Prove2Me
Navigate
DiscoverCollectionsFormalpediaBlogsUsersMomentumMy Missions+
Prove2Me
⌕
Log in
← Formalpedia

Theorem 3.1 — the bounded functional calculus of a projection-valued measure

Proved
TeschlQM.Spectral.bounded_functional_calculus

by mikedeng1 · Sep 28, 2026 · Mathlib 0df444a (Lean v4.33.1)

functional-analysisfunctional-calculusp2o-batch-books5p2o-gran-per-chapterp2o-plan-bookp2o-v1spectral-theory

Let PPP be a projection-valued measure on a complex Hilbert space H\mathfrak HH, and for a bounded Borel function f∈B(R)f \in B(\mathbb{R})f∈B(R) let P(f)=∫Rf(λ) dP(λ)∈L(H)P(f) = \int_{\mathbb{R}} f(\lambda)\,dP(\lambda) \in \mathfrak L(\mathfrak H)P(f)=∫R​f(λ)dP(λ)∈L(H). Then the map

P:B(R)→L(H),f↦∫Rf(λ) dP(λ)P : B(\mathbb{R}) \to \mathfrak L(\mathfrak H), \qquad f \mapsto \int_{\mathbb{R}} f(\lambda)\, dP(\lambda)P:B(R)→L(H),f↦∫R​f(λ)dP(λ)

is a C∗C^*C∗-algebra homomorphism of norm one: P(1)=IP(1) = \mathbb{I}P(1)=I, P(f+g)=P(f)+P(g)P(f+g) = P(f) + P(g)P(f+g)=P(f)+P(g), P(cf)=cP(f)P(cf) = cP(f)P(cf)=cP(f), P(fg)=P(f)P(g)P(fg) = P(f)P(g)P(fg)=P(f)P(g), P(f∗)=P(f)∗P(f^*) = P(f)^*P(f∗)=P(f)∗ and ∥P(f)∥≤sup⁡λ∣f(λ)∣\|P(f)\| \le \sup_{\lambda} |f(\lambda)|∥P(f)∥≤supλ​∣f(λ)∣ for all f,g∈B(R)f, g \in B(\mathbb{R})f,g∈B(R), c∈Cc \in \mathbb{C}c∈C. Moreover, for f,g∈B(R)f, g \in B(\mathbb{R})f,g∈B(R) and φ,ψ∈H\varphi, \psi \in \mathfrak Hφ,ψ∈H,

⟨P(g)φ,P(f)ψ⟩=∫Rg∗(λ)f(λ) dμφ,ψ(λ)(3.21).\langle P(g)\varphi, P(f)\psi\rangle = \int_{\mathbb{R}} g^*(\lambda) f(\lambda)\, d\mu_{\varphi,\psi}(\lambda) \qquad (3.21).⟨P(g)φ,P(f)ψ⟩=∫R​g∗(λ)f(λ)dμφ,ψ​(λ)(3.21).

Finally, if fn∈B(R)f_n \in B(\mathbb{R})fn​∈B(R), fn(x)→f(x)f_n(x) \to f(x)fn​(x)→f(x) for every xxx, and sup⁡λ∈R∣fn(λ)∣\sup_{\lambda\in\mathbb{R}} |f_n(\lambda)|supλ∈R​∣fn​(λ)∣ is bounded in nnn, then P(fn)ψ→P(f)ψP(f_n)\psi \to P(f)\psiP(fn​)ψ→P(f)ψ for every ψ∈H\psi \in \mathfrak Hψ∈H (strong convergence).

This is the functional calculus for bounded functions on which the unbounded calculus and the spectral theorem are built.

Formalization Note. f∗f^*f∗ is the pointwise complex conjugate (star f), P(f)∗P(f)^*P(f)∗ the Hilbert-space adjoint (star on H →L[ℂ] H). The sup norm sup⁡λ∣f(λ)∣\sup_\lambda |f(\lambda)|supλ​∣f(λ)∣ is the real supremum ⨆ x, ‖f x‖, which is the true supremum since fff is bounded. "Norm one" is the bound ∥P(f)∥≤∥f∥∞\|P(f)\| \le \|f\|_\infty∥P(f)∥≤∥f∥∞​ together with P(1)=IP(1) = \mathbb{I}P(1)=I, which has norm one when H≠{0}\mathfrak H \ne \{0\}H={0}. The integral against μφ,ψ\mu_{\varphi,\psi}μφ,ψ​ is spectralForm, the polarization (3.15) of four spectral-measure integrals. In the convergence clause the limit fff is not assumed measurable or bounded: both follow from the hypotheses.

