The Autodidactic Universe
arxiv.org
arxiv.org
EDIT: After skimming through the paper, the paper more or less agrees. They focus on a specific layer in a hierarchy of laws and are not concerned with the bottom of the hierarchy.
More generally, we adopt the useful idea that the world can be analyzed in terms of a hierarchy of levels; the degrees of freedom and the regularities described at each level appear to observers there to be stable and unchanging or slowly changing. At the same time, when expressed in terms of the degrees of freedom and laws of a lower level, they may seem emergent and variable. We can then be agnostics as to whether there are fundamental, timeless laws down at the smallest scale.
Nice hypotheses. How would you go testing it to check its validity?
[1] This makes no assumptions whether you can know or compute or whatever this function, just that there exists some function from one state to the next.
I'm not particularly agreeing or disagreeing with anything, mind you. Just that the universe being probabilistic wouldn't change the argument
That’s more of an axiom.
The rules themselves can be non-deterministic?
Then what rules govern the dice? What determines when they are rolled?
A lot of ideas can boil down to the universe is a static construct. Adding randomness in the "unfolding over time" might just mean replaying it could lead to different outcomes. But in both cases I think the key question is Why do we perceive the passage of time?
I think we are not actually perceiving the passage of time but things changing. So what does it take to perceive change? Let's say there is an object changing color from red to blue and back every second or so. To perceive this change one must extract two bits of information from the current state of the universe, if they are different a color change occurred.
The rest is in a certain sense a matter of belief. One has to believe that the first bit extracted from the current state of the universe reflects the current color of the object. So this bit could come from the wavelength of a photon hitting the eye or a camera sensor in a certain spot together with some belief of how the universe works with regard to the way light behaves and interacts with matter.
For the second bit one has to believe that it reflects the color of the object some time ago. This bit would probably come from your memory or in case of the camera from some memory in the connected computer. This essentially requires all the same believes as for the first bit in order to determine the color of the object. But in addition it requires the believe that things change over time and that the memory stored and preserved information about the color of the object from some time ago.
So somewhat interestingly the perception of change and in turn the passage of time doesn't fundamentally require something to change at all, it just requires a current state of the universe and a lot of assumptions or believes about the working of the universe and how this lead to the current state. This is good enough for philosophical zombies, conscious experience and qualia and so on might require some additional ingredients.
Unfortunately I am not sure what this tells us about the universe. I guess the laws must at least allow the existence of some kind of memory, but that seems a low bar, anything with a delay can make a form of delay memory. The working of the universe must be stable to a certain extend across certain timescales, otherwise one couldn't rely on the memory. One could still perceive change but the change might not be real if the memory is not reliable. More interesting is probably what kind of options this allows for the evolution of the state of the universe, but I have no real idea what to say about that. But I guess as the perception of time is only dependent on the current state of the universe, the evolution could be quite different from the linear passage of time we perceive.
If the former, then you are treating time as a primitive, whereas the representation or even existence of time can be part of state and vary over state changes, so this does not work.
If the latter, are you assuming S[n] is an input because a universe where that is not the case and state "transitions" are completely independent from each other is indistinguishable from a multiverse with inaccessible sub-universes and should be called as such? In the latter, you are also assuming linearity and discreteness of state changes, whereas it could be a DAG or more complex structure.
Edit: On further reflection, I think you are proceeding with the standard post-Renaissance assumption that a universe must always be governed by fundamental rule(s) at the "lowest level," rather than a more chaotic or unstructured universe that behaves more like a disjoint set than an orderly structure. I think it makes sense to constrain to that assumption as departing from it makes reasoning difficult but it is worth stating explicitly as an axiom.
> rather than being completely unaffected by S[n]? Is n here a discretization of time or a "meta-time" that captures state progression?
It would be remarkable if the "next state" of the universe did not depend or was so correlated to the "current state" that symbolically we represent it as a dependency. A ball only rolls downhill because it was once uphill, after all - one of the few things that unites various models of the universe was the existence of causality.
> If the former, then you are treating time as a primitive, whereas the representation or even existence of time can be part of state and vary over state changes, so this does not work.
This is interesting but I feel reductive. the "nth" state is not the state at time n, but the state that follows n-1 - it is dimensionless, but one might express it in spacetime if that is convenient. Nothing suggests it is uniformly sampled or even computable.
> completely independent from each other is indistinguishable from a multiverse with inaccessible sub-universes and should be called as such
No idea what this implies, but in a non-deterministic state machine there are multiple S(n + 1) for any given S(n).
> you are also assuming linearity and discreteness of state changes, whereas it could be a DAG or more complex structure.
GP does not suggest evolve(S) is linear or discrete (they suggest either continuous or discrete, actually). Neither of these imply anything about topology so I'm not sure what that has to do with a DAG. State transitions can encapsulate arbitrary topologies (and vice versa). They are duals.
