detail could be interpreted as not a simulation; that is, as something that isn't computed or doesn't run with some mix of mechanical and random rules. In the context of the point of view that everything is a 'simulation', if "actual reality" is fragmented or in any other way fundamentally really, really different from my subjective experience, I don't care. I care about understanding the reality of the simulation I'm in. I only care about any reality outside the simulation to the extent it affects my simulation. And I believe I'm thoroughly justified in this focus of interest. If everything is a simulation, why should simulation B nested within larger simulation A be less "real" than A? I have no evidence that simulation B is fragile or inconsistent, so that I need to be prepared for A tomorrow. If Nemo told me today that this world is The Matrix, I would be very excited at first, but I would also temper my excitement until he answered my question whether any of the rules are different inside the simulation than what we thought. If there's no magic, no way to get the simulation to do something cool, then ultimately it just wouldn't make a difference. Reality is as real as it ever was. A 'less real' reality is only a reality that is inconsistent, that provides evidence of a parent reality that is arbitrarily manipulating the simulation in some way. Thus finally, I would define a simulation as "subjectively unreal" only if the simulation is impossible for the subjects to model without a model of the parent.
Nemo? Are you referring to some kind of Matrix/Finding Nemo fanfic?
Made me think more of Jules Vernes' original Captain Nemo...
Haha! "Neo". Just one letter difference, jeez!
What's the meaning of meaningful? Do you mean that you literally cannot understand the opposite of simulationism? Or are using "meaningful" to mean "empirically confirmable"? The empirical indetectability of a simulation follows from simulations premises, right enough....but it cannot be used to argue for them.
I mean, roughly, that not only are the two empircally indistinguishable, but that I don't even see a reason to care about whether I'm "in a simulation" or not, and it's not even clear what would qualify as a simulation...
Don't you mean "have been indistinguishable up to time T" Simulations support counterfactuals, such as shutdowns. getting out into the real world, etc. If we're given assurances that things you might care about, such as being abruptly halted, aren't going to happen, then you might have nothing further to care about....but it is difficult to see what such assurances would consist of,
Care to share why you don't buy it? I mean, I don't either but I can't think of a rational reason not to. Also we should note that it doesn't even really matter if you are running in a computer simulation; your conscious state could be encoded in anything with enough detail structure to represent your entire state for a particular time slice. Whether there is an encoding for prior or latter time slices isn't even important. Its easiest to work up to this by starting with "mind uploading" and then playing games with the information such as pausing it, rewinding it, removing time slices etc. Thats pretty much the first couple chapters of the novel.

Note that the problem of "Why do I perceive order instead of disorder?" isn't unique to this metaphysical dust theory business. Given that I expect the universe to eventually settle into thermal equilibrium, why do I believe that I'm not a Boltzmann brain temporarily and randomly formed out of that great chaos?

Good point; I don't think there is a lot of difference between the two problem statements. Boltzmann brains may require a somewhat more complex structure, if that problem includes requirements for information processing while time passes. Dust theory doesn't really require such complex structures; just structures that have enough memory and state to perceive a passage of time and a processing of information.

Isn't this, along with so many other problems, a candidate for our sometime friend the anthropic principle? That is: only in a conscious configuration field which has memories of perceptions of an orderly universe is the dust theory controversial or doubted?

The Anthropic Principle conditions on the fact of our existence. But you seem to be conditioning on the fact that we don't accept dust theory. That makes no sense - you could explain absolutely any observation that way!

Maybe that's not what you meant, but I don't see how the A. P. is relevant here.

