The Copenhagen Interpretation

The Copenhagen interpretation is the framework almost every physicist uses without naming it: a system evolves unitarily under the Schrödinger equation until a measurement occurs, at which point the wavefunction "collapses" to an eigenstate with Born-rule probability. It is not a single codified doctrine but a family of closely related positions developed at Bohr's institute in Copenhagen and at Göttingen between 1925 and 1935, associated above all with Niels Bohr, Werner Heisenberg, and Max Born. Its operational core is a two-process dynamics — von Neumann's "Process 2," the smooth unitary evolution between measurements, and "Process 1," the discontinuous, probabilistic jump at measurement — together with a movable classical–quantum boundary, the Heisenberg cut, and Bohr's principle of complementarity. This article owns the faithful reconstruction of that core and one modern sharpening. The reconstruction disentangles what is often conflated: Bohr's complementarity (a statement about mutually exclusive experimental arrangements) is not Heisenberg's cut, and neither is the cartoon "shut up and calculate." A recurring textbook error — that Bohr regarded wavefunction collapse as a physical process — is corrected here with the historical record: the language of a physical "reduction of the state" is von Neumann's and Dirac's, and the projection postulate itself is developed in Von Neumann Measurement and the Projection Postulate by name. The modern sharpening is the cut-movability result. A perennial worry about Copenhagen is that the location of the classical–quantum boundary seems arbitrary. Decoherence (developed in Quantum Decoherence and the Density Matrix by name) explains why it does not matter: wherever you place the cut, the measured statistics come out identical, because tracing over the environment renders the pointer's reduced state diagonal regardless. Copenhagen thus survives as a consistent operational framework — but it brackets the measurement problem rather than solving it, which is exactly the tension the rival interpretations were built to relieve.

Key takeaways

The Copenhagen interpretation is the orthodox operational framework: unitary evolution ("Process 2") between measurements, and a probabilistic collapse to an eigenstate ("Process 1") at measurement, with Born-rule weights.
Orthodox-rules ledger (owned): the four commitments — the Heisenberg cut, unitary evolution, Born probabilities plus collapse, and Bohr's complementarity — reconstructed faithfully and kept distinct from one another (Bohr's complementarity ≠ Heisenberg's cut ≠ "shut up and calculate").
Cut-movability (owned): the Born statistics are identical wherever you place the classical–quantum cut — verified by treating the pointer quantum-mechanically and tracing out its environment (recalled from Quantum Decoherence and the Density Matrix).
History-as-physics repair (owned): Bohr never described collapse as a physical process; the "reduction of the state" is von Neumann/Dirac's language and mechanism (developed in Von Neumann Measurement and the Projection Postulate by name).
Copenhagen brackets the measurement problem — it prescribes when to apply collapse but gives no dynamics for it (developed in The Quantum Measurement Problem by name).
Worked: a spin measured by a pointer gives the same Born probabilities $(0.36, 0.64)$ whether the cut is drawn before or after the (decohered) pointer; complementarity is quantified by the duality relation $D^2 + V^2 \le 1$ .
Key idea: Copenhagen is a self-consistent recipe — unitary evolution plus a Born-rule collapse at a movable, ultimately statistics-independent cut — that works flawlessly as an operational framework while leaving the why of measurement open.

The core insight

The Copenhagen interpretation prescribes unitary evolution between measurements and a probabilistic collapse at measurement, across a movable Heisenberg cut, with Bohr's complementarity organizing which classical descriptions can coexist. Decoherence shows the cut's location does not affect predictions, so the framework is operationally consistent — but it stipulates when to collapse without explaining how or why, bracketing rather than solving the measurement problem.

