Notwithstanding Bohr, the Reasons for QBism


Without Niels Bohr, QBism would be nothing. But QBism is not Bohr. This paper attempts to show that, despite a popular misconception, QBism is no minor tweak to Bohr's interpretation of quantum mechanics. It is something quite distinct. 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." Finally, an explicit comparison is given between QBism and Bohr with regard to three subjects: a) The issue of the "detached observer" as it arose in a debate between Pauli and Bohr, b) Bohr's reply to Einstein, Podolsky, and Rosen, and c) Bohr's mature notion of "quantum phenomena." At the end, we discuss how Bohr's notion of phenomena may have something to offer the philosophy of William James: A physics from which to further develop his vision of the world---call it an ontology if you will---in which "new being comes in local spots and patches."
Submitted 9 May 2017 to Quantum Physics [quant-ph]
Published 11 May 2017
Author comments: 43 pages, 1 figure


Planat May 11 2017 09:03 UTC (2 points)

Dear Christopher,

1. Could you comment on the connection to the fine structure constant in footnote 15 in which you write "Implicit in it is the number 137!"?

2. Would the Qbism philosophy be destroyed by restricting to IC's instead of SICs as in


Edited May 11 2017 16:08 UTC by Steve Flammia

Christopher A. Fuchs May 12 2017 17:14 UTC (4 points)

Dear Michel,

1. It was just a goofy thing that I thought would get the readers to smile. But Wolfgang Pauli did have quite a mystical interest in 137 precisely because of its connection to the fine structure constant. This is documented in quite a number of places; the book by Suzanne Gieser, "The Innermost Kernel: Depth Psychology and Quantum Physics. Wolfgang Pauli's Dialogue with C. G. Jung," is quite a good source. I just quoted from something I could easily get my hands on at my desk. Preceding the quote I placed in the paper, Enz wrote, "The depth of Pauli's philosophical views incited Arthur Koestler to remark that Pauli `perhaps had a deeper knowledge of the limits of the natural sciences than most of his colleagues'. One of these limits which disturbed Pauli during all his scientific life was the duality between the electric field strength and the means to measure it by its action on a charge ... The disturbing aspect is that the precision of such a measurement is limited by the atomicity of electric charges as given by Sommerfeld's fine structure constant 1/137, which has not yet been explained."

2. That one might *think* or *get the impression* that QBism will be destroyed as an interpretation of quantum mechanics if the SICs do not exist in every finite dimension is a perennial worry of my colleagues Ruediger Schack and David Mermin. They don't want people to get that impression. They feel I place too much emphasis on the mathematical problem since, as an interpretation of QM, QBism need not depend on any technical innovations in the formalism. And they are correct: QBism, as a self-consistent interpretation, neither stands nor falls on the existence of SICs.

What the SICs would add is just a very pretty way of rewriting the Born Rule to be purely in terms of probabilities, and with a pretty way of expressing the idea, the hope is that it will facilitate philosophical conversations. For instance, the Oxford philosopher Harvey Brown wrote in a recent paper, "Powerful plausibility arguments have long been available, some since the birth of QM, to the effect that the quantum state is something real. They almost all have to do, in one way or another, with quantum phase, with the fact that the wavefunction, in its relation to probability, is strictly a (generally complex) probability amplitude: it has more structure than a probability distribution does." Well, it is just wrong that a quantum state has more structure than a probability distribution does. And a SIC representation of quantum states (if it exists) helps the Born Rule, written in terms of probabilities, look as simple as possible. The hope is that if one can make it look so pretty and simple, even a philosopher might take note. Probably fat chance, but I haven't given up yet.

As far as whether one needs the SICs even for that conceptual discussion, surely one doesn't. One can already make the conceptual point with a representation based on any minimal informationally complete POVM consisting of rank-1 elements. Schack and I make that point in arXiv:1412.4211 [quant-ph]. On the other hand, if one wants to take the Born Rule in probabilistic terms as a *fundamental axiom* of quantum theory, then I have the feeling that such an axiomatization will be most facilitated if the SICs exist: For, it would allow the Born Rule---as an axiom---to *sneak in* so much of the structure of quantum theory with one simple statement, it might just be God's hammer. Some baby steps in that direction can be found here arXiv:1612.03234 [quant-ph].

I hope that clarifies things for you.

With best wishes, Chris

Joel Klassen May 13 2017 19:20 UTC

At the end of page 19, section 2.2 you introduce the tenet

"My Probabilities Cannot Tell Nature What To Do"

Can you elaborate on why it is necessary to include this tenet in QBism? Or more precisely, in what way is QBism unique in having this tenet? Are there any serious thinkers that are proposing that our scientific theories dictate to nature what it can and cannot do, as is suggested by the example of the woman's beliefs about her husband?

"Using the Born Rule, the agent would then calculate the probability of a “yes” outcome to be exactly 1—i.e., she should believe a “yes” will occur with all her heart and soul. But that does not mean that the world must comply. The agent’s quantum state assignment does not mean that the world is forbidden to give her a “no” outcome for this measurement."