I suppose I am over-generalizing if not outright misusing the A.P. It seems that a similar principal does apply here though. Egan's rejection of Dust theory is based on the fact that we observe an orderly universe. I'm arguing if we didn't perceive an orderly universe we wouldn't even debate if dust theory was true; maybe observers in a chaotic universe wouldn't consider it a problem at all but a simple and obvious fact that they owe their existence to.
What you say is true. But you can't make the leap from that to "and therefore we cannot reject Dust theory". That's the leap that would require the AP, and you can't use the AP that way. We do observe an ordered universe, and therefore we reject Dust theory. That much is a fact. Things could be different, but they're not, so Egan is right to reject Dust theory. This is the simple, ordinary way in which we accept or reject theories based on observed facts. The AP says we can't do that with certain observed facts, if it would be impossible for us to observe their negation. For instance, we couldn't observe a universe where we didn't exist because intelligent life had not evolved; so we can't use the fact that intelligent life did evolve, to judge theories of the a priori likelihood of such evolution. But we could certainly observe a less-ordered universe, which would give (relatively) more support to Dust theory. So the AP does not apply.
Well, I agree we need to base theory on observable facts. Dust theory is more of a thought experiment or problem related to the nature of consciousness, which is not something we've been able to attack empirically. Nonetheless, Egan dismisses it with this "orderly universe" business. I'm arguing that there is a selection bias - if dust theory were true then only in an "orderly universe" do we argue about dust theory. But I'm starting to come around to the position I think most of you are taking, which is that this is just a "pink unicorn" what-if proposition that isn't worth contemplating - since we do in fact have an orderly universe which seems to account for things the way they are. My problem is that once I accept the information state theory of consciousness, "dust based" conscious entities seems like an inevitable result, unless the actual universe is sufficiently bounded in time and space that the statistical likliehood of this is prohibitive. But for anyone that adopts a many-worlds interpretation of QM or any variant of cosmology theory that yields an unbounded configuration space then I don't see a way around this problem.
That's just the Boltzmann's brain problem, right? It goes away if you assume our universe started with very low entropy for some reason other than pure Boltzmann-style chance. (I believe some theories, like String theories, provide this feature?) It seems to be an open problem in physics/cosmology. I don't understand QM. But intuitively I would expect that even in many-worlds, with an unbounded configuration space, the sum of probability over all configurations is (should be?) constant. Is that so, and would it solve the Boltzmann problem by reducing it to another (unsolved) problem, namely the origin and meaning of the Born probabilities?

Tegmark's level IV multiverse is the only explanation I've ever heard for why there is something rather than nothing. I intuitively lean toward it for that reason. Of course, I don't know how to put a measure on that space that explains my subjective experience, but that seems like a much smaller problem then the most fundamental problem of why anything exists in the first place.

Personally, I'd like to hear alternatives to Tegmark's theory more than I'd like to hear rebuttals.

Tegmark's level IV multiverse is the only explanation I've ever heard for why there is something rather than nothing.

You can tell when something has been explained because it no longer has the same air of mystery that it did at the start. The Level IV hypothesis might very well be true, but it's not an explanation.

