Complications in Physics Lend Support to Multiverse Hypothesis | Hacker News Reader

Complications in Physics Lend Support to Multiverse Hypothesis

quantamagazine.org

45 pointsghosh60 comments

I used to believe that the history of physics gave unwavering evidence that there is always a deep reason for the patterns and phenomena we observe: maybe not for the specifics of everyday life, but certainly for fundamental facts. (Think of how atomic theory made sense of the elements, or how nuclear physics explained how the sun could shine for so long, or how the unification of electricity and magnetism miraculously turned out to explain optics as well.) So I was convinced that we would someday find a fundamental theory that would be able to predict things like the fine structure constant or the charge to mass ratio of the electron.

But eventually someone pointed out that before we understood the history and nature of the solar system, natural philosophers looked for similarly deep explanations for the orbits of the planets. Kepler once proposed a deep connection between the (five) gaps between the six known planets and the five platonic solids (see, e.g. http://www.pbs.org/wgbh/nova/blogs/physics/2011/12/beautiful...), and Bode's law for planetary orbits was widely accepted until Neptune was discovered in the "wrong" place in 1846 (see http://en.wikipedia.org/wiki/Titius%E2%80%93Bode_law).

Today, we understand that there is no reason to expect any particular pattern of planetary orbits: solar systems are a dime a dozen in the galaxy, and their details are accidents of history. So while I don't like the idea, I've gradually come to accept that the "fundamental physical constants" of our universe could conceivably turn out to be just as arbitrary as the ratio of Jupiter's major orbital axis to Saturn's.

Not sure this analogy holds. Although the ratio of Jupiter's and Saturn's major orbital axis might seem arbitrary, given all the prior conditions to the orbit formations, current physics theories perhaps could have theoretically predicted this ratio long before it came to be. In the same way, physicists would like an elegant model that would predict or explain the constants in a way that makes more sense in a larger picture rather than having them as some magic numbers.

Exactly. That's pretty much what multiverse models suggest might be true for "fundamental constants": if you somehow knew the precise conditions of our part of the universe in the moments leading up to our big bang (and the parameters of whatever process triggered it: quantum fluctuations or whatever), you could in principle predict things like the charge to mass ratio of the electron. (E.g. if you knew the precise shape and fluxes of our 6-dimensional Calabi-Yau compactification manifold in string theory, you could work out the low-energy field content and coupling constants of the resulting universe.)

[String phenomenology isn't really my field, but I certainly haven't had a lot of luck deriving a unique background from pure theory. I did once think that I'd found a model that predicted the cosmological constant to within an order of magnitude or two, but that obviously didn't pan out.]

That’s the idea of an infinite number of onion layers [1]. What I find intriguing is that if you assume that everything is based on cause and effect, then (I think) you necessarily run into an infinite regress problem (the cause of a cause of a cause …) unless you define some "uncaused cause", or the concept of causation breaks down at some point.

[1] https://en.wikipedia.org/wiki/Theory_of_everything#Infinite_...

> What I find intriguing is that if you assume that everything is based on cause and effect, then (I think) you necessarily run into an infinite regress problem (the cause of a cause of a cause …) unless you define some "uncaused cause", or the concept of causation breaks down at some point.

Another possible resolution is a cyclic picture (something like https://en.wikipedia.org/wiki/Cyclic_model), in which the cause of the cause of the cause of … may be the original event.

Isn’t that an infinite regress problem?

> Isn’t that an infinite regress problem?

I can certainly see the sense in which it might be said to be, but I think that's not the sense in which the term is originally meant. I tried to come up with an explanation of why I think they were meaningfully different by an appeal to mathematical analogy, but I'm not sure that it's convincing.

Fermat argued as follows: a positive integer solution to `x^4 + y^4 = z^4` could always be converted into a smaller such solution, and you can't produce ever-smaller positive integers forever, so there weren't any such solutions in the first place. I'd say that this is clearly an infinite regress (though it was actually called infinite descent).

On the other hand, consider the problem of subtracting 1. In the system of ordinary (not-necessarily-positive) integers, this leads to a (non-paradoxical) infinite regress; whereas, in the system of integers modulo some fixed positive integer, it leads to a (non-paradoxical) circularity (where repeatedly subtracting 1 from a number will eventually yield that number itself). I think that it is reasonable to call these different phenomena.

Cyclic causality is even more troublesome than infinite regress.

