Randomness

[Polymath257]

I don't think you've quite gotten it. Science doesn't have a stance on the validity of untestable hypotheses. It may seem like we are splitting hairs here, but when you declare the hypothesis to be 'not valid' because of the absence of an experiment that tests for it, you simply are not doing science. Science does not assert the invalidity of hypotheses that it does not test for.

Actually, science does take a stand on untestable hypotheses: it declares them to be un-scientific. And such hypotheses *should* be considered with a great deal of skepticism: it is trivial to come up with multitudes of untestable ideas that should be immediately rejected. For example, perhaps angels are what drives the planets instead of gravity: they just do it in exactly the way gravity would.

We agree that it isn't included in science and the reason it isn't included is because a scientific hypothesis must be testable (or falsifiable). And yes, Ockham's Razor does say the hypothesis is unnecessary. But that is not an assertion about validity of the hypothesis. This is the point we are odds with each other about.

But the fact that it is unnecessary and untestable *should* put it under a cloud of suspicion. Again, it is just too easy to come up with untestable ideas.

When you say the hypothesis is invalid, then you've gone outside of science in order to make that claim. You're doing something else that isn't science. So you cannot claim your assertion is supported by science. This is what it seems to me that you are trying to do.

Anyway, to bring things back to the question of what is random. I think science uses the term random in reference to certain things beyond control (or prediction). And that intelligence, as understood by science, is a human phenomenon. Whether or not there is an intelligence (or supernatural intelligence), science is going to refer to the events such as mutations as random. Science can even refer to a car being driven through a an intersection by a conscious intelligent human as random if it is outside the scope of control and prediction, yet can influence the outcome of an experiment or the occurrence of an event. For example: someone is measuring the wind speed at the intersection and his results are subject to the random appearance of cars that pass through the intersection.

Randomness is a matter related to observation.

I think there are two types of 'randomness' that are common in science.

The first is where we don't know or don't care about the causes involved because the specific origin doesn't matter at the level of detail we are considering. So, in mutations, the cause may well be a chemical in the environment or a cosmic ray and the action on a specific DNA site may well be fairly specific. But only slight changes in conditions would lead to a very different mutation (different cell hit by the cosmic ray, different location on DNA changed, etc). This sensitive dependence on initial conditions means that it is impossible in practice to predict what mutation will happen, even though theoretically it might be possible if we had perfect knowledge and perfect models with accurate enough computability.

This seems to be the type of randomness you are focusing on and it *is* the typical aspect of randomness seen. Pretty much all of classical statistical mechanics is of this sort, for example. This is the type of randomness that happens when we toss a coin or throw dice. We *could*, given sufficient time, accurate information, and computing power, predict the results, but we don't want to go to that level of detail.

But, the type of randomness in quantum mechanics is of a different sort. Instead of an approximation obtained by ignoring aspects that are too difficult to compute, QM *is* the basic theory. And what the basic theory predicts is probabilities, not specifics. Even if we had perfect information about initial conditions and all the computing power we could want, the basic theoretical description *on a fundamental level* is probabilistic. There is randomness at the heart of the theory itself. And, based on all we know, in this regard any fundamental theory will have this aspect in it (yes, it is always a danger to make statements like this).

This is NOT a randomness because we are ignorant of the causes. It is not a randomness because we want to ignore the details for simplicity. It is not a situation where our instruments just aren't good enough. It is something about the basic structure of the universe at the quantum level. The randomness is written into the theory. And the theory works.

 

[Ponder This]

Actually, science does take a stand on untestable hypotheses: it declares them to be un-scientific. And such hypotheses *should* be considered with a great deal of skepticism: it is trivial to come up with multitudes of untestable ideas that should be immediately rejected. For example, perhaps angels are what drives the planets instead of gravity: they just do it in exactly the way gravity would.

Well, that's exactly the point, isn't it? They are unscientific as opposed to valid or invalid.
We shouldn't teach that angels drive the planets instead of gravity in a science class, because it isn't a scientific theory (as opposed to 'because it isn't true that angels drive the planets'). The theory isn't 'rejected' in the absence of an experiment that tests the hypothesis. You do understand that science requires that hypotheses be tested by experiment before being accepted or rejected, don't you?

