When you publish physics papers with quantum in the title, you get e-mail. Sometimes it’s from people who want to share their radical new theory of dark matter, consciousness and psychoflexitive powers. Other times, it’s from Nature, asking you to do their survey about the “interpretation” of quantum mechanics. Gentle Reader, I did their survey, and I have thoughts. In fact, I wrote in a freeform answer to say that their questions were badly designed, and I would be telling anyone who asked that the survey results should not be trusted.
I’ve heard on the grapevine that Nature consulted “experts” while making up their questions. I don’t know who these “experts” were, but I would be unsurprised if they are better at being loud than at being thoughtful, and more interested in imposing their own ideas of what counts as “respectable” than at actually paying attention to what people have written.
I figured I’d get a blog post up about this, so that I can point people to it and then not have to say anything else. And, in the interests of fairness, I’m posting this before the results are public, so that I will be telling people the survey is unsound even if positions I’d agree with turn out apparently popular.
Right away, we dive in with a first question about what the respondent’s preferred interpretation of quantum mechanics is. And right away, we have a problem.
The list of options to choose from is more or less what you’d get by yoinking Wikipedia’s sidebar. But on one and only one option, there’s a footnote which says that this choice is not a self-consistent interpretation, while physicists still affiliate themselves with it. It’s like a joke about bad survey design: “To clarify, this candidate hates apple pie.” Deliberately or not, it makes this questionnaire into a push poll.
It’s true that there are multiple Copenhagen-ish interpretations. But a single Copenhagen-type interpretation, taken by itself, is probably going to be self-consistent. Moreover, there are multiple Everettian interpretations, there are at least two Relational Quantum Mechanics-es, and so forth. Why aren’t each of those called “not a self-consistent interpretation” if Copenhagen is?
[EDIT TO ADD (30 July 2025): Another problem with their description of “Copenhagen” is that they define it to include “nonlocality is real”. But proponents of different Copenhagen-ish interpretations take different messages from the violation of Bell inequalities. Some endorse nonlocality, but others say we should reject another assumption, like “counterfactual definiteness”. I don’t know how they labeled Copenhagen as “not a self-consistent interpretation” while also ignoring a big distinction among Copenhageners! It’s as though they listened to somebody who wanted to slather Copenhagen with a dirty label, without bothering to read the actual literature.]
Lumping “consistent histories” in with many-worlds interpretations is an error. Some advocates of history-oriented interpretations sound more Everettian (e.g., Gell-Mann and Hartle at times), while others sound more Copenhagen-ish (e.g., Hohenberg, who calls his version of consistent histories “Copenhagen done right”).
Relational QM was broken out as a separate item, instead of being part of the “information-based” interpretations. But why? Rovelli’s original paper said, right there in the abstract, “I consider a reformulation of quantum mechanics in terms of information theory.” For Rovelli, a quantum state is the information that one system has about another. He’s never been adequately precise about what “information” means, but that’s a notion he’s stuck to.
For that matter, how are we supposed to divide “information-based” from “Copenhagen”? Plenty of authors who have advocated an informational interpretation have portrayed this as sharpening up old Copenhagen-style ideas with modern tools. (To pick a convenient example, see Kofler and Zeilinger.) Rather than being crafted in a principled way, the options in this survey seem geared to confuse.
In addition, being forced to pick a single interpretation as uniquely “the best” fails to do the issue justice. What about the physicists who switch interpretations six times before breakfast (recalling question 13 of Schlosshauer et al.’s survey)? What about those for whom “shut up and calculate” is by far the correct choice, with anything beyond that only a vague preference?
2. Which of the following do you believe best solves the quantum measurement problem (that is, the inconsistency between the well-described and smooth evolution of properties according to the Schrödinger wave equation and the probabilistic outcomes of measurements)?
I had to choose “other” here. If you ask me, a wavefunction is nothing more or less than an encoding of an individual agent’s expectations, and so a “collapse” is merely a change of expectation due to new experience. This is similar to the “wavefunction gives only information” option provided, but distinct in that the wavefunction is belief rather than knowledge — doxastic rather than epistemic.
3. Is action at a distance real?
That is, does measuring the properties of one particle instantaneously (faster than the speed of light) influence its entangled partner such that it affects the result of future measurements?
The fact that this question provides an option phrased in terms of “beliefs” while the previous question asked about “information” suggests that the distinction between these concepts was not taken seriously.
4. For a given system, does there exist a boundary between classical and quantum objects?
This question is not phrased clearly enough to answer.
5a. For single-particle quantum systems, the wavefunction can generally be understood as propagating in ordinary, 3D, physical space. But for a system with more than one particle, the wavefunction instead propagates in the abstract higher-dimensional ‘configuration’ space.
In such situations, what, if anything, exists in ordinary 3D physical space?
“3D physical space” is itself a mathematical contrivance. To say that anything “exists in” it is to reify an abstraction. Nature is not Euclidean, nor Minkowskian, nor Riemannian.
7. Does a measurement (as described in the measurement problem question) require an observer?
The answer “Yes but consciousness is not relevant (and an ‘observer’ can include interaction with a macroscopic environment)” conflates two very different positions. In my view, the quantum formalism is a “user’s manual” that any agent can pick up and employ to manage their expectations about the consequences to them of their actions upon the rest of the world. Emphasis on the “any agent”: They do not have to be “conscious” in the way we are, with our particular allotment of emotions, our human kind of memory, etc. I’d say that nearly all invocations of “consciousness” in quantum foundations have been distractions. But “an ‘observer’ can include interaction with a macroscopic environment” is the kind of talk you get from people who think that chanting “decoherence” solves everything, and I can’t endorse that either.
8. Is it important to attempt to interpret the mathematics of quantum mechanics in a physical or intuitive way?
Neither “physical” nor “intuitive” are defined clearly enough for me to answer this. Are they being treated as synonymous? (They shouldn’t be; one can develop a visual or kinesthetic sense for abstract mathematics.) What would possibly qualify as a non-“physical” interpretation of a physical theory?