I am the anti-Wolfram

For a long long time, I was very sympathetic towards the point of view that the universe is one gigantic digital computer. This point of view has been championed by many people, including Ed Fredkin and Stephen Wolfram. Seth Lloyd has recently writen a book in which he updates this point of view, taking the point of view that that the universe is one gigantic quantum computer. One of the essential reasons for believing this sort of things is that computers (classical or quantum) can be used to simulate the physics of the universe. Of course there are all sorts of issues with this idea, for example, the simulations are invariably done up to a certain fixed accuracy. So while you can certainly do the simulaton on a quantum computer, as the physics becomes more and more accurate, you will need to believe that the computer doing the simulation is larger and larger. There is nothing intrinsicaly wrong with this notion, and indeed it may be that at its bare base there is a final discrete unit, but until such a unit shoes up in experiment, the hypothesis seems to me indistinguishable from not believe that the universe is a quantum computer. But this isn’t why I’ve become more skeptical of the universe as a computer point of view.
The reason I’ve become more skeptical is that I do not believe in computers. Er, well, at least I do not believe in digital perfectly working computers, nor do I believe in exact perfectly working quantum computers. When we dig down into the bowls of our computers we find that in their most basic form, these computers are made up of parts which are noisy or have uncertainties arising from quantum theory. Our digital computers (like the coming quantum computers) are emergent phenomena, and, further, it seems that in this emergence, nothing resembling absolutely perfect fault-tolerant computation is possible. Fault-tolerance can only be achieved over a certain time horizon (in both classical, where we currently have very very low error rates, and in quantum, where we are struggling to get the emergence to give us very low error rates.) Now of course, one can always work with the model of a perfect computer (neglecting a meta-level of thinking about what “working” means in terms of you, who are also a computer.) But if you do this, you must admit that you are talking about an entity that does not exists, as far as we known, in our universe. So somehow we are supposed to be comfortable in looking at the universe as a perfect computer, when such objects don’t exist in our universe? This makes me feel uncomfortable.
So I guess this makes me the anti-Wolfram, not subscribing to the view of our universe as a computer. Does this mean that I think that computation or information theory doesn’t have anything to say about physics? Actually, the answer is no. I’m just unconfortable with a naive translation of what it means for the universe to be a computer where everything is perfect and error free. Somehow I think the real answer must, somehow, be much deeper.

34 Replies to “I am the anti-Wolfram”

  1. It seems to me that if you argue that any physical system (including eg. the visible universe) isn’t a computer then what you’re really saying is that you don’t believe in the existence of a set of objective prescriptions describing time evolution.
    In other words, if you choose some system, with some initial conditions, and let it go, it will evolve via whatever the fundamental laws of physical evolution are…I don’t see how “time evolution due to cranking forward a set of physical rules” and “computation” can be different concepts–they’re just different descriptions of the same thing.

  2. I think you’ve really failed to think this one through. Ultra-high-energy cosmic rays, the Pioneer anomaly, Bush becoming president in 2000 … clearly these are all cases where the noisy computer running the universe made a goof.

  3. Dave,
    You’re not claiming to have overcome the possibility discussed by Descartes that you might be living in a simulation, are you?
    The inevitable errors (resulting from the functioning of the physical computer) can always be hidden from the layer at which your consciousness resides inside it. Plenty of room for layer upon layer of error correction, all invisible to you! 🙂
    Also, as Scott notes, how can you be sure that small errors are not occurring all the time?

  4. No I’m not claiming any great breakthrough in philosophy or physics or computation 🙂
    My argument is more like arguments used by many athiests against the existence of God. Certainly one cannot prove to you that God does not exist, but if you examine the reasons for why you believe in God you become more and more skeptical because these reasons seem to be founded on a very shaky foundation. I’m making a similar argument. Sure you can believe the universe is a computer and that it is just hiding its errors, and I ain’t ever going to give you a reason to not believe this (unless of course, experimental evidence in favor of this emerges!) On the other hand, it feels like you might have been led to this belief (i.e. in the digital universe) via something which I believe will eventually end up in the historical dustbin: the concept of a perfectly robust classical or quantum computer.