Where this lives today

Copenhagen remains the de facto working framework of practical quantum mechanics — the rules in equation (1) are what every spectroscopist, quantum-optician, and device engineer applies, usually without comment. Its modern status has been reshaped, not overturned, by two developments. Decoherence theory turned the movable cut from an embarrassment into a near-triviality: the environment continuously measures macroscopic systems, so the classical–quantum boundary is not a mysterious posit but an emergent, robust feature, and the cut can sit anywhere in the decohered region without consequence (developed in Quantum Decoherence and the Density Matrix by name). Quantum information recast the Born rule and measurement in operational terms — POVMs, quantum channels, and information-theoretic reconstructions of the formalism — that are broadly Copenhagen-compatible in spirit while sharpening what "measurement" means. What has not changed is the residual conceptual debt: Copenhagen still does not say what physically distinguishes a "measurement" from any other interaction, which is why the interpretation is measured against rivals that each pay the debt differently — a real dynamical collapse (Objective Collapse Theories), no collapse at all (The Many-Worlds Interpretation, De Broglie–Bohm Pilot Wave Theory), or a reframing of the state's very meaning (Relational and QBist Interpretations) — all by name.

The physical picture

Copenhagen is best understood as a division of labor between two languages. The quantum system is described by a wavefunction that evolves smoothly and reversibly, encoding amplitudes and interference. The apparatus and the observer are described in classical language — pointer positions, clicks, marks on a screen — because, Bohr insisted, that is the only language in which we can unambiguously communicate an experimental result. Between these two worlds runs a boundary, the Heisenberg cut, and the rule is simple: everything on the quantum side evolves by the Schrödinger equation; the act of reading a definite result on the classical side is where Born's probabilities apply and the state "reduces" to the observed eigenstate. The subtle part is that the cut is movable. You may draw it between the atom and the photographic plate, or push it out to include the plate, the microscope, even the experimenter's retina, as quantum systems — and Copenhagen insists the predictions do not change. Bohr's complementarity supplies the other half of the picture: a single experiment can reveal position or momentum, wave or particle, but never both, because the experimental arrangements that would reveal each are mutually exclusive. The two aspects are not contradictory; they are complementary, each a legitimate but partial classical account, jointly necessary and never simultaneously available. What the framework deliberately does not provide is a mechanism for the reduction — and that silence is the seam along which every rival interpretation pulls.

Figure

The Heisenberg cut is movable: the same Born statistics wherever you draw it

The formalism: two processes and a cut

The operational content of Copenhagen is compact. Between measurements the state obeys the Schrödinger equation; at a measurement of $\hat A$ with eigenstates $|a_n\rangle$ , the state jumps to $|a_n\rangle$ with probability $|\langle a_n|\psi\rangle|^2$ . These are von Neumann's Process 2 and Process 1, and the interpretation's job is to say where the boundary between them lies and why the placement does not matter. The two-process rule, stated together:

\text{Process 2:}\quad i\hbar\,\partial_t|\psi\rangle = \hat H|\psi\rangle \qquad\qquad \text{Process 1:}\quad |\psi\rangle \xrightarrow{\text{measure }\hat A}\; |a_n\rangle \;\text{ w.p. } |\langle a_n|\psi\rangle|^2 \qquad (1)

where: Process 2 is deterministic, reversible, and continuous; Process 1 is stochastic, irreversible, and discontinuous. Read them physically: the wavefunction carries the full quantum description and evolves unitarily whenever the system is not being measured, and the Born rule (owned by Born Rule and Measurement Probabilities) supplies the only observable content — probabilities — with the eigenstate projection (the collapse, developed in Von Neumann Measurement and the Projection Postulate by name) applied at the cut. The interpretive weight is entirely in the word "measurement": Copenhagen tells you when to switch from Process 2 to Process 1, but not why the switch is physically necessary — the gap that is the measurement problem.

Derivation ACut-movability: the Born statistics do not depend on where you draw the line

We show that treating the pointer as a classical device (cut before the pointer) or as a quantum system that decoheres (cut after the pointer) gives identical outcome probabilities. This is the result that rescues Copenhagen from the charge that its movable cut is arbitrary: the arbitrariness is harmless because it is invisible to predictions. The engine is that tracing out an environmental record diagonalizes the pointer's reduced state, reproducing the Born-rule populations.

