Is randomness necessary?

Is true randomness necessary for our models of physics to work?

Say nuclear decay. It is thought the be a random event, completely unpredictable. Would our understanding of physics be thoroughly shaken if we discovered that it does somehow depend on the doings of a so far undiscovered particle that follows some predictable laws that we just ignored?

Say, just for the sake of an example, that there is this other subatomic particle, the randomino, that flies around and sometimes hits an atom’s nucleus and that is what causes nuclear decay. Would that wreck our models?

Bear in mind that I am truly out of my depth here. Respond in small words.

In other words: Could randomness be just a convenient shorthand for “stuff we don’t yet understand?” or is it somehow important that it be truly random? (and how is randomness explained if it is important?)

That’s actually a very good question… one that I don’t have an answer for.

Throughout history, man tries to explain the world around him. Firstly with gods and the like, and later on, with organized religion. Randomness might actually be a manifestation of our natural inclination towards explaining everything scientifically.

I’ve never actually thought of that before, thanks!

The short answer is yes. Quantum Mechanics requires true randomness.

We also know that QM and relativity, as correct and proven as they are, do not mesh. Some larger theory must be generated that encompasses both of them and adds a bit more. It’s possible that this larger theory may explain away the seeming randomness that QM requires. Some physicists think so, but then you can always find some physicists who think anything.

I wouldn’t bet on it, though. Physicists have been looking very hard for an underlying cause for a long time and have come up with exactly nothing.

You can actually construct models which don’t depend on randomness, but they inevitably end up being nonlocal: That is to say, what happens in one place depends on what happens in other places far away, quicker than any information should be able to propagate from one place to the other. It’s pretty much a matter of taste among physicists which is more distasteful: Nondeterminism or nonlocality, since no way is known to distinguish between them.

My Magic 8-Ball says “Ask again later.”

How is true randomness required for QM? (Again in small words) Aren’t all those probabilities just descriptive?

In other words, does it matter if behind the curtains God is actually rolling dice or instead reading from the decimal places of pi? (so to speak)

Take an electron. Its spin can be up or down. You measure it. In the long run 50% will be up and 50% will be down. For any individual electron, you can’t tell ahead of time. There is nothing we can think of that will give us this information. It is indeterminable.

This affects everything. Randomness is one reason why QM cannot be reconciled with Relativity, which is resolutely non-random, part of classical physics. In fact, it’s a definition of classical physics. Randomness is as important an aspect as QM as anything. It’s the aspect that Einstein could never accept but could never find a way around. Everything revolves around it.

Ok, I understand that. My question is, if we discovered that the spin of this electron is determined by some mechanism we didn’t know about and that yields that 50-50 split in a way that could not be predicted by someone not aware of the mechanism, would it in any way wreck the model or just add to it?

Some googling around tells me that I am asking for might be what is called pseudorandom. Something that is statistically random but is determned by some deterministic process. The example given being the decimals of pi. They appear to be random but come from an established procedure.

If we discovered that electrons pop into existence with spin up or down according to the decimals of pi (odd = up, even = down), they would still appear to be random and unpredictable to one not follwing the list of decimals and would stil have a 50-50 distribution, but they would not be truly random. Would it matter?

I believe randomness is also required for many forms of statistical analysis.

This is what’s called a hidden variable theory. Einstein was fond of such theories, as seen in his famous quote “God is subtle but not malicious. He may deal the cards where they cannot be seen, but I refuse to believe that He plays dice with the world.”. Unfortunately, there’s a mathematical result called Bell’s inequality that rigorously proves that any hidden variable theory, no matter how it’s constructed, would predict results inconsistent with what quantum mechanics predicts. And when you actually go and do the experiments, they end up matching with quantum mechanics, but not with hidden variables.

Clarification: it’s local hidden variables that are falsified. As stated earlier, nonlocal mechanisms are possible.

Obvious next question: What’s the difference between local and non-local hidden variables?

There are two ways randomness (i.e. statistics) is used in physics…

In some areas (e.g. Brownian Motion). All you are saying is “we don’t have to model each individual atom, we just know that in enough quantities they will tend to behave in random manner. And when the quantities approach the kind of numbers we are interested in we can take that random behavior for granted”. This is no different to predicting a coin toss, its incredibly hard (basically impossible given our current level of maths/science/computing knowledge) to predict a single coin toss, but a billion coin tosses will come out at close to 500,000,000 heads and 500,000,000 tails.

However when we are taking about quantum physics something much weirder is going one. When we use a probability function to describe your randomino’s position and velocity we are not saying “Or measurements are not good enough to predict where this particle will be, and what speed it will be travelling at, at a given time, but if we have enough of them their distribution will match this function”. We are saying that each INDIVIDUAL PARTICLE does not have a discrete position and velocity, instead their position and velocity in UNKNOWABLE until the wave function is collapsed by measurement (and which ever property you measure, you cannot measure the other) . No matter how good your measurements get, you can NEVER measure its position and velocity accurately.

Well, to give a very simplistic example of a nonlocal hidden variable, God could be hiding behind the curtain reading off digits of pi, or whatever, and then tweaking the spins of multiple electrons at once, in widely separated locations. The key is the “working in multiple locations at once”: It’s not good enough for God to just delegate an angel to every electron, who each only worry about their one electron and never communicate (or even communicate at the speed of light).

Ok. Then it would indeed wreck the models if it turned out that it was not truly random. Or at least leave us with a whole lot more questions than before.

Are there any more or less reasonable (and explainable to the layman) attempts at explaining this randomness? When a nucleus “decides” to decay, is there any known mechanisms for what is believed to happen?

I suspect that any interpretation which is explainable to the layman would end up being a nonlocal hidden variable interpretation, rather than a nondeterministic one. In the Many Worlds interpretation, for instance, one could take which universe one is in as the nonlocal hidden variable. Which is fine, since there’s no way to distinguish between the interpretations anyway, so it’s not like one interpretation is right and all the others are wrong.

By and large, most physicists prefer the “Shut up and do the calculations” interpretation, where you just worry about the final answer and not the intermediate “reality”, but of course the calculations are not generally accessible to the layman.

Short answer: magic. :wink:

Although I am sure I am missing a lot of the subtleties of it, I think I have a decent intuitive grasp of what non-local hidden variables might entail.

Thanks to all who answered without bludgeoning me to death with calculus.

If we’ve shown experimentally that entangled particles are indeed acting at great distances, isn’t that non-local? Or is it ok because no information is transferred? Does non-locality imply information transfer faster than light speed?