Though beware the trap of thinking that things must have one explanation rather than, say, 562 partial explanations, 46 decent explanations, 8 good explanations and 4 truly thorough explanations, all useful and all at different levels of abstraction or organization. Seriously, humans really suck at remembering this when thinking about things removed from their day-to-day experience or where signaling games dominate, like psychology, philosophy, theology, politics, politics, politics, politics...
It's unlikely that there are no few reasons much more powerful than all the rest, that everything is aligned exactly. So if one is tempted to explain using many weak reasons, or any reasons visibly weaker than other known reasons, that would suggest that the weak reasons are fake explanations obscuring what's really going on, even if all of these reasons are true.
That's a correct, non-obvious and useful consideration generally. Though in the situation I had in mind (explaining "akrasia": the lack of a thing) there are many "explanations" that are truly useful explanations---just not uniquely powerful ones. And in (a model of) a complex system each (overlapping, continuous...) level of abstraction has its own set of ways to fail. (There's probably some kind of relevant point exemplified by our back-and-forth here.) "Irrationality", like "akrasia", is another lack-of-thing with enticing "explanations", and "rationality" is a thing.
It's funny that the ideas humans spend most of their time thinking and debating about are implicitly ideas about thinking and debating and yet they never really get around to thinking and debating explicitly about thinking and debating because the thinking and debating is actually about signaling object-level thinking and debating skills and not the less desirable meta-level ones. It wasn't really until the Greeks' development of rhetoric and logic that this trend was slightly reversed; now explicit meta-level reasoning has leaked down somewhat into modern-day implicit object-level reasoning, and explicit meta-level reasoning is a huge financial sector in the form of consultancy. But even so explicit meta-level reasoning is rarely seen in the wild.
I don't have any alternatives. But Tegmark's four levels of multiverse have clarified a suspicion I've held for a while. I think it's a mistake to lean toward this explanation. This technique of positing multiverses to explain phenomena is just way too easy and too general of a strategy. Take any event for which we have not identified a causal regularity which produces it. Take the appearance of comets in the ancient world. The ancients interpreted these as signals from the gods. But it would be just as easy to say "Oh, there must be alternate earths with each which a different possible sky and we see this sky because we happen to be on this earth". "Perhaps there are alternate Egypts where the Nile flows over at different times of the year." Etc. It's just too powerful of a tool. You can explain anything this way. Now Tegmark says the first level multiverse has confirming evidence, I suppose I can take his word for it (though after reading some of his stuff I'm still a little confused about how the evidence could falsify all non-multiverse theories). And certainly nothing I've said can dismiss the possibility of a multiverse at some level. But the fact that any phenomenon for which we have no other explanation can be explained by claiming that all possibilities come to be means that such explanations shouldn't be singled out from that whole set of possible explanations that haven't even been formulated. So you shouldn't lean toward it. The fact that it is the only explanation you've ever heard just means that we are astonishingly far from being able to answer the question. And yes. I think this is a concern for QM too (that isn't an endorsement of Copenhagen, though). Anyway, I'm pretty sure "why is there something rather than nothing" is category error/wrong question territory.
But it doesn't explain why. If all possible worlds necessarily exist, that does explain why this world exists. But why does possibility imply necessity?
The whole point of the mathematical universe hypothesis is that the above is a Wrong Question. Possibility doesn't imply necessity, possibility is necessity.
That statement is just mystical dogma unless you can explain why they are the same thing. And just saying "all possible worlds exist" does not explain why they exist. Why not just some of them? Why any of them at all?
I won't say the truth of the mathematical universe hypothesis is self-evident or anything, but there's certainly nothing mystical about it. All it says is that there was never any reason in the first place to suppose there's a difference between possibility and existence. It's not like we have perfectly clear concepts of 'possibility' and 'existence' that clearly conflict with one another. The best analogy I can think of is the statement that time [] is the same thing as a coherent set of universe-configurations. Why are they the same thing? Um, because that's what time is. What makes you think it's anything else? Likewise for existence: What makes you think it's anything other than possibility?
That's good to know. So the next time someone tells me that something might be a cure for cancer, I won't have to think about it or research it, because if it is possibly a cure for cancer, then it is actually a cure for cancer. And the next possible theory of 9/11 that I hear, I don't need to wonder if it's true, because, being a possibility, it's also an actuality. And I don't need to worry about what you actually mean by your writing, because whatever possible interpretation I come up with, that must be the actual intended meaning... Wait - you're saying that's not what you meant... that all those possibilities which seem to simply not exist... you're saying that they do exist, but somewhere else, in some other world? Well, uh, that's an interesting idea... pretty radical... I don't see what evidence you could have for it... but I guess I have to admit it's possible, ha ha... But wait - isn't it also possible that these other worlds don't exist? So does that possibility - meta-possibility - "exist" somewhere too? Now please note: I am not disputing your right to build byzantine multiverse theories and to engage in abstruse logic-chopping which will allow your new interpretation of the word "possibility" to become consistent. But it is absurd to say that "there was never any reason in the first place to suppose there's a difference between possibility and existence". You may as well say there was never any reason to suppose that the Flying Spaghetti Monster doesn't exist, or that there was never any reason to suppose that black isn't secretly white as well. There are clear differences between the everyday original meanings of possible and actual, and proposing to negate them by supposing that all possible worlds are actual is radical metaphysical innovation, and before I believe it, it is reasonable to request some evidence, or at least an argument in its favor.
