Sources and References

• David Albert, Quantum Mechanics and Experience (Harvard University Press, 1992)
• Tim Maudlin, Quantum Non-Locality and Relativity, 3rd ed. (Wiley-Blackwell, 2011)
• R.I.G. Hughes, The Structure and Interpretation of Quantum Mechanics (Harvard University Press, 1989)
• John Bell, Speakable and Unspeakable in Quantum Mechanics, 2nd ed. (Cambridge University Press, 2004)
• Hugh Everett III, “Relative State Formulation of Quantum Mechanics,” Reviews of Modern Physics 29 (1957)
• David Deutsch, The Fabric of Reality (Penguin, 1997)
• David Wallace, The Emergent Multiverse (Oxford University Press, 2012)
• David Bohm, “A Suggested Interpretation of the Quantum Theory in Terms of ‘Hidden’ Variables,” Physical Review 85 (1952)
• G.C. Ghirardi, A. Rimini, and T. Weber, “Unified Dynamics for Microscopic and Macroscopic Systems,” Physical Review D 34 (1986)
• Michael Nielsen and Isaac Chuang, Quantum Computation and Quantum Information (Cambridge University Press, 2000)
• Christopher Fuchs, “QBism, the Perimeter of Quantum Bayesianism,” arXiv:1003.5209 (2010)
• Jeffrey Bub, Interpreting the Quantum World (Cambridge University Press, 1997)
• Carlo Rovelli, “Relational Quantum Mechanics,” International Journal of Theoretical Physics 35 (1996)
• Wojciech Zurek, “Decoherence, Einselection, and the Quantum Origins of the Classical,” Reviews of Modern Physics 75 (2003)
• Niels Bohr, Atomic Theory and the Description of Nature (Cambridge University Press, 1934)

---

Chapter 1: Quantum Mechanics — A Philosophical Introduction

1.1 Why Quantum Theory Demands Philosophy

Physics and philosophy have long been intertwined, but nowhere is this entanglement more intimate — or more philosophically productive — than in quantum mechanics. Since its mathematical consolidation in the 1920s, quantum theory has achieved extraordinary predictive success. It underlies the transistor, the laser, magnetic resonance imaging, and the emerging architecture of quantum computing. And yet, for all its empirical triumph, quantum mechanics resists the most basic demand we place on a physical theory: it resists interpretation.

To interpret a physical theory is to answer the question: if this theory is true, what is the world like? For classical Newtonian mechanics, the answer is relatively transparent — there are particles with definite positions and momenta, evolving deterministically under forces. For quantum mechanics, the question remains genuinely, controversially open after a century of reflection. Whether quantum states represent objective reality or subjective degrees of belief; whether measurement is a physical process or a primitive act; whether the world is deterministic at a deeper level — these are not settled questions. They are the questions this course addresses.

Interpretation (诠释) of a physical theory is the philosophical project of specifying what the theory’s formalism commits us to: what entities exist, what processes are real, and how theoretical terms relate to observable phenomena. The difficulty of interpreting quantum mechanics is not a gap in our mathematical understanding but a deep conceptual puzzle. The mathematics is precise; the ontology it encodes is contested.

1.2 Historical and Conceptual Setting

Quantum mechanics emerged from two crises in late-nineteenth and early-twentieth-century physics. The first was the ultraviolet catastrophe: classical electrodynamics predicted that a black body in thermal equilibrium would radiate infinite energy at high frequencies. Max Planck resolved this in 1900 by postulating that radiation is emitted in discrete packets — quanta (量子) — of energy \( E = h\nu \), where \( h \) is Planck’s constant and \( \nu \) is frequency. This was an ad hoc fix that Planck himself did not fully believe.

The second crisis centered on atomic spectra. Neils Bohr’s 1913 model of the hydrogen atom introduced quantized electron orbits but lacked any dynamical justification. It was a recipe, not an explanation. The mature theory — wave mechanics (Schrödinger, 1926) and matrix mechanics (Heisenberg, 1925), proved equivalent by Dirac and von Neumann — resolved these difficulties but introduced conceptual puzzles of its own.

1.3 Overview of the Philosophical Landscape

This course proceeds in three broad arcs. The first (Chapters 1–4) establishes the formalism and the classic problems: wave-particle duality, the measurement problem, and non-locality. The second arc (Chapters 5–10) examines the major interpretive programs: Copenhagen, hidden variables, many worlds, decoherence, spontaneous collapse, and Bohmian mechanics. The third arc (Chapter 11) turns to contemporary approaches inspired by quantum information theory (量子信息理论): QBism, relational quantum mechanics, and Jeffrey Bub’s information-theoretic interpretation.

A recurring theme is that interpretive choices are not merely aesthetic. They carry commitments about determinism, realism, locality, and the scope of physical law. Students are encouraged to evaluate arguments carefully and to hold provisional positions.

---

Chapter 2: The Mathematical Framework — Hilbert Spaces and Observables

2.1 Hilbert Space

The mathematical setting for quantum mechanics is a complex Hilbert space (希尔伯特空间) \( \mathcal{H} \). A Hilbert space is a complete inner product space over the complex numbers \( \mathbb{C} \). For finite-dimensional systems (such as a spin-\(\frac{1}{2}\) particle), \( \mathcal{H} \cong \mathbb{C}^n \). For continuous systems (position, momentum), the relevant space is \( L^2(\mathbb{R}) \), the space of square-integrable functions.

The Dirac notation (狄拉克符号) uses kets \( |\psi\rangle \) for vectors and bras \( \langle\phi| \) for dual vectors (linear functionals). The inner product of two vectors is written \( \langle\phi|\psi\rangle \in \mathbb{C} \). The norm of a vector is \( \| |\psi\rangle \| = \sqrt{\langle\psi|\psi\rangle} \).

Superposition (叠加原理) is the central mathematical feature of Hilbert space: if \( |\psi_1\rangle \) and \( |\psi_2\rangle \) are states in \( \mathcal{H} \), so is any linear combination

\[ |\psi\rangle = \alpha|\psi_1\rangle + \beta|\psi_2\rangle \]

for \( \alpha, \beta \in \mathbb{C} \) with \( |\alpha|^2 + |\beta|^2 = 1 \). The physical meaning of superposition is the central interpretive challenge.

2.2 Observables, Self-Adjoint Operators, and the Spectral Theorem

Physical quantities — position, momentum, spin, energy — are represented in quantum mechanics not by real-valued functions but by self-adjoint operators (自伴算符) on \( \mathcal{H} \). An operator \( \hat{A} \) is self-adjoint if \( \hat{A} = \hat{A}^\dagger \), where \( \hat{A}^\dagger \) is the Hermitian conjugate.

For infinite-dimensional systems with continuous spectra — such as position \( \hat{x} \) or momentum \( \hat{p} \) — the spectral theorem takes a more subtle form. Rather than a discrete sum, we have a spectral measure \( dE(\lambda) \) and the decomposition

\[ \hat{A} = \int_{\mathbb{R}} \lambda \, dE(\lambda) \]

where \( E(\Delta) \) is the projection onto the eigenspace associated with the measurable subset \( \Delta \subseteq \mathbb{R} \). This formalism, developed rigorously by von Neumann, ensures that observables like position and momentum — which have no normalizable eigenvectors in \( L^2(\mathbb{R}) \) — are still well-defined self-adjoint operators with real spectra. The spectral theorem is philosophically significant: it tells us that every observable has a complete, orthogonal family of projectors that partition the Hilbert space. Measurement outcomes correspond to which subspace the post-measurement state belongs to.

