Interpretations of quantum mechanics are not really something that a lot of physicists worry about and, in my humble opinion, they aren’t that interesting anyway. The idea of ‘interpretations’ is mostly a relic from the days of Bohr and Einstein. Their approach to physics was quite different from today. Virtually everyone agrees that the act of measurement is nothing more than the entanglement of the ‘experiment’ system with the ‘observer’ system, and that as far as physics is concerned, that’s the only thing that matters. You can call it ‘wavefunction collapse’ or ‘choosing a world’ or whatever you want, you’re talking about the same idea. In fact physicists often use these various metaphors interchangeably.
There are a few slightly more interesting concepts that are related to interpretations, such as if hidden variable theories are plausible (they’d have to be non-local and super-deterministic to be plausible) or if Bohmian pilot wave mechanics is an adequate mechanism for reality (it’s not yet known whether it can incorporate special relativity, and that’s really really important for modern physics since the standard model is intimately tied with special relativity).
I guess you could say that the many-worlds interpretation is the closest to a ‘consistent’ interpretation. But there’s a danger here. Just because it’s a consistent interpretation doesn’t mean that there must actually be an infinite number of ‘worlds’. At the end of the day, it could turn out that quantum mechanics is just the result of some unknown, deeper theory, and that there is just a single world.
Virtually everyone agrees that the act of measurement is nothing more than the entanglement of the ‘experiment’ system with the ‘observer’ system, and that as far as physics is concerned, that’s the only thing that matters.
As I pointed out, Hugh Everett does not agree.
He describes observation as being performed by an automatically functioning with sensory gear and a memory.
The observer has to create a memory record in memory for a measurement to occur.
Quantum theory was barely understood when Everett wrote his thesis. In particular, quantum mechanics as operator theory on Hilbert spaces was only starting to become understood, and Bell’s theorem had not yet been proven. An article written 50 years ago has little bearing on what physicists think today.
But anyway, the distinction of ‘creating a memory’ does not apply when you consider the observer and experiment together as a single quantum system. All quantum systems are reversible and follow unitary transformation laws. This means that no information can ever be lost or created within a quantum system.
But anyway, the distinction of ‘creating a memory’ does not apply when you consider the observer and experiment together as a single quantum system.
Why not?
Let’s say I am looking at a clock.
I’m a physical thing, interacting with another physical thing. You can consider me+clock to be a single physical system.
I still record the measurement in my memory. I still remember looking at the clock. That doesn’t magically go away.
Accroding to Everett, his idea is to “deduce the subjective appearance of phenomena” by looking that contents of my memory.
In other words, Everett’s model does not make a prediction until a measurement record is created. He then suggests these measurment records are consistent with our empirical observations, and also the equivalient to the predictions derived from a collapse.
If the memory of an observer is destroyed before its measurements can be deduced, then the model doesn’t have any measurement records (ie, no predictions).
If you have some actual reason for thinking that the Everett interpretation is bad, why don’t you give it here rather than just saying you don’t like it?
[EDITED to add: Ah, I see that what you dislike isn’t the Everett interpretation but de Witt’s formulation of it. Fair enough, but the same remark applies mutatis mutandis.]
Here is one reason, you say “Everett Interpretation” and “DeWitt’s Formulation”, but you’ve got that crossed.
Everett provided the Relative State Formulation. That’s what he called his paper, which is a bold and provocative attempt to mathematically model two ontologies, absolute and relative state, unlike one that is traditional in mathematics.
DeWitt provided the Many Worlds Interpretation, that was the name of their paper.
While the MWI lends itself to some interesting things, far and away, Everett’s models lend themselves to the stuff we actually work on today, computer vision and computer hearing. He left theoretical physics to work on machine learning. Because that’s how reality worked to him.
The type of thing he writes about on Page 9 is coming within our grasp, technologically, for the first time.
This idea seems to be anticipated by Leibniz, and Plato, and Kant, and a great many other traditions wherein our objective reality is a relative reality based on an absolute reality.
So Everett’s Formulation gives us new mathematical territory to explore, which we have been exploring independently as machine learning. And the result is model that contains an observer that produces a measurement record. Probably not far off.
Are all interpretations of QM equally wrong?
Is there not one that is less wrong?
Is there not one that is truer to the themes of rationalism?
What is the difference between rationalism and empiricism?
