5 ms·

What is need is less arguing and more experimenting. That's how this will be resolved.

willis936 7y ago
No no, arguing is important. It lubricates an essential step in the scientific process. No one can design experiments, let alone predict results or even understand results, without understanding.
- miltondts 7y ago
  Let me clarify. In this particular case there has been plenty of arguing already and no progress in more than 50 years. So I'm suggesting it is time to start experimenting more.
  - skosch 7y ago
    I've been partial to superdeterminism for a long time, but I have no idea how one would go about testing it. In fact, it seems as unfalsifiable as many-worlds or other theories. Do you have a suggestion?
    - miltondts 7y ago
      The kind of superdeterministic theories proposed by t'Hooft are falsifiable. From [0] "If engineers ever succeed in making such quantum computers, it seems to me that the CAT is falsified; no classical theory can explain quantum mechanics." By "such quantum computers" he means computers that can run Shor's algorithm. "...but factoring a number with millions of digits into its prime factors will not be possible – unless fundamentally improved classical algorithms turn out to exist." As for the author of the article I've never seen a clear proposal but it appears the idea is to do repeated measurements that display quantum effects while reducing noise as much as possible and check if there are deviations from quantum theory. EDIT: found a more concrete proposal http://backreaction.blogspot.com/2013/10/testing-conspiracy-theories.html http://backreaction.blogspot.com/2013/10/testing-conspiracy-... [0] - https://arxiv.org/abs/1405.1548 https://arxiv.org/abs/1405.1548
    - baddox 7y ago
      My understanding was that many-worlds could be tested experimentally if we were able to set up large objects in superpositions, and I thought that there is no reason to expect that it isn’t physically possible to do so (we just don’t know how at the moment).
      - akvadrako 7y ago
        You can’t get any information from that because the main interpretations all make the same predictions from an outside perspective, which is what you’ll expect to see. Besides to put a person in a superposition may require a machine as large as the universe, so you get issues with the speed of light. The only way to test MWI is multiverse immortality - many worlds means you should expect to always have some future experience - there is no real death.
        GoblinSlayer 7y ago
        Some version of you will expect it, some versions won't.
      - GoblinSlayer 7y ago
        >set up large objects in superpositions EPR experiment does exactly that. In fact, it's a good experiment to get intuitive understanding of MWI.
    - magicalhippo 7y ago
      Hossenfelder proposes[1] to measure non-commuting variables in identical systems in as noise-free an environment as possible. In standard QM says the measurements should be completely uncorrelated, but she argues that in a superdeterministic theory results somewhat correlated. [1]: https://arxiv.org/abs/1105.4326 https://arxiv.org/abs/1105.4326
  - qayxc 7y ago