  5. I agree with a lot of this, but not the part at the end. For the most part, calling the universe a computer goes in the category of “not even wrong”. A computer is an invention, not a natural phenomenon. So it almost does sound like belief in God; after all, if the universe is an invention, “someone” invented it.
    Let us interpret Seth Lloyd more charitably to say that the universe resembles a computer, evidently a quantum computer. That by itself is not predictive either, but it could in theory point in the direction of a predictive theory. What is missing is any reason why the universe should resemble a computer in any specific way.
    The only motivated resemblance that I can think of is that the laws of physics are computationally universal, and computers are by definition computationally universal (*). That doesn’t say a whole lot, because there are tons of ways to achieve universality, but it is at least a useful sanity check for physical theories. Any set of physical laws that are computationally too easy cannot be the whole truth.
    (*) Well, universal except for the fact that the universe might be finite or effectively finite.
    On the other hand, what people like Ed Fredkin have in mind is something like that spacetime is Cartesian-raster, because many features of computers are. This strikes me as a complete non sequitur. I will concede that there exist Cartesian-raster objects in nature — they are called crystals — but it seems to me to be sheer wishful thinking to suppose that just because you like cellular automata, the universe therefore is one. Even in solid-state chemistry, which is the one place where a raster model turned out to be the truth, sober scientists such as Pasteur waited until they had very good reasons to claim that it is so.
    Now there are good reasons to suppose that the laws of physics have an ultraviolet cutoff. It is mathematically problematic if the laws of physics have ever “smaller fleas that bite em”. But rasterization is only one way to make an ultraviolet cutoff. If you have seen basic quantum mechanics, then you can already guess that it is not a particularly creative solution. (Of course, Cartesian rasterization is even less creative than granularity of some kind.) You might suppose that coordinate measurements do not exactly commute. If they don’t, this can easily lead to an ultraviolet cutoff. (Note that this is also only the beginning of a solution; the specific way that commutation breaks down could be very complicated.)
    The place where I do not agree with Dave is the suggestion (which he may not have intended) that the laws of physics are inevitably noisy. If the laws of physics are noisy, then why can’t you expand the laws of physics to include all of the source of noise? Do you expect yet other noise that makes the noise that you see? It is almost back to “smaller fleas that bite ’em”.
    Or, more technically in the context of quantum information theory, we all learn Stinespring’s theorem that every quantum operation has a noiseless dilation. Is there any good reason to leave the laws of physics at the noisy level, instead of lifting them to laws that describe the dilation?

  6. The big problem with the simplistic idea that the universe is a giant computer made of stuff is the question: WHERE is this universe? If mass and other spatial observables are to be constructive in a computational logic sense, their mathematical description cannot rely on an a priori classical reality. So Dave is right again.

  7. For a long time, God was likened to the “Great Geometer.” Then he was the “Great Watchmaker.” Today he is the “Great Programmer.” It makes me wonder—what will God be likened to a millenia from now? Given this progression, surely it will be mind-boggling in our current context. What these comparisons make clear is that this discussion is more religion than science, as has been noted by prior comments. 🙂

  8. It sounds like God has always been the “Great” version of whatever humanity is doing at any given time. Maybe in massage parlors, God is the “Great Masseuse”.

  9. Actually, Seth Lloyd does claim that the universe _is_ a quantum computers. Isn’t like one, doesn’t resemble one, but *is*.
    The concept of a computer is certainly an invention, but so is the concept of a crystal.
    Under the straight-forward and objective definitions I’ve seen, it’s hard to argue he’s wrong. It’s just that the computations it’s doing aren’t terribly “useful” from our point of view, and most don’t “look like” what we think of computations as being.
    Is this a useful way of looking at the universe? Maybe, in some ways. Maybe not.

  10. @Aaron: Computation is NOT an invention, for the same reason that the laws of physics aren’t inventions. Both are discoveries about objective properties of nature. Both were around long before humans stumbled across them.
    Also your statement that the natural evolution of the stuff around us isn’t doing a useful calculation is certainly horribly wrong…our continued existence depends on this calculation continuing…

  11. From gr-qc/0404088 (Hendryk Pfeiffer, 2004):

    Scenario for quantum gravity. We have reached a first goal: the diffeomorphism gauge symmetry of general relativity on a closed space-time manifold has been translated into a purely combinatorial problem involving triangulations that consist of only a finite number of simplices, and their manipulation by finite sequences of Pachner moves. If not only the partition function, but also the full path integral of general relativity in d ≤ 5+1 is given by a PL-QFT, we know that all observables are invariant under Pachner moves.
    ………………
    If quantum general relativity in d = 3+1 is indeed a PL-QFT, the following two statements which sound philosophically completely contrary,
    • Nature is fundamentally smooth.
    • Nature is fundamentally discrete.
    are just two different points of view on the same underlying mathematical structure: equivalence classes of smooth manifolds up to diffeomorphism.

  12. The concept of a computer is certainly an invention, but so is the concept of a crystal.
    I didn’t just mean that the concept of a computer is an invention; I meant that the actual object is an invention. Saying that the universe is a computer is, at best, either a truism or an idle metaphor. It’s like saying that a sunrise is an alarm clock.

