Is anything truly random?

[Kyanzes]

Our universe seems to follow known laws etc. etc.
I've seen references to a "true random generator" that takes a seed value from atmospheric distortions to produce a very "random" number but is that even really random?

Is anything truly random?

Definition of Random for this topic: Lacking any consistent rules or order or purpose that can lead to any patterns.

Unpredictable doesn't automatically mean random. While random is always unpredictable.

Unpredictable: Incapable of being determined in advance whether by observation, experience or reason.

If something is unobservable, unmeasurable or can't be reasoned doesn't mean it is random (ex. quantum phenomena).

With zero serious scientific background my personal opinion is that very probably nothing is truly random.

"Definition of Random for this topic: Lacking any consistent rules or order or purpose that can lead to any patterns."

I think this should probably be changed into

"Definition of Random [snip]: Lacking any perceivable consistent rules or order or purpose that can lead to any patterns."

So, something could be considered random, for your purposes, as long as nobody is able to perceive any kind of consistent rule or order to the phenomenon.

Since causality seems to be inherent everywhere, I can only assume that everything has a cause. And a previous one. And so on.

Of course, the question always remains, was there an initiator cause [or effect] that had not been preceded by a cause? And if so, can it happen again? Does it happen occasionally, perhaps so rarely (or in a hidden way) that we can't quite find such occurrences? Dunno.

True randomness (not perceived randomness) in my understanding would require causality to not be consistent in some way. But even if one found such a phenomenon the question would still remain, is there some hidden cause behind it?

Some addition:

I cannot really believe in true randomness. I think that everything (every effect) has a cause behind it. And when there is cause, there's logic.

In my opinion, everything that we perceive as random is simply out of our reach observation-wise to show the logic.

My question: even if found a truly random phenomenon, what use could that be put to?

Perhaps it would prove something (like the debate about free will) I don't know.

(It's also a bit hard to even speak about a truly random phenomenon since I cannot really call a random occurrence an effect. Nor is it truly a cause. So what is it?)

 

[disappoint]

Isn't radiation caused by the instability in the atomic nucleus? Too many neutrons and protons packed together in one small spherical region. Just a possibility.

It's not just a possibility. Radioactive decay is well understood. Yes the larger the nucleus the more unstable.

 

[disappoint]

There _is_ no such thing as chance in _any_ universe. Through laws of physics may change from one reality to another, structure remains. The only freedom is that precived

All chance lies in what we do not know. God does not play dice with the universe.

My dear Einstein, stop telling God what to do. /Bohr

Let the arrogant people who pray to God do that.

 

[PrincessFrosty]

The flip side of the question is essentially, is there anything in physics that isn't deterministic? Most of the laws of classical physics are highly deterministic, things behave in predictable ways and follow fairly basic fundamental laws that govern their behaviour.

The only thing I'm aware of that could potentially be truly random is quantum mechanics, the problem is that quantum states are tricky to measure without collapsing the quantum system into a more classical/deterministic one.

I believe the best random sources we have these days is radioactive decay, you can get fairly decent sources of randomness from computers by measuring noisy deterministic systems. Read up on /dev/random.

 

[Puppies04]

Radiation causes radioactive decay.

And if you ask why atoms radiate I'll bitch slap you.

Thanks for the laugh.

 

[lenasmith]

Of course there are, such as a gamble.

 

[intx13]

Like most discussions of randomness, this one went straight to philosophy! Randomness is a pet interest of mine and I registered just to write this long post. If you're not interested in what randomness really means in computer science / computational theory, this giant wall of text is going to be very boring!

In mathematics there are varying definitions of "random", depending on what you care about. The first thing to note is that a number is not random. Only processes/infinite sequences/etc. are random. A selection from any finite sequence - whether one number or a million numbers - is never "random".

A random number generator, therefore, does not produce a random number. It produces the next number in a random sequence.