A general state can be expanded in the eigenbasis of any observable with discrete spectrum:

\[ |\psi\rangle = \sum_i c_i |a_i\rangle, \quad c_i = \langle a_i|\psi\rangle \]

2.3 The Uncertainty Principle: Derivation

The uncertainty principle (不确定性原理) is a theorem about quantum states, not merely an epistemological claim. For any state \( |\psi\rangle \) and any two self-adjoint operators \( \hat{A} \) and \( \hat{B} \), one can prove the Robertson inequality:

\[ \Delta A \cdot \Delta B \geq \frac{1}{2}|\langle[\hat{A},\hat{B}]\rangle| \]

where \( \Delta A = \sqrt{\langle\hat{A}^2\rangle - \langle\hat{A}\rangle^2} \) is the standard deviation and \( [\hat{A},\hat{B}] = \hat{A}\hat{B} - \hat{B}\hat{A} \) is the commutator. The derivation proceeds via the Cauchy-Schwarz inequality applied to vectors \( (\hat{A} - \langle\hat{A}\rangle)|\psi\rangle \) and \( (\hat{B} - \langle\hat{B}\rangle)|\psi\rangle \).

For position and momentum, the canonical commutation relation gives \( [\hat{x}, \hat{p}] = i\hbar \), so \( \langle[\hat{x},\hat{p}]\rangle = i\hbar \) and therefore:

\[ \Delta x \cdot \Delta p \geq \frac{\hbar}{2} \]

This is an exact, state-independent lower bound. The bound is saturated precisely for Gaussian (coherent) states — minimum-uncertainty wave packets of the form \( \psi(x) \propto e^{-(x-x_0)^2/(4\sigma^2)} e^{ip_0 x/\hbar} \), which have \( \Delta x = \sigma \) and \( \Delta p = \hbar/(2\sigma) \).

The philosophical significance is considerable. The uncertainty principle is not a statement about the limitations of measuring instruments — it is a statement about the structure of quantum states. A particle in a state of definite position (\( \Delta x = 0 \)) is in a position eigenstate, which is a Dirac delta and has \( \Delta p = \infty \): it literally has no determinate momentum whatsoever. This reflects a genuine feature of the quantum state, not our ignorance.

2.4 The Born Rule

The connection between the formalism and empirical predictions is made by the Born rule (玻恩规则). Given a system in state \( |\psi\rangle \) and an observable \( \hat{A} \) with eigenstates \( |a_i\rangle \) and eigenvalues \( a_i \), the probability of obtaining outcome \( a_i \) upon measurement is:

\[ P(a_i) = |\langle a_i|\psi\rangle|^2 = |c_i|^2 \]

2.5 The Schrödinger Equation

Between measurements, a quantum state evolves deterministically according to the Schrödinger equation (薛定谔方程):

\[ i\hbar\frac{\partial}{\partial t}|\psi(t)\rangle = \hat{H}|\psi(t)\rangle \]

where \( \hat{H} \) is the Hamiltonian operator representing the system’s total energy. This evolution is linear, unitary, and reversible — it preserves the norm of \( |\psi\rangle \) and never collapses the state. The tension between this smooth, unitary evolution and the abrupt, stochastic change associated with measurement is the formal expression of the measurement problem.

2.6 Composite Systems and Tensor Products

When a physical system is composed of two subsystems \( A \) and \( B \), their joint Hilbert space is the tensor product (张量积) \( \mathcal{H}_A \otimes \mathcal{H}_B \). If \( \{|a_i\rangle\} \) is an orthonormal basis for \( \mathcal{H}_A \) and \( \{|b_j\rangle\} \) for \( \mathcal{H}_B \), then \( \{|a_i\rangle \otimes |b_j\rangle\} \) is an orthonormal basis for the composite space, which has dimension \( \dim(\mathcal{H}_A) \times \dim(\mathcal{H}_B) \).

Entanglement (量子纠缠) arises precisely because not every vector in \( \mathcal{H}_A \otimes \mathcal{H}_B \) is a product state. A state is separable if it can be written as \( |\psi_A\rangle \otimes |\psi_B\rangle \); otherwise it is entangled. An entangled state assigns no definite pure state to either subsystem individually.

---

Chapter 3: Wave-Particle Duality and the Double-Slit Experiment

3.1 The Conceptual Paradox

Classical physics assigns definite natures to objects: a billiard ball is a particle; a sound wave is a wave. Quantum objects resist this categorization. Electrons, photons, and atoms exhibit interference (a wave phenomenon) in some experimental contexts and discrete impacts (a particle phenomenon) in others. This is wave-particle duality (波粒二象性), one of the founding puzzles of quantum mechanics.

3.2 The Double-Slit Experiment

The double-slit experiment is, as Richard Feynman said, the one experiment that “has in it the heart of quantum mechanics.” Consider electrons fired one at a time at a barrier with two slits, and a detection screen behind. When both slits are open and no which-path measurement is made, an interference pattern builds up on the screen — bands of high and low detection probability. When a detector is placed at one slit to determine which path the electron takes, the interference pattern disappears.

3.3 The Mach-Zehnder Interferometer

The Mach-Zehnder interferometer (马赫-曾德尔干涉仪) provides an especially clean setting for studying wave-particle duality and which-way information. A single photon enters a first beam splitter (BS1), which puts it in an equal superposition of two paths — an upper arm and a lower arm. The paths are recombined at a second beam splitter (BS2), and two detectors \( D_1 \) and \( D_2 \) register the output.

If the path lengths are balanced, quantum interference ensures that all photons exit at \( D_1 \) and none at \( D_2 \). Formally, if the beam splitters implement Hadamard-like transformations, the state after BS1 is

\[ |\psi\rangle = \frac{1}{\sqrt{2}}(|U\rangle + |L\rangle) \]

where \( |U\rangle \) and \( |L\rangle \) represent the upper and lower arms. After recombination at BS2 with no path-length difference, the interference is constructive toward \( D_1 \) and destructive toward \( D_2 \). The photon “goes both ways” — its probability amplitude traverses both arms and the resulting interference determines the output.

Now suppose a which-way detector is inserted in one arm. Even if the detector does not absorb the photon, merely recording which-path information collapses the superposition. The photon state becomes entangled with the detector:

\[ |\Psi\rangle = \frac{1}{\sqrt{2}}(|U\rangle|d_U\rangle + |L\rangle|d_L\rangle) \]

If \( \langle d_U|d_L\rangle = 0 \) (orthogonal detector states, full which-way information), the interference at BS2 is completely destroyed and \( D_1 \) and \( D_2 \) fire with equal probability. If \( \langle d_U|d_L\rangle = 1 \) (no which-way information), full interference is recovered. Intermediate cases give partial interference, quantified by the visibility \( V = |\langle d_U|d_L\rangle| \).

This is the quantitative complementarity relation: the visibility \( V \) of the interference pattern and the distinguishability \( D \) of the paths satisfy \( V^2 + D^2 \leq 1 \). You cannot simultaneously have full interference and full which-way knowledge. This is a theorem provable from the structure of the formalism.

3.4 Which-Way Information and Complementarity

The which-way / interference trade-off illuminates Bohr’s complementarity principle (互补性原理) more precisely than any purely verbal statement. The key insight is that which-way information need not be actively observed — it suffices that the information be, in principle, available in the physical world. A photon that has interacted with a which-way detector loses its interference capacity even if no one ever reads the detector.

This has a surprising corollary in the quantum eraser experiment (Scully and Drühl, 1982). If the which-way information stored in the detector is subsequently erased — by performing a measurement on the detector that destroys the path information — the interference can be recovered in post-selected subensembles of the data. The “ghost” of interference reappears not in the full ensemble but only in coincidence with the erasure events. This demonstrates that it is information-theoretic entanglement, not mechanical disturbance, that kills interference.

3.5 Bohr’s Como Lecture Interpretation

Bohr’s most sustained philosophical account of wave-particle duality appears in his 1927 Como lecture, “The Quantum Postulate and the Recent Development of Atomic Theory,” delivered at the Como conference in honor of Volta. This is the foundational text for what later became the Copenhagen interpretation.