This kind of thing "It is possible modal realism is false." is a nice example of a Godel-type statement that doesn't involve math. I'm not particularly familiar with Tegmark's position (beyond looking at the website) but modal realism (which afaict is at least a pretty similar position) just declares "actual" to be a kind of indexical, like "here" or "now". Saying "we actually have a cure for cancer" is like saying "We presently have a cure for cancer." So modal realism (and I don't see why Tegmark IV couldn't do the same) doesn't negate the difference in meaning, rather it just interprets words that we're already confused about. I'm similarly suspicious of the strategy as well (see the reply to Jordan I'm about to write). But I don't think the argument is dedicated to the negation of the difference between possible and actual.
Like what? The property of existence? How would a bit-by-bit description of an object with that property differ from a bit-by-bit description of an object without it? By the way, here's the link to Tegmark's articles, if you haven't read them yet: []
It would have some extra bits saying "this object exists".
Sounds like the Ontological Argument. God exists because he's defined to have the existence bit set to 1. Whether an object exists in some reality is a property of that reality, not of the object.
Okay, I guess it's my fault for not being precise enough. I meant a complete bit-by-bit description of an object, down to the subatomic level, or whatever level is necessary so that the description says everything that could ever be said about the object. Such a description of, say, an apple, would differ from the description of a pear by many bits. But how would it differ from the description of a non-existent apple? You could add an 'existence' bit to the description, but it would be meaningless, because the apple already exists: A complete description of an apple is an apple. A description of a non-existent apple isn't the description of an existent apple plus or minus a few bits, it's not a description at all, it's zero bits.
IAWYC, but let me point out that you are describing your position rather than supporting it. As a way of supporting it, imagine that there's some other universe with physics that encapsulate ours: the Dark Lords of the Matrix can cheaply run a faithful simulation of anything that happens in this universe on that universe's computers. It's clear to me that, given the setup you two are discussing, the Dark Lords would see that those extra bits aren't doing anything at all; they can be removed without altering anything that they could observe in the simulation. Now, anything we can observe about our conscious experience is assumed to have an effect on our brains as we think of it, and thus the Dark Lords could observe it as well. (Namely, if there's a distinction between a "physically real" world and another one that's just "mathematically possible" with an identical copy of you, there's no difference from the Dark Lords' perspective between the description of your brain thinking "But I really exist!" and the description of your copy's brain thinking the same thing.) Note that this is an instance of the GAZP [] in action. So by Occam's Razor, I don't think there's a justification for adding extra bits to the way the universe is described, when literally nobody within or outside the universe can be pointed to as having justification that those bits are one way or another.
You are really mixed up. What you are saying is nonsense and it should be obvious that it is nonsense. Suppose I have an incomplete description of an apple; in the form of words on paper, just to be specific. I have a few thousand words describing some hypothetical apple, its color, its taste, its size, and so on. Now suppose I add however many bazillion more words I need in order to make it a complete description. What are you saying - that at some point my stack of paper turned into an apple, even though it's still a stack of paper? The thing is, you don't need to think like this in order to have a multiverse theory. It's only this peculiar neoplatonic desire to believe that reality is mathematics (or is computation, entirely abstracted from substance) which leads to the nonsense.
Of course not. A description by itself is nearly meaningless. It only becomes meaningful when it's being interpreted. Make the complete description of an apple the input of the right computer program, and the pattern resulting from the sequence of all states of the computation will be the apple. Substance? What's that?
This is still nonsense. If I have a computer made only of naturally occurring atomic elements, and I use it to simulate a plutonium nucleus, are you saying I now actually have a plutonium nucleus there? It's what things are made of.
"Actually have" is vague and should probably be tabooed []. Yes, in that there is a causal structure there isomorphic to a plutonium nucleus; no, in that this causal structure has the wrong relationship to you.
Yes, that's exactly what I'm saying. Of course, the 'simulated' nucleus exists within the system that is the simulation, just as 'real' nuclei exist within the system that is our universe. Thus it would be silly to ask why we can't create simulated uranium and take it 'out' of the computer to generate electricity, or something, because that's exactly like asking why, if our universe exists within another universe, the aliens living in the other universe can't just 'reach out' and take the Sun into their universe. The only way to (sort of) do both of these things would be to convert the pattern/thing we wish to take out of the inner system into the equivalent pattern in the outer system. What would a mathematical description of substance look like? Not a description of a really simple object made of substance; a description of substance itself.
