Completeness relations, the workhorse identity of every calculation.
"Insert a complete set of states" is the most used sentence in quantum mechanics. Behind it sit orthogonal projectors, resolutions of the identity, and some surprisingly deep theorems about what a pair of projectors can look like.
What you’ll learn
Characterize orthogonal projectors algebraically (P2=P=P†), spectrally and metrically (∥P∥≤1), and distinguish them from oblique idempotents.
Use orthogonal and overcomplete resolutions of the identity: PVMs, path sums, frames, POVMs and coherent states.
Prove ∫dψ∣ψ⟩⟨ψ∣⊗k=Πsym(k)/(kd+k−1) with Schur's lemma and derive the average-fidelity formula.
State Jordan's lemma for two projectors and apply it to products of reflections.
Projectors, precisely
Oblique idempotents are the warning case. E=(1010) satisfies E2=E but projects along the slanted direction kerE=span{(1,−1)T}, and ∥E∥∞=2. It is not a measurement: ⟨ψ∣E∣ψ⟩ can be complex, or exceed 1 (it reaches 21+21 for real unit vectors).
Projectors form a lattice: P≤Q iff PQ=P iff ranP⊆ranQ. If P and Q commute, PQ projects onto ranP∩ranQ. If they do not, PQ is not even Hermitian, but von Neumann's alternating-projection theorem recovers the intersection: (PQ)n→P∧Q, the projector onto ranP∩ranQ.
Think first How fast does (PQ)n converge? Try two lines in R2 at angle θ: P=∣u⟩⟨u∣, Q=∣v⟩⟨v∣ with ⟨u∣v⟩=cosθ.
PQ=cosθ∣u⟩⟨v∣ and (PQ)n=cos2n−1θ∣u⟩⟨v∣, so ∥(PQ)n−0∥=cos2n−1θ. Convergence is geometric at a rate set by the angle between the subspaces; in general the exact rate is c2n−1 with c the cosine of the Friedrichs angle (Kayalar–Weinert). Nearly parallel subspaces converge slowly.
Resolutions of the identity
An orthogonal resolution of the identity is a family of projectors with PkPl=δklPk and ∑kPk=I. One object, three names: a projection-valued measure (PVM), a projective measurement with Pr(k)=Tr(Pkρ), and the spectral decomposition of the observable A=∑kakPk. Inserting resolutions is how almost every calculation proceeds:
Insert a computational-basis resolution ∑z∣z⟩⟨z∣ between consecutive layers of an n-qubit circuit U=UT⋯U1:
amplitude=paths∑t∏(local amplitude)⟹BQP⊆PSPACE
Each term is a product of gate matrix elements along one computational path. There are 2n(T−1) paths, but each term takes polynomial time and the sum can be accumulated term by term in polynomial space:
BQP⊆PP⊆PSPACE
A finer counting version of the same sum gives BQP⊆PP (Adleman, DeMarrais & Huang, 1997). Interference, the cancellation among exponentially many paths, is what classical samplers cannot easily reproduce.
Resolutions need not be orthogonal. Any family of positive operators summing to I is a POVM, and rank-one families {ck∣ψk⟩⟨ψk∣} are tight frames: overcomplete "bases" that still come with a reconstruction formula.
Resolution
Identity
Notes
Orthonormal basis
∑k=1d∣k⟩⟨k∣=I
a PVM with d outcomes
Trine (qubit)
32∑k=02∣ψk⟩⟨ψk∣=I
∣ψk⟩=cos32πk∣0⟩+sin32πk∣1⟩: Bloch vectors 120∘ apart on a great circle
SIC-POVM
d1∑j=1d2∣ψj⟩⟨ψj∣=I
∣⟨ψj∣ψk⟩∣2=d+11 for j=k; for a qubit, a regular tetrahedron
Coherent states
π1∫C∣α⟩⟨α∣d2α=I
overcomplete in infinite dimensions; ∣⟨α∣β⟩∣2=e−∣α−β∣2
Think first The three trine states satisfy ∑k∣ψk⟩⟨ψk∣=cI. Without computing any matrix, find c.
Take traces: ∑kTr∣ψk⟩⟨ψk∣=3=cTrI=2c, so c=23 and the POVM elements are Ek=32∣ψk⟩⟨ψk∣. That the sum is proportional to I at all is because ∣ψk⟩⟨ψk∣=21(I+rk⋅σ) and the three Bloch vectors, 120∘ apart, add up to zero.
The identity from symmetry: Haar averages
Average ∣ψ⟩⟨ψ∣ over all pure states with the unitarily invariant (Haar) measure. The result commutes with every unitary, so by Schur's lemma it is a multiple of I; its trace is 1, so ∫dψ∣ψ⟩⟨ψ∣=I/d. A uniformly random state is a resolution of the identity. The same argument on k copies is a workhorse for designs, de Finetti theorems and benchmarking:
DerivationProof sketch, and a payoff
[M,U⊗k]=0∀U∈U(d)
Call the left side M. Each ∣ψ⟩⊗k lies in Symk, so M=ΠsymMΠsym, and Haar invariance gives U⊗kMU†⊗k=M for all U∈U(d):
1=TrM=c(kd+k−1)
Symk(Cd) is an irreducible representation of U(d), so Schur's lemma gives M=cΠsym. Fix c by traces; dimSymk(Cd)=(kd+k−1) counts multisets of size k from d labels.
