Documentation

Books.Analysis2_Tao_2022.Chapters.Chap08.section02_part2

Helper for Theorem 8.3: every rational number fails a fixed-exponent Diophantine lower bound with a strictly positive constant.

Helper for Theorem 8.3: every rational number lies in the complement of the fixed-exponent Diophantine set.

Helper for Theorem 8.3: the embedded rationals form a subset of the fixed-exponent complement.

Helper for Theorem 8.3: every integer belongs to the complement of the fixed-exponent Diophantine set.

Helper for Theorem 8.3: the complement of the fixed-exponent Diophantine set is nonempty (for instance, it contains rational numbers).

Helper for Theorem 8.3: the complement of the fixed-exponent Diophantine set is infinite because it contains every integer.

Helper for Theorem 8.3: negating the fixed-exponent Diophantine lower-bound is equivalent to saying that every positive constant is violated by some rational approximation.

Helper for Theorem 8.3: if τ ≤ τ', then the complement of the τ'-Diophantine set is contained in the complement of the τ-Diophantine set.

Helper for Theorem 8.3: the complement-measure is antitone with respect to the Diophantine exponent.

Helper for Theorem 8.3: if a fixed exponent τ > 1 Diophantine lower bound holds Lebesgue-almost everywhere, then IsDiophantineReal also holds almost everywhere.

Helper for Theorem 8.3: an almost-everywhere fixed-exponent bound with τ > 1 forces the complement of IsDiophantineReal to have Lebesgue measure zero.

Helper for Theorem 8.3: a full-measure fixed-exponent Diophantine pair with τ > 1 implies that IsDiophantineReal holds almost everywhere.

Helper for Theorem 8.3: a full-measure fixed-exponent Diophantine pair with τ > 1 forces the non-diophantine set to have measure zero.

Helper for Theorem 8.3: at exponent τ = 2, the origin fails the fixed-exponent Diophantine lower bound.

Helper for Theorem 8.3: at exponent τ = 2, every integer lies in the complement of the fixed-exponent Diophantine set.

Helper for Theorem 8.3: at exponent τ = 2, the complement of the fixed-exponent Diophantine set is nonempty.

Helper for Theorem 8.3: a full-measure fixed-exponent pair at τ = 2 forces the non-diophantine set to have Lebesgue measure zero.

Helper for Theorem 8.3: at τ = 2, a full-measure fixed-exponent pair yields both almost-everywhere IsDiophantineReal and measure-zero non-diophantine complement.

Helper for Theorem 8.3: a universal fixed-exponent full-measure claim would imply IsDiophantineReal almost everywhere by specializing to τ = 2.

Helper for Theorem 8.3: a universal fixed-exponent full-measure claim would force the non-diophantine set to have Lebesgue measure zero.

Helper for Theorem 8.3: a universal fixed-exponent full-measure claim yields both almost-everywhere IsDiophantineReal and measure-zero non-diophantine complement.

Helper for Theorem 8.3: once the fixed-exponent complement has measure zero, one gets both the almost-everywhere exponent-τ bound and almost-everywhere IsDiophantineReal (for τ > 1).

Helper for Theorem 8.3: for a fixed exponent τ > 1, the theorem's target full-measure pair follows from the null-complement statement.

Helper for Theorem 8.3: for a fixed exponent τ > 1, proving the theorem's target pair is equivalent to proving the null-complement statement.

Helper for Theorem 8.3: from a fixed exponent τ > 1 null-complement statement, one obtains both the theorem's target full-measure pair and almost-everywhere IsDiophantineReal.

Helper for Theorem 8.3: for τ > 1, augmenting the theorem's target pair with the derived almost-everywhere IsDiophantineReal consequence is equivalent to the target pair itself.

Helper for Theorem 8.3: for a fixed exponent τ > 1, the theorem's target pair implies both almost-everywhere IsDiophantineReal and measure-zero non-IsDiophantineReal complement.

theorem helperForTheorem_8_3_positive_lower_bound_for_bounded_denominators_nonrat {τ x : } ( : 0 τ) (hx_nonrat : xSet.range Rat.cast) (N : ) :
∃ (c : ), 0 < c ∀ (p : ) (q : ), 0 < qq Nc / q ^ τ |x - p / q|

Helper for Theorem 8.3: a nonrational real has a uniform positive lower bound for |x - p/q| against all integers p and all positive denominators q ≤ N; with τ ≥ 0, this yields a bound of the form c / q^τ.

