QBism — originally short for Quantum Bayesianism, though its founders now treat "QBism" as a standalone term — is a radical interpretation of quantum mechanics that applies subjective Bayesian probability to the quantum domain. In the QBist view, the quantum state (wave function or density matrix) assigned to a system does not describe an objective property of that system. Instead, it encodes a particular agent's beliefs about what she will experience if she performs a measurement. Two agents may assign different quantum states to the same system, and neither need be wrong — just as two Bayesian agents may hold different priors.
This interpretation dissolves many of the conceptual paradoxes that have plagued quantum foundations — the measurement problem, wave function collapse, entanglement-at-a-distance — by relocating quantum states from the external world to the agent's epistemic state. The "collapse of the wave function" is not a physical process but a Bayesian update: the agent revises her beliefs in light of a measurement outcome.
Origins and Development
Carlton Caves, Christopher Fuchs, and Rudiger Schack begin exploring the idea that quantum probabilities are subjective, publishing a series of papers arguing that the quantum state is a state of belief, not a state of nature.
Fuchs's influential paper "Quantum Mechanics as Quantum Information (and Only a Little More)" articulates the core QBist position: quantum theory is a normative framework for agents, not a descriptive theory of objective reality.
Fuchs, Mermin, and Schack coin the term "QBism" and publish accessible expositions. N. David Mermin becomes a prominent advocate, writing a series of articles in Physics Today.
QBism is developed further with connections to symmetric informationally complete (SIC) measurements and the reformulation of quantum mechanics in terms of probabilities rather than amplitudes.
The QBist Interpretation of Core Quantum Concepts
Quantum States
In standard textbook quantum mechanics, a quantum state |ψ⟩ is often treated as an objective property of a physical system. QBism rejects this. The quantum state is an agent's tool for computing probabilities of measurement outcomes — it is a bookkeeping device for her beliefs, nothing more. Just as a Bayesian prior is personal, so is the quantum state assignment.
Measurement and Collapse
The measurement problem — how and why the wave function "collapses" upon observation — has been called the central mystery of quantum mechanics. QBism dissolves it entirely. There is no physical collapse. When an agent performs a measurement and obtains a result, she updates her quantum state assignment using the quantum analog of Bayes' rule (the Luders rule). The change in the quantum state is a change in her beliefs, prompted by a new experience, not a physical change in the system.
Entanglement and Nonlocality
Entanglement appears mysterious when quantum states are taken to be objective: measuring one particle seems to instantaneously affect a distant partner. But if quantum states are personal beliefs, the mystery evaporates. When Alice measures her particle and updates her quantum state for Bob's particle, she is revising her expectations — just as learning the outcome of a coin flip in New York changes your beliefs about a correlated coin flip in London without any physical signal being sent.
In standard quantum mechanics, the Born rule — P(outcome) = |⟨outcome|ψ⟩|² — is a postulate connecting the mathematical formalism to empirical predictions. In QBism, the Born rule is reinterpreted as a normative constraint on an agent's beliefs, analogous to the Dutch book coherence requirement in classical Bayesian epistemology. Just as classical coherence requires credences to satisfy the probability axioms, quantum coherence requires credences about measurement outcomes to satisfy the Born rule. Violating it leads to a quantum analog of a Dutch book — a set of gambles guaranteeing a sure loss.
Relation to Bayesian Epistemology
QBism is, at its core, the application of subjective Bayesian epistemology to the quantum domain. It inherits the strengths and faces the challenges of subjectivism. The key move is to treat quantum theory as an extension of probability theory — a "new kind of probability theory" as Fuchs describes it — where the Born rule supplements the classical axioms of Kolmogorov with additional structure specific to the quantum world.
This connection runs deep. De Finetti's dictum that "probability does not exist" as an objective feature of the world becomes, in QBism, the claim that the quantum state does not exist as an objective feature of the world. The representation theorem, which shows that exchangeability judgments imply a prior, has analogs in the quantum setting through the quantum de Finetti theorem. And the washing-out theorem — posteriors converge regardless of priors — has quantum analogs ensuring that agents who perform many measurements will come to agree on future predictions.
Classical Analog P(H | E) = P(E | H) · P(H) / P(E)
Criticisms
QBism's critics argue that it is excessively solipsistic — by making quantum states subjective, it seems to deny the existence of an objective physical world. QBists respond that they deny only that the quantum state is objective, not that the world is. The world exists and is the cause of the agent's experiences; quantum mechanics is a theory about how to navigate that world, not a literal description of its furniture.
Another criticism is that QBism seems to make quantum mechanics a theory about single agents rather than a universal physical theory. QBists embrace this: physics, they argue, is always for an agent — a tool that any agent can use to organize her experience. There is no "view from nowhere."
"Quantum mechanics is a tool that any agent can use to help manage her expectations. It is not a description of the world as it is independently of the agent." — Christopher Fuchs, N. David Mermin, and Rudiger Schack, "An Introduction to QBism" (2014)
Significance for Bayesian Foundations
Whether or not QBism is the correct interpretation of quantum mechanics, it demonstrates the extraordinary reach of Bayesian epistemology. If even the fundamental theory of physics can be recast as a normative framework for an agent's beliefs, then the Bayesian program extends far beyond its origins in the analysis of games of chance. QBism suggests that the boundary between epistemology and physics may be far less sharp than traditionally supposed — that the deepest physical theories may be, at root, theories of rational belief.
Example: A Physicist Measures a Photon
A quantum physicist sends a single photon toward a beam splitter — a device that reflects or transmits light with equal probability. A detector is placed at each output. The photon is detected at Detector A.
The Standard (Copenhagen) View
Under the standard interpretation, before measurement the photon is in a superposition of "going to A" and "going to B." Measurement collapses the wave function, and the photon is really at Detector A.
The QBist View
A QBist interprets this differently. The quantum state was never a description of the photon's reality — it was the physicist's credence assignment about what she would experience upon measurement.
These are personal Bayesian credences, not objective physical states.
After measurement: Agent experiences the outcome "Detector A."
She updates: P(Detector A) = 1.00
"Collapse" is not a physical process — it is Bayesian conditionalization.
The Born rule — quantum mechanics' formula for computing measurement probabilities — is recast as a normative constraint on an agent's credences, analogous to how the probability axioms constrain rational belief in classical Bayesian epistemology.
The "measurement problem" — why does observation cause collapse? — vanishes under QBism because there is no collapse. There is only an agent updating her beliefs. Schrödinger's cat isn't simultaneously alive and dead; the quantum state simply represents the observer's uncertainty. QBism trades the mystery of physical collapse for a radical philosophical commitment: quantum mechanics is about what agents should believe, not about what the world is. It is Bayesian epistemology extended to the deepest level of physical reality.