I find myself leaning towards causation breaking down at some point. Just like our concepts of "position" and "momentum" break down at the quantum level.

You say that causa sui or at least cyclic cause is troublesome without saying why. I'm curious for your reasoning.

I certainly didn't intend to suggest that an infinite number of theoretical layers might be necessary to explain nature (though I suppose I can't rule it out). Everything in my comment was meant to be compatible with (to take one example) the possibility that some particular well-defined refinement of string theory turns out to be the ultimate theory of nature, and only variations in local conditions are unpredictable. (This is meant in much the same way that the well established laws of gravitation, electromagnetism, and nuclear physics will give rise to very diverse solar systems.)

There are interesting questions in many multiverse models about the history and origin of the universe, and I've seen people express concerns about the possibility of "infinite history" in much the same words you've used here. But that's a different issue than Smolin's notion of infinitely many layers of theory being needed to keep the cosmic machinery running.

Personally, I'm not overly bothered by the notion of space and time existing forever in both directions, with new bubbles of interesting reality occasionally spawning themselves from the remnants of the old ones at random after uncounted eons of darkness. I don't know why that structure might exist rather than something else (or why there is something rather than nothing), but positing a "first cause" wouldn't make me any more satisfied (and would raise additional questions). I think modern math and science are more comfortable with the possibility of infinity than folks used to be, and that our notion of a "cause" is rather different. (When I think of "causality", all I think about are light cones and Cauchy problems: http://en.wikipedia.org/wiki/Cauchy_problem .)

The cosmology end of physics is kind of stuck right now. Many-worlds and the anthropic principle are back again. What they're theorizing about has way too much stuff that can't be observed or tested experimentally. Some people question whether this is even science.

"Science is prediction, not explanation." - Fred Hoyle.

It wouldn't surprise me at all if there are very hard limits to what science can prove when it comes to questions like these. After all some experiments can't be done from within the box and the box that we're in is made of the very laws that we're trying to escape from in order to do the experiments.

It's almost the physical equivalent of Gödel's incompleteness theorem.

As with Gödel's proofs, proving that it is impossible to come up with an experimentally verifiable demonstration that the multiverse theory is correct would be a major achievement. And Gödel didn't end mathematics with his theorems. It simply moved on, as would physics.

Along similar lines, I've wondered if it's possible to prove "any observing entity (e.g. a human) cannot possibly fully understand the system within which it is contained" .. any complete knowledge of a system should necessitate a larger system.

I've always wondered this as well.

Consider a finite collection of atoms consisting of a human and an environment. For simplicity, assume these atoms move classically. The human is a subsystem, and since the entire system involves deterministically, there is only a limited number of configurations that the human subsystem can realize within the environment. If a "human" is what is necessary to comprehend reality, then there are aspects of the system that cannot be realized by the subsystem, since there are far more states available to the total system.

Then again, I could just be rambling.

I would argue it depends on the nature of the system. If the system follows (simple) rules then you can likely compress the description of the system into something that fits into a (small) subsystem. If the system and its behavior is a uncompressible mess this will not work out. Up to now, I would say, our universe with all its relatively simple laws of physics looks pretty structured and easily compressible.

I don't think this works. You don't have to encompass everything about the universe in your mind. You can just comprehend a formula that allows you in principle to calculate any part of it that you need.

Think of a fractal. http://en.wikipedia.org/wiki/Mandelbrot_set

There's some relevant discussion on this old thread: https://news.ycombinator.com/item?id=7244476

It’s easy to prove "we don’t know that" since maybe it is possible to introspect our universe. For example, suppose someone interacts with some quantum system and it unexpectedly outputs a binary string. It turns out that this string can be interpreted as a command-line interface where one can access the states of individual particles and inspect the running process. Given the program code of our universe is reasonably short, we could possibly fully understand the system from within.

> Gödel didn't end mathematics with his theorems. It simply moved on

That's another way of saying Godel's theorems didn't end Mathematics, mathematicians just redefined what Mathematics was, presumably so they could keep their jobs.

The negation of the multiverse hypothesis can be taken to be, "no, there aren't multiple universes, let alone an infinity of them; there is only this one".

This negation is just as hard to prove as the multiverse hypothesis.

Lack of evidence of the existence of other universes doesn't prove their non-existence, nor does it rule out their existence.