> rather than a more chaotic or unstructured universe that behaves more like a disjoint set than an orderly structure
I feel GP covers this base by mentioning that state transitions may not be deterministic. Just because a system is chaotic or unpredictable does not imply we cannot formulate the nature of its chaos or unpredictability symbolically.
If different states would be independent, if there were not even statistical relations, it would just be a sequence of random states. It would not even really make any sense to say that S[n] evolved into S[n + 1].
Is n here a discretization of time or a "meta-time" that captures state progression?
I did not want to imply any discretization, I just could not come up with a better notation. Proper mathematical treatment would probably use some index set with some kind of order defined on it. This would probably also overcome the linearity of state evolution implied by my notation.
And I used n intentionally to not imply time so that time can be emergent. On the other hand the entire idea of the state of the universe evolving is very heavily influenced by our experience with things changing over time. So I am actually less worried about the details of the independent variable, whether it is time or true-time, how many dimensions it has, or what not. I think the real problem is that the entire idea of state evolution might be wrong, maybe there is no state or no evolution and our brains are just not capable of imagining the way the universe actually works
On the other hand we have quite a few very general results like Turing-completeness, incompleteness theorems, uniqueness of the reals, and whatnot. And all this is built on top of rather simple axioms of logic without assuming much so it seems not that easy that we are totally missing some fundamental structure that we are not only unable to imagine but where we also lack the necessary logical or mathematical tools to even describe them. But it surly might be that there is a third option besides x is true and x is not true that humans are incapable of understanding and therefore our logic and everything build on top of it is missing.
My point here is that scientists are stuck in a box where everything must be a law. Those questioning fixed laws are still substituting meta-laws by which the laws evolve. To me, it's wishful thinking. Of all the possible universes I can imagine, I contend that those with fixed laws OR meta-laws would both be in the minority. The most likely case is that there's nothing to discover but happenstance.
If it is not absolute randomness, then there is some law, at the very least something is correlated or distributed in some way and one can at least statistically say something about the universe. But even if it is absolute randomness, if every variable might take on every possible value at every time [1], doesn't this also imply a law, everything is uniformly distributed?
I really have a very hard time to imagining something about that you can essentially not make any true statement. Because that is what it kind of boils down to, if nothing is fixed in the universe, then more or less no statement about the universe is true. [2] On the other hand, if you can make a true statement about the universe, that implies some regularity, some law the universe follows.
[1] For simplicity I assume the universe has some state that evolves over time, which is certainly not the most general imaginable universe.
[2] This is a bit handwavy, one can make all kinds of true statements - the universe is red or it is not the case that the universe is red. But I hope the gist of the argument is clear, if nothing is fixed then also statements like the universe has three spatial dimensions or the universe contains electrons or opposite charges attract each other all become false.
[0]: https://archive.org/details/perfectvacuum00lems/
[1]: https://en.wikipedia.org/wiki/Cosmological_natural_selection
[2]: https://archive.org/details/lifeofcosmos00smol/
Edit: fix link order
EDIT: To be clear, I mean, that they would interact to produce something that could in the future, reproduce them back again. Sort of like how life doesn't immediately create duplicates, it creates babies or immature stem cells, that then eventually mature.
[1] Like a Poincare disk: https://fineartamerica.com/shop/posters/poincare+disk
You certainly could make some analogies here, and it's an interesting topic. The mass of humankind as a substrate on which memes are selected for fitness, where more viral memes outcompete the less viral ones, now there's a fun one. But a single superorganism... I find it hard to see.
Yes ultimately we obey our genes ("programming") in a way determined by our environment... but so does cancer tissue. Plus there's such a great variability in how humans might respond to the same stimuli that we're not like well-behaved cells at all, plus much of the environment surrounding each human is not determined by other humans, which throws a lot of indeterminacy into the mix.
Like how we imprison criminals and fight ISIS, etc.
Others believe humans have as much free will as a rock rolling down the side of a mountain.
It’s either an act of faith as to which you believe (if humans have free will) or a random event (if humans don’t and the laws of the universe are set up to appear to contemplate this in some clouds of atoms).
Since these are axioms, it’s impossible to prove one way or the other.
A proper discussion by someone who does not believe in free will would have to be entirely passive voice, i.e. describing what a flame does when it is exposed to gusts of wind or describing what a cloud of human atoms does when it is exposed to its environment.
To your second point, millions of people do still and have believed this: https://en.m.wikipedia.org/wiki/Animism
‘I am’ is a fun one to think about :)
Just because someone doesn't give an argument for something doesn't mean that they don't have an argument for something. It seems you may be unaware of the study of language use and metaphysics that went on for most of the 20th century in analytic philosophy. I recommend looking into that.