Try as I might, I still find it very difficult to take this seriously as more than a way of speaking about something which in reality is not a plutonium nucleus, but just a simulation of a plutonium nucleus. One reason is that most actual simulations are not complete. If I simulate the global economy on my PC, the simulation is assuredly not complete in the sense of picking out a single possible world or specifying where every atom of every product is located. Most likely it will consist of a few macroeconomic variables evolving over time according to plausible causal relationships. The common-sense understanding of a simulation is that it is not the thing simulated; it is a computational process designed to mimic aspects of the thing simulated. But you ask me to believe that, in those rare cases where the simulation is somehow "complete", the thing itself - the object of simulation - is actually present, albeit "within the system that is the simulation". Every actual computer on Earth is a structure of atoms. Every actual computation ever carried out within these computers is a physical process which has also been given a semantic interpretation, explicitly or implicitly, by human users. I will leave to one side for the moment the question of how to accurately describe what is going on in "biological computers" like the human brain; but I would expect you to at least agree that the previous description is accurate when applied to the technological artefacts we call computers. So there is the common-sense analysis of computers and computation: computers are physical objects, passing through a sequence of physical states which by design admit elaborate semantic interpretations, such as "representing" or "simulating" this or that (and perhaps inaccurately). To insist that these physical states in a computer intrinsically have one true computational interpretation is already highly dubious. The closest a computer comes to having an intrinsic computational pro
Right. The usual dodge is that all possible computational interpretations are simultaneously happening.
If you learned that we were in fact in a simulations, would you feel that you were not real? Would you have an overwhelming desire to escape, in order to become real?
Cogito ergo sum. If I am learning or feeling anything, I do actually exist (though my perceptions and beliefs may be false). It is impossible for me to not be real in the sense of not presently existing, however it is that the alleged simulation works.
Do you agree that it would be possible to feel real inside a simulation (whether the simulation is of one mind or a whole universe)?
"Inside a simulation" means several things. A brain in a vat, plugged into Second Life 2.0, is "inside a simulation". But you want the brain itself to be a simulation too, yes? I certainly don't believe that every simulation of a brain is conscious. A sign with a smiley face on it is a crude "simulation of happiness", but I think we will agree there's no actual happiness around in that situation. There is a continuum of possibilities between a smiley face and a happy person, and at some point on that continuum you get the real thing, but certainly not everywhere on the continuum.
Exactly. In fact, if possibility and existence are the same, the Flying Spaghetti Monster does exist.
The empirical inaccessibility of counterfactual worlds.
I'm working on a post about this. It actually turns out to be surprisingly believable.
Many scientists and rationalists won't offer a "why" alternative because we hit an information boundary at the unique cosmological singularity of the big bang. And most scientists think we should have evidence which we can use to build models that make accurate, verifiable predictions before we claim to understand the "why" of anything.
I understand that objection, but I disagree. We do have at least two pieces of relevant evidence: 1. The universe we find ourselves in seems to be a rather simple mathematical object. The final verdict on this, of course, is still out; but if the bottom were something extremely complicated, it would seem to be an unbelievable coincidence that the approximations we've found (QM and GR) are so tidy and so very very accurate. 2. We can usually think of good reasons that simpler mathematical objects wouldn't contain a great proportion of observers. The particular cosmology of our universe appears to be one of the simplest ways to make evolution of sentient life plausible— for example, QM without gravity results in a dispersive wavefunction, and remotely brain-sized interacting configurations become vanishingly unlikely. I mean, you could build a conscious agent in Conway's Game of Life, but it seems unlikely that any not-too-complicated starting configuration would result in one (well, a chaotic one would result in very occasional Boltzmann brains, but those would be much rarer than conscious life is in our universe, more than enough of a difference to outweigh a reasonable complexity penalty on universes). From these facts, if we expect to be "typical" elements of our reference class (i.e. conscious agents, perhaps with some additional conditions), something like the Level IV Multiverse is strongly supported, and predicts that 1 and 2 will continue to be more strongly validated. (2, in particular, is a pretty decent prediction; it says that we can't come up with a simpler cosmology-leading-to-sustained-evolution than the one we live in).
Well my comment wasn't an objection to Tegmark's mutliverse hypothesis but rather an explanation as to why its the only explanation you've ever heard. But if may object to your objection, I disagree that QM is so very tidy. The standard model has what - 18 free parameters with values assigned as necessary to fit the experimental data? I don't know that anyone considers this tidy, or that many people think particle physics is "done". What we have for particle physics is a useful mathematical model but it isn't an elegant one. The expectation that we should find an elegant model is not unreasonable but it is not yet accomplished.
Yes, but compare that to the number of free parameters implicit in chemistry before QM and QED came along.
Well there is a difference between saying x is more tidy than y and saying x is very very tidy.