However, note that the multiverse hypothesis is simpler in its content. The statement that this universe exists, while all the other ones do not, goes one step beyond the statement that they all exist.

Is there no evidence that there are no other universes?

Mathematics shows us that there are: it shows us that entities exist which are not located anywhere in this universe.

For instance, the number pi. We accept pi as real, but where in this universe is it? It is not computable, so that even if we use all the particles in this universe as the beads of an abacus, it will not represent pi.

If this universe is all there is, then there is no pi; we must stop talking in terms such as "there exists a real number that gives the ratio of the circumference to the diameter of a circle".

(Of course, the existence of this one is shared between the multiverse hypothesis and its negation.)

PI is a concenpt. Of course it doesn't exist physically* in the universe.

* Well, various "physical" representations/encodings of this concept must exist in different people's brains.

Whether the existence of pi is different from the existence of a given stone is a philosophical question! Maybe you and I exist exactly in the same sense that pi exists: no more, no less.

Concepts can be reshaped. If pi is merely a concept, why can't we conceive it to be something else, like 4.0?

In a curved spacetime Pi could easily be 4.0.

Pi is an idealised approximation of a certain kind of curvature in a certain kind of spacetime. Pi appears to have infinite precision only if you assume the curvature and the spacetime are perfectly smooth and have infinite resolution.

Given there's a Planck limit, it's unlikely the universe really is perfectly smooth. So Pi in the real universe is not very likely to have infinite precision.

That's where the concept of Pi as a Platonic example of mathematical perfection trips over the reality of physics as a messy thing where traditional math structures can only ever provide approximate descriptions.

Beyond a certain level of detail Other Stuff is happening, and you can't assume smoothness or infinite precision.

We don't know a whole lot about the Other Stuff yet. One reason is that math was built to work with smooth functions with infinite precision, and there's no equivalent math that describes the lumps of whatever spacetime is made of, or the relationships between the lumps.

Pi isn't an approximation but a real number which is uncomputable, but otherwise precisely identified. On the other hand, a rational number like 3.14159 serves as an approximation of pi.

pi doesn't help you find the distance around a circle in some curved space-time, but pi is what it is.

The multiverse hypothesis by definition asserts that such a thing is impossible. The universes are separate. Any entanglement between universes means they are in fact just constituents of a larger universe.

I think that the hypothesis will be quasi-proven when science hits the ultimate limit beyond which there is no explanation. This will show that the world just proceeds by arbitrary rules that are traceable to axioms. Once you have that, you have shown that it's indistinguishable from any other such a formal system. That is to say, that it exists in the same way: the world is no more or less real than, say, the plane of the complex numbers. After that, although you still don't have empirical evidence of other worlds, there is no short supply of imaginable worlds (such as in mathematics) which are just as real.

In other words, once it becomes obvious that the world is just "math all the way down", it will be as foolish to deny other existences as to deny the square root of -1.

[...] some experiments can't be done from within the box [...]

Only true if the box is not everything.

https://en.wikipedia.org/wiki/Anthropic_principle

I can find nothing unnatural about the existence of a single universe with particular constants that give rise to life that then studies the universe. There is nothing in this that leads inevitably to the conclusion someone has been rolling the dice many times until this came about.

This is exactly the way I think about it too. it doesn't seem reasonable to assume vastly more complicated universal structure based on nothing more than our own existence.

It sounds like you're thinking of the "strong anthropic principle" where the constants have to be set so as to produce life at some point in the future. It leaves the question as to how that worked unless you cite devine intervention. I find the multiverse idea easier to believe.

Most large religions today require not a weak or strong anthropic principle, but a "doubly strong" one, so that not only the initial constants of the Universe and the subsequent random events leading to intelligent life on Earth are somehow fine-tuned, but also various historical events after intelligence emerged. Which events would depend on which particular religion, e.g. parting the Red Sea for Christianity, the revelations to Mohammed for Islam, etc.

The funny thing is that the Antrophic principle believes in some kind of 'true randomness' in this way - even if it is just about the constants for the otherwise deterministic universe :-)

We developed highly accurate instruments that measure incredibly small and large things. Yet all of our scientific progress is biased, because we have evolved in a very specific environment that we handle well. Virtually all of our scientific measurements had been done in a relatively small time frame[1] and on one very specific planet with certain conditions [2].

