Indistinguishability

Identical quantum particles are not distinguishable particles whose labels are merely unknown. If two electrons, two photons, or two atoms of the same isotope have the same intrinsic properties, there is no physical observable that says “this one is particle 1” and “that one is particle 2.”

The labels used in wavefunctions and tensor products are bookkeeping labels. They name argument slots in a mathematical representation, not persistent particle identities. This distinction is the entry point to exchange symmetry, the Pauli exclusion principle, occupation-number notation, and the many-particle language used in quantum matter and QFT.

Classical Labels Versus Quantum Labels

For two distinguishable classical particles, one can imagine tracking worldlines. Even if the particles are otherwise similar, particle $A$ and particle $B$ can be followed through time in principle. A configuration might be written as

$$(x_A,p_A;\ x_B,p_B),$$

and the labels $A$ and $B$ refer to individual systems.

For two identical quantum particles, the formal two-particle wavefunction may be written as

$$\Psi(x_1,x_2),$$

but the labels $1$ and $2$ do not name two observable individuals. They label the first and second coordinate arguments. Exchanging the arguments gives

$$\Psi(x_2,x_1),$$

which is not a new physical situation obtained by moving named particles around. It is the same unlabeled configuration described with the argument slots interchanged.

This is sharper than ignorance. Ignorance would mean there is a fact about which particle is which, but we do not know it. Identical-particle quantum mechanics instead says that the individual labels are not physical observables in the first place.

Particle Labels as Bookkeeping

The distinguishable two-particle Hilbert space built from a one-particle Hilbert space $\mathcal h$ is

$$\mathcal h\otimes\mathcal h.$$

This tensor product is useful even for identical particles, but the two tensor factors are not automatically two observable particles. They are slots on which exchange operations act.

For one-particle states $\lvert\alpha\rangle$ and $\lvert\beta\rangle$, define the exchange operator $P_{12}$ by

$$P_{12} \bigl( \lvert\alpha\rangle_1\otimes\lvert\beta\rangle_2 \bigr) = \lvert\beta\rangle_1\otimes\lvert\alpha\rangle_2.$$

In the position representation for spinless particles,

$$(P_{12}\Psi)(x_1,x_2) = \Psi(x_2,x_1).$$

The operation swaps the formal slots. It does not reveal a hidden identity tag attached to either particle.

Applying the exchange twice returns the original slot ordering:

$$P_{12}^2=I.$$

Therefore the eigenvalues of $P_{12}$ are $\pm1$ on two-particle exchange sectors. The next structural step is the symmetrization postulate: ordinary identical bosons occupy symmetric sectors, while ordinary identical fermions occupy antisymmetric sectors. This page prepares that statement by clarifying why exchange is a physical constraint rather than a naming convention.

Physical Observables and Exchange

An observable for identical particles must not depend on an arbitrary slot label. For two identical particles, a physical observable $O$ must be invariant under exchange:

$$P_{12}OP_{12}^{-1}=O.$$

Equivalently,

$$[O,P_{12}]=0.$$

For two identical spinless particles on a line, the operator $\hat x_1$ alone is not a physical “position of particle 1” observable. The label $1$ has no observable identity. Exchange sends it to $\hat x_2$:

$$P_{12}\hat x_1P_{12}^{-1} = \hat x_2.$$

By contrast, the center-of-mass coordinate is invariant:

$$\hat X = \frac{\hat x_1+\hat x_2}{2}, \qquad P_{12}\hat X P_{12}^{-1} = \hat X.$$

The squared separation is also invariant:

$$(\hat x_1-\hat x_2)^2 \quad\longmapsto\quad (\hat x_2-\hat x_1)^2 = (\hat x_1-\hat x_2)^2.$$

A one-body potential for two identical particles must treat both slots in the same way:

$$V(\hat x_1)+V(\hat x_2),$$

not $V(\hat x_1)$ alone as a complete two-particle observable. Interactions likewise depend on symmetric combinations such as the separation between the two particles.

