7 ms·

I've always been mystified at why people take superdeterminism as a good explanation for anything, when it actually acts as a stopsign for thought. For example

by knzhou 7y ago

I've always been mystified at why people take superdeterminism as a good explanation for anything, when it actually acts as a stopsign for thought. For example, suppose I am betting with you on the outcomes of a coin toss. You call heads, the coin comes up tails, you give me some money. You call tails, the coin comes up heads, you give me some more money. This repeats a number of times. The common sense response is to start suspecting that I am rigging the game, i.e. that the simple model of the coin flips being independent and random is wrong. The superdeterminist response is to say, "no, your guesses have been superdetermined to be all wrong by the initial conditions of the universe -- keep on playing!" You can use superdeterminism to explain literally anything (not merely a lot, but actually, literally everything) while gaining precisely zero insight into what is actually going on, and precluding any future insight. Sure, it fixes a few philosophical problems, but so does saying you're a brain in a vat; in both cases you throw out all of science as a side effect. While their paper does address some philosophical objections, it doesn't even mention this scientific objection, which seems far more severe to me.

sysbin 7y ago
Your example given with the conclusion to follow doesn't make sense. I'm a hard determinist, I've known others like myself and nobody leaps to the thought, "the initial conditions of the universe means this is all meaningless" and we do try to observe every possibility such as the game being rigged. I don't necessarily think superdeterminism can explain literally anything or that if it did explain something there would be no gain of insight. That just seems like you're postulating.
- knzhou 7y ago
  You're not distinguishing between determinism and superdeterminism. Determinism is perfectly compatible with science, as it admits the idea of controlled and reproducible experiments. (For example, if you believe in determinism, you could grab the coin out of my hand and run some tests on it yourself.) The whole point of superdeterminism is to go further and deny this.
  - perl4ever 7y ago
    I'm definitely not distinguishing between determinism and superdeterminism. The wikipedia page on superdeterminism starts out: "In quantum mechanics, superdeterminism is a hypothetical class of theories that evade Bell's theorem by virtue of being completely deterministic" Logically, something that is "completely deterministic" would have to be...deterministic, no? Does superdeterminism go to 11?
    - knzhou 7y ago
      Superdeterminism is determinism, plus a radically different way of thinking about experimental results. To continue with the coin flip analogy, consider a classical world where a coin flip is perfectly deterministic, given the initial conditions. In the usual way we think about determinism, it's still possible to test whether a coin is fair. Just flip it a bunch of times, a little differently each time, and see if you get a ratio close to 50/50. That's standard science. In the superdeterministic view, you don't admit this kind of explanation. Instead you say that whatever outcome happened was determined by the initial conditions of the entire universe. If you get all heads, you can't say anything about the coin -- we merely conclude that your hands were predetermined to give it the kinds of initial conditions that would have led to that outcome. Now imagine repeating an explanation like this for every experiment ever performed. It's on the level of answering "God did it" to every question. We don't do that in science because it kind of defeats the whole point.
      - username90 7y ago
        > It's on the level of answering "God did it" to every question. And saying that it is random with "God rolled a dice" isn't? The entire universe could have used a fixed random seed and our experiments and thoughts would have been exactly the same for all we know, determinism doesn't have the effects you talk about. We perform trials and do statistics when we don't know the exact workings of things, not because we believe that the thing is inherently random.
        knzhou 7y ago
ReadEvalPost 7y ago
I don't really understand what your coin flip metaphor is supposed to mean in the context of the paper. That Superdeterminism could be used in an entirely facile way doesn't really suggest anything about the theory itself. When choosing Superdeterminism as an ontology it suggests different paths for doing research, paths that are unexplored and potentially full of insight. That's what is exciting about it to me.
- knzhou 7y ago
  While I can think of exciting questions sparked by many different interpretations of quantum mechanics, I can't think of any for superdeterminism, for exactly the reasons I said -- it's scientifically dead from the start. Other ideas have limits, superdeterminism doesn't. Do you have a good example?
  - npo9 7y ago
    In Superdeterminism the obvious line of research is to look for hidden variables when we find things that are seemingly probabilistic or random. If we found hidden variables to control when a radioactive isotope will decay we could possibly have much safer handling of fission materials, depending on what these hypothetical hidden variables are. * I don’t have an opinion on Superdeterminism, I’m just thinking about hypotheticals out loud.
    - knzhou 7y ago
      Hidden variables would be cool. But there are plenty of hidden variable theories that aren't superdeterministic. My point is that if we all genuinely adopted the superdeterministic way of thinking and explaining, we wouldn't even do experiments in the first place.
      - npo9 7y ago
        Wouldn’t we? If Superdeterminism is correct than it wouldn’t be up to us if we continue to experiment. Maybe we are compelled to try to understand the world.
millstone 7y ago
