Does superdeterminism undermine the scientific method?

[PeterDonis]

That's what the superdeterminists would have you believe: that they have an irrefutable theory that has as much scientific validity as any other.

Something that is just a personal preference is not a theory. A theory makes testable predictions. So if they are claiming it's a theory, I would say they are incorrect.

Everything, of course, is personal choice.

I didn't say personal choice. I said personal preference. As contrasted with a theory that actually makes testable predictions. Of course you can ignore testable predictions that have been tested and found to be true (or false), as creationists, flat earthers, etc. do), but that's not the kind of "personal preference" I was talking about.

Let's do all science from now on assuming SD is correct. There is nothing to do.

But no superdeterminists actually do this. Sabine Hossenfelder hasn't stopped doing science because she believes superdeterminism is correct. You might not think she is doing "hard scientific work", but from your description I would say most theoretical (as opposed to experimental) physics is not "hard scientific work" by your criterion.

 

[PeterDonis]

That's what the superdeterminists would have you believe: that they have an irrefutable theory that has as much scientific validity as any other.

Note, btw, that in the post of mine that you were responding to here, I was not responding to superdeterminists; I was responding to the OP, who was asking for a "real reason" to believe that superdeterminism is wrong. Since superdeterminism makes no testable predictions, there isn't one. That was my point.

 

[WernerQH]

If that is the only reason superdeterminism is rejected by modern physics then superdeterminism should not be rejected. Is there another reason?

Sabine Hossenfelder argued that there is no evidence for the statistical independence "assumption". She might as well have said that there is no evidence for probability theory. Common sense demands an explanation for correlations, not for their absence. Superdeterminists claim a special license to disregard probability if they dislike the results. It is logically possible to eschew probability theory, but it means giving up most of science.

Free will is a bit of a red herring. (It's a meaningful concept, but not in physics.) One could base the settings in Bell-type experiments on a play by Shakespeare, or on Bach's music. The experiments would involve rather long causal chains, and superdeterminists haven't explained why deviations from statistical independence would have to be taken into account in this case, but can be ignored in drug trials, for example. Hossenfelder offers only a rather vague criterion that the effects show up only in the "quantum world"; obviously she doesn't want to jettison probability theory altogether.

Superdeterminism is an attempt at restoring locality and determinism to the description of one single class of experiments. But it fails badly with respect to locality, because in principle one would have to consider possible correlations of any experiment with the entire universe. In my view it makes more sense to give up locality and determinism, which are preconceptions firmly rooted in classical physics.

 

[PeroK]

Note, btw, that in the post of mine that you were responding to here, I was not responding to superdeterminists; I was responding to the OP, who was asking for a "real reason" to believe that superdeterminism is wrong. Since superdeterminism makes no testable predictions, there isn't one. That was my point.

Hossenfelder has written a paper where she claims (among other things) that SD is not an interpretation of QM and if enough effort went into SD research (!), then a new theory would emerge with testable predictions.

https://arxiv.org/pdf/2010.01324.pdf

To be fair to her, she is not trying to hide behind an unchallengeable theory that is impossible to test and cannot be disproved by definition.

 

[PeterDonis]

To be fair to her, she is not trying to hide behind an unchallengeable theory that is impossible to test and cannot be disproved by definition.

I agree she is not trying to, since she says that if enough effort were put into research in this area, it would result in a new theory with testable predictions. But that's an "if". Unless and until the "if" gets realized, the claimed new theory with testable predictions does not exist.

 

[Jarvis323]

The main issue, I believe, is what happens if we accept SD and ask what we do next? Let's do all science from now on assuming SD is correct. There is nothing to do. There is no substance behind the theory - not even a hint of how the laws of physics actually work. SD is indistinguishable from believing that everything is controlled by an unknowable and capricious deity. And, therefore, SD is (IMO) fundamentally anti-scientific. At best it's pseudo-science.

The thing is though that SD doesn't say anything at all could be explained with SD, it just says that correlations between the measurement device settings and the particle's state would be. And that would imply some kind of repeatable deterministic pattern. Maybe it would be a hopelessly complex and seemingly implausible one, but at least a consistent one. That's different from a deity being able to just change or control things at a whim at any instant.

