Existing in Mathlib · Measure theory and analysis
Hölder's Inequality
For two almost-everywhere measurable nonnegative extended-real functions, conjugate powers control the lower integral of their pointwise product by the product of two separate power-integral scales.
- Hölder's inequality
- lower Lebesgue integral
- extended nonnegative reals
- Holder-conjugate exponents
- power-integral scales
- Young's inequality
- almost-everywhere measurability
Exact theorem
Exact Mathlib statement
theorem ENNReal.lintegral_mul_le_Lp_mul_Lq {α : Type*} [MeasurableSpace α] (μ : Measure α) {p q : ℝ} (hpq : p.HolderConjugate q) {f g : α → ℝ≥0∞} (hf : AEMeasurable f μ) (hg : AEMeasurable g μ) : (∫⁻ a, (f * g) a ∂μ) ≤ (∫⁻ a, f a ^ p ∂μ) ^ (1 / p) * (∫⁻ a, g a ^ q ∂μ) ^ (1 / q)The theorem at a glance
Hölder's inequality for lintegrals at a glance
A visual guide to the theorem's hypotheses, structure, and conclusion; the exact statement gives the formal detail.
Hölder's inequality for lintegrals at a glance

Detailed visual description
Two independently AEMeasurable ℝ≥0∞-valued functions share one measured domain. Their pointwise product feeds the left side of Hölder's inequality, while separate p- and q-power lintegrals supply the two multiplied scales on the right. The checked route splits off vanishing and top-valued cases, then normalizes the finite nonzero branch, applies Young's inequality pointwise, integrates, and rescales.
Statement structure
From hypotheses to conclusion
This map summarizes the statement's structure; the exact Mathlib declaration remains authoritative.
Overlapping profiles feed one product accumulation

Detailed visual description
Two nonnegative profiles share one measured domain and are simultaneously positive on a common central interval. Their pointwise product vanishes outside that overlap and accumulates below a paired support fed by the two separate power-scale vessels. The construction communicates the upper bound without asserting equality, an optimizer, finite integrals, or a probability-only specialization.
Why it matters
A mathematical landmark
Hölder's inequality is one of the central bridges between multiplication and integrability, underpinning duality, norm estimates, and much of modern analysis. Mathlib's selected declaration is especially instructive because its extended-real form keeps zero and infinite cases explicit while the finite branch exposes the normalization-and-Young-inequality mechanism.
ProofAtlas record
What has been checked
Mathlib is the source of the theorem; the local Lean replay and page review are separate.
Claim boundary
No new theorem is claimed
This page indexes Mathlib's Hölder inequality for two AEMeasurable functions f and g from a measured type α to the extended nonnegative reals ℝ≥0∞. For Holder-conjugate real exponents p and q, it bounds the lower Lebesgue integral of the pointwise product by the product of the corresponding p- and q-power lintegral scales. The declaration includes zero and top-valued cases; it does not state the signed real- or complex-valued absolute-value form, assume finite power integrals, characterize equality, or assert existence or uniqueness of an optimizer.
- The selected declaration concerns AEMeasurable functions α → ℝ≥0∞ and the lower Lebesgue integral; it is not the familiar signed real- or complex-valued absolute-value formulation.
- The theorem does not assume that either power integral is finite or nonzero; zero and top-valued cases are part of the source proof.
- It gives an upper bound only and does not characterize equality or assert a unique extremizer.
- The selected endpoint is a two-function inequality, not a finite-family, discrete-sum, probability-only, or interpolation theorem.
- Proof Atlas did not originate Hölder's inequality or Mathlib's declaration, and the generated explanations and visuals are not proof evidence.
Source and local evidence
Where the theorem comes from
- Existing declaration
ENNReal.lintegral_mul_le_Lp_mul_Lqin mathlib- Relationship
- The checked artifact is the upstream declaration itself
- Local Lean evidence
artifact.library.mathlib.holder-inequality.v001