This, in the most simplest terms, is why you shouldn't change your bets even when the dice are "hot". Whatever pattern of dice rolls you just saw isn't what we call random. The process of rolling dice is a random process. What you saw was just a selection from that process.

The biggest, baddest definition of randomness is an algorithmically random sequence, also called Martin-Lof randomness. There are several definitions, all of which rely on fairly abstract computational theory. But basically it comes down to...

You wrote a program (in whatever language you like) that takes a finite sequence of numbers. When run, it will tell you the minimum number of lines of Basic code it would take to display the input sequence. For example, I choose 10 numbers at "random" and your program reports "You need at least 30 lines of Basic to print those numbers." I give you a different set of numbers, maybe 25 this time, and your program reports "You need at least 40 lines of Basic to print those numbers."

I starting giving you sets of numbers to plug into your program. I made up these sets, who knows where they come from. Some of the sets I give you are short and sweet: 1 2 3. Here's a three line Basic program that prints that out:

print 1
print 2
print 3

Other sets I give you are long, but simple: 1 2 3 ... 100. There are 100 numbers in that set, but it only takes a 5-line Basic program with a loop to print it out.

let x = 1
while x <= 100
print x
x = x + 1
loop

These Basic programs are compressed versions of the input. In other words, the Basic program is the same size or shorter (in lines of code) than the input. n the first case, it was the same size. In the second, the Basic code only had 5 lines, but the input set had 100 numbers.

The difference between the size of the set and the size of the smallest program that will print out that set is called c. In the first example, c = 3 - 3 = 0. In the second, c = 100 - 5 = 95. Intuitively, c measures how well a set of numbers can be compressed in a specific programming language. If you can't do better than a particular c value, then the set you tried to compress is called c-incompressible.

Now, rather than just making up sets of numbers, you challenge me to use some infinite sequence of numbers and pull sets from that. I can give you short sets, long sets, massively long sets, but so long as they come from some magic sequence in my pocket, you'll type them as input into your program and find out how short of a Basic program we can write to output them.

As an example, let's use the infinite sequence 1, 2... If I pull out the first 3 numbers (like our first example) your program reports c = 0. If I pull out the first 100 numbers (like our second example) your program reports c = 95. Because my sequence is so simple, it doesn't take much code even if I pull out really long sets.

Now I challenge you. "Is there anything in my sequence that you can compress with a c bigger than 500?" You say "Sure! Take the first 600 numbers from your set. My program reports that you can write a 5 line Basic program - in fact, the same program we used above - to output that set. c = 600 - 5 = 595, that's bigger than 500."

Clearly, no matter what c I challenge you with, you can find a set of numbers taken from my sequence that you can compress better than c.

Now I'm annoyed so I throw away that simple sequence and I come up with something ludicrously complicated. In fact, it's so complicated that you can't see any discernible pattern. Good thing you have your magic program that will tell you how short a Basic program we can write. So I challenge you with c = 100,000 and you start looking for really, really long sets from the sequence that can, according to the program, be outputted with programs short enough that the difference between the set size and the program size is greater than 100,000.

Unfortunately, there's an infinite number of sets you can pull from my infinite sequence. If you happen to find one that results in c > 100,000 then we're done, you beat my challenge. I can pick an even bigger c and try again. But if you run this thing for all eternity and we hire a mathematics professor to prove, on paper, that it impossible to beat my challenge, then my infinite sequence is algorithmically random.

To restate it, if you're allowed to compress any starting set from my infinite sequence and you're allowed to go at it for all eternity and it is impossible to ever find a single sequence for which a Basic program can be written in fewer than 100,000 lines of code less than the length of the set, the sequence is random.

-

Now some small notes.

The choice of Basic was arbitrary. If an infinite sequence is algorithmically random for some programming language and some c value specific to that programming language, then it's algorithmically random for all programming languages. Each will have a different c value.

In reality, we don't have handy little programs that figure out the smallest Basic program to output some set of numbers. Instead we use our big monkey brains and computer science degrees to prove on paper the limits on compression.