    The authors don't argue that there's not enough experiments, though. They argue it's wrong type of experiments that are conducted. To quote Dr. Hossenfelder directly: "In standard quantum mechanics the measurement outcomes will be non-correlated. In a superdeterministric hidden variables theory, they'll be correlated - provided you can make a case that the hidden variables don't change in between the measurements." [1] That last sentence is the catch here: in case the experiment fails to show any correlation, it can always be argued that the hidden variables changed for whatever reason. If the calculated theoretical boundaries (e.g. temperature & measurement time) are insufficient, there's still no way of telling systematic errors from falsifying the initial hypothesis. It's little details like this that theorists can hide behind while still shouting "Foul!" from the peanut gallery. Since experiments cost time, money, and pin down talent, research facilities need to be picky about what they test. "Because I like it." [2] is not the most compelling argument when trying to make your case ;) I wonder whether crowd-funding would work in this case... [1] https://backreaction.blogspot.com/2013/10/testing-conspiracy-theories.html https://backreaction.blogspot.com/2013/10/testing-conspiracy... [2] https://www.fetzer-franklin-fund.org/wp-content/uploads/2014/12/Hossenfelder-EmQM131.pdf https://www.fetzer-franklin-fund.org/wp-content/uploads/2014...
  - normalnorm 7y ago
    This has been the mainstream position for most of those 50 years. Working on the "foundations of physics", which is what you call "arguing" was considered disreputable and career-destroying for a long time. Read, for example, "Something Deeply Hidden" by Sean Caroll for more about this. A lot of really really smart people tried to solve this by experimenting more. It didn't work. It's time for philosophy again, and in my view, also to accept that the weirdness is not going away. Nature doesn't care about what we find weird or not.
  - GoblinSlayer 7y ago
    The progress is stiffled by belief that this is not a problem and there should be no arguing and no experimenting and everybody should shut up and calculate.
  - de_watcher 7y ago
    Superdeterminism says that arguing is a part of the experiment anyway.
nwallin 7y ago
You can't. Not just the way it's arbitrarily hard from an engineering perspective to design an experiment to disprove string theory. It is theoretically impossible, the way it's impossible to distinguish between Copenhagen and many worlds.
- vilhelm_s 7y ago
  No, you actually can. The authors of this article discuss the issue more in their longer paper here: https://arxiv.org/abs/1912.06462 https://arxiv.org/abs/1912.06462 The idea is that superdeterminstic theories are deterministic, while quantum mechanical measurements are random, so you should be able to set up an experiment where QM predicts you would just get random results, but actually you get the same result each time.
  - bosswipe 7y ago
    There are many classical physics experiments you could do where you understand all the physics and formulas to excruciating detail and you control every variable but you still can't get the same results everytime. Things like throwing dice, as mentioned in the article, or the chaos generated by flowing fluids. Given that any superdeterministic mechanisms would be even even more complex and weird and unknown it seems impossible to disprove. Any failed experiment could be excused by saying that there are more unknown uncontroled variables.
    - zelphirkalt 7y ago
      Well, the determinist would argue, that, if you cannot predict the result reliably every time, you do not actually know all the variables, even if you think you do.
  - GoblinSlayer 7y ago
    Setting up such an experiment means to control hidden variables, but then they aren't hidden.
    - 7y ago