Step 2. Cut after the pointer (quantum pointer). Now treat the pointer — and an environment that records it — quantum-mechanically. A faithful measurement plus environmental monitoring produces the entangled state $a|0\rangle|P_0\rangle|E_0\rangle + b|1\rangle|P_1\rangle|E_1\rangle$ with orthogonal pointer and environment records. Tracing out the system and the environment (partial trace recalled from The Reduced Density Matrix and Partial Trace) leaves the pointer in the reduced state $\rho_P = \mathrm{diag}(|a|^2, |b|^2)$ — diagonal, with populations equal to the Born probabilities.

Step 3. Compare and conclude. The pointer populations from Step 2 equal the system Born probabilities from Step 1, and the reduced state is diagonal (no coherence to distinguish a "genuine" collapse from a decohered mixture at the level of statistics). Pushing the cut further out — adding more apparatus stages correlated with the pointer — leaves the pointer marginal fixed at $(|a|^2, |b|^2)$ . Why this matters: the location of the Heisenberg cut is a matter of description, not of physics; every placement yields the same measured statistics, so Copenhagen's movable cut is operationally well-defined even though it is conceptually uncomfortable. *Verified numerically: for $a = 0.6, b = 0.8$ (so $|a|^2 = 0.36, |b|^2 = 0.64$ ), the quantum-pointer reduced state is $\mathrm{diag}(0.36, 0.64)$ with off-diagonal $0$ , matching the direct Born probabilities exactly; moving the cut outward leaves the marginal unchanged.*

Go deeper · Derivation B · 1 minComplementarity, quantified: the wave–particle duality relation

Derivation BComplementarity, quantified: the wave–particle duality relation

We make Bohr's complementarity precise for a two-way interferometer as a quantitative trade-off between path knowledge and interference visibility, $D^2 + V^2 \le 1$ . The point is that complementarity is not vague philosophy but a sharp inequality: the more which-path information an arrangement provides, the less interference it can show, with a hard bound saturated by pure states. The engine is that path predictability and fringe visibility are read from the diagonal and off-diagonal of the same $2\times2$ density matrix.

Step 2. Combine them. Directly, $D^2 + V^2 = \cos^2 2\chi + \sin^2 2\chi = 1$ for this pure state. Why the trade-off is forced: $D$ and $V$ are $|\cos2\chi|$ and $|\sin2\chi|$ , so as one rises the other must fall — full path knowledge ( $D=1$ , $\chi=0$ ) kills the fringes ( $V=0$ ), and an equal superposition ( $D=0$ , $\chi=\pi/4$ ) gives maximal fringes ( $V=1$ ). Wave and particle behavior are the two ends of one dial, exactly Bohr's mutually exclusive arrangements.

Step 3. Include mixedness. For a partially decohered path state with reduced coherence $\gamma$ , the visibility drops to $V = 2\times\tfrac12\gamma = \gamma$ while $D$ stays fixed, so $D^2 + V^2 < 1$ — the inequality becomes strict. Why: coupling to an environment (which-path marking) lowers the visibility without necessarily raising the predictability, moving the state inside the unit circle. The bound is saturated only by pure states. *Verified numerically: $D^2 + V^2 = 1$ at $\chi = 0, \pi/8, \pi/4, \pi/3$ ; $\chi=0$ gives $(D,V)=(1,0)$ (particle-like), $\chi=\pi/4$ gives $(0,1)$ (wave-like); a mixed path state with $\gamma = 0.5$ gives $D^2 + V^2 = 0.25 < 1$ .*

Figure

Complementarity as a hard bound: pure states ride the unit circle, decoherence pushes inside