But the fact that it is unnecessary and untestable *should* put it under a cloud of suspicion. Again, it is just too easy to come up with untestable ideas.

Indeed, I wonder if anyone has come up with a way to test the many worlds theory. If it is untestable and unnecessary, we should be skeptical. If it is testable, then it might be worthy of a footnote in a science text (as opposed to untestable hypotheses which are not worth a mention).

I think there are two types of 'randomness' that are common in science.

The first is where we don't know or don't care about the causes involved because the specific origin doesn't matter at the level of detail we are considering. So, in mutations, the cause may well be a chemical in the environment or a cosmic ray and the action on a specific DNA site may well be fairly specific. But only slight changes in conditions would lead to a very different mutation (different cell hit by the cosmic ray, different location on DNA changed, etc). This sensitive dependence on initial conditions means that it is impossible in practice to predict what mutation will happen, even though theoretically it might be possible if we had perfect knowledge and perfect models with accurate enough computability.

This seems to be the type of randomness you are focusing on and it *is* the typical aspect of randomness seen. Pretty much all of classical statistical mechanics is of this sort, for example. This is the type of randomness that happens when we toss a coin or throw dice. We *could*, given sufficient time, accurate information, and computing power, predict the results, but we don't want to go to that level of detail.

You seem to be confused about what randomness is in the 'classical' sense. 'Classical' Randomness isn't a separate 'type' of randomness. Gathering accurate (classical mechanical) information about things is not required to satisfy randomness. In fact, the opposite is true: we don't require that accurate information be gathered and predictions calculated - regardless of whether or not such predictions are possible, in theory. What does gathering information and calculating accurate results have to do with the 'classical' theory of randomness? Nothing.

For example, the computer generated random number is based upon perfect classical information. We know all the initial conditions and if we want we can calculate exactly what the outcomes will be. Why would such a thing ever be considered 'random'? Because classically perfect information is irrelevant to the notion of what it means for a thing to be 'random'.

But, the type of randomness in quantum mechanics is of a different sort. Instead of an approximation obtained by ignoring aspects that are too difficult to compute, QM *is* the basic theory. And what the basic theory predicts is probabilities, not specifics. Even if we had perfect information about initial conditions and all the computing power we could want, the basic theoretical description *on a fundamental level* is probabilistic. There is randomness at the heart of the theory itself. And, based on all we know, in this regard any fundamental theory will have this aspect in it (yes, it is always a danger to make statements like this).

Perfect classical mechanical information is not relevant to the notion of being random. You are confusing the simple distinction between classical and quantum mechanics with the notion of what it means for something to be random. Once upon a time, no one had the capability to classically predict the outcomes of dice rolls. It was simply impossible to do so.

This is NOT a randomness because we are ignorant of the causes. It is not a randomness because we want to ignore the details for simplicity. It is not a situation where our instruments just aren't good enough. It is something about the basic structure of the universe at the quantum level. The randomness is written into the theory. And the theory works.

Randomness has always been written into the theory without the need to make speculations about the nature of reality.

 

[Polymath257]

Well, that's exactly the point, isn't it? They are unscientific as opposed to valid or invalid.
We shouldn't teach that angels drive the planets instead of gravity in a science class, because it isn't a scientific theory (as opposed to 'because it isn't true that angels drive the planets'). The theory isn't 'rejected' in the absence of an experiment that tests the hypothesis. You do understand that science requires that hypotheses be tested by experiment before being accepted or rejected, don't you?

Yes, that is the point: we test to be sure our hypotheses are valid. We reject those ideas that cannot be tested as not even true or false. As such they are invalid as truth claims.

Indeed, I wonder if anyone has come up with a way to test the many worlds theory. If it is untestable and unnecessary, we should be skeptical. If it is testable, then it might be worthy of a footnote in a science text (as opposed to untestable hypotheses which are not worth a mention).

Most of the theories that involve a many-world hypothesis are testable in other particulars. But I agree: we should be skeptical of those aspects that cannot be tested.