Moreover, for a superdeterminist, there is no hard scientific work to be done. Hossenfelder has a paper where she plays about with a few concepts from QM and some statistics. But, it's no more than that. It bears no relation to the hard science and complex experiments that underpin conventional physics - such as the muon anomaly saga.

She does propose an approach for looking for evidence of the hypothesized correlations. Since those correlations would probably be highly complex and non-linear, and probably beyond our ability to de-tangle on paper, she thinks that big data and deep learning might be able to discover some hidden correlations linking the measurement device and the particles state somehow. But if you suppose we one day did find some of these hypothesized hiding non-linear correlations, however, I don't think we could use that as very strong evidence of SD specifically, because they wouldn't necessarily imply locality or determinism. Maybe the consciousness causes collapse people could somehow make the same kind of claim from that evidence for example. But it does seem potentially possible that evidence against the assumption of statistical independence of the measurement device settings and the particle could be found, if they truly are dependent. And if we did find that they are not independent, then it would beg further work to try to figure out how that happens.

 

[wittgenstein]

I feel a little intimidated by the expertise here. So please accept my questions even tho they may seem stupid to you.
1. Do you agree that Zellinger's objection is silly? He basically says that if we have no free will there can be no science. Computers solve problems all the time and they have no free will. Also, that implies that our thoughts are not the result of a biological computer (our brains). Zellinger implies that if our thoughts do not come from a source outside nature, they cannot be rational.
2. Please explain why superdeterminism requires a cosmic conspiracy. That seems to me to be based on a misunderstanding of the anthropic principle. Some people think that the anthropic principle puts humans at the center of reality. It does not. The fact that our Earth is in the goldilocks zone (not too hot and not to cold) is why there is life on earth. Did physics "conspire" to create us? NO! Also, evolution shows that order can arise from disorder. The fact that I can make rational decisions does not mean that some super intelligent agency at the big bang designed and conspired to create the rationality of some of my decisions.
I am sure I am misunderstanding something. Please show me where I am misunderstanding. This site has many interesting debates in it , but it is also about answering questions from intelligent but less informed people (such as myself) . To be honest I am not a fan of superdeterminism. I want someone to prove it wrong to me or at least make it seem outrageously improbable. Obviously, I am open to the possibility (though reluctantly) that superdeterminism is true.

 

[Fra]

Some additional reflections:

1. I personally disagree with Sabines analysis to use superdeterminism to escape Bells inequality: I don't think statistical independence is the problem, for me the more likely problem is our incomplete understanding of the nature of causality.

But Einsteins quest for a casaul mechanism is IMO rational. Correlations begs an explanation. But IMO the partitioning of probabilities into the sum over hidden variable is contains implicitly preconceptions on the nature of causality, not sure why, but this is rarely mentioned. It sticks out to me. But the alternative causal mechanism is not yet understood. But I think it's the place to look for progress. Superdeterminism itself has to me no mechanisms with explanatory value, as has been mentioned by others as well, it's more like an excuse to keep a loophole open, that will not be useful. I think it's the wrong hole.

2. Also generally, not having statistical independence does not imply superdeterminism. They other way is true though. There is a middle path, where the choices are partly free (ie guided) but not deterministically determined; this is how I see it. And agents "freedom" is to me similar to a random walk; or making a decision based on incomplete information. One can think of it as a free decision, but I see it just as a random choice - but a guided one. So it's not black or white for me.

/Fredrik

 

[PeterDonis]

Do you agree that Zellinger's objection is silly?

No.

He basically says that if we have no free will there can be no science.

No, that's not what he's saying. What he's saying is that if what Hossenfelder calls the assumption of Statistical Independence is false, we cannot draw inferences about what the laws of nature are from experimental results. By "freedom of the experimentalist" Zeilinger does not mean any mystical concept of "free will" but the straightforward assumption that there are no hidden correlations between measurement settings and preparations of systems to be measured that affect the measurement results. Superdeterminism denies that assumption; and if you deny that assumption, then you can't, for example, infer that QM is correct from experimental results that violate the Bell inequalities, because hidden superdeterministic correlations between the measurement settings and the preparations of the measured systems could have affected the results, so that they appear to us to support QM but are actually produced by a completely different underlying theory.