Limiting cases

case	result	reading
no measurement	pure Process 2 (unitary)	Schrödinger evolution, full coherence
projective measurement	Process 1: collapse to $\|a_n\rangle$ , prob $\|\langle a_n\|\psi\rangle\|^2$	Born rule applied at the cut
cut moved outward	same statistics	cut-movability; decoherence diagonalizes the pointer
full which-path ( $\chi=0$ )	$D=1, V=0$	particle-like: which-path known, no fringes
equal superposition ( $\chi=\pi/4$ )	$D=0, V=1$	wave-like: no path info, maximal fringes
decohered path ( $\gamma<1$ )	$D^2+V^2<1$	mixedness lowers visibility, strict inequality

Common trapBohr did not describe wavefunction collapse as a physical process — that is a later conflation

The textbook shorthand "Copenhagen says the wavefunction physically collapses" misrepresents Bohr and blurs a real historical distinction. Bohr's own writing centers on complementarity and the necessity of classical concepts for describing experiments; he treated the quantum state instrumentally and did not speak of a physical "reduction of the wave packet" as an added dynamical event. The language and mechanism of state reduction — Process 1 as a projection — are due to von Neumann (1932) and Dirac, and belong to Von Neumann Measurement and the Projection Postulate by name. Heisenberg's emphasis was different again: a transition from "potentiality" to "actuality" tied to the observer's knowledge. Lumping these together as one "collapse" doctrine erases the fact that Copenhagen is a family of positions, and it manufactures a straw target — a physical collapse that Bohr never asserted. The reliable statement is: Copenhagen prescribes when to apply the Born rule and the eigenstate update, and is deliberately silent on any underlying mechanism, which is precisely why the measurement problem (developed in The Quantum Measurement Problem by name) remains open within it.

Worked examples

Example 1 — the cut is invisible to the data (confidence builder). A spin is prepared in $|\psi\rangle = 0.6\,|0\rangle + 0.8\,|1\rangle$ and measured in the $\{|0\rangle, |1\rangle\}$ basis by a pointer. Show the outcome probabilities are the same whether the pointer is treated classically or quantum-mechanically.

Step 1. Classical pointer (cut before). Apply Born directly: $p_0 = |0.6|^2 = 0.36$ , $p_1 = |0.8|^2 = 0.64$ (and $0.36 + 0.64 = 1$ , as required).

Step 2. Quantum pointer (cut after). Entangle system, pointer, and environment: $0.6\,|0\rangle|P_0\rangle|E_0\rangle + 0.8\,|1\rangle|P_1\rangle|E_1\rangle$ . Trace out system and environment; the pointer reduced state is $\rho_P = \mathrm{diag}(0.36, 0.64)$ with off-diagonal $0$ (machine-verified). Its populations are the outcome probabilities: $0.36$ and $0.64$ .

Step 3. Read the moral. The two accounts agree exactly, and moving the cut further out (more apparatus stages) leaves the pointer marginal at $(0.36, 0.64)$ . Whether you call the pointer "classical" or "a decohered quantum system" changes the story you tell, not any number you measure. Copenhagen's cut is a bookkeeping choice; the physics is cut-independent.

Example 2 — complementarity as a hard inequality (exam level). In a two-path interferometer with path state $\cos\chi\,|1\rangle + \sin\chi\,|2\rangle$ , compute the path predictability $D$ and fringe visibility $V$ at three settings, and check the duality bound.

Step 1. Full which-path, $\chi = 0$ . Then $D = |\cos0| = 1$ and $V = |\sin0| = 0$ : complete path knowledge, zero fringes. $D^2 + V^2 = 1$ (machine-verified) — pure particle behavior.

Step 2. Equal superposition, $\chi = \pi/4$ . Then $D = |\cos(\pi/2)| = 0$ and $V = |\sin(\pi/2)| = 1$ : no path information, maximal fringes. $D^2 + V^2 = 1$ (machine-verified) — pure wave behavior. The same interferometer shows opposite faces at opposite settings, never both at once.

Step 3. A decohered path, $\gamma = 0.5$ . If the environment partially marks the path, the reduced coherence drops to $\gamma = 0.5$ , so $V = \gamma = 0.5$ while $D = 0$ (balanced populations), giving $D^2 + V^2 = 0.25 < 1$ (machine-verified). The state has slid inside the unit circle: neither cleanly wave nor cleanly particle, with the missing certainty carried off into the environment. The bound $D^2 + V^2 \le 1$ is Bohr's complementarity made numerical.