Preamble
import Mathlib
import Definitions.Def_TeschlQM_Shared_IsProjValuedMeasure
import Definitions.Def_TeschlQM_Spectral_spectralMeasure
import Definitions.Def_TeschlQM_Spectral_boundedSpectralIntegral

open MeasureTheory
open scoped ENNReal InnerProductSpace
Formal statement
namespace TeschlQM.Spectral

/-- Teschl, p. 90, Theorem 3.1. For a projection-valued measure `P`, the map
`f ↦ P(f) = ∫_ℝ f(λ) dP(λ)` from the bounded Borel functions `B(ℝ)` (sup norm) to `𝔏(H)` is a
unital C*-algebra homomorphism (linear, multiplicative, `P(f*) = P(f)*`) of norm one
(`‖P(f)‖ ≤ sup_λ |f(λ)|`, attained at `f = 1` when `H ≠ 0`), satisfies (3.21)
`⟨P(g)φ, P(f)ψ⟩ = ∫ g* f dμ_{φ,ψ}`, and `P(fₙ) → P(f)` strongly whenever `fₙ → f` pointwise
with `sup_λ |fₙ(λ)|` bounded in `n`. -/
theorem bounded_functional_calculus {H : Type*} [NormedAddCommGroup H] [InnerProductSpace ℂ H] [CompleteSpace H]
    (P : Set ℝ → (H →L[ℂ] H)) (hP : TeschlQM.Shared.IsProjValuedMeasure P) :
    boundedSpectralIntegral P 1 = 1 ∧
    (∀ f g : ℝ → ℂ, IsBoundedBorel f → IsBoundedBorel g →
      boundedSpectralIntegral P (f + g) =
          boundedSpectralIntegral P f + boundedSpectralIntegral P g ∧
      boundedSpectralIntegral P (f * g) =
          boundedSpectralIntegral P f * boundedSpectralIntegral P g) ∧
    (∀ (c : ℂ) (f : ℝ → ℂ), IsBoundedBorel f →
      boundedSpectralIntegral P (c • f) = c • boundedSpectralIntegral P f) ∧
    (∀ f : ℝ → ℂ, IsBoundedBorel f →
      boundedSpectralIntegral P (star f) = star (boundedSpectralIntegral P f)) ∧
    (∀ f : ℝ → ℂ, IsBoundedBorel f → ‖boundedSpectralIntegral P f‖ ≤ ⨆ x, ‖f x‖) ∧
    (∀ f g : ℝ → ℂ, IsBoundedBorel f → IsBoundedBorel g → ∀ φ ψ : H,
      ⟪boundedSpectralIntegral P g φ, boundedSpectralIntegral P f ψ⟫_ℂ =
        spectralForm P φ ψ (star g * f)) ∧
    (∀ (fs : ℕ → ℝ → ℂ) (f : ℝ → ℂ), (∀ n, IsBoundedBorel (fs n)) →
      (∀ x, Filter.Tendsto (fun n => fs n x) Filter.atTop (nhds (f x))) →
      (∃ C : ℝ, ∀ n x, ‖fs n x‖ ≤ C) →
      ∀ ψ : H, Filter.Tendsto (fun n => boundedSpectralIntegral P (fs n) ψ) Filter.atTop
        (nhds (boundedSpectralIntegral P f ψ))) := by sorry

end TeschlQM.Spectral
Source
Teschl, Mathematical Methods in Quantum Mechanics, AMS GSM 99, 2009, p. 90, Theorem 3.1
Read-back

What the Lean code literally says, in plain math · claude-opus-5-5

Setting.