It turns out to be really, really tough to find an algorithmically random sequences. I know exactly one, and I know that at least two others were discovered. Algorithmically random sequences are really tough to describe - even vaguely in English. It's part of the fact that they're incompressible, and describing something in English is, in a way, compression.

At the same time, however, it turns out that almost every infinite sequence that can exist in the world of math is algorithmically random. The basic foundations of mathematics support an uncountably infinite number of algorithmically random sequences, but we barely know what any of them are.

-

If anyone is interested in an example I can describe the "halting sequence" - a relatively simple example of a truly random sequence. I can also talk about sequences that pass this test not for Basic, but for systems with less computational "power". Not what we think of as computers, that is. I can provide lots of examples of those. We call them "relatively random" compared to whatever weak "computer" we're comparing against, like an abacus.

Alternatively, if I just blew up half your thread with my first post on this forum and you're angry about it, I can keep quiet

But that, philosophy geeks, is real randomness! There are sequences of numbers that are so complex that they don't contain even a single sub-sequence that can be compressed by some fixed amount, no matter what, by any computer, ever.

(Finally, this is a hobby area of interest for me, and if you are a real expert feel free to correct the many mistakes I probably made!)

 

[jaqie]

Isn't radiation caused by the instability in the atomic nucleus? Too many neutrons and protons packed together in one small spherical region. Just a possibility.

Yes. This, exactly. Anyone with even a physics 101 passing grade knows this... and exactly when each radio event will happen is completely random within a defined probability set of time. Basically, we know how much U238 radiates and how long its half-life is, but there is no way at all to tell when it is going to radiate, just the probabilities of how often. This is actually a reason why radiation sounds like static (ex. white noise) instead of a steadily decreasing wave formation. More atoms make more cases of randomness within the predefined likelihoods, which is in a way distantly similar to the "shotgun on the side of a barn" principle talked about with statistical maps and research with enough (thousands) of points of data on a single object without another reference point.

 

[_Rick_]

I fail to see a reason why radioactive decay would be truly random. While current observation makes it appear that way, the decay is still based on an interaction of sub-atomic forces, which should have deterministic values. These forces may even be constant (not where they act upon, but as a force field), but I'm no high-energy physicist. It's possible that protons and neutrons are not immutable, but interchange sub-particles, charges, etc.
Now, I perceive a U238 nuclear core as a slow moving maelstrom of neutrons and protons (much like a 238-large sphere of buckyballs), where eventually a configuration is reached, that ejects a neutron. Now if we could look close enough at the core geometry, radiation could become non-random. (i.e. there is a finite set of ejection scenarios, that are themselves preceded by a limited number of more-or-less deterministic geometric configurations)

Of course, I could be horribly wrong.

As for quantum effects introducing randomness, yes and no!
The limited ability to observe quantum-states, combined with the characteristic of superposition, makes it physically impossible to observe the state of a quantum signal, and only allows us to obtain a result. Of course, quantum state observation is slowly advancing, to the point where now it's possible to determine certain meta-information of a superposed signal. Whether this is helpful to determine the result of an observation of a quantum signal, is beyond my knowledge.

On a philosophical level, it doesn't matter much whether or not determinism holds at all levels of the physical world, as we humans are very bad at determining whether something is random or not (see intx13's post). Plus, even if an event were truly random, by observing its result, this reality is changed accordingly. Whether it was random or not changes very little (except if you rely on random numbers for the purpose of encryption).

 

[Paul98]

The problem people have is that in our daily lives everything seems to be cause and effect. Even things we think of as random such as a flip of a coin or roll of a dice we can see the cause and effect and that we can actually control the outcome.

This is much different in the quantum world, the average person thinks we just don't have a way to measure everything that is going on so we get random results. Where in fact where something is exactly or how it's moving doesn't exist until we actually "look" at it. It's that these particles are a just a sea of possibilities till it's "measured". Now we can calculate these probabilities so we know the percentage chance something has of happening. But for each particle we won't know what it is doing till we measure it, and that act removes the randomness.