I've tried to formalize the Dust argument, though I don't know if it actually succeeds at being deductively valid. FYI, I don't believe the conclusions are true, necessarily, I'm just working through the argument. There is for sure plenty wrong with it.

  1. Subjective experience is found anytime there are mental states where relation K obtains. (K is something like the psychological or memorial continuity relation that gets used in personal identity theory, but I'm not sure I can define it except by appealing to the brain: where m1 and m2 are mental states an

... (read more)
Objection 1: many difficulties (Dust theory being one) are avoided if you simply do not use the term 'subjective experience'. Don't try to define it. Don't assume something exists that should be called that. What is the discussion of 'subjective experience' needed for? What is the problem with discarding the entire concept? (I'm aware there are some problems, but I'm interested in your take on it because I think most of them can be explained away.) Objection 2, to your item 3: the mapping of a 'mental state' to the configuration of some physical system is purely a matter of interpretation. The problem here is that you ascribe to physical configurations, some properties that are normally reserved for causal sequences of physical states, i.e. outright simulations. Suppose I have a model of your 'mentality' - that thing whose states are your mental states. Since it's embodied in a physical system, I can enumerate all possible mental states. Suppose there are countable many states (I don't know physics that well, but at the very least this is valid if you accept arbitrarily precise descriptions of the physical system as mapping to arbitrarily precise specifications of your corresponding mental state). Now, I can write down a few (very large) numbers that map to some mental states of yours. Do you think my act of writing down these numbers literally brings into existence a subjective experience that did not otherwise exists? (You may object because the actual numbers involved are so big they probably can't be literally written down even on a Universe-sized piece of paper. But I can use any physical system, not much more complex than the one I'm modelling (that's you), to encode the numbers. Pen and paper aren't privileged media.) Suppose I find a number, or a series of numbers, that corresponds to a state of extreme suffering on your part. How many real-life resources would you invest to prevent me, an AI, from storing that number in my memory where no-one's loo
It sounds like I should clarify that I don't actually endorse the argument. I'm just trying to make the argument explicit so that we can stop all the hand-waving. I successfully referred to something with the phrase. I know I did because your response wasn't "Huh? What does that word mean?" I'm more than open to the suggestion that subjective experience is a illusion or an error-- but it is the constitutive feature of our existence. Curious people aren't going to just stop talking about it without a very good reason. The burden is on those who don't think it should not be discussed, to explain why. Agreed. This is a good line of attack. Egan's response in the FAQ is: I don't really know where he is coming from. If that is "the point" of Dust theory I don't see how he as made that argument. It looks to me like brains and genuine simulations are indeed causal but arbitrary patterns are not. That said, it isn't obvious to me why causation would be necessary for consciousness. Say we simulate your brain and record the simulation. We then divide the recording into 100,000 pieces, scramble them, and put them back together. Then we play the recording. The Dust theory says that the recording will be conscious just in no way proceeding along the arrow of time the way we are. Is the recording still a causal system? Regarding the rest of the objection: First, Obviously the argument is counter-intuitive. Second, as I understand the argument, your mental intention has nothing to do with it. What matters is that there be multiple structural patterns that relate to one another as sequential mental states do. Thats it. If happiness are suffering are symmetrical it might ever be the case that every time you experience suffering there is another you that is happy and vice versa. You definitely can't just say of some set of patterns "These are Jack suffering" and make them that way. With patterns that we are sure are sufficiently complex that they include a series of mental states
How about "consciousness is computation, and causation is necesary for computation". Yes. But who ever said that any old causation is necesary for consciousness? Both computationalism and phsyicalism say a particular kind of causal process. The causation of playing back a simulation is not the causation of generating it in the first place.
Just because the encoding of the different states are scattered about the universe doesn't mean the conscious experience does not appear to be contiguous and linear to the observer; while they'd be in the minority in an infinite configuration space it is impossible that there won't be states without memories of contiguous experiences. Could either of you explain how you would expect your current state of consciousness with its memories of experiences to be any different from how it is now if it were a dust-mind?
Of course it wouldn't be different at all. But what matters is that my current state of consciousness would be extremely unlikely for a dust mind. This doesn't totally rule out the possibility but it basically puts it in the same category as every other skeptical thesis. And actually it is probably worse than the other skeptical theses since it includes some really weird assumptions about information and causation, as far as I can tell.
It is extremely unlikely, but in an unbounded configuration space it simply has to happen, and to happen many times.
That's also true of things like the Christian Trinity and immortal souls and consciousness and acausal free will. All these words refer to things that are untestable and unobservable, or are described in internally inconsistent ways (logical impossibilities); those of them that could potentially exist, don't exist as a matter of fact; and some of them are just meaningless strings of words. The real referent in these cases is just the sum of everything people tend to say or feel about these supposed concepts. I certainly agree that experience exists - I know I have mine, everyone else says the same about themselves. But if we insist on treating it as purely subjective experience, then we'll never be able to say anything about it, pretty much by definition. In my experience all those curious people are talking about badly-understood notions deriving from beliefs of body-mind duality. No matter how much we learn about objective experience, even if we can manipulate somebody's experience in any way we like, people can still say they don't understand subjective experience. It's easy to think that because we experience things (as a verb), there must be some subjective experience to talk about. But my position is that if we can't formulate a question about subjective experience - a question that will make us behave differently depending on the answer - then there's nothing to talk about. We can go searching for answers, but there's no such thing as searching for a question. Are we supposed to one day think of a question and be enlightened? But that question exists in its own right, and can be answered if we ever think it's important in the way that any question may be important. Meanwhile, if we have some kind of psychological drive to look for The Question, we might as well ignore that drive or look for a way to suppress it - just as we do with other unprofitable, un-fullfillable drives. That's my position, anyway... In a timeless view, causality is just (regular) c
I was actually going to remark in the original comment and my previous one that I thought "subjective experience" was redundant. I truly have no idea what non-subjective experience could possibly be. "Subjective experience" isn't something that is contrasted from other kinds of experience. It isn't my coinage, as far as I know it is a legacy term but helpful in that it combines 'the subject' with 'experiencing'. If that makes you uncomfortable by all means replace every instance of 'subjective experience' with 'experience'. I think you can safely do the same with 'consciousness' or 'qualia' but I imagine you don't like those terms either. The mercury in my barometer always drops before a thunderstorm. My barometer has never caused a thunderstorm. Thus, I prefer a counterfactual theory of causation []. If you think Egan is right then how is a dust mind different from "causal sequences of physical states, i.e. outright simulations"> I think the argument requires that there be more than one mental state, though one can skip mental states. But lets say you had three detailed scans from a period of sadness. Whether or not it is immoral would depend on whether or not we distinguish identical and simultaneous copies of persons in our utility function. But if you do care about such copies then yeah, it wouldn't be the nicest thing to do. The concepts of 'consciousness' and 'causality' describe features of the way we relate to the external world. I would like a coherent picture of this relation. Cause and effect, in particular is a huge part of how we experience the world. How this concept relates to what is actually going on is a really interesting question. If a system needs to be the kind of system we recognize as a causal system in order to produce a subject that experiences the world that would be something interesting to know. Getting a really precise definition about what consciousness is would be really c