Helper for Theorem 8.3: for nonrational x and nonnegative exponent τ, failure of the fixed-exponent Diophantine lower bound implies LiouvilleWith τ x.

Helper for Theorem 8.3: the complement of the fixed-exponent Diophantine set is contained in the union of the embedded rationals and the fixed-exponent Liouville set (for nonnegative exponent).

Helper for Theorem 8.3: for τ > 2, the complement of the fixed-exponent Diophantine set has Lebesgue measure zero.

Theorem 8.3: if τ > 2, then the set of real numbers x for which there exists c(x) > 0 such that |x - p/q| ≥ c(x) / q^τ for all p ∈ ℤ and q ∈ ℕ with q > 0 has full Lebesgue measure; equivalently, this Diophantine bound holds for Lebesgue-almost every real x.

Helper for Proposition 8.7: the shifted geometric series with ratio 1/2 is finite in ℝ≥0∞.

theorem helperForProposition_8_7_superlevel_measure_bound (f : ) (hmeas : ∀ (n : ), AEMeasurable (f n) MeasureTheory.volume) (hintegral : ∀ (n : ), 1 n∫⁻ (x : ), ENNReal.ofReal (f n x) (1 / 4) ^ n) (n : ) :
MeasureTheory.volume {x : | (1 / 2) ^ (n + 1) ENNReal.ofReal (f (n + 1) x)} (1 / 2) ^ (n + 1)

Helper for Proposition 8.7: Markov/Chebyshev gives the superlevel-set measure bound m({x | 2^{-(n+1)} ≤ f_{n+1}(x)}) ≤ 2^{-(n+1)}.

Helper for Proposition 8.7: the limsup of superlevel sets has Lebesgue measure zero.

theorem helperForProposition_8_7_eventually_upper_bound_outside_limsup (f : ) (hnonneg : ∀ (n : ) (x : ), 0 f n x) (x : ) (hx : xFilter.limsup (fun (n : ) => {y : | (1 / 2) ^ (n + 1) ENNReal.ofReal (f (n + 1) y)}) Filter.atTop) :
∀ᶠ (n : ) in Filter.atTop, f (n + 1) x (1 / 2) ^ (n + 1)

Helper for Proposition 8.7: outside the limsup superlevel set, one eventually has the pointwise bound f_{n+1}(x) ≤ 2^{-(n+1)}.

theorem helperForProposition_8_7_tendsto_shifted_then_unshift (f : ) (x : ) (hnonneg : ∀ (n : ) (y : ), 0 f n y) (hbound : ∀ᶠ (n : ) in Filter.atTop, f (n + 1) x (1 / 2) ^ (n + 1)) :
Filter.Tendsto (fun (n : ) => f n x) Filter.atTop (nhds 0)

Helper for Proposition 8.7: an eventual geometric bound on the shifted sequence implies f_n(x) → 0.

theorem proposition_8_7_tendsto_zero_outside_small_exceptional_set (f : ) (hmeas : ∀ (n : ), AEMeasurable (f n) MeasureTheory.volume) (hnonneg : ∀ (n : ) (x : ), 0 f n x) (hintegral : ∀ (n : ), 1 n∫⁻ (x : ), ENNReal.ofReal (f n x) (1 / 4) ^ n) (ε : ) :
0 < ε∃ (E : Set ), MeasurableSet E MeasureTheory.volume E ENNReal.ofReal ε xE, Filter.Tendsto (fun (n : ) => f n x) Filter.atTop (nhds 0)

Proposition 8.7: if (f_n)_{n ≥ 1} are nonnegative Lebesgue measurable functions on with ∫⁻ f_n dm ≤ 4^{-n} for all n ≥ 1, then for every ε > 0 there exists a Lebesgue measurable set E with m(E) ≤ ε such that f_n x → 0 for all x ∈ ℝ \ E.