The linked paper addresses this objection in a fun way: perhaps the tobacco company lawyers argue in court that "smoking doesn't cause cancer; every experiment was super-determined to assign those pre-disposed to cancer into the smoking group!" The paper replies with the following argument: in super-determinism, statistical independence is an emergent property which holds at macro scales, so classical experiments do enjoy statistical independence. In your hypothesis - yeah, the game is rigged! But superdeterminism comes into play at the quantum scale where we know things are already kind of fucked, which is why we're talking about this at all. The idea is that God places Her thumb on the quantum scale alone. You might try to foil Her with an experiment exquisitely sensitive to a quantum measurement: the cat lives or dies according to the millionth decimal place of some measurement. But actually "cat lives" and "cat dies" might be very close in state-space - in the sense that two arbitrary real numbers might have some arithmetic relationship which cannot be proven (Chaitin's incompleteness theorem).
- knzhou 7y ago
  > The paper replies with the following argument: in super-determinism, statistical independence is an emergent property which holds at macro scales, so classical experiments do enjoy statistical independence. In your hypothesis - yeah, the game is rigged! > But superdeterminism comes into play at the quantum scale where we know things are already kind of fucked, which is why we're talking about this at all. But experiments already do use classical (i.e. macroscopic) sources of randomness. For example, to oversimplify a little, you might literally flip a coin to decide the specific measurements done in a run of a Bell test experiment. Such experiments still give puzzling results, so if superdeterminism is to work as an explanation of anything at all, it must hold all the way up. Superdeterministic explanations have to apply even to things like coin flips, which, again, leads to the end of science, wailing and gnashing of teeth, etc.
  - trevyn 7y ago
    Perhaps all forms of randomness are pseudorandomness of various quality levels, which turns out to be good enough? It would be the end of the current interpretation of what science means, but the process and results of science could carry on as they always have. There is also the glimmer of post-science, in which we might be able to find a process/method that is more effective/efficient than the scientific in some domains. And you’re right — there will certainly be wailing and gnashing of teeth. :-)
mycall 7y ago
A coin can have 51%/49% probabilities based on weight of one side.
coldtea 7y ago
>The common sense response is to start suspecting that I am rigging the game, i.e. that the simple model of the coin flips being independent and random is wrong. The superdeterminist response is to say, "no, your guesses have been superdetermined to be all wrong by the initial conditions of the universe -- keep on playing!" Err, whether you "keep on playing" or not, would also part of it. Besides, lacking a source of actual randomness (even non superdeterminist QM doesn't provide one), plain on determinism seems plausible even without needing to explain quantum mechanics. Besides if we did had a source of randomness it would still be just randomness. Where would "free will" come from (it being something different than randomness, essentially an agent non-determined by the prior state of the universe nor one's prior history, making decisions inside one's head).
jayd16 7y ago
Is this a proper understanding of superdeterminism? I may be totally off base but my understanding is the analogy is more like Craps. Two dice with fair odds create an uneven distribution such that 7 is more common. The dice have "conspired" to bring 7 up more often than 2 because their states interact in a non-linear way.
- caf 7y ago
  Ahem, there isn't anything more linear than the sum of two variables...
  - jayd16 7y ago
    Its just an analogy. Something to show that a distribution might appear interesting but is actually not. The point is that Bell's theorem relies on the assumption that a linear distribution must be expected and the paper is arguing that that has not been sufficiently proven. At least that's my understanding of the paper's argument.
    - akvadrako 7y ago
      Bell's theorem does not rely on that.
      - jayd16 7y ago
        Can you explain how? Wikipedia states: >With the measurements oriented at intermediate angles between these basic cases, the existence of local hidden variables could agree with/would be consistent with a linear dependence of the correlation in the angle but, according to Bell's inequality (see below), could not agree with the dependence predicted by quantum mechanical theory, namely, that the correlation is the negative cosine of the angle.
        akvadrako 7y ago
        What that sentence says is: Bell's theorem predicts the linear dependence of the correlations angles - or at least it's consistent with such a model. It's not an assumption of the theory. Also keep in mind we don't observe linear correlations - at least one of the assumptions of Bell's theorem must be wrong.
Rury 7y ago
>I've always been mystified at why people take superdeterminism as a good explanation for anything, when it actually acts as a stopsign for thought. What? It's like the opposite... Currently we don't have a perfect model for quantum physics, so what do we do when we try to create a model of something where we have data for, but we don't fully understand? We use statistics and probabilities to fit/approximate the behavior behind the model. Doing this can lead to a useful but imperfect model, but that doesn't prove one way or another that the behavior of what we're trying to understand is inherently random, just that it exhibits random behavior at some level or in some way. Now the opposite: saying things like Heisenberg's uncertainty principle is full stop proven, is itself a stop sign for thought IMO. Whereas opposing such principles, and assuming behavior isn't inherently random (i.e. the idea behind determinism), opens for testable ideas on what might actually happening behind the apparent randomness of the model. You know... one which might lead to a perfect model fully explaining the behavior (which hey isn't that the purpose of science??) To put it backwards... the ideal in science is a model that explains how everything works, if this is ever realized it'll prove superdeterminism. To say stuff is inherently random, to me is equivalent to giving up and settling for less than the ideal...
- knzhou 7y ago
  I must not have put this clearly, because I've gotten piled on by 50 comments, but: there is nothing wrong with trying to find a deterministic explanation for QM. Lots of people are working on that. Superdeterministism is very very very very very different!