What perspective in the philosophy of science or in quantum foundations is this woman's flawed reasoning meant to represent?

Finally, in what way does introducing this tenet, that the credence we assign to the behaviour of nature not dictate to nature what it can or cannot do, unhinge the Born rule from being a law of nature in a way distinct from how such a tenet operates on other laws of nature. When a person invokes a law of nature, such as the law of gravity, they are not saying "Masses attract each other because I believe really strongly that they will." They are saying, in accordance with this tenet, "I believe really strongly that masses attract each other." But that doesn't make the law of gravity any less a law of nature.

Ruediger Schack May 15 2017 14:00 UTC (2 points)

Dear Joel,

Thank you for this question about Fuchs's paper. As you suggest, if taken out of context, the tenet "My probabilities cannot tell nature what to do" is a little mystifying. No serious thinker should believe that *his* probabilities tell nature what to do. The actual content of the tenet is that in QBism probabilities are not taken to be real properties. Serious thinkers, or positions in philosophy, contradicting this tenet are those that say, e.g., that there exists a law of nature that fixes a probability x and tells a radium atom to decay with probability x in the next hour.

Instead of probabilities that are real properties, what you have in QBism are my probabilities, your probabilities, his probabilities, etc. If these probabilities tried to tell nature what to do, she would get pretty confused. They don't. QBist probabilities are also not what philosophers of science call "credences", i.e., beliefs as to the truth of something. QBist probabilities are exclusively for decision making. My probabilities help me decide how best to act, your probabilities help you decide how best to act, etc.

A more explicit form of the tenet can be found in the abstract: "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." A key use of this understanding about probability-1 can be found in Section 3.2, where it is deployed in the QBist analysis of the EPR argument, as well as in Footnote 20, where it is used in explaining how QBism evades the recent no-go theorem of Frauchiger and Renner arXiv:1604.07422 [quant-ph].

Best regards,


Joel Klassen May 15 2017 22:01 UTC

Dear Ruediger,

Thanks for your prompt and cordial response. I hope you'll forgive the absence of address and signoff in my previous comment, my excitement got the better of my internet etiquette.

I think I understand what you are saying. The notion is that by making a statement like "a rubidium atom decays with probability x" a potential misapprehension may be that the calculated probability is a property of the atom which dictates the future behaviour of the rubidium atom. I think I also understand why QBism needs to address this point, I can see why some people might be inclined to think this way about quantum probabilities.

However no misapprehension of this type is made when we consider the statement "the coin comes up heads with a probability 1/2." In this case we understand it to mean that we know a sufficient amount about the structure of the coin, and the nature of the flipping procedure to make a prediction of heads with probability 1/2, absent more detailed information. In common usage, the assignment of probability to the coin flip does commit us to properties of the coin, even if those properties are not one and the same with the probability assignment of coming up heads. It seems to me that merely claiming that probabilities are not properties does not preclude the existence of properties which consequently dictate probabilities. Realists can be bayesians too.

I suppose I am trying to understand the distinction between QBism and instrumentalism. Both classical and quantum states can be represented as probability functions on the set of propositions, where propositions can be formulated in terms of measurement procedures. When QBism makes a statement like "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. " it seems to imply that not only do quantum states not have any "ontic hold", but neither do classical states. If the claim is that no state assignment carries "ontic hold" then is this not simply instrumentalism? If not, how is it that quantum states are subject to more rigorous standards than classical states?

For example in section 3.2, the central proposition QBism takes issue with in the EPR paradox is Einstein's assertion:

"if, upon freely choosing to [measure something on A without interfering with B],I am able to predict something [on B], then this something must exist in reality.”

To which QBism responds "my probabilities can not tell nature what to do". But this seems a bit unfair to Einstein. He is positing a metaphysical assumption, which is that science's capacity to predict the behaviour of nature is made possible by the real structure of nature. If QBism is claiming to reject this metaphysical assumption, why confuse things by speaking in particular about quantum states, why not just say QBism is instrumentalism?

I imagine you've probably fielded questions like this a hundred times over. I am not especially well versed in QBism, and the literature is not terse. So I greatly appreciate your help.

With gratitude,

Christopher A. Fuchs May 15 2017 22:36 UTC (1 points)

Dear Joel,

We are indeed "fielded questions like this a hundred times over." That's why I try to write some papers to allay it: It never works. Anyway, here's one example that's relevant for your queries: My own view is that taking first-person elements seriously within physics (as QBism does) *adds* to the notion of reality, rather than taking away from it or ignoring it (as full-blown instrumentalism does). Also QBism retains an element of "something like" a structural realism that is usually ignored (namely, in the structure of the Born rule as a normative condition that's good for anybody). These things are discussed in the paper just cited. I hope that helps. Maybe Ruediger will want to add something.

Best wishes, Chris