Ancestors

Written by Daniel Lakeland on 2025-01-19 at 04:59

https://statmodeling.stat.columbia.edu/2025/01/12/string-theory-wars-an-opinion-field-inversion/#comment-2387954

@david_j_marcus_phd

Not being expert in QM, I don't see what's the big deal about the series of Stern Gerlach experiments mentioned by Anonymous... Split the beam along the z axis, then split the z-up beam along the x axis.

Now we could do one of two experiments... Either split the x-up beam along the z axis (and get 50-50 z-up/down), or restore the split x-up/down beams to a single beam and measure z up as 100% z-up.

Anon says ...

=> More informations about this toot | More toots from dlakelan@mastodon.sdf.org

Toot

Written by Daniel Lakeland on 2025-01-19 at 04:59

"You cannot say that “different experiments are different” now. It’s measuring the same quantity of the same object (the z spin)."

But that's just wrong. Clearly the experiment where I don't combine the x-up/down beams is different from the one where I do combine the x-up/down beams, as there are very different apparatus in the path. It's mysterious to me why physicists think like this.

=> More informations about this toot | More toots from dlakelan@mastodon.sdf.org

Descendants

Written by CEDunn on 2025-01-19 at 05:46

@dlakelan

Are you asking rhetorically, or do you want to discuss?

=> More informations about this toot | More toots from cdunnpasadena@sfba.social

Written by Daniel Lakeland on 2025-01-19 at 06:00

@cdunnpasadena

I'm up for discussion yeah.

=> More informations about this toot | More toots from dlakelan@mastodon.sdf.org

Written by CEDunn on 2025-01-19 at 06:22

@dlakelan

The reason this bunch of experiments shows up in QM classes is to demonstrate that the behavior of subatomic particles requires new physics to describe.

They are supposed to show how electrons don’t behave the way that a bunch of tiny charged spheres (or any other shape) would behave according to classical electrodynamics, and to show that, mathematically, the departure is pretty simple and can be modeled with 2-dimensional, complex, linear algebra that tracks the “quantum state” of the electrons.

You are right that if you put the beam back together, it is a new experiment, but the point is that the result is not one you would expect from classical physics.

=> More informations about this toot | More toots from cdunnpasadena@sfba.social

Written by Daniel Lakeland on 2025-01-19 at 06:32

@cdunnpasadena

Right, I don't deny that QM makes different predictions from Newton + Maxwell.

The context of all of this though is the original statement where it was asserted that "the usual rules of conditional probability fail". And this David and I and Carlos at the original blog have repeatedly argued is simply not true. And these Stern-Gerlach experiments don't change any of that. For context it may be useful (or just annoying) to review a variety of comments in the thread.

=> More informations about this toot | More toots from dlakelan@mastodon.sdf.org

Written by Daniel Lakeland on 2025-01-19 at 06:32

@cdunnpasadena

But the core point is, we shouldn't "expect" the split vs the combined beams to be somehow the same, just as in my totally classical example, we shouldn't necessarily expect the plastic bead to float on the water across surface waves the same in an experiment with one weir vs one with two weirs. Or expect the "rules of arithmetic" to apply to piping 10 acre feet into a reservoir for 7 days implying 70 acre feet increase in the reservoir (because evaporation or withdrawal or etc)

=> More informations about this toot | More toots from dlakelan@mastodon.sdf.org

Written by Daniel Lakeland on 2025-01-19 at 06:32

@cdunnpasadena

That is, in a Bayesian context there's some probabilities for unknown inputs, there's some predictor apparatus, and there's some set of probabilities for outcomes. If the observed outcomes don't agree with the predicted ones it's either because your assumptions are wrong about the unknown inputs, or your predictor is wrong, it's not because "probability doesn't work"

=> More informations about this toot | More toots from dlakelan@mastodon.sdf.org

Written by CEDunn on 2025-01-19 at 06:52

@dlakelan

I haven’t read all of the background, so I apologize if I haven’t understood. I think that you are right.

My comments are based on grading QM as a PhD student in the 80s, and the thing that would be key is that the probabilities in QM behave differently than initial state dispersion or instrumental error in CM. I don’t know Bayesian analysis, but once you have allowed for the internal states of the particle the interpretation of the experiment relies on probabilistic analysis the same as a classical experiment. I can say this with confidence as I got my PhD in particle physics and spent my career doing spacecraft.

=> More informations about this toot | More toots from cdunnpasadena@sfba.social

Written by Daniel Lakeland on 2025-01-19 at 07:27

@cdunnpasadena

Thanks, that's helpful. Because of the apparently straightforward way in which Bayes can apply, and because the particles are "be-ables" as Bell would call them, I confess to having a preference for Bohmian mechanics. And I think this sometimes gets peoples hackles up because almost no physics textbooks teach it. But basically, you'd start with particles entering an apparatus at an unknown location, you'd have a wave on configuration space that moves the particles through ...

=> More informations about this toot | More toots from dlakelan@mastodon.sdf.org

Written by Daniel Lakeland on 2025-01-19 at 07:27

@cdunnpasadena

The apparatus. The wave would be different if you installed different magnets and detectors and beam splitters and things, and so the behavior of the particle would be different for the different experiments. Then in the end it'd hit your detector. Whether you choose to split the beam and detect, or split the beam, recombine the beam, and detect would of course result in potentially different outcomes.

=> More informations about this toot | More toots from dlakelan@mastodon.sdf.org

Written by Daniel Lakeland on 2025-01-19 at 07:27

@cdunnpasadena

But from a probability perspective this is no different from say shooting a billiard ball directly into a pocket, or first having it hit some elliptical obstruction... Putting physically different apparatus leads to different outcomes. We are only led astray if we think the behavior of the particle is somehow dependant on the particle alone, rather than the particle and it's entire environment within which the wave operates.

=> More informations about this toot | More toots from dlakelan@mastodon.sdf.org

Written by David J. Marcus, Ph.D. on 2025-01-19 at 13:53

@dlakelan Bohmian mechanics is the least nonlocal of the alternatives.

=> More informations about this toot | More toots from david_j_marcus_phd@mastodon.social

Written by David J. Marcus, Ph.D. on 2025-01-19 at 14:50

@dlakelan Yes, you certainly can say that different experiments are different. You may be trying to measure what you think is the same thing, but the world may not accommodate you. What does "philosophically controversial" mean? It seems to mean "many people are confused and I'm about to add more confusion". Anonymous said the correct statement is "no local hidden variable model”, but the correct statement (as Bell clearly stated) is "no local theory”.

=> More informations about this toot | More toots from david_j_marcus_phd@mastodon.social