You seem to be confused about what randomness is in the 'classical' sense. 'Classical' Randomness isn't a separate 'type' of randomness. Gathering accurate (classical mechanical) information about things is not required to satisfy randomness. In fact, the opposite is true: we don't require that accurate information be gathered and predictions calculated - regardless of whether or not such predictions are possible, in theory. What does gathering information and calculating accurate results have to do with the 'classical' theory of randomness? Nothing.

For example, the computer generated random number is based upon perfect classical information. We know all the initial conditions and if we want we can calculate exactly what the outcomes will be. Why would such a thing ever be considered 'random'? Because classically perfect information is irrelevant to the notion of what it means for a thing to be 'random'.[/QTE]

A computer generated 'random number' is usually very far from being *actually* random. if we had perfect knowledge, we *could* predict the results of most computer random number generators. That is why they are merely chaotic (at best).

Perfect classical mechanical information is not relevant to the notion of being random. You are confusing the simple distinction between classical and quantum mechanics with the notion of what it means for something to be random. Once upon a time, no one had the capability to classically predict the outcomes of dice rolls. It was simply impossible to do so.

Nobody had the *computational power* to predict the dice rolls. But the basic equations governing the motion of the dice are deterministic in classical mechanics. if you know the initial conditions and have enough computing power, you could predict the results. That is different than what happens in QM, where even if you have perfect knowledge of the initial state of a particle, you *cannot* predict, even with infinite computing power, what will happen. That *is* a different type of randomness.

Randomness has always been written into the theory without the need to make speculations about the nature of reality.

Simply not true. Newtonian dynamics is deterministic: if you had perfect knowledge of the state at one time (positions and velocities of everything), then all future behavior is determined. it would even be computable given perfect accuracy of the initial conditions.

That is NOT the case in QM.

 

[Ponder This]

Yes, that is the point: we test to be sure our hypotheses are valid. We reject those ideas that cannot be tested as not even true or false. As such they are invalid as truth claims.

Okay, I think we agree here. The claim that 'no intelligence is at work' is not validated (affirmed or rejected) by science because there is no experiment that tests it.

A computer generated 'random number' is usually very far from being *actually* random. if we had perfect knowledge, we *could* predict the results of most computer random number generators. That is why they are merely chaotic (at best).

You aren't making a statement about randomness. You are making a statement about prediction from perfect knowledge of an initial state.

Nobody had the *computational power* to predict the dice rolls. But the basic equations governing the motion of the dice are deterministic in classical mechanics. if you know the initial conditions and have enough computing power, you could predict the results. That is different than what happens in QM, where even if you have perfect knowledge of the initial state of a particle, you *cannot* predict, even with infinite computing power, what will happen. That *is* a different type of randomness.

Your statement about the capability to predict in a classical, deterministic system has nothing to do with randomness. Prediction of a deterministic system from perfect knowledge is a statement about knowability.
Your assertion about predictions in QM from perfect knowledge is also a statement about knowability.

Simply not true. Newtonian dynamics is deterministic: if you had perfect knowledge of the state at one time (positions and velocities of everything), then all future behavior is determined. it would even be computable given perfect accuracy of the initial conditions.

That is NOT the case in QM.

Again you keep talking about perfect knowledge as if this was important to the concept of randomness. It isn't. Randomness has been present in the theory of science for as long as the theory has described errors in measurement (error in measurement is always present, classically and quantum mechanically).

The distinction between 'apparent randomness' and 'inherent randomness' is a qualification of randomness in terms of what the theory could predict if we had perfect complete knowledge and is not a distinction that defines what it means to be 'random'.

I appreciate that you want to make a distinction between things that are apparently random and things that are inherently random. But I don't accept it as proven that QM represents inherent randomness. I'm not even sure if it matters if QM is inherently random.

For example, Lets say Bob and Alice want to use quantum entanglement to exchange private messages. The idea is that because of quantum entanglement, only Bob and Alice can know what the messages they exchange say. What does this really mean?
Well, if we know the states of the quantum particles, then we can decode the message. This is how Bob and Alice decode the messages. The secrecy between Bob and Alice is a consequence of the knowability of the quantum states. It's no different from rolling dice where only Bob and Alice have private control of the dice.