I think it is way to early to call universe unnatural. Sure there is a chance that it might be true, but we are not going to find out until we have way more data. For one there is so much in cosmology that we cannot explain yet, such as dark matter and accelerating expansion.

[1] Last 250 years or so.

[2] Gravity, magnetic field, stable trajectory around the sun (sun's gravity is more or less unchanging)... I do know there are variations and experiments in microgravity, but all high-energy particle acceleration experiments had been done in relatively same conditions.

WRT HEP experiments, I'm hoping the developments in laser acceleration of charged particles leads to accelerator-like experiments being much cheaper and thus more easily doable for smaller groups and labs, it would at least increase the number of variations in conditions.

But, you're right about scientists assuming a lot about the universe. In cosmology, this is known as the Cosmological Principle[0], which is expressed in the language of the field as homogeneity of the matter distribution of the universe. Implicit hypotheses with this is that fundamental constants are constant and the laws of physics are the same. The basic idea is the assumption that the Earth is not special, therefore, experiments done here, while accounting for particular effects like earth's gravity, being a non-inertial frame, etc., would be representative of any point other point of space.

And it's not that this is a bad assumption. For example, spectra analysis of elements on earth has given us good ways to figure out the elements of stars far out of reach, and it is one hell of a coincidence that the spectra we see in stars is so similar to elements here like hydrogen and helium that one can't assume that earth was just gifted with these elements while other parts weren't. This is just one example of the success of this assumption.

[0] https://en.wikipedia.org/wiki/Cosmological_principle

"Yet all of our scientific progress is biased, because we have evolved in a very specific environment that we handle well."

I often get discouraged thinking about this. Our entire body of scientific knowledge consists of inferences drawn from human observation, but our sensory apparatus is only sensitive to a few different things/processes. And the mental processes we apply to our observations evolved to simply help us survive our physical environment. We could be fatally under-equipped to ever come close to "understanding" the universe, right?

Also, I'm WAY out of my depth here, but is it at least possible to imagine a system where there are multiple sets of elementary particles, with each set not interacting with the rest? I'm imagining an arbitrary number of sets of particles that are each identical in function to the particles comprising our physical world, but the states of the particles in Set 1 have no effect on the states of the particles in Set 2, and so on. The result could be many intertwined or overlaid physical universes, all occupying the same "space." Could something like this ever be disproved?

That is basically string theory^^

The deeper we go, the stranger things get. But I doubt scientists will "lose the desire to continue looking for new physics" as is warned by this article! We'll go deeper and things will get stranger, and that's just the way of human exploration.

Check out The Reference Frame blog if you want to read more (and more in-depth) about naturalness. For example: http://motls.blogspot.cz/2012/07/diphoton-higgs-enhancement-... http://motls.blogspot.cz/2013/12/doubly-protected-higgs-is-n...

Otherwise: https://www.google.com/?q=site:motls.blogspot.com+naturalnes...

End of science? Hardly. just because we may be in a multiverse doesn't mean we should stop trying to figure out how our particular universe works, and this info may yield insight into the other universes. There is still so much to learn

"Parallel universes cannot be tested for" The multiverse idea just sounds like a more respectable version of "turtles all the way down"

Why? "Turtles" is a bad hypothesis because it's so arbitrary. It is no more likely than "ducks all the way down", etc. The fact that it's not testable is the least of it's problems.

Whereas a multiverse is a much higher prior probability. In fact believing we are the only universe is just as untestable a hypothesis, and even more silly.

Please stop with the probabilities.

We can demonstrate, our universe exists. Claims about the number of universes (>= 1) are untestable (for now) and should therefore be considered speculation.

Per your statement, (number of universes >= 1) is testable and true ("demonstrated").

Maybe you mean (number of universes > 1)?

I'm not saying that there isn't any evidence our universe exists. That would be silly. I'm saying there is no evidence that our universe is alone and it would be incredibly surprising if it was. I don't know why "number of universes = 1" should be a default hypothesis.

There is no evidence for unicorns, but I don't know why everyone assumes they don't exist. Isn't it equally likely that they are just really good at hiding?

If the multiverse hypothesis is true, then beasts which resemble unicorns most likely do exist among the infinitude of universes. I mean, it would be astonishing if that could be ruled out: that there is an infinity of universes of all imaginable kinds, but none of them have anything that fits the description of unicorns.

Also, many universes probably have a being which serves, essentially, as a god: that is, these worlds are partitioned into an intelligent part, and a remainder which the intelligent part created and controls.

