4 ms·

Superdeterminism needs some kind of simplicity, for lack of a better word. To take her vaccine trial analogy further, imagine that you randomly assign a group o

Superdeterminism needs some kind of simplicity, for lack of a better word. To take her vaccine trial analogy further, imagine that you randomly assign a group of 100 people to a treatment and control cell (50-50). It could happen that all 50 people in the control cell just happened to be the 50 most a priori healthy people in the group. But it’s extremely unlikely. Following the analogy, superdeterminism says it happens anyways. But the level of coincidence involved invokes such an absurdly complex series of coincidences, that your coin flip comes up heads on these particular people, that occam’s razor might as well spontaneously cease to exist. Are there any superdeterministic theories that maintain the level of simplicity that the universe is posited to have without it?

candiodari 5y ago
The point is that it's only unlikely (or other situations likely) if the events, even from patient to patient, are random. If whether a patient is sick or not determined if they're in the placebo group or the healthy group, you would be testing exactly nothing. The randomness in placing the patient is a required part of the experiment. And more than random, the patient placement must be random and independent from their disease status (misdiagnosed healthy patients get treated too, after all. Generally it works quite well, as there isn't any problem to solve) and whether the treatment will work. And of course superdeterministic theories are "simple". In the same sense that you can model any sequence as a function: just list all observed values for all time. Very simple, very little predictive power when you need it. That also gives you the criticism of superdeterminism: if we do things that way, nothing needs an explanation, as it just can't be simplified from that full specification that we observe ... so there's no point to science. Note that it can still be more complex than we'll ever usefully realise, so whilst it means everything is determined, it does not necessarily (in fact quite unlikely imho) make it predictable for us limited beings. But, in at least one meaning of the word, it's certainly simple. Every function in physics, whether quantum gravity or electroweak forces or the number of puppies your cat will have next year: f(x) = the x'th entry in the table ... (tables, of course, that you generally don't have access to) Which isn't to say this is a strategy that can't work and perform useful functions. Take a robot for example, and look at inverse kinematics. That comes with 3d calculus, volume intersection (robot mustn't self-intersect), and there's not just limits in position but in speed, acceleration and torque as well. You can fully develop this theory. Or you can move the robot, observe all variables and save them to disk, then use nearest neighbour to figure out robot movement. Works like a charm, and doesn't even need to know how many arms the robot has. And while it takes a while, I bet it's a hell of a lot faster than building the theory for even the simplest of robots, never mind multi-armed or nonlinear robots. Plus it's trivial to make a machine learning algorithm do it for you, and the same can't be said of developing the correct theory for a new robot model.
- dfilppi 5y ago
- 6gvONxR4sf7o 5y ago
  The randomness here is a stand in for complexity. Substitute some other volume measure and the “probability” logic still works out, meaning that you can interpret “unlikely” as meaning “only happens in a teensy tiny sliver of phase space.” And with any reasonable way of trying to do randomness (coin flipping, PRNGs) it also means that it’s not just any sliver of phase space, it’s an extraordinarily complex sliver of it. Hence the original statement about how superdeterminism theories can potentially be. And hence the question about whether any of today’s superdeterministic theories avoid this extraordinary complexity issue.
naasking 5y ago
> But the level of coincidence involved invokes such an absurdly complex series of coincidences, that your coin flip comes up heads on these particular people, that occam’s razor might as well spontaneously cease to exist. Sure, but the experiments you have to run to see these absurd coincidence are also absurdly contrived, so is that really surprising? It's worth noting that Sabine's take on superdeterminism is also a bit different than past takes that are rife with "conspiracy", so I suggest reading her papers.

candiodari 5y ago

The point is that it's only unlikely (or other situations likely) if the events, even from patient to patient, are random. If whether a patient is sick or not determined if they're in the placebo group or the healthy group, you would be testing exactly nothing. The randomness in placing the patient is a required part of the experiment. And more than random, the patient placement must be random and independent from their disease status (misdiagnosed healthy patients get treated too, after all. Generally it works quite well, as there isn't any problem to solve) and whether the treatment will work. And of course superdeterministic theories are "simple". In the same sense that you can model any sequence as a function: just list all observed values for all time. Very simple, very little predictive power when you need it. That also gives you the criticism of superdeterminism: if we do things that way, nothing needs an explanation, as it just can't be simplified from that full specification that we observe ... so there's no point to science. Note that it can still be more complex than we'll ever usefully realise, so whilst it means everything is determined, it does not necessarily (in fact quite unlikely imho) make it predictable for us limited beings. But, in at least one meaning of the word, it's certainly simple. Every function in physics, whether quantum gravity or electroweak forces or the number of puppies your cat will have next year: f(x) = the x'th entry in the table ... (tables, of course, that you generally don't have access to) Which isn't to say this is a strategy that can't work and perform useful functions. Take a robot for example, and look at inverse kinematics. That comes with 3d calculus, volume intersection (robot mustn't self-intersect), and there's not just limits in position but in speed, acceleration and torque as well. You can fully develop this theory. Or you can move the robot, observe all variables and save them to disk, then use nearest neighbour to figure out robot movement. Works like a charm, and doesn't even need to know how many arms the robot has. And while it takes a while, I bet it's a hell of a lot faster than building the theory for even the simplest of robots, never mind multi-armed or nonlinear robots. Plus it's trivial to make a machine learning algorithm do it for you, and the same can't be said of developing the correct theory for a new robot model.

dfilppi 5y ago
6gvONxR4sf7o 5y ago
The randomness here is a stand in for complexity. Substitute some other volume measure and the “probability” logic still works out, meaning that you can interpret “unlikely” as meaning “only happens in a teensy tiny sliver of phase space.” And with any reasonable way of trying to do randomness (coin flipping, PRNGs) it also means that it’s not just any sliver of phase space, it’s an extraordinarily complex sliver of it. Hence the original statement about how superdeterminism theories can potentially be. And hence the question about whether any of today’s superdeterministic theories avoid this extraordinary complexity issue.

naasking 5y ago

> But the level of coincidence involved invokes such an absurdly complex series of coincidences, that your coin flip comes up heads on these particular people, that occam’s razor might as well spontaneously cease to exist. Sure, but the experiments you have to run to see these absurd coincidence are also absurdly contrived, so is that really surprising? It's worth noting that Sabine's take on superdeterminism is also a bit different than past takes that are rife with "conspiracy", so I suggest reading her papers.