by knzhou 7y ago

sysbin 7y ago

Your example given with the conclusion to follow doesn't make sense. I'm a hard determinist, I've known others like myself and nobody leaps to the thought, "the initial conditions of the universe means this is all meaningless" and we do try to observe every possibility such as the game being rigged. I don't necessarily think superdeterminism can explain literally anything or that if it did explain something there would be no gain of insight. That just seems like you're postulating.

knzhou 7y ago
You're not distinguishing between determinism and superdeterminism. Determinism is perfectly compatible with science, as it admits the idea of controlled and reproducible experiments. (For example, if you believe in determinism, you could grab the coin out of my hand and run some tests on it yourself.) The whole point of superdeterminism is to go further and deny this.
- perl4ever 7y ago
  I'm definitely not distinguishing between determinism and superdeterminism. The wikipedia page on superdeterminism starts out: "In quantum mechanics, superdeterminism is a hypothetical class of theories that evade Bell's theorem by virtue of being completely deterministic" Logically, something that is "completely deterministic" would have to be...deterministic, no? Does superdeterminism go to 11?
  - knzhou 7y ago
    Superdeterminism is determinism, plus a radically different way of thinking about experimental results. To continue with the coin flip analogy, consider a classical world where a coin flip is perfectly deterministic, given the initial conditions. In the usual way we think about determinism, it's still possible to test whether a coin is fair. Just flip it a bunch of times, a little differently each time, and see if you get a ratio close to 50/50. That's standard science. In the superdeterministic view, you don't admit this kind of explanation. Instead you say that whatever outcome happened was determined by the initial conditions of the entire universe. If you get all heads, you can't say anything about the coin -- we merely conclude that your hands were predetermined to give it the kinds of initial conditions that would have led to that outcome. Now imagine repeating an explanation like this for every experiment ever performed. It's on the level of answering "God did it" to every question. We don't do that in science because it kind of defeats the whole point.
    - username90 7y ago
      > It's on the level of answering "God did it" to every question. And saying that it is random with "God rolled a dice" isn't? The entire universe could have used a fixed random seed and our experiments and thoughts would have been exactly the same for all we know, determinism doesn't have the effects you talk about. We perform trials and do statistics when we don't know the exact workings of things, not because we believe that the thing is inherently random.
      - knzhou 7y ago