Assembled at Bohr's institute

Origins · 1 mincopenhagen interpretation and how this was discovered

copenhagen interpretation

1925–1935 · Copenhagen, Göttingen

The "orthodox" reading of quantum mechanics has no single author and, at first, no name. It was assembled at Niels Bohr's Institute for Theoretical Physics in Copenhagen and at Göttingen in the decade after 1925, out of several distinct strands: Max Born's 1926 discovery that $|\psi|^2$ is a probability density; Werner Heisenberg's uncertainty relations and his emphasis on a movable classical–quantum boundary; and, most distinctively, Bohr's principle of complementarity, unveiled at the 1927 Como conference, holding that mutually exclusive experimental arrangements yield complementary, jointly necessary classical descriptions. The participants did not agree on everything — Bohr's complementarity and Heisenberg's "potentiality-to-actuality" reading pull in different directions — which is why "the Copenhagen interpretation" is best understood as a family rather than a creed. The name itself came late: Heisenberg popularized "Copenhagen interpretation" only around 1955, partly to demarcate the orthodox view from the rising alternatives of Bohm and Everett. The framework's completeness claim was pressed hardest by the 1935 EPR paper, which forced the debate that Bell would later make experimental.

History

1926

— Born: the wavefunction gives probabilities,

P = |\langle a_n|\psi\rangle|^2

1927

— Heisenberg's uncertainty relations; Bohr unveils complementarity at Como

1932

— von Neumann formalizes measurement (Process 1) and the projection postulate

1935

— EPR challenge the completeness of the quantum description

1955

— Heisenberg coins and popularizes the label "Copenhagen interpretation"

Going further

References. N. Bohr, Nature 121, 580 (1928) (Como lecture on complementarity) · M. Born, Z. Phys. 37, 863 (1926) · J. von Neumann, Mathematische Grundlagen der Quantenmechanik (1932), Ch. V–VI · W. Heisenberg, Physics and Philosophy (1958) · Englert, Phys. Rev. Lett. 77, 2154 (1996) (duality relation) · Griffiths & Schroeter, Introduction to Quantum Mechanics, 3rd ed., Ch. 12 (Afterword).

Provenance. All results are standard. The Born rule and measurement probabilities are owned by Born Rule and Measurement Probabilities; the projection postulate and the von Neumann measurement chain by Von Neumann Measurement and the Projection Postulate; the partial trace and reduced states by The Reduced Density Matrix and Partial Trace; decoherence and the diagonalization of the pointer's reduced state by Quantum Decoherence and the Density Matrix — all recalled by name. Owned here: (1) the faithful orthodox-rules ledger (Heisenberg cut, Process 1/2, Born plus collapse, complementarity) kept distinct; (2) the cut-movability result, verified numerically — a spin measured with $a=0.6, b=0.8$ gives the same $(0.36, 0.64)$ Born statistics whether the pointer is classical or a decohered quantum system; (3) the complementarity duality relation $D^2 + V^2 \le 1$ , verified numerically (saturated at $\chi=0,\pi/8,\pi/4,\pi/3$ for pure states, and $0.25<1$ for a decohered path with $\gamma=0.5$ ); and (4) the history-as-physics repair that Bohr did not treat collapse as a physical process. The measurement problem itself is developed in The Quantum Measurement Problem, and the rival interpretations in Objective Collapse Theories, The Many-Worlds Interpretation, De Broglie–Bohm Pilot Wave Theory, and Relational and QBist Interpretations — all by name.

Builds on: Born Rule and Measurement Probabilities · The Density Operator and Mixed States · Properties and Time Evolution of the Density Matrix · Quantum Decoherence and the Density Matrix

Leads to: The Quantum Measurement Problem · The Many-Worlds Interpretation · De Broglie–Bohm Pilot Wave Theory

Is "shut up and calculate" the Copenhagen interpretation?