Configuration-Space Wavefunctions Need Symmetry Constraints

The formal space $L^2(\mathbb R^3)\otimes L^2(\mathbb R^3)$ can be identified with square-integrable functions $\Psi(\mathbf x_1,\mathbf x_2)$. For distinguishable particles, all normalized functions in that space are possible pure states, subject to the Hamiltonian and boundary conditions.

For identical particles, not every vector in this slot-labeled tensor product represents a physical state. The physical state must transform consistently under exchange. For two ordinary identical particles in three dimensions, the allowed sectors are

$$P_{12}\lvert\Psi\rangle = \eta\lvert\Psi\rangle, \qquad \eta=\pm1.$$

In the coordinate representation this becomes

$$\Psi(\mathbf x_2,\mathbf x_1) = \eta\Psi(\mathbf x_1,\mathbf x_2).$$

The value $\eta=+1$ gives symmetric states and $\eta=-1$ gives antisymmetric states. The assignment of particles to those sectors is not optional: ordinary bosons use the symmetric sector and ordinary fermions use the antisymmetric sector.

Spin must be included in the same exchange. If $q=(\mathbf x,s)$ denotes spatial and spin labels together, the exchange rule acts on the full argument:

$$\Psi(q_2,q_1) = \eta\Psi(q_1,q_2).$$

Thus an antisymmetric fermion state may have a symmetric spatial part and an antisymmetric spin part, or the reverse, as long as the combined state has the correct exchange behavior. This is why spin singlets and triplets matter in atomic and molecular applications.

Effective Distinguishability

Identical particles can sometimes be treated as effectively distinguishable because physical modes or regions supply operational labels. Two identical atoms held in well-separated traps can be described by left and right trap modes. If the wave packets have negligible overlap and the dynamics does not exchange the traps, calculations may look almost like a distinguishable-particle tensor product.

The important point is what carries the label. “Left trap” and “right trap” are physical modes or regions. “Particle 1” and “particle 2” are not physical names. When the modes overlap, particles tunnel, or exchange processes matter, the slot-label approximation can fail.

This distinction is also central in optics. Photons are identical particles, but modes are physically meaningful: polarization modes, spatial modes, frequency modes, cavity modes, and wave-packet modes can be prepared and measured. Occupation-number notation describes how many photons occupy each mode, not which photon has which private identity.

Relation to Entanglement

For distinguishable subsystems, entanglement is defined relative to a tensor-product split such as

$$\mathcal H_A\otimes\mathcal H_B.$$

For identical particles, the slot-labeled split

$$\mathcal h_1\otimes\mathcal h_2$$

is not automatically an operational subsystem split. An antisymmetrized two-fermion state may look nonfactorizable in the formal slots even when it describes a single Slater determinant. Calling that nonfactorization “particle entanglement” without further qualification can be misleading.

Operational questions usually refer instead to modes, regions, species, internal degrees of freedom, or algebras of observables. Once those are specified, entanglement questions become meaningful again. The caution is not that identical particles cannot be entangled; it is that the subsystem decomposition must be physically stated. See Identical-Particle Entanglement Cautions for the dedicated guide.

Examples

Two electrons in an atom are identical fermions. Their individual labels are not observable, and the total electronic state must be antisymmetric under exchange of all electronic degrees of freedom. Pauli exclusion follows from this antisymmetry when two fermions are assigned the same one-particle state.

Two photons entering a beam splitter are identical bosons. The meaningful labels are input and output modes, polarization modes, frequencies, or wave packets. Interference effects depend on whether the photons are indistinguishable in all degrees of freedom relevant to the experiment.

The optical unitary, phase conventions, and coincidence cancellation are developed at Beam Splitters.

Two identical atoms in widely separated traps can often be labeled by the traps. This is effective distinguishability by physical localization, not a restoration of hidden particle identities. If the atoms are released and their wave packets overlap, exchange symmetry becomes experimentally relevant.

