Tuesday, May 16, 2017

QBism: Fuchs vs Bohr+Motl

Quantum Bayesianist Christopher Fuchs wrote a laughable, 43-page-long reply (titled Notwithstanding Bohr, the Reasons for QBism) to two blog posts of mine,
Bohr, Heisenberg, Landau wouldn't find QBism new [116]

Is quantum reality "personal"? [117]
which tries to claim that he has found something that Bohr and other founders of quantum mechanics didn't know about the meaning of the laws of quantum mechanics and the probabilities that they predict. Fuchs thanks two people who live in "time portals to our history", several other uninteresting names,
and Luboš Motl for showing off just how poor the scholarship on this subject can be in some corners of physics [116, 117].
Because of an extreme time and sleeping deficit (days of hosting Richard Lindzen and his wife, including a rather intense yesterday's trip to Prague where Lindzen gave a wonderful talk masterminded by your humble correspondent, hosted by Czech ex-president Václav Klaus, and we ate in two expensive restaurants and meeting with a top archaeologist at noon and Václav Klaus and his aides in the evening, new phone I just received, and many other things), I won't read this preprint carefully and I think that credible physicists won't read it, either, but the abstract will be enough for them to be rather certain that I am right and Fuchs is wrong: He just hasn't added anything on top of Bohr that would make sense.

OK, Niels Bohr famously said
It is wrong to think that the task of physics is to find out how Nature is. Physics concerns what we say about Nature.
It's very important because the sentence explains the actual difference between classical and quantum mechanics. Classical physics allows you to assume that some things objectively exist. You may make true statements about the objects in Nature but there are underlying objects and all the true statements are just reflections of something that is out there.

Quantum mechanics – but to some extent, it was already true in classical statistical physics as well – allows you to assign truth values or probabilities (a continuous version of the truth value) to propositions about Nature, too. However, you can no longer assume that the true statements that you may derive from quantum mechanics are reflections of the objective reality.

That makes sense. Quantum mechanics is an analogy of mathematical logic that allows you to prove and derive new valid propositions out of some assumed ones, the axioms. To make the story short, Fuchs would like to go further and tell you what you should do with the derived truths or probabilities. What you should do is to gamble.

The key part of the abstract tries to define QBism as follows:
Along the way, we lay out three tenets of QBism in some detail:
  1. The Born Rule—the foundation of what quantum theory means for QBism—is a normative statement. It is about the decision-making behavior any individual agent should strive for; it is not a descriptive “law of nature” in the usual sense.
  2. All probabilities, including all quantum probabilities, are so subjective they never tell nature what to do. This includes probability-1 assignments. Quantum states thus have no “ontic hold” on the world.
  3. Quantum measurement outcomes just are personal experiences for the agent gambling upon them. Particularly, quantum measurement outcomes are not, to paraphrase Bohr, instances of “irreversible amplification in devices whose design is communicable in common language suitably refined by the terminology of classical physics.”
Look at the first slogan that is supposed to describe the most important difference between Fuchs and Bohr. Bohr accepts Born's rule, of course, and says that physics may produce true statements about Nature. Fuchs also claims to embrace Born's rule but his new "contribution" is the principle that Born's rule is a normative statement.

In other words, it's a statement expressing a value judgement, telling you what is desirable, what is not, what you should do, and perhaps, what is morally right, too.

Please, give me a break. Natural science just doesn't discuss normative statements. Period. The only quantitative "science" that discusses normative statements is economics – it is doing so because economics needs to analyze humans and their desires (i.e. what is desirable for them). Instead, in modern physics, if you calculate that assuming some knowledge that you have because you have observed it, the probability of the event AB is the number XY, science obviously doesn't tell you what you should do with it. Science doesn't tell you to become a missionary or spread your wealth or regulate carbon dioxide or (and it's what Fuchs tells you) become a gambler. In general, the word "desirable" is simply ill-defined in natural science.

(BTW concerning gambling, on Friday, when it became pretty clear to everybody that I was right and most market analysts and even board members of the national bank were wrong about the safety of the bet on the strengthening Czech currency, my broker forced me to close my position – by abruptly and unjustifiably increasing the required margin by a factor of twenty – so I had to go away with a 5 times smaller profit than what I would have otherwise.)

The second proposition among the three says that probabilities are subjective – probabilities believed to be adequate by someone are always subjective (Bayesian) and this fact has been known for a very long time. Probabilities may also be "measured" by a repetition of the same experiment – in that way, we obtain the "frequentist" probabilities. These two aspects of probabilities and their relationships have been known and discussed for centuries and Fuchs isn't adding anything new here, either. To say the least, he is adding nothing that would be both new and coherent.

Indeed, probabilities in quantum mechanics don't tell Nature what to do. Not only this statement wouldn't be new for Bohr. Bohr has made it. He urged Einstein to "[s]top telling God what to do with his dice." But just like the laws of physics don't tell Nature what to do, they don't tell humans what to do, either.

In the third principle above, "Quantum measurement outcomes just are personal experiences for the agent gambling upon them." Up to the word "gambling", it's the standard Copenhagen talk once again. What Fuchs is adding is that the observer should be "gambling". How is this verb exactly defined? What activities of the observer count as gambling? And why the difference between gambling and other acts should matter? It makes no sense. Probabilities are continuous generalizations of truth values and both truth values and probabilities represent unemotional knowledge (in the case of probabilities, it is uncertain knowledge). It is utterly stupid to connect this quantitative knowledge with normative statements or recommended activities.

OK, I don't want to read the rest of the paper – and I don't even want to read the rest of the abstract carefully – because it seems like a pure waste of time. It's very clear that everything by which Fuchs tries to differentiate from Bohr is either ill-defined or downright anti-scientific or both.

Sorry, Mr Fuchs, but you are not Bohr. And you are not even myself. Your frantic efforts to place yourself above the two of us are childish and no credible physicist will take you seriously. You would have been a rather rare member of the "interpretation community" if you understood why Bohr was right about the principles sketched above. But you're as foolish as they are if you think that you have found something important and true that the founding fathers didn't know.

No comments:

Post a Comment