No — that slogan (attributed to Mermin, describing a common attitude) is an instrumentalist minimalism that declines to interpret at all, whereas Copenhagen makes substantive claims: the completeness of $\psi$ , the necessity of a classical description of the apparatus, and Bohr's complementarity. One can use the calculational rules while refusing Copenhagen's interpretive commitments (that is what "shut up and calculate" recommends), or one can adopt the full Bohr–Heisenberg framework. Conflating them flattens a real distinction: Copenhagen has a philosophy of measurement; the minimalist stance is the deliberate refusal of one.

Where exactly is the Heisenberg cut supposed to be?

Nowhere in particular — and that is the point of the cut-movability result. Copenhagen requires a boundary between the quantum-described system and the classically-described apparatus, but it does not fix its location: you may include more or fewer degrees of freedom on the quantum side and the predictions are unchanged, because decoherence diagonalizes the pointer's reduced state wherever the cut sits. Historically this movability was seen as a weakness (the cut looks arbitrary); the decoherence account reframes it as harmless, since the boundary is an emergent, prediction-independent feature rather than a fundamental posit. What Copenhagen cannot do is push the cut all the way out to include the observer without confronting the measurement problem.

Does decoherence rescue Copenhagen or replace it?

It rescues the operational framework and sharpens its silence. Decoherence explains why macroscopic pointers behave classically and why the cut can be placed anywhere without affecting statistics — both long-standing Copenhagen puzzles. But decoherence produces an improper mixture (the global state stays pure), so it does not by itself select a single definite outcome; it explains the appearance of collapse, not collapse as a fact. Copenhagen still needs to stipulate that a definite outcome occurs (Process 1). So decoherence is best seen as making Copenhagen's operational rules well-founded while leaving its central conceptual gap — why a definite result — exactly where it was, in the measurement problem.

How does complementarity relate to the uncertainty principle?

They are close cousins but not identical. The uncertainty principle is a quantitative bound on the spreads of conjugate observables in a given state ( $\Delta x\,\Delta p \ge \hbar/2$ ). Complementarity is a statement about experimental arrangements: the apparatus that would sharply measure position excludes the one that would sharply measure momentum, so the two descriptions cannot be jointly realized. The duality relation $D^2 + V^2 \le 1$ is complementarity made quantitative for a two-path interferometer, and it implies a which-path/visibility trade-off that is logically distinct from (though related to) the position–momentum uncertainty relation. Bohr regarded complementarity as the deeper, more general principle.

If Copenhagen brackets the measurement problem, why is it still the default?

Because it works — flawlessly, for every practical prediction — and because bracketing a foundational problem is often the pragmatic choice for people building lasers and computers rather than debating ontology. The two-process rule plus the Born rule reproduces all data, and the interpretive gap (what is a measurement, physically?) never bites in the laboratory, since apparatus are unambiguously classical thanks to decoherence. The gap matters when you ask about the wavefunction of the universe (no external observer), or push quantum coherence to ever-larger systems, or care about realism — which is exactly when physicists reach for the alternatives. Copenhagen is the default not because it answers the deep question but because, for most purposes, you never have to ask it.

Did the EPR paper refute Copenhagen?

It did not refute it, but it forced the issue that Copenhagen would rather have left implicit: whether the quantum description is complete. EPR argued that if quantum mechanics is local, then $\psi$ cannot be a complete description of reality — there must be "elements of reality" it omits. Bohr's reply defended completeness by denying the EPR criterion of reality (the elements of reality depend on the whole experimental context, a complementarity move). The debate was philosophical until Bell turned it into an experimental question. The upshot: EPR did not overturn Copenhagen's predictions (nothing has), but it exposed the tension between completeness and locality that motivates the hidden-variable and other rival programs, developed in The EPR Paradox and Local Realism and Bell's Theorem and the CHSH Inequality by name.