Now I wish this wasn't true I would love to see some experiment to prove these things wrong. But so far even the strange predictions that are made from quantum mechanics have turned out true.

 

[Eureka]

I fail to see a reason why radioactive decay would be truly random. While current observation makes it appear that way, the decay is still based on an interaction of sub-atomic forces, which should have deterministic values. These forces may even be constant (not where they act upon, but as a force field), but I'm no high-energy physicist. It's possible that protons and neutrons are not immutable, but interchange sub-particles, charges, etc.
Now, I perceive a U238 nuclear core as a slow moving maelstrom of neutrons and protons (much like a 238-large sphere of buckyballs), where eventually a configuration is reached, that ejects a neutron. Now if we could look close enough at the core geometry, radiation could become non-random. (i.e. there is a finite set of ejection scenarios, that are themselves preceded by a limited number of more-or-less deterministic geometric configurations)

Of course, I could be horribly wrong.

As for quantum effects introducing randomness, yes and no!
The limited ability to observe quantum-states, combined with the characteristic of superposition, makes it physically impossible to observe the state of a quantum signal, and only allows us to obtain a result. Of course, quantum state observation is slowly advancing, to the point where now it's possible to determine certain meta-information of a superposed signal. Whether this is helpful to determine the result of an observation of a quantum signal, is beyond my knowledge.

On a philosophical level, it doesn't matter much whether or not determinism holds at all levels of the physical world, as we humans are very bad at determining whether something is random or not (see intx13's post). Plus, even if an event were truly random, by observing its result, this reality is changed accordingly. Whether it was random or not changes very little (except if you rely on random numbers for the purpose of encryption).

As a particle-physics-layman, it's because those sub-atomic forces are still dependent on quantum forces at the end. And until we know otherwise, quantum forces are still random.

 

[intx13]

As a particle-physics-layman, it's because those sub-atomic forces are still dependent on quantum forces at the end. And until we know otherwise, quantum forces are still random.

Some subatomic features, like the location of an electron, are random in the sense that we cannot specify the "answer" for a single event, but rather can specify a statistical distribution to which the event adheres. Those statistical distributions can, through various mechanisms, be combined in ways that transforms them. The result is still a statistical distribution, but maybe most of the probability is narrowed down to one or two different "answers". Sure, it's still random, but that doesn't mean we don't have any information on it.

That's why you have to be really careful in building a random number generator based on something like radioactive decay. You don't want to build a system that can, through some outside influence like time or heat or whatever, transform into a distribution with all the probability centered around the number "2". Not so random anymore!

Usually you constantly run a battery of statistical tests on your random number generator to make sure it isn't degrading. You also apply "whitening" transformations to take whatever distribution it produces (Gaussian, for example) and turn it into the uniform distribution that we really want when we call the "rand" function.

 

[Biftheunderstudy]

First, let me preface that I'm by no means an expert on quantum (even with the 5 or 6 courses I've taken all the way to graduate student level).

Also, the OP has now basically said, "is there anything other than quantum that is random".

So, I will simply leave you with a few tantalizing things to look up.

1) Quantum determinism. Quantum mechanics is a unitary theory, as such it is completely deterministic..yet still has truly mathematical randomness in it... weird right?

2) Bell's Theorem: for everyone that says "Well maybe if we just look closer we'll see more things and can then predict radioactivity etc", seriously, look into this theorem. Things don't appear to be so clear cut. This example is called the hidden variable problem, and it gets really messy, really quickly. Basically, there are a few possibility: local vs. non-local and real vs. not real. I can't really describe these very well, anyone take up the sword here?

3) The wavefunction collapse terminology has sort of fallen out of vogue now-a-days, people like to talk about de-coherence now..though again, I've only ever heard this from people giving talks at my university physics department. To completely muddy the quantum waters, everything you have learned in quantum mechanics is kind of a lie. Look up weak measurement. Seriously, this will blow your brain if you know a little bit about QM.