Payoff (k=2): the average fidelity of a unitary error U over Haar-random inputs, using Tr[(A⊗B)SWAP]=Tr(AB):
Predict before you look
A qubit suffers an unwanted Z. Averaged over Haar-random pure inputs, what is ∫dψ∣⟨ψ∣Z∣ψ⟩∣2, in percent?
Show the answer
33 %
TrZ=0, so the formula gives 2⋅30+2=31. Check on the Bloch sphere: ⟨ψ∣Z∣ψ⟩=z, and the average of z2 over the unit sphere is 31. Even a full Pauli error leaves average fidelity 31, not 0: states near the z-axis are barely affected.
Two projectors: Jordan's lemma
DerivationProof sketch
PQP∣vj⟩=cj2∣vj⟩,0≤cj≤1
PQP is Hermitian and maps ranP to itself. Diagonalize it there with orthonormal ∣vj⟩∈ranP:
Different blocks are orthogonal, because every overlap reduces to a matrix element of PQP:
H=j⨁Bj⊕(common eigenvectors of P and Q)
Vectors with cj∈{0,1}, and the part of kerP orthogonal to all blocks (where Q restricts to a projector), split into common eigenvectors. Writing cj=cosθj gives the normal form.
Jordan's lemma turns any problem about two projectors into many independent qubit problems. The reflections RP=2P−I and RQ=2Q−I act on block j as reflections about two lines at angle θj, so their product is a rotation by 2θj, with eigenvalues e±2iθj. That single fact drives:
Grover search and amplitude amplification: one block, a rotation by 2θ per iteration;
Szegedy quantum walks: walk spectra from principal angles, and quadratically faster hitting times;
Marriott–Watrous in-place amplification of QMA verifiers, using a single copy of the witness;
QSVT, which applies polynomials to exactly the numbers cosθj, the singular values of PQ.
Check your understanding
P=∣u⟩⟨u∣ and Q=∣v⟩⟨v∣ with ∣⟨u∣v⟩∣=cosθ. What is ∥P−Q∥∞?
sinθ
1−cosθ
2sinθ
cosθ
Show the answer
A.sinθ
On their two-dimensional block, P−Q is traceless Hermitian with Tr(P−Q)2=2−2cos2θ, so its eigenvalues are ±sinθ. Hence also the trace distance between the pure states, 21∥P−Q∥1=sinθ=1−∣⟨u∣v⟩∣2.
Key takeaways
Orthogonal projectors: P2=P=P†, equivalently idempotent with ∥P∥∞≤1; oblique idempotents are not measurements.
Orthogonal resolutions are PVMs and spectral decompositions; inserting them gives matrix products, traces, basis changes and Feynman path sums (BQP⊆PSPACE).
Overcomplete resolutions (trine, SIC, coherent states) are POVMs and tight frames; SIC existence in every dimension is open.
Haar averages: ∫dψ∣ψ⟩⟨ψ∣⊗k=Πsym/(kd+k−1), so the average fidelity of U is (∣TrU∣2+d)/(d(d+1)).
Jordan's lemma: two projectors split into 1- and 2-dimensional blocks, and products of reflections rotate by twice the principal angles.
Check yourself
In the interactive lesson these are scored with your confidence, and what you miss comes back for review.
Question 1
Which of these is an orthogonal projector?
(1010)
X
21(I+X)
21(I+Z)
Show the answer
C.21(I+X)
21(I+X)=∣+⟩⟨+∣: it is Hermitian, and (21(I+X))2=41(I+2X+I)=21(I+X).
Question 2
For a Haar-random pure state in Cd and any operator A, what is ∫dψ⟨ψ∣A∣ψ⟩?
TrA
Tr(A)/d
∥A∥∞
Tr(A)2/d2
Show the answer
B.Tr(A)/d
∫dψ⟨ψ∣A∣ψ⟩=Tr(A∫dψ∣ψ⟩⟨ψ∣)=Tr(AI/d)=Tr(A)/d.
Question 3
In a Jordan block where P and Q project onto lines at angle θ, what is (2Q−I)(2P−I)?
A reflection about the bisector
A rotation by θ
A rotation by 4θ
A rotation by 2θ
Show the answer
D. A rotation by 2θ
Composing reflections about two lines at angle θ gives a rotation by 2θ, with eigenvalues e±2iθ. In Grover search sinθ=1/N, so about 4πN iterations rotate the start state onto the marked one.
Question 4
Inserting a computational-basis resolution of the identity between each pair of consecutive layers of an n-qubit, T-layer circuit writes one amplitude as a sum of how many terms, in general?
nT
2n(T−1)
2n
Tn
Show the answer
B.2n(T−1)
Each of the T−1 insertions sums over 2n basis states independently: (2n)T−1 paths. Exponentially many terms, each cheap to compute, which is exactly why the path sum proves BQP⊆PSPACE rather than giving an efficient simulation.
Further reading
P. R. Halmos, "Two subspaces," Trans. Amer. Math. Soc. 144, 381 (1969).
A. W. Harrow, "The church of the symmetric subspace," arXiv:1308.6595 (2013).
J. M. Renes, R. Blume-Kohout, A. J. Scott & C. M. Caves, "Symmetric informationally complete quantum measurements," J. Math. Phys. 45, 2171 (2004).
C. Marriott & J. Watrous, "Quantum Arthur–Merlin games," Computational Complexity (2005).
L. M. Adleman, J. DeMarrais & M.-D. A. Huang, "Quantum computability," SIAM J. Comput. (1997).