Complementarity

Complementarity was Niels Bohr’s proposal for how to speak responsibly about quantum phenomena when no single classical picture is adequate. Wave-like and particle-like descriptions, for example, may both be needed, but the experimental arrangements that make each description applicable are mutually exclusive.

This page treats complementarity historically and conceptually. It is not a mathematical theorem. Modern quantum mechanics expresses many of the same constraints through states, noncommuting observables, measurement context, entanglement, and decoherence.

Bohr’s Motivation

By the late 1920s, quantum theory had accumulated several pressures against classical description:

light behaved particle-like in energy exchange but wave-like in interference;
electrons behaved particle-like in localized detection but wave-like in diffraction;
atomic spectra required discrete transitions rather than continuous orbital radiation;
matrix mechanics made observables noncommuting;
wave mechanics made probability amplitudes central.

Bohr’s complementarity was an attempt to organize this situation without pretending that one classical model could simply replace all the others. The word “classical” mattered to him because experiments are described with ordinary macroscopic apparatus: slits, screens, clocks, rods, photographic plates, magnets, and counters. Yet the quantum system under study cannot always be assigned the full set of classical properties suggested by those descriptions.

The principle can be summarized cautiously:

an experiment defines the conditions under which particular concepts are applicable;
some experimental arrangements are mutually exclusive;
descriptions obtained from those arrangements can be jointly necessary for a complete account;
one should not combine them into a single classical picture when the conditions for applying them conflict.

This was Bohr’s answer to a specific historical problem: how to use indispensable classical language without reintroducing an inconsistent classical ontology for atoms, electrons, and radiation.

Wave-Particle Experimental Arrangements

The double slit is the standard example. If a coherent beam passes through two open slits and no path record is produced, the amplitudes for arrival at a screen point add:

\psi(x) = \psi_1(x)+\psi_2(x), \qquad P(x) = \lvert\psi_1(x)+\psi_2(x)\rvert^2.

The resulting distribution contains an interference term:

P(x) = \lvert\psi_1(x)\rvert^2 + \lvert\psi_2(x)\rvert^2 + 2\operatorname{Re} \left[ \psi_1^*(x)\psi_2(x) \right].

This is the wave-like arrangement. It reveals relative phase through fringes.

If the apparatus records which slit the particle used, the relevant state includes detector records:

\Psi(x) = \psi_1(x)\lvert D_1\rangle + \psi_2(x)\lvert D_2\rangle.

When the detector records are ignored, the screen probability contains the overlap

\begin{aligned} P(x) = {}& \lvert\psi_1(x)\rvert^2 + \lvert\psi_2(x)\rvert^2 \\ &+ 2\operatorname{Re} \left[ \psi_1^*(x)\psi_2(x) \langle D_1\vert D_2\rangle \right]. \end{aligned}

If $\langle D_1\vert D_2\rangle=0$ , the path records are perfectly distinguishable and the interference term vanishes. This is the particle-like arrangement. It reveals which alternative occurred, but it destroys the phase-sensitive interference pattern.

Bohr’s point was not that the quantum object literally flips between being a tiny classical particle and being a classical wave. The point was that the experimental conditions for applying particle-like path language and wave-like interference language cannot both be fully realized in the same arrangement.

Complementarity and Measurement Context

Complementarity is closely related to measurement context. A quantum state does not by itself answer every possible question in classical terms. The apparatus selects what question is being asked.

In modern projective language, an outcome family is represented by projectors $P_a$ , and the probability is

P(a) = \langle\psi\vert P_a\vert\psi\rangle.

A different incompatible measurement uses a different set of projectors. If two observables do not commute,

[A,B]\ne0,

then they generally do not share a complete set of eigenstates. Preparing or measuring one sharp property may preclude assigning the other property the same classical status in that context.

This algebraic fact is not identical to Bohr’s complementarity, but it is one reason complementarity became plausible. The historical principle says: do not merge the results of incompatible arrangements into one classical picture as though all the corresponding properties were jointly displayed.

Measurement context should not be confused with subjective choice. The apparatus is a physical system. Slits, magnets, polarizers, detector pixels, and environmental records determine which amplitudes remain coherent and which alternatives become distinguishable. The observer’s later knowledge is not the mechanism that removes interference.

Modern Views

Modern presentations usually do not take complementarity as an independent postulate. They use more precise tools:

noncommuting observables and incompatible bases;
POVMs and generalized measurement models;
entanglement between system and apparatus;
decoherence and environmental records;
quantitative wave-particle duality relations.