 

[Jarvis323]

1. Do you agree that Zellinger's objection is silly?

Here is my take, and I should prefsce it with the fact thst I'm not an expert on this and might have misconceptions.

Yes and no, depending on how you read it. If you read it literally, I think it's a bad argument. But in context, it makes some sense. It's just one person's philisophical opinion though.

My understanding is this: in science you usually can assume independence even though it's only approximately independent. For example, you might ignore Pluto's effect on the Earth's orbit around the Sun, or ignore the air pressure's effect on the fairness of a dice roll.

When we assume statistical independence of the measurement settings and the quantum particle, we think we can do so because we can try to decide those settings through as random of a process as possible. And we even have something called deterministic chaos which can be indishinguishable from true randomness. So you would think even if the universe were fully deterministic, our best attempt to randomize the measurement settings would lead to independence. If not, you have to figure out how the particles state can depend on a persons decisions, the flip of a coin miles away, etc.

Superdeterminism is saying, maybe we don't know how that is possible, but let's assume that the particle's future state can depend on those kinds of things in the events they are used to set up the device settings, even though we don't know how. And that would violate the assumption of the Bell test and invalidate it, opening up new possibilities to explain the results.

Doing this kind of thing in general seems problematic, because you could apply the same kind of thing to any case where an important unexplained scientific result that depends on what seems we would commonly think of as a valid assumption, in order to invalidate the unexplained result. So if you use ans accept superdeterminism argument frivilously, you can invalidate any scientific result.

So IMO this isn't necessarilly an argument just about determinisn or free will. It's about using these things as loopholes.

That said, questioning an assumption isn't such a bad thing, even if it seems valid. And thst is at the heart of what Sabine us advocating I think.

 

[wittgenstein]

I am confused. Zellinger said, " "We always implicitly assume the freedom of the experimentalist... This fundamental assumption is essential to doing science. If this were not true, then, I suggest, it would make no sense at all to ask nature questions in an experiment, since then nature could determine what our questions are, and that could guide our questions such that we arrive at a false picture of nature" It seems to me that he is saying that if our questions are predetermined (by nature/ our brains) they lack validity because they are not objective. You wrote, " but the straightforward assumption that there are no hidden correlations between measurement settings and preparations of systems to be measured that affect the measurement results. Superdeterminism denies that assumption; " I sincerely apologize but can you explain that to me, a university educated person, but physics is not my specialty. I value " Physics forum" because it is an opportunity for an interested layman to ask questions.

 

[wittgenstein]

I do have some speculations as to what you meant but it would be presumptuous for me to include them.

 

[Jarvis323]

I am confused. Zellinger said, " "We always implicitly assume the freedom of the experimentalist...

Right, but it doesn't mean that making the assumption necessitates it being completely true. Pluto does affect the Earth's orbit, but we can usually assume it doesn't. Even if we can't say that the experimentalist can do completely unbiased science, we can make the assumption (even if it is an approximation) that they can do unbiased enough science to get reliable results.

SD is claiming that in some particular case, that assumption isn't valid, and that no matter what we do, our results for this particular experiment become biased, and that invalidates the Bell test result, which opens up the possibilities the Bell test result would otherwise rule out.

So the problem here is using such a loophole could theoretically be done to invalidate any scientific result. But we would have to decide when is appropriate to use the loophole.

In the QM case, though, we are struggling to understand the results of the Bell test. And it makes some sense to just consider the possibility the assumption of independence isn't valid and explore the possibilities. Why not. But it isn't really usefull for making any solid conclusions. Unless at some point, some proof of that dependence springs up (even though some would seriously doubt that).