Interpretations of quantum mechanics are not really something that a lot of physicists worry about and, in my humble opinion, they aren’t that interesting anyway. The idea of ‘interpretations’ is mostly a relic from the days of Bohr and Einstein. Their approach to physics was quite different from today. Virtually everyone agrees that the act of measurement is nothing more than the entanglement of the ‘experiment’ system with the ‘observer’ system, and that as far as physics is concerned, that’s the only thing that matters. You can call it ‘wavefunction collapse’ or ‘choosing a world’ or whatever you want, you’re talking about the same idea. In fact physicists often use these various metaphors interchangeably.
There are a few slightly more interesting concepts that are related to interpretations, such as if hidden variable theories are plausible (they’d have to be non-local and super-deterministic to be plausible) or if Bohmian pilot wave mechanics is an adequate mechanism for reality (it’s not yet known whether it can incorporate special relativity, and that’s really really important for modern physics since the standard model is intimately tied with special relativity).
I guess you could say that the many-worlds interpretation is the closest to a ‘consistent’ interpretation. But there’s a danger here. Just because it’s a consistent interpretation doesn’t mean that there must actually be an infinite number of ‘worlds’. At the end of the day, it could turn out that quantum mechanics is just the result of some unknown, deeper theory, and that there is just a single world.
Have you actually talked to physicists about this? I have and that is the opposite of my conclusion.
As I pointed out, Hugh Everett does not agree.
He describes observation as being performed by an automatically functioning with sensory gear and a memory.
The observer has to create a memory record in memory for a measurement to occur.
It’s on page 9 on his thesis:
http://philosophyfaculty.ucsd.edu/faculty/wuthrich/PhilPhys/EverettHugh1957PhDThesis_BarrettComments.pdf
Quantum theory was barely understood when Everett wrote his thesis. In particular, quantum mechanics as operator theory on Hilbert spaces was only starting to become understood, and Bell’s theorem had not yet been proven. An article written 50 years ago has little bearing on what physicists think today.
But anyway, the distinction of ‘creating a memory’ does not apply when you consider the observer and experiment together as a single quantum system. All quantum systems are reversible and follow unitary transformation laws. This means that no information can ever be lost or created within a quantum system.
Why not?
Let’s say I am looking at a clock.
I’m a physical thing, interacting with another physical thing. You can consider me+clock to be a single physical system.
I still record the measurement in my memory. I still remember looking at the clock. That doesn’t magically go away.
Accroding to Everett, his idea is to “deduce the subjective appearance of phenomena” by looking that contents of my memory.
In other words, Everett’s model does not make a prediction until a measurement record is created. He then suggests these measurment records are consistent with our empirical observations, and also the equivalient to the predictions derived from a collapse.
Yes, but what if a memory is created and then destroyed?
If the memory of an observer is destroyed before its measurements can be deduced, then the model doesn’t have any measurement records (ie, no predictions).
You haven’t seen that one big sequence that mostly argues for one of the QM interpretations, have you?
http://lesswrong.com/lw/r8/and_the_winner_is_manyworlds/
Ugh. DeWitt strikes again.
If you have some actual reason for thinking that the Everett interpretation is bad, why don’t you give it here rather than just saying you don’t like it?
[EDITED to add: Ah, I see that what you dislike isn’t the Everett interpretation but de Witt’s formulation of it. Fair enough, but the same remark applies mutatis mutandis.]
Indeed, I should.
Here is one reason, you say “Everett Interpretation” and “DeWitt’s Formulation”, but you’ve got that crossed.
Everett provided the Relative State Formulation. That’s what he called his paper, which is a bold and provocative attempt to mathematically model two ontologies, absolute and relative state, unlike one that is traditional in mathematics.
DeWitt provided the Many Worlds Interpretation, that was the name of their paper.
While the MWI lends itself to some interesting things, far and away, Everett’s models lend themselves to the stuff we actually work on today, computer vision and computer hearing. He left theoretical physics to work on machine learning. Because that’s how reality worked to him.
The type of thing he writes about on Page 9 is coming within our grasp, technologically, for the first time.
This idea seems to be anticipated by Leibniz, and Plato, and Kant, and a great many other traditions wherein our objective reality is a relative reality based on an absolute reality.
So Everett’s Formulation gives us new mathematical territory to explore, which we have been exploring independently as machine learning. And the result is model that contains an observer that produces a measurement record. Probably not far off.