Bohr’s central claim in the Como lecture is that quantum theory requires a renunciation of the classical idea of a unique space-time description combined with the law of causality. The quantum postulate — that every atomic process involves a discrete and finite amount of energy exchange (the quantum of action \( h \)) — implies that any observation of atomic phenomena involves an interaction with the observing agency that cannot be neglected. This irreducible interaction forbids us from assigning definite classical attributes to the system independent of the experimental arrangement.

Bohr introduces complementarity as a logical relationship between descriptions: a particle description (definite trajectories in space-time) and a wave description (definite frequency/momentum) are mutually exclusive — an experimental arrangement that reveals one precludes the other. Neither description is more fundamental. Together they are exhaustive.

A crucial feature of Bohr’s Como position is its holism: the property being measured does not belong to the system alone but to the whole experimental arrangement (system plus apparatus). Asking what the electron’s momentum is “really” — independent of any measurement — is, for Bohr, a category mistake. Properties are always arrangement-relative.

3.6 Philosophical Significance

The double-slit experiment and the Mach-Zehnder interferometer drive home several points that resist classical intuition. First, probability in quantum mechanics is not merely epistemic (due to ignorance); the interference pattern is destroyed even if we simply acquire which-path information and then discard it, before checking the screen. Second, the electron does not “go through one slit.” Any story that assigns a definite path eliminates the interference. Third, the experiment does not teach us that electrons are “really” waves — they still produce discrete point-like impacts. The lesson is that neither the wave picture nor the particle picture is the full story.

---

Chapter 4: The Measurement Problem

4.1 The Formal Statement: Three Incompatible Claims

The measurement problem (测量问题) is the central interpretive crisis of quantum mechanics. Albert’s formulation in Quantum Mechanics and Experience is canonical: quantum mechanics gives us two dynamical laws that cannot both be universally applicable. But Maudlin’s three-horned formulation is the cleanest logical statement.

The formal structure of the problem:

• Dynamical Evolution (Rule 1): When a physical system is not being measured, its state evolves according to the Schrödinger equation — linearly, unitarily, continuously.
• Collapse Postulate (Rule 2): When a measurement is made, the quantum state “collapses” instantaneously to one of the eigenstates of the measured observable, with probability given by the Born rule.

Rule 1 is deterministic; Rule 2 is stochastic. Rule 1 preserves superpositions; Rule 2 eliminates them. And if everything is physical — including measuring devices and observers — Rule 1 should apply to them too.

4.2 The Measurement Interaction in Detail

Consider a system \( S \) in a superposition of spin eigenstates:

\[ |\psi_S\rangle = \alpha|{\uparrow}\rangle + \beta|{\downarrow}\rangle \]

A measurement apparatus \( A \) begins in a neutral “ready” state \( |r\rangle_A \). According to Rule 1, the combined state after interaction is:

\[ |\Psi\rangle = \alpha|{\uparrow}\rangle_S|{\uparrow}\rangle_A + \beta|{\downarrow}\rangle_S|{\downarrow}\rangle_A \]

This is an entangled superposition of the apparatus pointing to spin-up and spin-down. But we never observe measuring devices in superpositions. We always find a definite outcome. So either Rule 1 is not universally valid (collapse really occurs), or there is no collapse and the appearance of definite outcomes must be explained by other means.

4.3 Why Projection Is Ad Hoc

The collapse postulate is widely acknowledged — even by proponents of its practical use — to be theoretically unsatisfactory. Several features make it appear ad hoc:

First, it is discontinuous and instantaneous, in stark contrast to the smooth, continuous Schrödinger evolution. No physical mechanism is specified for the transition from Rule 1 to Rule 2. Second, it is observer-dependent in the sense that the postulate is triggered by “measurement,” a term that has no precise definition within the theory. What counts as a measurement? Must there be a conscious observer? A macroscopic apparatus? A thermodynamically irreversible event? The theory provides no answer.

Third, the collapse postulate is non-local: if a composite system is spread over a large region, the collapse of the wavefunction occurs instantaneously everywhere. This sits uneasily with special relativity. Fourth, and most fundamentally, if the measuring apparatus is itself a quantum system — which it surely is — then the Schrödinger equation applied to the composite system produces the entangled superposition above, not a collapse. The postulate must be applied by hand, from outside the theory’s own dynamics.

Von Neumann himself noted this problem in Mathematical Foundations of Quantum Mechanics (1932), distinguishing “Process 1” (measurement, non-unitary) from “Process 2” (evolution, unitary) and acknowledging that the boundary between them could be pushed arbitrarily in either direction — the “von Neumann chain” of entanglements extends through apparatus, observer, and beyond.

4.4 The Preferred Basis Problem

Even if one accepts collapse, a further problem arises: the preferred basis problem (首选基矢问题). The collapse postulate says the state collapses to an eigenstate of the measured observable. But in what basis does collapse occur? The measurement interaction Hamiltonian determines which observable is being measured, but quantum mechanics does not specify a distinguished “measurement basis” in which collapse preferentially occurs.

Consider the state \( |\Psi\rangle = \alpha|{\uparrow}\rangle_S|{\uparrow}\rangle_A + \beta|{\downarrow}\rangle_S|{\downarrow}\rangle_A \). This can equally be written as a superposition in a rotated basis:

\[ |\Psi\rangle = \gamma|+\rangle_S|+\rangle_A + \delta|-\rangle_S|-\rangle_A \]

for appropriately chosen \( |\pm\rangle \) states. Each decomposition suggests a different “definite outcome” basis. A complete account of measurement must specify why the apparatus-pointer variable picks out the spin-up/spin-down basis rather than some rotated alternative. Decoherence (Chapter 8) provides the most physically motivated answer to this question, but as we shall see, it does not fully resolve it.

4.5 Schrödinger’s Cat

Schrödinger’s famous cat thought experiment dramatizes the measurement problem at the macroscopic scale. A radioactive atom, a Geiger counter, a vial of poison, and a cat are enclosed in a box. If the atom decays, the Geiger counter triggers, shattering the vial and killing the cat. After one half-life, the state is (by Rule 1):

\[ |\Psi\rangle = \frac{1}{\sqrt{2}}(|\text{decayed}\rangle|\text{dead}\rangle + |\text{not decayed}\rangle|\text{alive}\rangle) \]

Is the cat literally in a superposition of alive and dead? Our macroscopic experience emphatically says no. The cat thought experiment shows that the measurement problem is not confined to the microscopic: it infects the macroscopic world if quantum mechanics applies universally.

---

Chapter 5: The Copenhagen Interpretation

5.1 Historical Origins and the Bohr-Heisenberg Framework

The Copenhagen interpretation (哥本哈根诠释) is not a single, unified doctrine but a cluster of views associated primarily with Niels Bohr and Werner Heisenberg, and institutionally with the Copenhagen school of the 1920s–30s. It became the de facto standard taught in physics courses for decades, though philosophers have increasingly challenged it as internally inconsistent or simply evasive.

Bohr’s central commitments, as expressed in Atomic Theory and the Description of Nature and elaborated through decades of correspondence and debate, include:

1. The quantum state (wavefunction) is not a description of independent physical reality. It is a tool for calculating probabilities of outcomes of experiments.
2. The wave and particle descriptions are complementary: mutually exclusive in application but jointly required.
3. The classical world — including measuring instruments — cannot itself be given a quantum description. The classical-quantum cut (Heisenberg cut) is a pragmatic boundary, not a physical one.
4. It is meaningless to ask about properties of quantum systems that are not being measured.

5.2 Bohr’s Complementarity Principle in Depth

Bohr’s complementarity principle is both more subtle and more philosophically substantial than a simple wave-or-particle dichotomy. In its mature form (developed through the late 1920s and 1930s), complementarity is a general epistemological principle about the conditions of unambiguous communication in quantum physics.