Isn't this, along with so many other problems, a candidate for our sometime friend the anthropic principle? That is: only in a conscious configuration field which has memories of perceptions of an orderly universe is the dust theory controversial or doubted? In the vastly more numerous conscious configuration fields with memories of perceptions of a chaotic and disorderly universe lacking a rational way to support the observer the dust theory could be accepted a priori or at least be a favored theory.

Ahem, this is circular logic. "The vast majority... (read more)

I don't think that really works. The problem is "I perceive far more order than would be needed for me to (briefly) exist from this moment"

ie, I observe that my surroundings seem to match my recent memories, etc etc..

QM arguably IS dust theory, just that one adds on a few rules for how the arrangements relate to each other, and then assign to each a complex number and have various rules for those relate to each other, how that changes, etc...

Yes I perceive nearly infinitely more order than is necessary. But in an infinite configuration space some observers would remember or experience a perception of that. What are the odds its me? Well if I weren't perceiving that - I wouldn't even be troubled by this problem! Only entities that remember perceptions of a universe of far greater complexity than needed to support their information state perceive a problem.
Not really. I right now notice far more order than needed. I STILL am noticing that. I do not observe my bookshelf spontaneously turning into a pizza, or, better yet, into just random noise. Not even slowly. Stuff is WAY more ordered than needed to merely support me existing for several moments or even several hours. Raw dust hypothesis would SEEM to be refuted with high probability. I admit, anthropics has some stuff that still confuses be, but it sure seems like raw dust is unlikely to be true.
When you say you are still noticing it, what you are saying is that you have a memory of noticing it before, and you perceive it now. Talking to me about experiencing the passage of time amounts to just saying "dust theory is wrong". I agree with that conclusion, but you haven't given me any information I don't already seem to have 33 years worth of.