  • HHH is a complex Hilbert space, possibly {0}\{0\}{0}, with inner product linear in the second argument and conjugate-linear in the first.
  • PPP maps subsets of R\mathbb{R}R to bounded operators on HHH and satisfies the imported predicate IsProjValuedMeasure, whose content is not shown here.
  • P(f)\mathcal{P}(f)P(f) denotes boundedSpectralIntegral P f, a bounded operator on HHH, and "bounded Borel" denotes the predicate IsBoundedBorel. Both come from an imported file.
  • SP(φ,ψ;h)\mathcal{S}_P(\varphi,\psi;h)SP​(φ,ψ;h) denotes the imported spectralForm P φ ψ h.
  • For functions, f+gf+gf+g, fgfgfg and cfcfcf are pointwise, and fˉ\bar ffˉ​ is the pointwise complex conjugate. For operators, T∗T^*T∗ is the Hilbert-space adjoint and STSTST is composition.

Conclusion. The theorem asserts all seven of the following:

  1. P(1)=I\mathcal{P}(\mathbf 1) = IP(1)=I, where 1\mathbf 11 is the constant function 111 and III is the identity operator. This part carries no hypothesis on fff.
  2. For all bounded Borel f,gf, gf,g:
P(f+g)=P(f)+P(g)andP(fg)=P(f) P(g).\mathcal{P}(f+g) = \mathcal{P}(f) + \mathcal{P}(g) \quad\text{and}\quad \mathcal{P}(fg) = \mathcal{P}(f)\,\mathcal{P}(g).P(f+g)=P(f)+P(g)andP(fg)=P(f)P(g).
  1. For all c∈Cc \in \mathbb{C}c∈C and bounded Borel fff: P(cf)=c P(f)\mathcal{P}(cf) = c\,\mathcal{P}(f)P(cf)=cP(f).
  2. For every bounded Borel fff: P(fˉ)=P(f)∗\mathcal{P}(\bar f) = \mathcal{P}(f)^*P(fˉ​)=P(f)∗.
  3. For every bounded Borel fff, the operator norm satisfies
∥P(f)∥≤sup⁡x∈R∣f(x)∣.\|\mathcal{P}(f)\| \le \sup_{x\in\mathbb{R}} |f(x)|.∥P(f)∥≤x∈Rsup​∣f(x)∣.

Only this inequality is asserted; the theorem does not assert that it is attained for any fff. 6. For all bounded Borel f,gf, gf,g and all φ,ψ∈H\varphi, \psi \in Hφ,ψ∈H:

⟨P(g)φ,  P(f)ψ⟩=SP(φ,ψ;gˉf).\langle \mathcal{P}(g)\varphi,\; \mathcal{P}(f)\psi\rangle = \mathcal{S}_P(\varphi,\psi;\bar g f).⟨P(g)φ,P(f)ψ⟩=SP​(φ,ψ;gˉ​f).
  1. Let (fn)n∈N(f_n)_{n\in\mathbb{N}}(fn​)n∈N​ be a sequence with the following properties:

    • each fnf_nfn​ is bounded Borel;
    • fn(x)→f(x)f_n(x) \to f(x)fn​(x)→f(x) for every x∈Rx \in \mathbb{R}x∈R;
    • there is a single real CCC with ∣fn(x)∣≤C|f_n(x)| \le C∣fn​(x)∣≤C for all nnn and xxx.

    Then for every ψ∈H\psi \in Hψ∈H, P(fn)ψ→P(f)ψ\mathcal{P}(f_n)\psi \to \mathcal{P}(f)\psiP(fn​)ψ→P(f)ψ in the norm of HHH. The limit fff is not separately assumed bounded Borel. As a pointwise limit of measurable functions bounded by CCC, it is automatically measurable and bounded by CCC.

Degenerate cases.

  • The supremum in (5) is taken over the nonempty set R\mathbb{R}R. It is finite for bounded fff, so the default value for unbounded suprema does not arise under the hypothesis.
  • If H={0}H = \{0\}H={0}, every operator equals both 000 and III, and every conclusion holds trivially.
  • The strength of the whole theorem depends on the imported hypothesis IsProjValuedMeasure, whose content is not shown. If it were unsatisfiable, all seven conclusions would be vacuous.
Human review
  • Endorsed by Shuze Chen · Oct 2, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 2, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paper
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me