For an ideal two-path interferometer, a modern complementarity-style inequality is often written

\mathcal V^2+\mathcal D^2\le1,

where $\mathcal V$ measures fringe visibility and $\mathcal D$ measures path distinguishability. This relation is a later quantitative statement, not Bohr’s original 1920s formulation. It makes precise a lesson that Bohr expressed qualitatively: more path information leaves less room for interference visibility in a fixed two-path arrangement.

Decoherence also clarifies part of the story. Environmental degrees of freedom can carry which-path information even when no person reads a detector. When the environmental records associated with alternatives become nearly orthogonal, local interference is suppressed. This gives a physical mechanism behind many complementarity examples, while still leaving broader interpretation questions open.

Complementarity remains useful as a warning label. It tells the reader to ask: what arrangement defines the question, what alternatives remain coherent, what records distinguish them, and which classical words are justified by that setup?

Common Overstatements

Complementarity is often stretched beyond its responsible use. The following claims are too strong:

“A quantum object is both a classical wave and a classical particle at the same time.”
“Conscious observation creates the outcome.”
“The uncertainty principle proves complementarity.”
“Complementarity solves the measurement problem.”
“Quantum mechanics says reality does not exist.”
“Any two different descriptions are complementary.”
“One may freely combine data from incompatible arrangements into a single classical trajectory.”

The careful statement is narrower. Certain classical descriptions are tied to mutually exclusive experimental conditions. Quantum mechanics supplies a single formalism for predicting the outcomes, but the resulting evidence cannot always be compressed into one classical image.

What This Page Does Not Claim

This page does not claim that Bohr’s wording is the last word on foundations. Complementarity was historically influential, but modern measurement theory and foundations use sharper distinctions than Bohr’s original language.

It also does not claim that complementarity is a theorem equivalent to noncommutativity, uncertainty, or decoherence. Those are formal or dynamical structures with precise definitions. Complementarity is a historical interpretive principle that helped physicists avoid inconsistent classical pictures.

Finally, complementarity should not be used to block analysis. When a problem can be formulated with Hilbert spaces, projectors, POVMs, density matrices, or open-system dynamics, those tools should be used.

Cross-Links

References

N. Bohr, “The quantum postulate and the recent development of atomic theory,” Nature 121, 580-590, 1928, DOI: 10.1038/121580a0.
N. Bohr, Atomic Theory and the Description of Nature, Cambridge University Press, 1934.
W. Heisenberg, “Über den anschaulichen Inhalt der quantentheoretischen Kinematik und Mechanik,” Zeitschrift für Physik 43, 172-198, 1927, DOI: 10.1007/BF01397280.
N. Bohr, “Discussion with Einstein on epistemological problems in atomic physics,” in P. A. Schilpp, ed., Albert Einstein: Philosopher-Scientist, Open Court, 1949.
W. K. Wootters and W. H. Zurek, “Complementarity in the double-slit experiment: Quantum nonseparability and a quantitative statement of Bohr’s principle,” Physical Review D 19, 473-484, 1979, DOI: 10.1103/PhysRevD.19.473.
B.-G. Englert, “Fringe visibility and which-way information: An inequality,” Physical Review Letters 77, 2154-2157, 1996, DOI: 10.1103/PhysRevLett.77.2154.
M. Jammer, The Philosophy of Quantum Mechanics, Wiley, 1974.
M. Schlosshauer, Decoherence and the Quantum-to-Classical Transition, 2nd ed., Springer, 2019.

Exercises

In a two-slit setup, suppose $\langle D_1\vert D_2\rangle=1$ . What happens to the interference term?

Solution

If $\langle D_1\vert D_2\rangle=1$ , the detector states are identical for the two alternatives. The interference term remains fully present:

2\operatorname{Re} \left[ \psi_1^*(x)\psi_2(x) \right].

The apparatus has not recorded which-path information.

In the same formula, what changes if $\langle D_1\vert D_2\rangle=0$ ?

Solution

If $\langle D_1\vert D_2\rangle=0$ , the path records are orthogonal and therefore perfectly distinguishable. The interference term is multiplied by zero, so the screen probability reduces to

P(x) = \lvert\psi_1(x)\rvert^2 + \lvert\psi_2(x)\rvert^2.

The probabilities add as an incoherent mixture.

Why is complementarity not the same thing as the uncertainty relation?

Solution

Uncertainty relations are mathematical inequalities about the spreads of noncommuting observables in a quantum state. Complementarity is a broader historical principle about mutually exclusive experimental arrangements and the classical concepts applicable to them. They are related, but one is not simply a proof or restatement of the other.

Give one example of a misleading overstatement of complementarity and correct it.

Solution

One misleading statement is: “The electron is both a classical wave and a classical particle.” A better statement is: electron experiments can require wave-like amplitude descriptions in interference arrangements and particle-like localized detection descriptions in counting arrangements, but those classical pictures should not be fused into one literal model.