No no, arguing is important. It lubricates an essential step in the scientific process. No one can design experiments, let alone predict results or even understand results, without understanding.

miltondts 7y ago
Let me clarify. In this particular case there has been plenty of arguing already and no progress in more than 50 years. So I'm suggesting it is time to start experimenting more.
- skosch 7y ago
  I've been partial to superdeterminism for a long time, but I have no idea how one would go about testing it. In fact, it seems as unfalsifiable as many-worlds or other theories. Do you have a suggestion?
  - miltondts 7y ago
    The kind of superdeterministic theories proposed by t'Hooft are falsifiable. From [0] "If engineers ever succeed in making such quantum computers, it seems to me that the CAT is falsified; no classical theory can explain quantum mechanics." By "such quantum computers" he means computers that can run Shor's algorithm. "...but factoring a number with millions of digits into its prime factors will not be possible – unless fundamentally improved classical algorithms turn out to exist." As for the author of the article I've never seen a clear proposal but it appears the idea is to do repeated measurements that display quantum effects while reducing noise as much as possible and check if there are deviations from quantum theory. EDIT: found a more concrete proposal http://backreaction.blogspot.com/2013/10/testing-conspiracy-theories.html http://backreaction.blogspot.com/2013/10/testing-conspiracy-... [0] - https://arxiv.org/abs/1405.1548 https://arxiv.org/abs/1405.1548
  - baddox 7y ago
    My understanding was that many-worlds could be tested experimentally if we were able to set up large objects in superpositions, and I thought that there is no reason to expect that it isn’t physically possible to do so (we just don’t know how at the moment).
    - akvadrako 7y ago
      You can’t get any information from that because the main interpretations all make the same predictions from an outside perspective, which is what you’ll expect to see. Besides to put a person in a superposition may require a machine as large as the universe, so you get issues with the speed of light. The only way to test MWI is multiverse immortality - many worlds means you should expect to always have some future experience - there is no real death.
      - GoblinSlayer 7y ago
        Some version of you will expect it, some versions won't.
    - GoblinSlayer 7y ago
      >set up large objects in superpositions EPR experiment does exactly that. In fact, it's a good experiment to get intuitive understanding of MWI.
  - magicalhippo 7y ago
    Hossenfelder proposes[1] to measure non-commuting variables in identical systems in as noise-free an environment as possible. In standard QM says the measurements should be completely uncorrelated, but she argues that in a superdeterministic theory results somewhat correlated. [1]: https://arxiv.org/abs/1105.4326 https://arxiv.org/abs/1105.4326
- qayxc 7y ago
  The authors don't argue that there's not enough experiments, though. They argue it's wrong type of experiments that are conducted. To quote Dr. Hossenfelder directly: "In standard quantum mechanics the measurement outcomes will be non-correlated. In a superdeterministric hidden variables theory, they'll be correlated - provided you can make a case that the hidden variables don't change in between the measurements." [1] That last sentence is the catch here: in case the experiment fails to show any correlation, it can always be argued that the hidden variables changed for whatever reason. If the calculated theoretical boundaries (e.g. temperature & measurement time) are insufficient, there's still no way of telling systematic errors from falsifying the initial hypothesis. It's little details like this that theorists can hide behind while still shouting "Foul!" from the peanut gallery. Since experiments cost time, money, and pin down talent, research facilities need to be picky about what they test. "Because I like it." [2] is not the most compelling argument when trying to make your case ;) I wonder whether crowd-funding would work in this case... [1] https://backreaction.blogspot.com/2013/10/testing-conspiracy-theories.html https://backreaction.blogspot.com/2013/10/testing-conspiracy... [2] https://www.fetzer-franklin-fund.org/wp-content/uploads/2014/12/Hossenfelder-EmQM131.pdf https://www.fetzer-franklin-fund.org/wp-content/uploads/2014...
- normalnorm 7y ago
  This has been the mainstream position for most of those 50 years. Working on the "foundations of physics", which is what you call "arguing" was considered disreputable and career-destroying for a long time. Read, for example, "Something Deeply Hidden" by Sean Caroll for more about this. A lot of really really smart people tried to solve this by experimenting more. It didn't work. It's time for philosophy again, and in my view, also to accept that the weirdness is not going away. Nature doesn't care about what we find weird or not.
- GoblinSlayer 7y ago
  The progress is stiffled by belief that this is not a problem and there should be no arguing and no experimenting and everybody should shut up and calculate.
- de_watcher 7y ago
  Superdeterminism says that arguing is a part of the experiment anyway.

nwallin 7y ago

You can't. Not just the way it's arbitrarily hard from an engineering perspective to design an experiment to disprove string theory. It is theoretically impossible, the way it's impossible to distinguish between Copenhagen and many worlds.

vilhelm_s 7y ago
No, you actually can. The authors of this article discuss the issue more in their longer paper here: https://arxiv.org/abs/1912.06462 https://arxiv.org/abs/1912.06462 The idea is that superdeterminstic theories are deterministic, while quantum mechanical measurements are random, so you should be able to set up an experiment where QM predicts you would just get random results, but actually you get the same result each time.
- bosswipe 7y ago
  There are many classical physics experiments you could do where you understand all the physics and formulas to excruciating detail and you control every variable but you still can't get the same results everytime. Things like throwing dice, as mentioned in the article, or the chaos generated by flowing fluids. Given that any superdeterministic mechanisms would be even even more complex and weird and unknown it seems impossible to disprove. Any failed experiment could be excused by saying that there are more unknown uncontroled variables.
  - zelphirkalt 7y ago
    Well, the determinist would argue, that, if you cannot predict the result reliably every time, you do not actually know all the variables, even if you think you do.
- GoblinSlayer 7y ago
  Setting up such an experiment means to control hidden variables, but then they aren't hidden.
  - 7y ago