  13. I guess part of my feelings towards this are that I think the amazing thing is that computers exist at all! (It didn’t have to be that way.) If I were of a particular religious bent, in every computer, like every piece of life, I would see a miracle. Go ahead, turn off your computer, you killer 😉

  14. As background for this discussion, it’s important to remember that Ernst Mach regarded atomism itself as at best a useful fiction right up until his death in the early 20th century. This is notwithstanding the fact that the notion of spacetime discreteness is much more subtle; one must somehow avoid the surreptitious reintroduction of a background manifold, or of fixed structure that leaves a semi-classical imprints inconsistent with various observational limits.
    By the way, the objection that “the universe as a computer” leads to an infinite regress because computers are, after all, to be regarded merely as abstractions from physical devices constructed of effectively classical components strikes me as, well, dumb. Physical fields were once regarded as being, necessarily, mere manifestations of an underlying medium with certain mechanical properties governed by Newtonian laws. In the end, the untenability of this assumption didn’t force the abandonment of the notion of a physical field.
    The objection stemming from the noisiness inherent in real working computers begs the question—which Greg Kuperberg appears to be addressing—of how such noisiness is to be accommodated in any theoretical foundation for physics. Wolfram has insisted on the power of certain kinds of cellular automata to reproduce stochastic features of physical processes. Whatever the merits of his specific assertions (especially with regard to quantum mechanics), the distinction between fundamental determinism and stochasticity becomes difficult to specify at the level of the observable phenomena one hopes to explain.
    In the end, Fredkin’s and Wolfram’s ideas haven’t as yet contributed to a clear and compelling formulation of a heretofore vexingly obscure physical question, and pointed the way to further crucial insights. Atomism ultimately did do that, and was well on its way to acceptance by the time Einstein finally pushed Mach’s skepticism into a sterile corner. (The profound irony of this, in light of Mach’s influence on Einstein, is perhaps not mentioned often enough.)
    —————
    PS: In this context one really should mention the work of Gerard ‘t Hooft and Stephen Adler over the last 15 years, which shows a much better grasp of the core issues in fundamental physics than the writings of Fredkin and Wolfram.

  15. “By the way, the objection that “the universe as a computer” leads to an infinite regress because computers are, after all, to be regarded merely as abstractions from physical devices constructed of effectively classical components strikes me as, well, dumb.”
    Thanks!

  16. But seriously, it’s not the infinite regress that bothers me at all. It’s the fact that digital computers are merely a good approximation. We like to think that digital information exists in an incorruptable form, but it doesn’t as far as I can tell. Indeed I think our attachment to digital error free information is only a historical anomoly. As we build smaller and smallear computers which have to deal with noise more and more. I suspect that at that point the postulate that the universe is a noisy computer will be written about. 🙂

  17. I’ve always been a big fan of topological discreteness! Indeed I once gave a group meeting trying to convince anyone in the room that this was a key to understanding quantum theory. And failed miserably!

  18. Agreeing with Greg Kuperberg, I have always thought that saying “the universe is a computer” is an obvious truism. These days, I think that saying “the universe is a quantum computer” is almost equally truistic. (Therefore, I don’t get what Dave is rejecting, but that’s not the point of my post.) I think things start to get interesting if you say something like “the universe is a giant finite state machine”, (Is this Fredkin’s and/or Wolfram’s point?) or that “there is an even larger, faster classical computer simulating the quantum computer” or something like that. The possibility of going on to more detailed (possibly refutable??) speculations is what makes me think that saying “the universe is a computer” is an interesting starting point, as well as an irrefutable, but possibly also uninteresting, truism. Just my $.02 worth.
    Jim Graber

  19. Apparently Dave was writing his post at the same time I was writing mine:
    ” I suspect that at that point the postulate that the universe is a noisy computer will be written about.”
    Oh, absolutely. The current most obvious truism is that the universe is a very noisy quantum computer. No?
    Jim Graber