Bohr distinguishes two kinds of complementary descriptions. Kinematic complementarity refers to the mutual exclusivity of space-time description (position, time) and dynamical description (momentum, energy). Wave-particle complementarity refers to the mutual exclusivity of wave-like behavior (interference, diffraction) and particle-like behavior (definite trajectories, discrete impacts).

The principle has a logical structure: if an experimental arrangement is set up to reveal the position of a quantum system, then no arrangement in that same experiment can simultaneously reveal the momentum. This is not a practical limitation but a logical feature of how experimental arrangements are described in classical terms — the only terms available to us.

Crucially for Bohr, complementarity is not about the properties of the quantum system in isolation. It is about the entire experimental context. A quantum system does not have wave-like or particle-like nature simpliciter; it manifests these properties relative to specific experimental arrangements. This context-dependence — sometimes called holism (整体论) — is the core of Bohr’s philosophy.

5.3 Heisenberg’s Disturbance Interpretation vs. Bohr’s

Heisenberg and Bohr agreed on the uncertainty relations but disagreed on their physical interpretation, and this disagreement reveals important philosophical distinctions within the Copenhagen school.

Heisenberg’s gamma-ray microscope argument (1927) interpreted uncertainty as a disturbance caused by measurement. To measure the position of an electron, one must illuminate it with a photon (the gamma-ray microscope). The photon’s recoil transfers an uncontrollable momentum to the electron, disturbing its momentum. This is an epistemological picture: before measurement, the electron may have had definite position and momentum; measurement disturbs it, introducing irreducible uncertainty. The uncertainty principle, on this reading, is about measurement disturbance.

Bohr’s objection was decisive: Heisenberg’s picture smuggles in classical concepts inconsistently. To speak of the electron’s momentum “before” measurement, in a context where it is not being measured, is precisely the kind of language Bohr held to be meaningless. For Bohr, the uncertainty principle is not about disturbance but about the limits of simultaneous definition. An experimental arrangement that defines position precisely is logically incompatible with one that defines momentum precisely — not because they disturb each other, but because these are mutually exclusive descriptions.

This distinction matters for the interpretation of Bell’s theorem (Chapter 6) and for assessing information-disturbance trade-offs in quantum information theory. Bohr’s position is more consistently anti-realist about unmeasured properties; Heisenberg’s implies a quasi-classical hidden reality that measurement disturbs. Modern treatments, following Ozawa and Busch, have formalized Heisenberg’s intuition in terms of measurement-error / disturbance inequalities, showing they have a different structure from the state-dispersion uncertainty principle.

5.4 The Einstein-Bohr Debates and EPR

The most celebrated debate in the philosophy of physics is the extended exchange between Bohr and Einstein, conducted partly in person at the Solvay Conferences (1927, 1930) and partly through the EPR paper (1935) and Bohr’s reply.

At the 1927 Solvay Conference, Einstein proposed thought experiments designed to show that quantum mechanics was incomplete: by using a double-slit apparatus with a movable screen, one could simultaneously infer the particle’s momentum (from the screen’s recoil) and which slit it passed through (from the detection mark). If so, uncertainty would be violated. Bohr’s response: analyzing the apparatus quantum mechanically, the uncertainty in the screen’s momentum prevents the simultaneous determination of which slit was used. The apparatus itself obeys uncertainty; the classical-quantum cut does not protect Einstein’s argument.

At the 1930 Solvay Conference, Einstein proposed the “clock-in-box” thought experiment: weigh a box before and after emitting a photon to determine the photon’s energy exactly, while allowing arbitrarily long time for the measurement to occur. This would violate the time-energy uncertainty relation. Bohr’s famous overnight reply appealed to the gravitational red shift from general relativity: a clock in a gravitational field runs at a rate that depends on its position, and the very mechanism by which the energy is measured introduces time uncertainty, preserving \( \Delta E \cdot \Delta t \geq \hbar/2 \).

The 1935 EPR paper (Einstein, Podolsky, Rosen) escalated the challenge. In its modern spin formulation: prepare two particles in the singlet state and separate them to space-like separation. Measure particle A’s spin along the z-axis. Depending on the result, particle B’s state is immediately determined — either \( |\uparrow\rangle \) or \( |\downarrow\rangle \) — without any physical disturbance to B. EPR’s criterion of reality: “if, without in any way disturbing a system, we can predict with certainty the value of a physical quantity, then there exists an element of physical reality corresponding to this physical quantity.” B’s spin along z is determined without disturbing B; so it is an element of reality. But quantum mechanics does not assign a definite value to B’s spin before measurement. Therefore quantum mechanics is incomplete.

Bohr’s reply denied the EPR criterion of reality in the quantum context. He argued that “without any disturbance” cannot be applied to quantum systems in the way EPR assumed; the experimental arrangements for A and B are part of a single quantum phenomenon even when A and B are separated. The EPR correlations do not demonstrate incompleteness; they demonstrate the inappropriateness of applying classical locality intuitions to quantum systems. This reply was widely found unsatisfying, and the debate remained unresolved until Bell’s theorem (Chapter 6) gave it a precise mathematical form.

5.5 Anti-Realism and the Limits of Description

Bohr drew a sharp distinction between the quantum domain and the classical domain of instruments and observers. For him, quantum mechanics does not describe a microscopic reality that exists independently of measurement contexts. The “measurement problem” does not arise in a threatening way because the wavefunction was never a description of an objective reality to begin with — it is a predictive tool.

5.6 Heisenberg’s Uncertainty Principle

Heisenberg’s uncertainty principle (不确定性原理) is often presented as an epistemological claim — that we cannot simultaneously know position and momentum. But its mathematical content is a constraint on the dispersion of quantum states, not on knowledge:

\[ \Delta x \cdot \Delta p \geq \frac{\hbar}{2} \]

where \( \Delta x = \sqrt{\langle \hat{x}^2\rangle - \langle\hat{x}\rangle^2} \) is the standard deviation of position outcomes and similarly for \( \Delta p \). This is a theorem provable from the non-commutativity of position and momentum operators: \( [\hat{x}, \hat{p}] = i\hbar \). The question is what to make of it ontologically: is it a statement about the disturbance caused by measurement (Heisenberg’s original thought), or a statement that particles do not have simultaneously definite position and momentum (a more radical reading)?

5.7 Assessment

The Copenhagen interpretation’s virtue is that it avoids the measurement problem by refusing to apply quantum mechanics to the measurement process itself. Its philosophical defect is precisely this refusal — it is inherently dualistic, positing an unanalyzed classical-quantum divide without physical justification. As Bell acerbically noted, “the word ‘measurement’ has had such a damaging effect on the discussion that it should be banned from discussions of fundamental quantum mechanics.”

---

Chapter 6: Hidden Variables and Bell’s Theorem

6.1 The EPR Argument and Its Legacy

In 1935, Einstein, Podolsky, and Rosen published the argument examined in section 5.4. Their conclusion: the quantum state is incomplete — there exist elements of physical reality not represented by \( |\Psi\rangle \). The natural response is that a more complete theory could supply additional variables — hidden variables (隐变量) — whose values determine measurement outcomes deterministically. Quantum mechanics would then be a statistical theory, like thermodynamics, whose probabilities reflect our ignorance of the underlying hidden variables.

The EPR argument rests on two key assumptions: locality (a measurement on A does not physically disturb B when A and B are space-like separated) and completeness (the quantum state provides a complete physical description). EPR shows these two assumptions together imply that B’s spin is predetermined by hidden variables, contradicting quantum completeness. The argument is valid; the question is which premise to deny. Bohr denied completeness (in the sense that quantum states do not describe context-independent properties). Bell’s theorem (1964) showed that the locality assumption itself is empirically testable — and fails.