Common Mistakes

• Treating $x_1$ and $x_2$ as names painted onto two particles.
• Saying indistinguishability means we are merely ignorant of which particle is which.
• Confusing exchange symmetry with ordinary mirror symmetry or rotation symmetry in space.
• Forgetting that exchange acts on spin and spatial variables together.
• Calling every antisymmetrized wavefunction entangled without specifying a mode or observable split.
• Using a one-particle observable such as $\hat x_1$ as if it were physical for identical particles without symmetrizing it.

Cross-Links

• Notation and Subsystem Labels
• Tensor Products of Hilbert Spaces
• Symmetrization Postulate
• Bosons
• Fermions
• Exchange Operators
• Symmetric and Antisymmetric Wavefunctions
• Pauli Exclusion Principle
• Identical-Particle Entanglement Cautions
• Entangled States
• Entanglement Depends on a Decomposition
• Schmidt Decomposition
• Two Spin-1/2 Particles
• Reference Bridge: Second Quantization

References

• P. A. M. Dirac, The Principles of Quantum Mechanics, 4th ed., Oxford University Press, 1958.
• W. Pauli, “The Connection Between Spin and Statistics,” Physical Review 58, 716-722, 1940.
• A. Messiah and O. W. Greenberg, “Symmetrization Postulate and Its Experimental Foundation,” Physical Review 136, B248-B267, 1964.
• L. E. Ballentine, Quantum Mechanics: A Modern Development, 2nd ed., World Scientific, 2014.
• J. J. Sakurai and J. Napolitano, Modern Quantum Mechanics, 3rd ed., Cambridge University Press, 2020.
• A. L. Fetter and J. D. Walecka, Quantum Theory of Many-Particle Systems, McGraw-Hill, 1971.

Exercises

1. In the wavefunction $\Psi(x_1,x_2)$ for two identical spinless particles, why should $x_1$ and $x_2$ not be read as names of two particles?

Solution

They are coordinate arguments, or slots, in a representation of the two-particle state. Exchanging the slots gives $\Psi(x_2,x_1)$, which describes the same unlabeled physical configuration with the arguments interchanged. There is no physical observable that follows a private identity tag called particle $1$ through the experiment.

2. Use the definition of $P_{12}$ on product vectors to show that $P_{12}^2=I$.

Solution

Start with a product vector:

$$\lvert\alpha\rangle_1\otimes\lvert\beta\rangle_2.$$

One exchange gives

$$P_{12} \bigl( \lvert\alpha\rangle_1\otimes\lvert\beta\rangle_2 \bigr) = \lvert\beta\rangle_1\otimes\lvert\alpha\rangle_2.$$

A second exchange gives

$$P_{12} \bigl( \lvert\beta\rangle_1\otimes\lvert\alpha\rangle_2 \bigr) = \lvert\alpha\rangle_1\otimes\lvert\beta\rangle_2.$$

By linearity this holds for arbitrary two-slot states, so $P_{12}^2=I$.

3. Which of the following can be a physical observable for two identical particles on a line: $\hat x_1$, $\hat x_1+\hat x_2$, or $(\hat x_1-\hat x_2)^2$?

Solution

$\hat x_1$ alone is not exchange-invariant, because exchange maps it to $\hat x_2$. The sum $\hat x_1+\hat x_2$ is exchange-invariant. The squared separation $(\hat x_1-\hat x_2)^2$ is also exchange-invariant because exchanging the two slots changes the sign of $\hat x_1-\hat x_2$ but not its square.

4. Two identical atoms are trapped far apart in a left well and a right well. Why can the labels “left” and “right” be useful even though “particle 1” and “particle 2” are not physical names?

Solution

Left and right refer to physical trap modes or spatial regions that can be prepared and measured. If the wave packets remain well separated and exchange processes are negligible, those modes supply an effective subsystem description. The particle labels are still formal slot labels; what is physically meaningful is the occupation and state of each trap mode.