Millions of people believe many things that are false and their belief doesn't make any of those false things true. That's a fallacy called argumentum ad populum.
E.g.
Some assert there is free will (by definition, can’t be proven).
Some assert there is no free will (by definition, can’t be proven).
It is presumably a sliding scale, though, in terms of some things more actively consuming order than others. If you want to define a proton as slightly alive, fine, though it's not very proactive in its consuming of things, it just waits around to bump into the right particle at the right speed. An amoeba is considerably more proactive about it, a fly even more so. I'm not trying to retreat to the old notion of "movement" as a criterion for life here, there is something else that defines this scale (given the known correspondence between information and entropy - computation?)
I'm guessing Schrodinger had thought about the atomic mass thing, though. I'd check out his very short and readable "What is life?" if you haven't already.
Still, this all begs the question of whether the particles that dominate our universe are species or are inherent from the physical laws. Then when you consider that the laws themselves may be species ..well..shoot me already. My head has begun to hurt.
2. I think using the term reproduction for this is stretching the term to the point of uselessness. I mean, for example in a chemical solution you have a redox reaction where molecules on one side of the "equation" turn into the molecules on the other side and back in a cyclical loop, but I wouldn't say that they are "reproducing", as I don't know what the point of saying so would be. The redox reaction and the way organisms reproduce are two of many kinds of cyclical processes. At least that seems to me an useful distinction.
1. Life looks like an optimization problem because living things die often and have to compete for limited resources. Electrons however stick around much longer and don't need oxygen to survive.
2. If this was true, we'd expect to see varying properties for the same type of elementary particle (especially those created in ex. the LHC), but we don't - all electrons act the same, even newly created ones.
For there to be some sort of "evolution" mechanic for particles (like there is for life) there has to be competition between particles of different properties. Then the particles that are better able to capture energy or whatever wins, and those fill the universe. However, to my knowledge, there is no experimental evidence that there can be elementary particles with differing properties, which suggests this is impossible.
Not to mention, there are unstable, rare, and unseen configurations like pentaquarks and such. Some particles are unstable _because_ of their environment. In a different universe, they might be the populous, stable particles.
P.S "Elementary" particle doesn't mean much since its definition shifts as we discover more about the structure of the universe.
Electrons sticking around longer is probably questionable. Do we stick around for 60-80 years? All our cells get replaced in the interim, multiple times. What if the same is happening in electrons?
We know for a fact that it's happening in protons and neutrons. There's no such thing as a "proton" that exists for a period of time. It's a configuration of colored quarks that are constantly in flux, replacing each other.
Also the idea that electrons act the same is incorrect. Many people don't realize, but things like the weight of an electron, charge of an electron... they're APPROXIMATE, and they actually vary a little. We only know the average properties of an electron.
But we also know the average behavior of a voter from an exit poll. Does this mean people are 100% identical? To someone who only analyzes our average behavior it would seem so, but they'd be completely wrong.
Incorrect, we have never experimentally found an electron with a different mass/charge/behavior/etc. We don't know the exact value of these properties because they are very hard to measure, but there is no evidence they differ between electrons. To my knowledge this is the same for all elementary particles.
The rest of your comment seems to be latching on to my "stick around for much longer" point which was honestly weak and I probably shouldn't have made it. I consider point 2 to be much more relevant: for some sort of "evolution" mechanic to exist for particles there needs to be particles that have differing properties from the common set. I am not aware of any experimental evidence that has found non-identical elementary particles.
Additionally, it seems to be near universally accepted by particle physicists that fundamental particles (like electrons) are identical, and I trust they have good reasons for that belief, so while it is not something we can definitively prove I am pretty confident about it.
In my book, the jury is out. We take the scientific approach, yes?
As for what particle physicists accept conventionally, versus what they claim is the actual nature of reality are two things. Again, a scientists to another, when questioned specifically on a topic, won't say "oh it's like that because we've universally accepted it's like that". Because they know history is full of things science accepted, and then was disproven.
Either there's sufficient evidence, or there isn't. One of the top ways we get stuck in science is when we confuse model and convention with proven objective reality.
My claim is that no evidence exists to support this claim. Obviously, I cannot provide proof for the absence of evidence, the burden of proof is on you to provide _any_ reliable evidence to support this claim.
While the absence of evidence does not 100% prove the theory is untrue, we generally treat theories with no evidence as untrue/very likely to be untrue (ex. https://en.wikipedia.org/wiki/Russell%27s_teapot).
This is not a Russell's Teapot situation at all because this revised middleground statement isn't just something arbitrary I decided. Instead it's literally the most specific statement either of us can make in face of currently established facts.