  20. Dave, regarding your follow-up (comment #140039), my remark [“..as, well, dumb”] was not directed at you. I should also emphasize that the significant thing to me about abstract computers is not their perfection as expressions of idealized, error-free computation, or more generally, as a way back to a lost Laplacian determinism. I don’t believe in determinism, or more precisely, in a deterministic theory that is anything more than sterile metaphysical wish fulfillment.
    However, I think it could be very fruitful to ask anew why the single-minded pursuit of a deterministic explanation is misguided as a matter of principle. Quantum mechanics appears to many people to provide the definitive answer to this question. I’m inclined to agree with John Stachel’s position that a major objection of Einstein to the indeterminism of quantum mechanics was not so much “why?”, as “why this much, and no more?”. That is, if one is to admit some elementary events ungoverned by causality, why not admit total, uncontrolled chaos? (Of course, we wouldn’t exist in such a universe, but let’s not get into that cul-de-sac.)
    I think a combination of the following considerations could gain some significant new insight into this above question:
    — Where does an attempt at a deterministic and discrete basis for physics lead, if one supposes that spacetime structure grows out of this basis, rather than being assumed at the outset?
    — How would a discrete substratum be detected, if it constitutes spacetime, and assumes no background of its own? Indeed, should we assume that it must be detectable in anything like the sense that the atomic structure of matter was ultimately detected? Doesn’t this notion assume something like “detection of discrete elements (somewhat) sparsely distributed within (or against) a background manifold”?
    — How does one get differential geometry, and the physical processes that ostensibly occur within it, out of nothing but the combinatorial dynamics of a collection of discrete elements?
    [After that, I think I should shut the f*** up—for a day or two, at least. :)]

  21. I like the list Chris W. I would add:
    – How does one get quantum theory out of nothing but the combinatorial dynamics of a collection of discrete elements.
    It’s interesting that, from what I can tell, this later question was what Penrose was after in his original invention of spin networks.
    And no shutting up! That would only make this place boring and full of me writin really boring stuff from only my point of view.

  22. It seems to me that it does not make much sense to talk about a computer without (at least implicitly) assuming the existence of its user.
    Thus the talk of the “universe as computer” cannot have much meaning in my opinion, except for the hidden religious meaning.

  23. I didn’t just mean that the concept of a computer is an invention; I meant that the actual object is an invention. Saying that the universe is a computer is, at best, either a truism or an idle metaphor. It’s like saying that a sunrise is an alarm clock.

    Sure, any particular instance you point at it is a human artifact. So? Almost any nuclear reactor you point at is a human artifact as well. But so is the sun.
    Yes, it is a truism. If you pick the definition right, out it pops. By itself, it really doesn’t say much. Following the implications of adopting this point of view might, and that I think is Seth Lloyd’s point. It could just as well be a dead end.

  24. Stephen Wolfram’s right and you and Freeman Dyson are wrong.
    Not everyone is one hundred percent right all the time.
    In his thousand page “A New Kind of Science”, Stephen Wolfram may not be one hundred percent right either but I am going to bet that he will be proven mostly right and that you are wrong for doubting him and that Freeman Dyson will be proven absolutely wrong for saying Stephen Wolfram’s book is worthless.
    Like Stephen Wolfram, I too have been saying that there is indeed a simple model and a few simple rules behind the way this entire universe works.
    Like Stephen Wolfram, I too have been warning about this unwarranted belief that our math will give us the answer as to how this entire universe of ours works.
    I’ve given my opinion here after reading Stephen Wolfram’s entire book.
    You may have read the entire book as well.
    But it is my understanding that Freeman Dyson gave his opinion after only reading a small part of “A New Kind of Science”.
    If Stephen Wolfram and I are both correct then binary stars of the same mass will be found to have opposite spins and this is something that should shock Freeman Dyson and knock him from his lofty position as self appointed high priest of science.
    You and I and Stephen Wolfram only have to wait and see.
    When astronomers find this to be true then all of us will know that Stephen Wolfram’s right and that you and Freeman Dyson are wrong.
    http://www.amperefitz.com/lawrm.htm

  25. This is an extremely simple universe providing you look at it as a scalar, standing wave universe.
    As Dr. Milo Wolff showed, each electron is a scalar standing wave entity giving and receiving energy to and from other surrounding electrons out to the Hubble limit (a finite amount).
    But each electron has spin, which IS also scalar in respect to the TOTAL of the surrounding electrons but spin is NOT scalar to individual electrons and herein lies the rub.
    A greater difference in TIME is simply more out of phase with the principal scalar frequency and a greater distance (more space) is merely more out of phase with the spin frequency.
    Repulsive force equates with more space just like the tensor math in GR.
    Attractive force equates with less space like the tensor math as well.
    Believe it or not it is a simple as that.
    Stephen Wolfram is absolutely right, it turns out.

  26. The quantum computer, Dave mentions, depends on the present concept of quantum interference being correct.
    I suspect instead that the real reason one electron only reacts with other electrons on a certain path and never with others on a so called identical path is because the detector on this other path places its detecting electrons at a phase difference where their detection of the initial electron is impossible.
    The reason for this phase difference has to be properly worked out before we can have Feynman’s quantum computers.

  27. Dave Bacon,
    WOW, I wrote that last post just for you.
    Then I put it on the internet in different spots.
    In one week it pulled in over 9,000 people to my site.
    I guess quantum computers is what they want to read about.
    Fitz

Leave a Reply

Your email address will not be published. Required fields are marked *