ReadEvalPost 7y ago

I don't really understand what your coin flip metaphor is supposed to mean in the context of the paper. That Superdeterminism could be used in an entirely facile way doesn't really suggest anything about the theory itself. When choosing Superdeterminism as an ontology it suggests different paths for doing research, paths that are unexplored and potentially full of insight. That's what is exciting about it to me.

knzhou 7y ago
While I can think of exciting questions sparked by many different interpretations of quantum mechanics, I can't think of any for superdeterminism, for exactly the reasons I said -- it's scientifically dead from the start. Other ideas have limits, superdeterminism doesn't. Do you have a good example?
- npo9 7y ago
  In Superdeterminism the obvious line of research is to look for hidden variables when we find things that are seemingly probabilistic or random. If we found hidden variables to control when a radioactive isotope will decay we could possibly have much safer handling of fission materials, depending on what these hypothetical hidden variables are. * I don’t have an opinion on Superdeterminism, I’m just thinking about hypotheticals out loud.
  - knzhou 7y ago
    Hidden variables would be cool. But there are plenty of hidden variable theories that aren't superdeterministic. My point is that if we all genuinely adopted the superdeterministic way of thinking and explaining, we wouldn't even do experiments in the first place.
    - npo9 7y ago
      Wouldn’t we? If Superdeterminism is correct than it wouldn’t be up to us if we continue to experiment. Maybe we are compelled to try to understand the world.

millstone 7y ago

The linked paper addresses this objection in a fun way: perhaps the tobacco company lawyers argue in court that "smoking doesn't cause cancer; every experiment was super-determined to assign those pre-disposed to cancer into the smoking group!" The paper replies with the following argument: in super-determinism, statistical independence is an emergent property which holds at macro scales, so classical experiments do enjoy statistical independence. In your hypothesis - yeah, the game is rigged! But superdeterminism comes into play at the quantum scale where we know things are already kind of fucked, which is why we're talking about this at all. The idea is that God places Her thumb on the quantum scale alone. You might try to foil Her with an experiment exquisitely sensitive to a quantum measurement: the cat lives or dies according to the millionth decimal place of some measurement. But actually "cat lives" and "cat dies" might be very close in state-space - in the sense that two arbitrary real numbers might have some arithmetic relationship which cannot be proven (Chaitin's incompleteness theorem).

knzhou 7y ago
> The paper replies with the following argument: in super-determinism, statistical independence is an emergent property which holds at macro scales, so classical experiments do enjoy statistical independence. In your hypothesis - yeah, the game is rigged! > But superdeterminism comes into play at the quantum scale where we know things are already kind of fucked, which is why we're talking about this at all. But experiments already do use classical (i.e. macroscopic) sources of randomness. For example, to oversimplify a little, you might literally flip a coin to decide the specific measurements done in a run of a Bell test experiment. Such experiments still give puzzling results, so if superdeterminism is to work as an explanation of anything at all, it must hold all the way up. Superdeterministic explanations have to apply even to things like coin flips, which, again, leads to the end of science, wailing and gnashing of teeth, etc.
- trevyn 7y ago
  Perhaps all forms of randomness are pseudorandomness of various quality levels, which turns out to be good enough? It would be the end of the current interpretation of what science means, but the process and results of science could carry on as they always have. There is also the glimmer of post-science, in which we might be able to find a process/method that is more effective/efficient than the scientific in some domains. And you’re right — there will certainly be wailing and gnashing of teeth. :-)

mycall 7y ago

A coin can have 51%/49% probabilities based on weight of one side.