6.2 Bell’s Derivation

Bell considered a source producing particle pairs in the singlet state \( |\Psi^-\rangle = \frac{1}{\sqrt{2}}(|{\uparrow}\rangle_A|{\downarrow}\rangle_B - |{\downarrow}\rangle_A|{\uparrow}\rangle_B) \), sent to two distant detectors. Alice measures spin along direction \( \hat{a} \) and Bob measures along \( \hat{b} \). The outcomes are \( \pm 1 \).

A local hidden variable (LHV) theory posits variables \( \lambda \) (distributed according to some measure \( \rho(\lambda) \)) that predetermine the outcomes: \( A(\hat{a},\lambda) = \pm 1 \) for Alice and \( B(\hat{b},\lambda) = \pm 1 \) for Bob. The locality condition requires that Alice’s outcome depends only on her setting \( \hat{a} \) and the shared \( \lambda \), not on Bob’s distant setting \( \hat{b} \).

The correlation is:

\[ E(\hat{a},\hat{b}) = \int \rho(\lambda)\, A(\hat{a},\lambda)\, B(\hat{b},\lambda)\, d\lambda \]

Bell showed that this functional form, subject to the locality constraint, implies:

\[ |E(\hat{a},\hat{b}) - E(\hat{a},\hat{b}')| + |E(\hat{a}',\hat{b}) + E(\hat{a}',\hat{b}')| \leq 2 \]

This is the CHSH inequality (CHSH不等式), due to Clauser, Horne, Shimony, and Holt (1969), which is the most experimentally useful form of Bell’s result.

6.3 The Quantum Violation

Quantum mechanics predicts, for the singlet state:

\[ E(\hat{a},\hat{b}) = -\hat{a} \cdot \hat{b} = -\cos\theta_{ab} \]

where \( \theta_{ab} \) is the angle between the measurement directions. Choosing \( a = 0^\circ \), \( a' = 90^\circ \), \( b = 45^\circ \), \( b' = 135^\circ \):

\[ S_{QM} = |-\cos 45° - (-\cos 135°) + (-\cos 45°) + (-\cos 45°)| = 2\sqrt{2} \approx 2.83 \]

The quantum value \( 2\sqrt{2} \) is the Tsirelson bound — the maximum possible quantum violation, greater than the LHV bound of 2. No local hidden variable theory can reproduce this prediction.

6.4 Loophole-Free Bell Tests

Experimental tests of Bell inequalities have a long history. Aspect’s 1982 experiments at Orsay were the first to close the detection loophole and the locality loophole (using rapidly switching analyzers). However, various experimental loopholes remained:

• Detector loophole (detection efficiency loophole): if detectors are inefficient, the detected subensemble may not be representative of all pairs.
• Locality loophole: the measurement settings must be chosen and the results recorded while A and B are space-like separated, so no signal can carry information from one wing to the other.
• Freedom-of-choice loophole: the choice of measurement settings must be independent of the hidden variables.

Hensen et al. (2015), working with nitrogen-vacancy centers in diamond at Delft, performed the first loophole-free Bell test — closing both the detector loophole and the locality loophole simultaneously. Using entanglement swapping via photons and local measurements on electron spins, they achieved detection efficiencies above the threshold required to close the detector loophole, with measurement stations 1.3 km apart. They found \( S = 2.42 \pm 0.20 \), violating the CHSH bound by more than three standard deviations. Subsequent loophole-free experiments (Giustina et al., Shalm et al., also 2015) confirmed these results with larger sample sizes.

The significance of these experiments: local hidden variable theories are not just theoretically disfavored — they are empirically refuted, with the standard loopholes closed.

6.5 What Bell’s Theorem Actually Proves

Bell’s theorem rules out local hidden variable theories. It does not rule out:

• Non-local hidden variable theories (Bohmian mechanics is one)
• Non-realist interpretations (Copenhagen, QBism)
• Many-worlds (which has a different account of correlations)

The theorem establishes that quantum non-locality (量子非定域性) — in the sense of correlations that cannot be explained by shared pre-existing local properties — is a genuine feature of the world. Maudlin argues in Quantum Non-Locality and Relativity that this non-locality is an irremovable feature of any empirically adequate theory, and that it requires us to reconsider the relationship between quantum mechanics and special relativity.

---

Chapter 7: The Many-Worlds Interpretation

7.1 Everett’s Relative State Formulation

Hugh Everett III’s 1957 dissertation, published as “Relative State Formulation of Quantum Mechanics,” offers a radical solution to the measurement problem: abandon the collapse postulate entirely. If the Schrödinger equation is universally valid, then measurements do not collapse superpositions — they entangle the observer with the system, creating a superposition of observer-system pairs.

Everett’s key insight is that the apparent collapse, from the perspective of a particular branch of the wavefunction, is an artifact of the observer’s correlations with the system. The “relative state” of subsystem \( A \) relative to a specific outcome in subsystem \( B \) is always definite. What appears to be a unique outcome is, in fact, one of many branches, each equally real.

The Many-Worlds Interpretation (多世界诠释), as the view came to be called (the label is DeWitt’s, not Everett’s), holds that all branches of the quantum state exist. Measurement does not select one branch; it creates new branches, one for each possible outcome.

7.2 The Probability Problem

The most serious philosophical challenge for many-worlds is the probability problem (概率问题). If all branches exist and all outcomes occur, what is the status of the Born-rule probabilities? When a physicist says “there is a 70% probability of spin-up,” what can this mean in a universe where both the spin-up and spin-down branches exist?

This is sharpened by the problem of branch counting: if we simply count branches, a 70%-30% superposition has the same number of branches as a 50%-50% superposition (one of each), so naive branch counting does not reproduce the Born rule. Moreover, the notion of “number of branches” depends on the basis in which we decompose the wavefunction, which is not fixed by the physics.

Several responses have been attempted. The self-locating belief approach (Vaidman, Saunders) argues that the Born-rule probabilities are probabilities of which branch “you” find yourself on after a measurement — a matter of self-locating uncertainty. The decision-theoretic approach (Deutsch, Wallace) argues that probability talk in many-worlds is best understood not as self-locating belief but as rational preference.

7.3 Deutsch’s Decision-Theoretic Approach to the Born Rule

David Deutsch’s 1999 paper “Quantum Theory of Probability and Decisions” and Wallace’s subsequent elaboration attempt to derive the Born rule from rationality alone within the Everett framework. The strategy is to show that a rational agent who knows they are in an Everett universe and who accepts a handful of decision-theoretic axioms must act as if Born-rule probabilities are the correct probabilities for outcomes.

The argument proceeds as follows. An agent faces a “quantum game” — an experiment with a known quantum superposition and payoffs assigned to each outcome. Rationality is cashed out via axioms: transitivity, dominance, and certain symmetry conditions (notably that the agent is indifferent between outcomes that have the same payoff structure, regardless of the “labeling” of branches). Deutsch argues that these axioms uniquely force expected utilities to be computed using the Born-rule weights \( |c_i|^2 \).

Wallace’s formalization introduces the equivalence principle (branches that are related by permutation symmetries of the state must be treated equally) and continuity (utilities vary continuously with amplitudes). From these, the Born rule is derived as a theorem. The key move is showing that any departure from Born-rule weighting is ruled out by the rationality axioms.

Critics have questioned whether the axioms are truly non-circular (do they smuggle in the Born rule covertly?), whether the decision-theoretic framework is appropriate for self-locating uncertainty, and whether rationality axioms can determine probabilities in a context where all outcomes are certain to occur. The debate is ongoing and technically demanding.

7.4 Wallace’s Decoherence-Based Development

Wallace’s The Emergent Multiverse integrates decoherence into the Everett program. The problem is that quantum mechanics does not, in the raw formalism, identify “branches” — the Hilbert space admits infinitely many decompositions. Wallace argues that decoherence (Chapter 8) provides the physical mechanism that makes certain decompositions quasi-classical and robust, thereby selecting the preferred branch structure.