(Not this one, though, as far as anyone credible has been able to tell.)

You are making my argument for me. The hypothesis "unicorns don't exist" has zero proof. You can never ever prove that unicorns don't exist. Therefore, the statement "unicorns don't exist" is just wild speculation and outside the realm of science. Therefore, until someone shows proof we will assume that unicorns do exist.

The idea that our universe is the only universe is just as crazy to me as saying unicorns exist. The laws of physics are incredibly arbitrary, and the probability of a random set of laws of physics supporting intelligent life is ridiculously small.

More formally, via Solomonoff induction, the hypothesis of multiple universes should be more likely by several orders of magnitude, since it can necessarily be expressed with fewer bits of complexity. E.g. "all computable programs" vs "one specific program", or "all possible laws of physics", vs "one very specific set of laws of physics."

I see you missed this:

http://phys.org/news/2014-10-interacting-worlds-theory-scien...

The thing is, nothing you can discover in physics can refute the multiverse hypothesis. It is not falsifiable in that way.

Even if you find that you're in a very special universe, you need an explanation of why it is that way. The multiverse hypothesis helps, by giving a plausible answer which avoids belief in gods.

> The multiverse hypothesis helps, by giving a plausible answer which avoids belief in gods.

But as long as there is no proof of the existence of these multiple universes it's still magical thinking.

Worse: Why do you consider the multiple universe hypothesis as more probable than the God hypothesis?

On the basis of data? No. (If you think otherwise, show us the data.)

On the basis of philosophical prejudice? Yes, I think that's the reason.

But if you say that, given no evidence for God, one should be an atheist, then given no evidence for multiple universes...

Philosophical prejudice in favor of "no hidden variables".

Even the hypothesis that this is the only universe requires hidden variables (to explain what snuffs out the other universes, setting up this one as the one and only).

But aren't other universes also "hidden variables"?

Even more, why do you say "to explain what snuffs out the other universes"? That means you're already assuming that there are (or were or should be) other universes. What is your basis for assuming that?

Other universes aren't hidden variables because they have no effect. That is to say, no process in one universe is postulated to explain a process in another universe (because any such a bridge would mean that we are really talking about the same universe, right?)

It's the same basis like for the notion that, for instance, there is no greatest integer.

It would be surprising if suddenly the integers stopped at a particular value. Such an extraordinary situation would require an extraordinary explanation.

Any set of rules you can imagine gives you a universe. If you say that there is only this universe and nothing else, then you're denying the existence of things like the complex plane.

Or else you're asserting that the physical world exists in a different sense.

But so far, all evidence points toward the world being nothing but math.

Not a single discovery in physics has ever been anything other than math: a mathematical relationship (equations) or a quantity (constant).

Maybe that's just a "streak" of a particular kind of luck: maybe the next discovery will be of something non-mathematical. I'm betting on it being "math all the way down".

> Any set of rules you can imagine gives you a universe.

That's like saying that any supernatural being you can imagine gives you a god. Reality is not the consequence of math and it is absurd to believe that things exist just because you had some results from an equation.

> it is absurd to believe that things exist just because you had some results from an equation.

Okay, so then: what has advancement in physics ever given us, other than yet another equation or a numeric constant?

> Even if you find that you're in a very special universe, you need an explanation of why it is that way.

Emotionally "need", perhaps. Scientifically, not really.

> The multiverse hypothesis helps, by giving a plausible answer which avoids belief in gods.

The multiverse conjecture is -- by your own description -- not a hypothesis in the scientific sense, because, as you say, it cannot be falsified empirically. Like any belief in God that is compatible with science, its simply -- from a scientific perspective -- superfluous. It has no advantage, except the emotional advantage it provides to someone with an emotional bias to (1) have an explanation even if it is untestable, and (2) to avoid that explanation relying on belief in God.

>Physicists reason that if the universe is unnatural, with extremely unlikely fundamental constants that make life possible, then an enormous number of universes must exist for our improbable case to have been realized.

somewhat reminds about what physics must have been before Newton's laws - the cannon ball flies this strange seemingly unnatural trajectory what it would sound reasonable to suppose that "enormous number of universes must exist for our improbable case to have been realized." Though they were lucky back then to have God's will as an easy always available tool for explanation. We're not that lucky today. We have to continue digging :)

Reply on news.ycombinator.com