coldtea 7y ago

>The common sense response is to start suspecting that I am rigging the game, i.e. that the simple model of the coin flips being independent and random is wrong. The superdeterminist response is to say, "no, your guesses have been superdetermined to be all wrong by the initial conditions of the universe -- keep on playing!" Err, whether you "keep on playing" or not, would also part of it. Besides, lacking a source of actual randomness (even non superdeterminist QM doesn't provide one), plain on determinism seems plausible even without needing to explain quantum mechanics. Besides if we did had a source of randomness it would still be just randomness. Where would "free will" come from (it being something different than randomness, essentially an agent non-determined by the prior state of the universe nor one's prior history, making decisions inside one's head).

jayd16 7y ago

Is this a proper understanding of superdeterminism? I may be totally off base but my understanding is the analogy is more like Craps. Two dice with fair odds create an uneven distribution such that 7 is more common. The dice have "conspired" to bring 7 up more often than 2 because their states interact in a non-linear way.

caf 7y ago
Ahem, there isn't anything more linear than the sum of two variables...
- jayd16 7y ago
  Its just an analogy. Something to show that a distribution might appear interesting but is actually not. The point is that Bell's theorem relies on the assumption that a linear distribution must be expected and the paper is arguing that that has not been sufficiently proven. At least that's my understanding of the paper's argument.
  - akvadrako 7y ago
    Bell's theorem does not rely on that.
    - jayd16 7y ago
      Can you explain how? Wikipedia states: >With the measurements oriented at intermediate angles between these basic cases, the existence of local hidden variables could agree with/would be consistent with a linear dependence of the correlation in the angle but, according to Bell's inequality (see below), could not agree with the dependence predicted by quantum mechanical theory, namely, that the correlation is the negative cosine of the angle.
      - akvadrako 7y ago
        What that sentence says is: Bell's theorem predicts the linear dependence of the correlations angles - or at least it's consistent with such a model. It's not an assumption of the theory. Also keep in mind we don't observe linear correlations - at least one of the assumptions of Bell's theorem must be wrong.

Rury 7y ago

>I've always been mystified at why people take superdeterminism as a good explanation for anything, when it actually acts as a stopsign for thought. What? It's like the opposite... Currently we don't have a perfect model for quantum physics, so what do we do when we try to create a model of something where we have data for, but we don't fully understand? We use statistics and probabilities to fit/approximate the behavior behind the model. Doing this can lead to a useful but imperfect model, but that doesn't prove one way or another that the behavior of what we're trying to understand is inherently random, just that it exhibits random behavior at some level or in some way. Now the opposite: saying things like Heisenberg's uncertainty principle is full stop proven, is itself a stop sign for thought IMO. Whereas opposing such principles, and assuming behavior isn't inherently random (i.e. the idea behind determinism), opens for testable ideas on what might actually happening behind the apparent randomness of the model. You know... one which might lead to a perfect model fully explaining the behavior (which hey isn't that the purpose of science??) To put it backwards... the ideal in science is a model that explains how everything works, if this is ever realized it'll prove superdeterminism. To say stuff is inherently random, to me is equivalent to giving up and settling for less than the ideal...

knzhou 7y ago
I must not have put this clearly, because I've gotten piled on by 50 comments, but: there is nothing wrong with trying to find a deterministic explanation for QM. Lots of people are working on that. Superdeterministism is very very very very very different!