The central claim: branches are not fundamental features of the wavefunction but emergent structures, defined by patterns in the universal wavefunction that are sufficiently stable, quasi-classical, and decohered to qualify as autonomous worlds. Just as classical tables and chairs are real emergent entities that supervene on molecular physics, worlds are real emergent entities that supervene on the universal wavefunction.

This is a functionalist account of branches: a branch exists if and only if there are functional structures within the wavefunction that instantiate the relevant properties. The apparent vagueness of branch individuation is no worse, Wallace argues, than the apparent vagueness in how many clouds there are — it reflects the inherently approximate nature of emergent descriptions, not a fundamental deficiency.

7.5 Philosophical Assessment

Many-worlds has several philosophical attractions. It takes quantum mechanics completely seriously — no additional variables, no collapse. It is fully deterministic at the level of the universal wavefunction. And it is, in principle, compatible with relativistic quantum field theory.

The costs are equally serious. The ontology is profligate: every quantum branching creates a new “world.” What is the metaphysics of these worlds? Are they spatiotemporally distinct? Do they share matter? The probability interpretation remains controversial. And many philosophers find the view that all outcomes occur — including every catastrophe and atrocity, each with equal ontological dignity — deeply counterintuitive.

---

Chapter 8: Decoherence and the Classical Limit

8.1 The Mechanism of Decoherence

Decoherence (退相干) is the process by which quantum coherences — superpositions — become effectively invisible due to the entanglement of a system with its environment. It is a physical, dynamical mechanism, not an interpretive posit.

Consider a system \( S \) in a superposition \( |\psi_S\rangle = \alpha|s_1\rangle + \beta|s_2\rangle \) interacting with an environment \( E \) in initial state \( |e_0\rangle \). After interaction, the combined state is:

\[ |\Psi\rangle = \alpha|s_1\rangle|e_1\rangle + \beta|s_2\rangle|e_2\rangle \]

If the environment states \( |e_1\rangle \) and \( |e_2\rangle \) are orthogonal — which happens rapidly in realistic environments with many degrees of freedom — then the reduced density matrix of \( S \) becomes:

\[ \hat{\rho}_S = \text{Tr}_E(|\Psi\rangle\langle\Psi|) = |\alpha|^2|s_1\rangle\langle s_1| + |\beta|^2|s_2\rangle\langle s_2| \]

The off-diagonal interference terms have vanished. The system looks classical: it appears to be in a statistical mixture of definite states with the Born-rule probabilities.

8.2 Zurek’s Einselection and Pointer States

Wojciech Zurek introduced the concept of einselection (环境选择) — environmentally induced superselection — to explain why decoherence picks out specific classical states (pointer states) rather than arbitrary superpositions. The pointer states (指针态) are those that are most stable under interaction with the environment: they are the eigenstates of the interaction Hamiltonian between system and environment.

Einselection explains why we see position eigenstates (and not energy eigenstates or momentum eigenstates) for macroscopic objects: position couples most strongly to most environments. A dust grain scattering photons and air molecules will have its position measured continually by the environment; energy superpositions are more stable. This is the physical explanation of the classical limit.

Quantum Darwinism is Zurek’s further development: information about pointer states gets redundantly encoded in the environment. Multiple observers can independently measure fragments of the environment and agree on the system’s properties — this is why macroscopic objects seem to have objective, intersubjectively accessible properties. The environment acts as a communication channel, broadcasting information about pointer states.

8.3 Decoherence Timescales

Decoherence is extraordinarily fast for macroscopic objects. For a dust grain of mass \( 10^{-15} \) kg in a superposition of two positions 1 nm apart at room temperature, the decoherence time due to air molecule scattering is of order \( 10^{-23} \) seconds — vastly shorter than any humanly accessible timescale. This explains why macroscopic superpositions are never observed in practice: the environment destroys them faster than any measurement apparatus could detect them.

For microscopic systems, decoherence is much slower. A carefully isolated atom in an optical trap can maintain coherence for seconds or more. This is the regime where quantum computers and quantum communication devices operate.

8.4 What Decoherence Does and Does Not Solve

Decoherence is a crucial physical insight, but it does not, by itself, solve the measurement problem. The distinction between what decoherence does and does not accomplish must be stated carefully.

What decoherence does:

• Suppresses off-diagonal elements of the reduced density matrix, making the system appear classical.
• Selects a preferred basis (the pointer basis) for the quasi-classical description.
• Explains the extreme rarity of observable macroscopic superpositions.
• Provides a dynamical account of the quantum-to-classical transition.
• Gives many-worlds a principled account of branch structure (Wallace’s program).

What decoherence does not do:

• It does not select a single branch — all branches remain present in the global wavefunction.
• It does not explain why this particular outcome occurs rather than another.
• It does not collapse the quantum state in the physical sense — the global pure state \( |\Psi\rangle \) remains a superposition.
• It does not by itself yield the Born rule — one still needs an interpretive framework to extract probability from amplitude.

The confusion arises because the reduced density matrix \( \hat{\rho}_S = |\alpha|^2|s_1\rangle\langle s_1| + |\beta|^2|s_2\rangle\langle s_2| \) looks like a classical mixture. But a proper mixture and an improper mixture (the reduced state of an entangled system) are empirically indistinguishable from the perspective of the subsystem alone, yet ontologically distinct. Decoherence produces improper mixtures; a genuine solution to the measurement problem requires an explanation of why they are as if proper mixtures — which is precisely what each major interpretation attempts in its own way.

---

Chapter 9: GRW — Spontaneous Collapse Theories

9.1 The GRW Proposal

In 1986, Ghirardi, Rimini, and Weber proposed a modification of quantum mechanics that makes collapse a physical, objective process. The GRW theory (GRW理论) adds a stochastic, spontaneous localization mechanism to the Schrödinger equation. Each particle has a small probability per unit time of undergoing a spontaneous “hit” — a localization in position. After a hit, the particle’s wavefunction is multiplied by a Gaussian of width \( r_C \approx 10^{-7} \) m, centered at a random position with probability given by the Born rule. The result is then renormalized.

The key parameters are:

• \( \lambda \approx 10^{-16} \) s\(^{-1}\): the spontaneous collapse rate per particle
• \( r_C \approx 10^{-7} \) m: the localization distance (collapse width)

9.2 Why GRW Works for Macro-Objects

For a single particle, the collapse rate \( \lambda \sim 10^{-16} \) s\(^{-1}\) is negligible — collapses are extremely rare. But a macroscopic object has \( N \sim 10^{23} \) particles. The composite collapse rate is \( N\lambda \sim 10^{7} \) s\(^{-1}\) — macroscopic superpositions collapse essentially instantaneously. Thus, GRW is designed to preserve the empirical success of quantum mechanics for microscopic systems while ensuring definite outcomes for macroscopic ones.

9.3 The Tails Problem

GRW faces the tails problem (尾问题), identified by Albert and Loewer. After a GRW hit localizes a wavefunction to a Gaussian of width \( r_C \), the wavefunction does not vanish outside that region — it has tails that extend, with rapidly diminishing amplitude, to all of space. Strictly speaking, a particle never has a perfectly definite position in GRW; its wavefunction always has tails.

For a macroscopic system, the tails have negligibly small amplitude — the probability of finding the cat’s constituent atoms far from the localized region is astronomically small. But the wavefunction is not zero there. If a physical property is defined by whether the wavefunction is identically zero in a region, then GRW never yields strictly definite properties; if it is defined by whether the amplitude is negligibly small, then some threshold must be specified — reintroducing vagueness.