The issue is an issue in the philosophy of science, which is really more about, in this case I think, how we make decisions what is worth pursuing or not. I'm not sure belief or disbelief itself should be the real goal, just a proxy for confidence we're on the right track or not in future work.

We can certainly use the loophole here, to motivate a new look, and it doesn't break science. It's not an all or nothing type of thing, so yes the criticism is a bit silly if taken too seriously, but there is a point to it.

 

[PeterDonis]

It seems to me that he is saying that if our questions are predetermined (by nature/ our brains) they lack validity because they are not objective.

He's not saying superdeterminism says our brains are predetermined. (That would just be determinism.) He's saying that superdeterminism says that whatever it is that determines the measurement settings has some hidden connection with whatever it is that determines the state preparation of the system being measured. And according to superdeterminism, this must be true no matter what precautions we take to keep those two things independent. For example, even if we use random noise from distant pulsars to determine the measurement settings (which, note carefully, is a process that does not involve human brains at all, and can be set up to depend entirely on events outside the past light cone of the state preparation event, so there is no possibility of signals of any kind passing between the two), according to superdeterminism, the measurement settings still have hidden connections to the state preparation process. When he talks about "the freedom of the experimentalist", he means we assume that such outlandish hidden connections don't exist. But superdeterminism says they do.

 

[PeroK]

No, that's not what he's saying. What he's saying is that if what Hossenfelder calls the assumption of Statistical Independence is false, we cannot draw inferences about what the laws of nature are from experimental results. By "freedom of the experimentalist" Zeilinger does not mean any mystical concept of "free will" but the straightforward assumption that there are no hidden correlations between measurement settings and preparations of systems to be measured that affect the measurement results. Superdeterminism denies that assumption;

It's also to be noted that these outlandish hidden connections fail to manifest themselves anywhere else! The whole concept of SD is designed purely to avoid accepting QM. It serves no other purpose, or sheds any light on any aspect of physical phenomena.

It may also be worth contrasting SD with simple determinism - for example in a comparison between Earth and Mars. In simple determinism, the laws of physics determine the evolution of both planets largely independently. In SD, every particle on Earth is correlated with every particle on Mars. It very difficult to see how any amount of research could produce a mathematical basis for that. If the Earth and Mars have no independent physical processes, it's difficult to see how evolution of any plant or animal life is possible, let alone the apparently rational behaviour of intelligent life.

This is why I believe it's an empty theory that serves no purpose other than to allow some people to say that they don't believe QM.

 

[Lord Jestocost]

1. Do you agree that Zellinger's objection is silly? He basically says that if we have no free will there can be no science.

He is completely right. Scientists who make a lot of noise about a “superdeterministic world” often don’t realize that they themselves are an included part of this world. When using the term "superdeterminism", one should know how it is defined. Gerard 't Hooft in "The Cellular Automaton Interpretation of Quantum Mechanics":

"If a theory is deterministic all the way, it implies that not only all observed phenomena, but also the observers themselves are controlled by deterministic laws. They certainly have no ‘free will’, their actions all have roots in the past, even the distant past. ... The notion that, also the actions by experimenters and observers are controlled by deterministic laws, is called superdeterminism." [Italics in original, LJ]

At the end, one can stop here if one assumes that we have superdeterminism. All what one does, writes or thinks has at the end no influence on the future course of events. Superdetermism actually tells you that what you are doing is futile anyway. And even when you believe that what you are doing isn’t futile, you are simply pre-determined to believe in this way.

 

[wittgenstein]

So you are saying that we cannot be rational if we lack free will. I disagree. Computers are rational and lack free will. Also that requires that we do not get our ideas from our brains since they are part of nature and not supernatural. It is a common mistake to believe that a lack of free will means that nothing matters. Many things act on the world that have no free will. I hope you are wrong because that makes Zellinger's objection silly.

 

[wittgenstein]

You wrote," He is completely right. Scientists who make a lot of noise about a “superdeterministic world” often don’t realize that they themselves are an included part of this world" I disagree. That is exactly what scientists know. Zellinger however, implies that we are separate from nature. Evolution is not based on free will and yet it "designed" our brains a problem solving computer.