I'm not sure if it counts as "purely logical", but another problem with Dust theory is that it assumes that conscious states supervene on brain states instantaneously. There is no evidence for that. We seem to be conscious moment-by-moment, but the "moments" in question are rather coarse-grained, corresponding to the specious present of 0.025-0.25 second or so. It's quite compatible with the phenomenology that it requires thousands or millions of neural events or processing steps to achieve a subjective instant of consciousness. Which ... (read more)

I am confused by the model of spontaneously occurring Boltzman-consciousnesses underlying this comment. If I am a spontaneously occuring Boltzman-consciousness, the fact that I manifested with a memory of having read a Wikipedia article that reports studies which indicate that I exhibit a specious present of 0.025-0.25 seconds just means that I have a memory which implies something false. Presumably, if I am a spontaneously occurring Boltzman-consciousness, all of my memories imply falsehoods. I don't see how the specious present is a distinct problem with Dust Theory, over and above the basic WTF? factor of Dust Theory. Put more simply: just because I've hallucinated something that, if true, would be incompatible with my hallucinating, doesn't mean I'm not hallucinating. FWIW, I agree with Egan and not Huffman here. The space of beliefs-about-the-world that Dust Theory allows for me to have includes so many that are less coherent than the ones I currently have that I shouldn't expect so much believed-in coherence, which is suggestive but not overwhelming evidence against Dust Theory. The anthropic principle doesn't apply, because nothing precludes my spontaneously occurring with more incoherent beliefs.
It's not that having an eperience of the specious present implies Dust Theory is false. It is rather that the computational theory of mind doens't strongly imply Dust Theory. Under the reasonable assumption that substantive stretches of computation are necessary for consciousness.
Ah. Well, I certainly agree with that.
[-][anonymous]12y 0

Isn't this, along with so many other problems, a candidate for our sometime friend the anthropic principle? That is: only in a conscious configuration field which has memories of perceptions of an orderly universe is the dust theory controversial or doubted?

How is this not a fully general argument against ever trying to explain anything?

[-][anonymous]12y 0

I say we are in a simulation. I'm not sure what the precise definition of 'simulation' is, but it should be a broad enough concept to include the universe, whatever the universe is. The universe may not be a directed simulation, it may not be a simulation that has a beginning and an end, and even the continuity of it may be a complete illusion. But I cannot imagine how anything at a sufficient level of detail could be interpreted as not a simulation; that is, as something that isn't computed or doesn't run with some mix of mechanical and random rules.

In th... (read more)

If we take the possibility of dust scenarios as a given, then perhaps our observation of a coherent universe can be explained by some idea of the "measure" of different possible universes/simulations. That is, if all possible universes are some Turing machine, then perhaps the simplest Turing machines have in some sense higher probability/measure, and the simplest machines that contain observers are still on the relatively simple and lawful side.

edit: or perhaps it's that lawful simulations have a higher density of observers.

I tend to ignore the dust theory simply because entities which are implemented as scattered states throughout spacetime can't be interacted with. Even just inverting the order of the states is enough to make interaction impossible - two observers with opposite time directions don't see each other as having any memories of past interactions.

What makes such interactions a requirement?
Thanks for the response! One way of looking at whether to take implementations of minds scattered across disconnected dust seriously is to look at the sets of minds we do know about and extrapolate from there. All existing minds that we know of (human, animal - even including any computation that responds to the world - even down to a thermostat) consists of causally connected states. "Dust" minds have at least the problems that: since we can't interact with them, what would constitute an experiment to demonstrate that they are really there? Are they observable in any way? since the causal processes in the world can't interact with them, they can't be tuned by the evolutionary processes that created us or other minds, which again puts them outside the set of minds we would extrapolate from those we've observed Do they survive Occam's razor?
That's not data. That's a definition.
I'm not convinced. For instance, I can point to plenty of examples of logic inverters that respond to causally to changes in their input logic states by making the inverse changes in their outputs. How does one slice and dice the states of the physical world to label some disconnected set of them as a "dust" inverter? In other words, if one explicitly enlarges the definition of a computing system to include "dust" systems, can one point to a correspondingly enlarged set of data on working examples?
I don't know, but it's not relevant. The fact that you can't usefully define "mind" or "computing system" to include dust, doesn't tell us anything about the ordinary definition of "mind" that does not include dust.
Your response is irrelevant to the central topic of this discussion. The entire point of the whole discussion is whether it makes sense to include dust in the class of places we consider as possible computing systems or not.
What I'm saying is this. You're faced with various ways in which dust-based systems differ from ordinary ones (e.g., you can't interact with them). Then you have two options: either say dust-based systems don't perform computation; or say computation doesn't require interaction. The difference between the two is solely in the definition of "computation". Discussing the appropriate definition of "computation" is a worthwhile discussion to have. But you have to be aware that you won't get any more information out of this deliberation than you put into it.