Albert and Loewer’s fuzzy link proposal: a system has the property \( P \) if the probability assigned by the wavefunction to the complement of \( P \) is below a threshold. This introduces a new primitive (the threshold), but it is argued to be no worse than other vagueness in physics. Lewis’s critique: the fuzzy link, applied carefully, implies that macroscopic pointers are never definitely in any position, since the tails from all GRW collapses contribute.

9.4 Ontological Options: Flash, Mass Density

GRW as a mathematical theory does not immediately specify what the physical ontology is. Two main options have been defended:

• GRWf (flash ontology): The fundamental physical events are the collapse “flashes” — point-like events in spacetime at which collapses occur. The macroscopic world is constituted by patterns of flashes. Bell championed this view because it offers a primitive ontology that is local in spacetime, even though the wavefunction on configuration space guides the flash distribution non-locally.
• GRWm (mass density ontology): The fundamental ontology is a continuous matter density field \( m(x,t) = \sum_k m_k |\psi_k(x,t)|^2 \) derived from the wavefunction. The cat is alive if the mass density is concentrated in the “alive” configuration.

Each ontology handles the tails problem differently. In GRWm, the tails represent a vanishingly small but nonzero matter density in the “wrong” region — the cat has a negligible but nonzero matter-density tail in the “dead” configuration. Whether this is problematic depends on how we read the ontology.

9.5 CSL: Continuous Spontaneous Localization

The Continuous Spontaneous Localization (CSL) model (Pearle; Ghirardi, Pearle, Rimini) is a continuous-time version of GRW that replaces the discrete “hits” with a continuously evolving stochastic field coupled to the mass density. The Schrödinger equation is modified to:

\[ d|\psi\rangle = \left[-\frac{i}{\hbar}\hat{H}\,dt + \sqrt{\lambda}\int d^3x\, (\hat{M}(\mathbf{x}) - \langle\hat{M}(\mathbf{x})\rangle)\,dW_t(\mathbf{x})\right]|\psi\rangle \]

where \( \hat{M}(\mathbf{x}) \) is the smeared mass density operator and \( dW_t(\mathbf{x}) \) is a Wiener noise field. CSL produces continuous collapse toward localized states without the abrupt hits of GRW, and it is more naturally formulated in a way that admits relativistic generalization.

Relativistic generalizations of spontaneous collapse theories are technically challenging because the stochastic noise field must be Lorentz-invariant. Tumulka’s relativistic GRW theory (rGRWf) succeeds for non-interacting particles in the flash ontology. A fully relativistic, interacting spontaneous collapse theory remains an open problem in foundations.

GRW/CSL makes predictions that differ from standard quantum mechanics: there are tiny violations of energy conservation (each localization event slightly broadens the energy distribution) and a background “collapse noise” that could in principle be detected. Current experiments constrain the parameter space but have not yet definitively tested or ruled out the theory.

---

Chapter 10: Bohmian Mechanics

10.1 The Pilot Wave Theory

Bohmian mechanics (玻姆力学), also called the pilot wave theory, was proposed by Louis de Broglie in 1927 and rigorously developed by David Bohm in 1952. It provides a deterministic, realist completion of quantum mechanics by adding particle trajectories to the wavefunction description.

In Bohmian mechanics, there are two ontological primitives:

1. The wavefunction \( \psi(q, t) \) defined on configuration space, evolving by the Schrödinger equation.
2. Actual particle positions \( Q_1(t), \ldots, Q_N(t) \) evolving by the guidance equation (引导方程):

\[ \frac{dQ_k}{dt} = \frac{\hbar}{m_k} \text{Im}\left(\frac{\nabla_k \psi}{\psi}\right)\bigg|_{\mathbf{Q}} \]

Writing \( \psi = R e^{iS/\hbar} \) in polar form (where \( R \) and \( S \) are real), this becomes:

\[ \frac{dQ_k}{dt} = \frac{\nabla_k S}{m_k}\bigg|_{\mathbf{Q}} \]

The guidance equation says that each particle’s velocity is determined by the gradient of the phase of the wavefunction evaluated at the actual particle configuration. The wavefunction guides the particles — hence “pilot wave” — but is not itself directly observable.

10.2 The Quantum Potential

An alternative formulation, closer to Bohm’s original 1952 presentation, introduces the quantum potential (量子势):

\[ U_Q = -\sum_k \frac{\hbar^2}{2m_k}\frac{\nabla_k^2 R}{R} \]

The equation of motion for particle \( k \) then takes a Newtonian form:

\[ m_k \frac{d^2 Q_k}{dt^2} = -\nabla_k (V + U_Q) \]

where \( V \) is the classical potential. The quantum potential encodes the “wave” aspects of quantum behavior: it is responsible for interference and tunneling. Crucially, \( U_Q \) depends on the global form of \( R = |\psi| \) — it is non-local for entangled systems, as the quantum potential for particle \( k \) depends on the positions of all other particles through the many-body wavefunction.

10.3 Equivariance and the Born Rule

In Bohmian mechanics, the quantum statistics arise because we have a quantum equilibrium distribution: the probability density for particle positions is \( |\psi(q,t)|^2 \), the Born-rule distribution.

Dürr, Goldstein, and Zanghì have argued that quantum equilibrium is not merely an assumption but is typical in a Bohmian universe: for almost all initial conditions of the universe’s particles (with respect to the natural measure on the space of initial conditions), the statistical behavior of subsystems converges to the Born rule. This is analogous to how the second law of thermodynamics emerges as typical from classical mechanics.

10.4 How Bohmian Mechanics Solves the Measurement Problem

In Bohmian mechanics, the measurement problem dissolves. There is no collapse: the wavefunction always evolves by the Schrödinger equation. The apparent randomness of measurement outcomes is not fundamental — it reflects our ignorance of the particles’ actual initial positions. The wavefunction develops branches corresponding to different outcomes; the actual particles follow one branch (selected deterministically by their initial conditions, which we do not know precisely). The other branches become “empty waves” that carry no particles and play no further physical role.

The resolution of Schrödinger’s cat: after interaction, the wavefunction of the universe contains a superposition of the “cat alive” and “cat dead” branches. But the actual particles are either in the “alive” configuration or the “dead” configuration — there is a fact of the matter about which one, determined by the initial particle positions. We see a definite outcome because we are made of particles with definite positions.

10.5 Non-Locality in Bohmian Mechanics

Bohmian mechanics is explicitly non-local. The guidance equation for particle \( k \) depends on the positions of all other particles simultaneously — even when they are spacelike separated. For a two-particle entangled state, the velocity of particle 1 at position \( \mathbf{q}_1 \) depends on the current position \( \mathbf{q}_2 \) of particle 2, no matter how far apart they are.

This is how Bohmian mechanics accommodates Bell non-locality. When Alice measures her particle and “collapses” (in the effective-collapse sense) the wavefunction, the guidance equation for Bob’s particle instantaneously changes. Crucially, this non-locality cannot be used to send signals: since we do not know the actual particle positions (only their statistical distribution \( |\psi|^2 \)), we cannot exploit the non-local influences for signaling. The no-signaling result is guaranteed by equivariance.

Maudlin argues that Bohmian mechanics’s explicit non-locality should be taken as evidence that quantum mechanics demands a fundamentally non-local metaphysics — and that attempts to restore locality (through many-worlds, relational mechanics, etc.) are less honest about this implication.

10.6 Philosophical Virtues and Difficulties

Bohmian mechanics is a genuinely realist theory: particles exist, have definite positions, and follow deterministic trajectories. The wavefunction is a real physical field on configuration space. The Born rule is derived, not postulated. The theory is empirically equivalent to standard quantum mechanics (for all experiments that fit within its domain).

The difficulties are also significant. The wavefunction lives on configuration space (a \( 3N \)-dimensional manifold for \( N \) particles), not on ordinary three-dimensional space — what is the ontological status of this high-dimensional entity? Extension to relativistic quantum field theory is technically difficult. And many find the “empty waves” metaphysically extravagant.