 

[kith]

@wittgenstein: Could you briefly outline your understanding of what superdeterminism is and why physicists are talking about it?

 

[Jarvis323]

Whether determinism excludes free will or not is a matter of philosophy and there are two schools, one which thinks it does and one whoch thinks it doesn't.

Anyway, some people have a personal philosophical issue with determinism, which should be separated from the debate in SD.

Many people would be very fine with, actually even strongly prefer, pure determinism and maybe only very reluctantly accept non-determinism, but still think SD is too much of a stretch.

 

[PeroK]

He is completely right. Scientists who make a lot of noise about a “superdeterministic world” often don’t realize that they themselves are an included part of this world. When using the term "superdeterminism", one should know how it is defined. Gerard 't Hooft in "The Cellular Automaton Interpretation of Quantum Mechanics":

"If a theory is deterministic all the way, it implies that not only all observed phenomena, but also the observers themselves are controlled by deterministic laws. They certainly have no ‘free will’, their actions all have roots in the past, even the distant past. ... The notion that, also the actions by experimenters and observers are controlled by deterministic laws, is called superdeterminism." [Italics in original, LJ]

I don't see how that definition on SD is enough to explain QM results. The key missing element is correlation between seemingly independent physical systems. For what it's worth, Wikipedia agrees with me on this:

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

"The term superdeterminism is misleading. Superdeterministic models are deterministic in the usual sense. But in addition to being deterministic, they also postulate correlations between the state that is measured and the measurement setting. "

For example, suppose experimenter A chooses something to do with English football results and experimenter B chooses Scottish football results to obtain data for an Bell-type experiment. It's one thing to say that these results are predetermined. It's quite another to say that they are correlated - in just the right way to make QM appear to be true.

Not least, because data cannot be correlated with all other data in every possible way. SD (to explain QM) requires that a very specific set of correlations emerge. For example, if we do the experiment one way it might require that Celtic and Chelsea have results that are correlated in some way. But, then, we change the way we use the data and from that moment on, Celtic and Chelsea have results that must be correlated in some different way.

SD says that some law of physics maintains this bizarre interrelationship where, for example, Celtic's winning streak comes to an end at precisely the time that it needs to in order that the data when used by the experimenters is correlated with photons in an experiment.

It's not that football results, experimental decisions and photon polarizations are all pre-determined; it's that they are all correlated in order to make QM appear true. It's not clear what form the laws of nature would have to take to ensure such correlated interrelationships.

 

[wittgenstein]

PeroK wrote "It may also be worth contrasting SD with simple determinism - for example in a comparison between Earth and Mars. In simple determinism, the laws of physics determine the evolution of both planets largely independently. In SD, every particle on Earth is correlated with every particle on Mars. " Pardon the imprecise metaphor. Suppose at the big bang God (I am once again using a metaphor. It's just an easy way to imply whatever caused the initial conditions.) God puts 2 gloves in separate boxes, one left-handed and the other right-handed) 13 billion years later I open a box and instantly know what the other glove is. Is your point that the odds of those 2 gloves being in the same vicinity are astronomically unlikely?

 

[Jarvis323]

It may also be worth contrasting SD with simple determinism - for example in a comparison between Earth and Mars. In simple determinism, the laws of physics determine the evolution of both planets largely independently. In SD, every particle on Earth is correlated with every particle on Mars. It very difficult to see how any amount of research could produce a mathematical basis for that. If the Earth and Mars have no independent physical processes, it's difficult to see how evolution of any plant or animal life is possible, let alone the apparently rational behaviour of intelligent life.

But if this were what SD was saying, then SD would be easily disproved and be incompatible with chaos theory wouldn't it?

I think SD, at least Sabine's version of it, is trying to be minimally presumptuous. It's just saying the correlation exists when we measure the particle, and technically that could just extend to some hidden connection with whatever went into those specific measurements and the particle. We would need to examine the union of all factors that have gone into Bell tests, and consider a minimal common denominator that needs a connection to produce the bias. Or something of that sort.

