Mermin replies: Robert Griffiths takes my joke about an undergraduate collapsing a wavepacket to be an attempt to solve the measurement problem. The actual QBist solution—versions of which predate QBism (see, for example, reference 1)—is that a quantum state is not an objective property of the world but a compendium of probabilities constructed by an agent. When an agent updates her state assignment, nothing changes in her external world. The only change is in the agent’s expectations for her subsequent experience of that world, and therefore there is no measurement problem. Consistent historians agree that collapse is not a physical process, but for entirely different reasons.
QBism takes the “mysterious whatever-it-is” that Griffiths criticizes in quantum orthodoxy to be the experience induced in an agent by the world. That, and the fact that different agents may have different experiences, drains the swamp, slays the dragon, and quiets the maddening paradoxes.
QBists hold that any agent can indeed apply quantum mechanics to her own external world, from quarks to quasars. Only an agent’s personal, directly perceived experience is boxed off from the domain of applicability. That exclusion disagrees with most versions of the many-worlds interpretation, but it does not limit the scope of science: Alice can apply quantum mechanics to the experience of another agent, Bob, as represented through his words, writings, or drawings, all of which do belong to her external world.
The way in which Bob’s experience impinges on Alice resonates with Niels Bohr’s emphasis on the need to state results of experiments in ordinary language. It also addresses Ching Hung Woo’s question of how different agents can come to agreement on their state assignments.
Griffiths correctly notes that QBism violates the first desideratum he quotes from my 15-year-old “Ithaca interpretation.” Like Barack Obama’s view of marriage, my thinking about quantum foundations has evolved. But Griffiths wrongly claims that QBism rejects the other desideratum he quotes: locality. Actually, QBism insists on it. The absurdity of quantum nonlocality provides an independent argument that quantum states cannot be objective internal properties of the systems they describe, whether or not one regards probabilities as subjective. Griffiths is probably right, however, that John Bell would not have been pleased with how QBists fix his shifty split.
I would not call QBism an “antiquated approach.” It is younger than Griffiths’ consistent histories and, as I have remarked upon elsewhere,2 taking reality itself to be relative to what he calls a framework, is vastly more radical than taking quantum state assignments to be relative to an agent.
But foreshadowings of QBism can indeed be found in antiquity. Since writing my commentary, I came across another striking anticipation in a 1931 letter from Erwin Schrödinger to Arnold Sommerfeld:3 “Quantum mechanics forbids statements about what really exists—statements about the object. It deals only with the object–subject relation. Although this holds, after all, for any description of nature, it appears to hold in a much more radical and far-reaching sense in quantum mechanics” (translated by N. D. Mermin and Ruediger Schack).
When Bell and Michael Nauenberg wrote their delightful article in 1966, there was indeed much less interest in quantum foundations than there is today. Most textbooks have wisely steered clear of the many contemporary interpretive positions, since none of them are held by more than a fraction of those physicists with an interest in such issues.
QBists do not reject the Born rule for calculating probabilities, but they do reject the objective frequentist interpretation of those probabilities presented by Nauenberg and held by most physicists. The frequentist view is notoriously circular. It defines probability using such notions as “equally probable,” “unlikely,” or (with Nauenberg) “identically prepared,” none of which make sense without a prior definition of probability. “Identically prepared” might seem safe from circularity, but two different preparations cannot be strictly identical. It would be more accurate to say they must differ in unimportant ways. Unimportant for what? For the probabilities of the outcomes.
When I started to learn about subjective probability, I was surprised to discover that most of the books were not in Cornell University’s physics or mathematics libraries but in the business school library. In our recent election, Americans have been told that business experience is necessary for being president; I would suggest that it may be even more helpful for understanding quantum mechanics.
I cannot tell from Nauenberg’s remarks about the Liouville equation whether he agrees with the view of QBists (and many others) that the collapse of the wavefunction is no more than the updating of a collection of probability distributions.
Nauenberg’s example of the Rydberg atom does not contradict the QBist resolution of Bell’s shifty split. The Rydberg atom is part of the world external to the agent, and therefore on the quantum side of her split. There are indeed circumstances under which fully quantum mechanical behavior can look very classical.
Art Hobson repeats the common view that only mixed states can be associated with subjective probabilities. QBism takes that to be the very mistake that got us lost for so long in the quantum swamp. Hobson also believes that wavefunctions were collapsing before there were physicists. Does he believe that probabilities were updating before there were statisticians?
Hobson wrongly claims that Chris Fuchs “might dispute the notion that some kind of physical reality actually exists.” What he and other QBists do deny is that there are properties of the external world—“elements of reality”—that underlie quantum state assignments. (The argument of Matthew Pusey and coauthors that Hobson cites, implying that quantum states are physically real, relies on such hidden variables.) An agent’s state assignment rests entirely on her experience of the external world. It is neither solipsistic nor superstitious to maintain that any agent’s experience comes fully into being only at the moment it is experienced.
Hobson’s treatment of Eugene Wigner and his friend relies on decoherence producing an objective mixed state, which both of them must agree on. But since quantum mechanics holds all the way out to infinity (as Griffiths puts it), even from an objective view of quantum states, that switch from superposition to either/or is at best a FAPP solution, to use Bell’s wonderfully sardonic abbreviation of “for all practical purposes.”
QBism does not, as Hobson concludes, give up on a realist interpretation of nature. But it does warn us not to confuse nature with the abstractions we have ingeniously constructed to help any agent deal with the very real impact of nature on his or her own internal experience (see my Reference Frame, Physics Today, May 2009, page 8).
I’m glad that Blake Stacey has called attention to some of the more applied spinoffs of QBism. I’m also pleased that he mentions Bruno de Finetti, one of the great 20th-century pioneers of subjective probability. Indeed, since there are objective as well as subjective Bayesians, if I had my way, the B in QBism would stand not for Thomas Bayes, but for Bruno de Finetti, who put the crucial point like this: “The abandonment of superstitious beliefs about the existence of Phlogiston, the Cosmic Ether, Absolute Space and Time … , or Fairies and Witches, was an essential step along the road to scientific thinking. Probability, too, if regarded as something endowed with some kind of objective existence, is no less a misleading misconception, an illusory attempt to exteriorize or materialize our [actual (vero)] probabilistic beliefs.”4