---

Chapter 11: Quantum Information and Its Interpretive Implications

11.1 The Rise of Quantum Information Theory

The late 1980s and 1990s saw the emergence of quantum information theory (量子信息理论) as a discipline at the intersection of physics, computer science, and mathematics. Nielsen and Chuang’s textbook Quantum Computation and Quantum Information (2000) established the standard framework. Key resources of quantum information include:

• The qubit (量子比特): a two-state quantum system, the quantum analog of the classical bit, with state space \( \alpha|0\rangle + \beta|1\rangle \).
• Quantum entanglement: a non-classical correlation resource enabling teleportation, super-dense coding, and quantum key distribution.
• No-cloning theorem: there is no operation that can produce a perfect copy of an unknown quantum state, a fundamental asymmetry with classical information.
• Quantum error correction and fault-tolerant computation.

The philosophical question this course addresses is: what, if anything, does the quantum-information framework tell us about the interpretation of quantum mechanics?

11.2 Jeffrey Bub’s Information-Theoretic Interpretation

Jeffrey Bub, in Interpreting the Quantum World and subsequent papers (including work with Pitowsky), argues that quantum mechanics is fundamentally a theory of information (信息) — specifically, a theory about what sorts of information-processing tasks are possible in our world. Quantum mechanics selects a particular non-classical probability theory from a space of possible theories (defined by convex sets of states).

On Bub’s view, quantum mechanics is not primarily a theory of matter and forces but a theory of information structure. The wavefunction is not a description of an underlying physical reality; it is a calculational device encoding probabilistic constraints on possible informational events.

The Bub-Pitowsky theorem is the technical core: quantum mechanics, characterized information-theoretically, can be derived from two fundamental information-theoretic principles — (i) there is a maximum of one bit of information accessible from a two-outcome measurement on any elementary system, and (ii) maximal information about a composite system does not uniquely determine the information about its subsystems (entanglement). These two principles, framed within the structure of non-Boolean event algebras, uniquely characterize the quantum probability calculus.

11.3 QBism: Quantum Bayesianism

QBism (量子贝叶斯主义), developed by Christopher Fuchs, Carlton Caves, and Rüdiger Schack, takes the subjectivist interpretation of probability further than any previous view. In QBism, the quantum state \( |\psi\rangle \) is an agent’s personal probability assignment — not a description of the external world, but a summary of the agent’s degrees of belief about future experiences.

The Born rule, in QBism, is a normative constraint on how agents ought to update their beliefs in response to actions on the world. It is a consistency condition, analogous to the Dutch-book coherence constraints of classical Bayesian probability theory, not a fundamental law of nature. When a QBist agent “measures” a quantum system, the outcome is a new experience that prompts belief revision. There is no wave collapse because the wavefunction never represented the world to begin with.

Fuchs characterizes QBism as a participatory view: the agent is not a passive observer but an active participant whose actions bring new facts into being. This has affinities with pragmatism and with William James’s radical empiricism, which Fuchs has explored explicitly. The name “quantum Bayesianism” was later contracted to “QBism” to signal that it is not merely a Bayesian interpretation of probability applied to quantum mechanics, but a deeper philosophical position about the nature of experience, agency, and reality.

Critics of QBism press several objections. First, the intersubjectivity problem: if each agent has their own quantum state assignment, what explains the fact that different agents, performing independent measurements on the same system, tend to agree on outcomes? QBism’s answer — that agents are embedded in a shared world that constrains which beliefs are reasonable — seems to smuggle in the objective world QBism was supposed to dissolve. Second, the reference class problem: if the Born rule is a normative constraint on agent degrees of belief, which agents? There is no natural boundary between “agents” and “non-agents” in physics. Third, Wigner’s friend: when Wigner assigns a quantum state to his friend in the box, and his friend assigns a different state to her own experience, whose state is correct? QBism says both are valid for their respective agents — but this threatens to undermine the project of physics as the search for a single, objective description of nature.

11.4 Relational Quantum Mechanics

Carlo Rovelli’s relational quantum mechanics (关系量子力学) (1996) proposes that quantum states are always relative to a system — there is no absolute, observer-independent state. What is real is not quantum states per se but the values of physical quantities that are actualized relative to a given reference system.

The motivating analogy is special relativity: velocity is not an absolute property of an object but a relation between two systems. Similarly, for Rovelli, the quantum state of system A is not an intrinsic property of A but a relation between A and some other system B (the “observer” — which need not be conscious). When B interacts with A and “measures” a property, that property takes a definite value relative to B. It may not take a definite value relative to some other system C that has not interacted with A.

This relational structure explains the Wigner’s friend scenario naturally: Wigner’s friend assigns a definite outcome relative to herself; Wigner assigns a superposition relative to himself. Both assignments are correct in their respective relational contexts. There is no contradiction because “the state” is not an absolute property but a relation.

Rovelli argues that the quantum formalism is best understood as a language of relations between physical systems, not a description of intrinsic properties. The correlations that quantum mechanics describes are real — they are the facts about what values systems take relative to one another — but the quantum state itself has no mind-independent physical correlate.

Criticisms of relational QM include: (i) The consistency problem — if system C subsequently interacts with both A and B, can the relative facts be made consistent? Rovelli asserts yes but the formal proof is contested. (ii) The ontology problem — if quantum states are relational, what is the fundamental ontology of the world? Rovelli speaks of “events” (actualizations of values relative to some system), but the nature of these events and how they ground the relational structure is underdeveloped. (iii) The comparison with special relativity may be misleading: in relativity, there is a Lorentz-invariant structure (the metric) that underlies the frame-dependent descriptions; the analog in relational QM is not obvious.

11.5 Evaluating Quantum Information Interpretations

Quantum information interpretations share a common impulse: to take seriously the idea that information — rather than matter or fields — is the primary ontological category. This represents a significant departure from traditional scientific realism, which holds that physical theories describe mind-independent entities.

Several objections press against these approaches:

1. The reference class problem: If quantum states encode beliefs, whose beliefs? What determines the appropriate reference class of observers?
2. The ontology question: If the wavefunction does not describe reality, what does? What is the physical world made of, on these views?
3. The normativity gap: The Born rule, on QBist and related views, is presented as a norm for rational agents. But why should the structure of rational belief match the structure of quantum correlations? This seems to require an independent explanation.
4. Wigner’s friend: In Wigner’s friend scenarios, two agents assign incompatible quantum states to the same system. Quantum information interpretations must give an account of how these incompatible assignments can both be valid.

---

Concluding Remarks: The Interpretive Landscape

After a century of reflection, quantum mechanics remains uninterpreted in any consensus sense. Each of the major programs offers genuine insights alongside genuine costs:

Interpretation	Collapse?	Hidden Variables?	Non-local?	Key Cost
Copenhagen	Yes (effective)	No	Ambiguous	Measurement dualism
Many-Worlds	No	No	No	Ontological proliferation; probability
Bohmian Mechanics	No	Yes	Yes	High-dim. configuration space
GRW	Yes (objective)	No	No	New parameters; relativistic extension
QBism	No (belief update)	No	No	Anti-realism; agent-dependence
Relational QM	No (relational)	No	No	Consistency between observers

The philosophy of quantum theory is not merely a historical curiosity or a problem for physicists to resolve. It touches the deepest questions in metaphysics (what is real?), epistemology (what is a measurement?), and philosophy of science (what is a physical theory for?). Students of philosophy are particularly well positioned to contribute to these debates — not by doing physics, but by bringing conceptual clarity, careful argumentation, and philosophical breadth to questions that physics alone cannot resolve.

The best contemporary work in foundations of quantum mechanics is genuinely interdisciplinary: it requires knowing the formalism, reading the experiments, and reasoning carefully about ontology, probability, and explanation. This course aims to equip students to engage with this literature at the frontier.