 

[Lord Jestocost]

I don's see how that definition on SD is enough to explain QM results.

@PeroK

To avoid any misunderstanding:

As a former experimental physicist, I regard "superdeterminism" as an irrelevant idea. I am in line with Zeilinger: “[W]e always implicitly assume the freedom of the experimentalist... This fundamental assumption is essential to doing science.”

 

[PeroK]

PeroK wrote "It may also be worth contrasting SD with simple determinism - for example in a comparison between Earth and Mars. In simple determinism, the laws of physics determine the evolution of both planets largely independently. In SD, every particle on Earth is correlated with every particle on Mars. " Pardon the imprecise metaphor. Suppose at the big bang God (I am once again using a metaphor. It's just an easy way to imply whatever caused the initial conditions.) God puts 2 gloves in separate boxes, one left-handed and the other right-handed) 13 billion years later I open a box and instantly know what the other glove is. Is your point that the odds of those 2 gloves being in the same vicinity are astronomically unlikely?

That wasn't my point at all. SD, if true, has nothing directly to do with QM entanglement. As per the Wikipedia page, it's not the determinism that is the issue, it's the correlations.

To take an example from mathematics. Take the digits of $\sqrt 2$ and $\sqrt 3$. These sets of digits are completely determined in advance. But, the two sets of digits are uncorrelated. Suppose you pick a digit somewhere in the decimal expansion or $\sqrt 2$ and it is $5$, say. You cannot then say that the correspoding digit of $\sqrt 3$ is more likely to be one digit than any other.

The analogy in the physical universe is that initial conditions produce two complex systems: Earth and Mars. Their respective evolutions might be completely pre-determined, but you wouldn't expect the data to be correlated. We imagine, therefore, that we may obtain statistically independent data from separate complex systems. We assume that the data in each system has been effectively randomised.

SD (e.g. the Hossenfelder paper) challenges this and claims that all physical data is statistically dependent. This is an incredible claim. And, goes way beyond any simply detereministic theory (whether or not that theory treats humans as ultimately deterministic systems).

 

[wittgenstein]

So we do not get our thoughts from our brains? Or are you saying that our brains transcend matter? Can computers perform rational operations or are they also required to have free will?

 

[wittgenstein]

Are you saying that our brains do not follow the laws of cause and effect?

 

[wittgenstein]

That was in response to Lord Jestocost. I disagree with him and believe that free will is not required for rationality.

 

[PeroK]

@PeroK

To avoid any misunderstanding:

As a former experimental physicist, I regard "superdeterminism" as an irrelevant idea. I am in line with Zeilinger: “[W]e always implicitly assume the freedom of the experimentalist... This fundamental assumption is essential to doing science.”

SD claims way more than loss of freedom. It claims that the decision-making is not just predetermined, but pre-determined in such a way that human decision making is correlated with all other data in the universe.

To take my example. You might be predetermined to choose "Chelsea" and when they score their first goal of the match. And I might be predetermined to choose "Celtic" and when they score their first goal. You get $17$ and I get $64$ say. But, what the hell has 17 to do with 64? It appears completely random. And, why would that be correlated with the polarization on a pair of photons? (That works quite well for a choice of angles, given there are 90 minutes in a game of football!)

To persevere with this line. We generate violations of Bell's inequalities. So, we decide that from now on it should be when Celtic first concede a goal. And now, lo and behold, this data is correlated in just the right way. Nature is not just predetermining our decision making, but somehow aligning it (in some mysterious, hidden way) with anything and everything else in the universe.

In short, it's not clear how data from multiple complex systems could possibly be correlated in this way.

 

[CoolMint]

What your brain does, is the 'free will'. At the cellular level, brain cells send electrical signals via ligands attaching to cell receptors. A ligand can induce a cell reaction that may involve production of specific hormones that drive physiological functions(sleep, appetite, etc.) Or it may create a specific human behavior - e.g. irritability, impaired cognition, depression, joy.
As for higher thoughts - these things also have their roots in cell signaling. It is a learned(acquired) behaviour.
Your brain is your free will. It drives you.