Once again, our everyday model is that electrons are little pinballs of exact charge and size, and that's a useful model. A more detailed model sees them as sizeless points of approximately measured charge that propagate in a probability quantum wave. That's also a useful model. Who knows what an even more detailed model would show?
I prefer to keep my mind open because, as I said, if you stick to convention, you lose track of what was proven and what wasn't, and you get stuck on the wrong path to truth.
In any case what I said about neutrons and protons is actually accepted scientific fact. They're not particles at all. They're a configuration of quarks in constant flux and exchange with the environment. And here we were not long ago, imagining they were persistent little pinballs as well. Go figure I guess.
Sounds a lot like a middlegrounds fallacy. My claim is "electron charge could possibly vary, or not, but all existing evidence suggests it does not, and no existing evidence suggests it does".
> A more detailed model sees them as sizeless points of approximately measured charge that propagate in a probability quantum wave.
I _think_ this is supposed to be a reference to quantum field theory, which explicitly assumes that particles represent a fixed, minimum amount of energy called a quanta (that's where the "quantum" in "quantum physics" comes from by the way).
It's so fundamental to the theory it's the beginning of the second paragraph on the wikipedia page for QFT (https://en.wikipedia.org/wiki/Quantum_field_theory):
"QFT treats particles as excited states (also called quanta) of their underlying quantum fields, which are more fundamental than the particles."
> In any case what I said about neutrons and protons is actually accepted scientific fact. They're not particles at all.
I'm referring to elementary particles, of which protons and neutrons are not (but quarks are).
But your claim is a falsehood. I asked you to provide evidence that it doesn't vary and you said you can't but you assume the variance comes solely from the measurement equipment. That's not what "evidence" means.
We have variance. Either there's proof that it's the measurement, or there isn't. Assuming is just that.
I'll tell you another assumption we often take for granted: speed of light. Actually we don't know if the speed of light is as measured, C. Do you know why? We can only measure reflected light. So we measure it going to some place, and then get reflected back.
Maybe it goes there at C/2, and then comes back instantly. OR maybe it goes there instantly, and comes back at C/2. Or even more likely, it goes there at some ratio of C, and comes back at the inverse ratio of C. We can't know that. In fact Einstein explicitly mentions this in his work, and calls C, the "average speed" just that - a convention.
Either you can prove it, or you can't. But how many people you hear, including scientists, talk about the speed of light this way? Not many. We take the model for granted to preserve our sanity. But from time to time we also need to remember: it's just a model.
> I'm referring to elementary particles, of which protons and neutrons are not (but quarks are).
Protons and neutrons are not. But until relatively recently, they were. What are we going to find out tomorrow?
I actually Googled this to see how elementary particles are defined, and very first thing you see is this sentence (emphasis mine): "Particles CURRENTLY THOUGHT to be elementary..."
See, when they need to be specific, they're quite aware this is just the current assumption. Nothing else.
You know, the only place it pays to be so sure about current models being objectively right and the law is school. In school you can propagate this myth as if you have the gospel in your hands. In the lab, this same PoV is frowned upon. "What if" is king. As long as it doesn't contradict existing evidence.
My claim is literally true. No evidence exists that suggests the mass of an electron varies. Feel free to prove me wrong by providing such evidence. All existing measurements of the mass/charge of an electron are consistent.
I really don't have the time to look up papers measuring the mass of an electron and figure out what degree of accuracy they're claiming and compare the measurements, but I guarantee you if you bother doing this you'll see the measurements are all the same.
> We have variance. Either there's proof that it's the measurement, or there isn't. Assuming is just that.
Variance is the expected outcome. Human-made measurement devices are imperfect. If the mass/charge of an electron varies, it either would have to do so at a level where our tools can detect it (provably untrue) or it would do so at a level below where our tools could detect it (unprovable until our tools improve obviously).
> I'll tell you another assumption we often take for granted: speed of light. ...
The fact we cannot measure the 1-way speed of light is well known. Using 'c' as the speed of light is a common convention. On the other hand, we _can_ measure the mass/charge of an electron, and that value is consistent to the degree we can measure it accurately.
All right, how many digits of precision do we have on that "consistent measurement"?
The particles we see are in constant flux of a number of their parameters, and we can see how environmental changes disturb this balance. Aggregate states of matter represent this very neatly. Just like flocks of birds fly together to protect themselves form harsh phenomena, gas condenses into liquids when temperature drops enough that standalone gas molecules become unstable. It's a long topic, and I'm kind of brushing over it superficially.
Particles may reproduce themselves in even more literal sense than you describe, because we've some evidence to believe no particle truly "exists" in time, it's a phenomena that keeps recreating itself through interaction of lower level parts and forces, much like we're in constant air/food/water exchange with our environment. Our cells change, but seemingly we stay. Is an electron really the "same electron" from one moment to the next, or is it more like a swirl in a constantly moving river water?