theorem helperForTheorem_8_4_stronglyMeasurable_seq_on_Icc (f : (Set.Icc 0 1)) (hmeas : ∀ (n : ), Measurable (f n)) (n : ) :

Helper for Theorem 8.4: measurable functions on [0,1] are strongly measurable.

theorem helperForTheorem_8_4_ae_tendsto_zero_on_Icc (f : (Set.Icc 0 1)) (hlim : ∀ (x : (Set.Icc 0 1)), Filter.Tendsto (fun (n : ) => f n x) Filter.atTop (nhds 0)) :
∀ᵐ (x : (Set.Icc 0 1)), Filter.Tendsto (fun (n : ) => f n x) Filter.atTop (nhds ((fun (x : (Set.Icc 0 1)) => 0) x))

Helper for Theorem 8.4: pointwise convergence to zero implies almost-everywhere convergence to the zero function on [0,1].

theorem helperForTheorem_8_4_apply_mathlib_egorov_finite_measure (f : (Set.Icc 0 1)) (hmeas : ∀ (n : ), Measurable (f n)) (hlim : ∀ (x : (Set.Icc 0 1)), Filter.Tendsto (fun (n : ) => f n x) Filter.atTop (nhds 0)) {ε : } ( : 0 < ε) :
∃ (E : Set (Set.Icc 0 1)), MeasurableSet E MeasureTheory.volume E ENNReal.ofReal ε TendstoUniformlyOn (fun (n : ) => f n) (fun (x : (Set.Icc 0 1)) => 0) Filter.atTop E

Helper for Theorem 8.4: Egorov's theorem on the finite-measure space [0,1].

theorem theorem_8_4_egoroff_on_unit_interval (f : (Set.Icc 0 1)) (hmeas : ∀ (n : ), Measurable (f n)) (hnonneg : ∀ (n : ) (x : (Set.Icc 0 1)), 0 f n x) (hlim : ∀ (x : (Set.Icc 0 1)), Filter.Tendsto (fun (n : ) => f n x) Filter.atTop (nhds 0)) (ε : ) :
0 < ε∃ (E : Set (Set.Icc 0 1)), MeasurableSet E MeasureTheory.volume E ENNReal.ofReal ε TendstoUniformlyOn (fun (n : ) => f n) (fun (x : (Set.Icc 0 1)) => 0) Filter.atTop E

Theorem 8.4 (Egoroff on [0,1]): if (f_n) are measurable and nonnegative on [0,1] and satisfy f_n(x) → 0 for every x ∈ [0,1], then for every ε > 0 there exists a measurable E ⊆ [0,1] with m(E) ≤ ε such that f_n → 0 uniformly on [0,1] \ E.

theorem helperForProposition_8_8_shiftedDiagonal_term_eq_ite_eq (n m : ) :
↑(if m + 1 = n + 1 then 1 else 0) = ↑(if m = n then 1 else 0)

Helper for Proposition 8.8: rewrite the shifted diagonal term into Kronecker-delta form.

theorem helperForProposition_8_8_row_hasSum_one (n : ) :
HasSum (fun (m : ) => ↑(if m = n then 1 else 0)) 1

Helper for Proposition 8.8: each row of the diagonal indicator has sum 1.

Helper for Proposition 8.8: each fixed column converges pointwise to 0.

Helper for Proposition 8.8: the row sums converge to 1.

Helper for Proposition 8.8: the sum of the pointwise limit function 0 is 0.

theorem proposition_8_8_exists_bounded_nonnegative_function_with_noncommuting_limit_and_series :
∃ (f : × NNReal), (∃ (C : NNReal), ∀ (n m : ), f (n, m) C) (∀ (n : ), Summable fun (m : ) => (f (n, m + 1))) (∀ (m : ), ∃ (l : ), Filter.Tendsto (fun (n : ) => (f (n, m + 1))) Filter.atTop (nhds l)) ∃ (g : ), (∀ (m : ), Filter.Tendsto (fun (n : ) => (f (n, m + 1))) Filter.atTop (nhds (g m))) Summable g ∃ (L : ), Filter.Tendsto (fun (n : ) => ∑' (m : ), (f (n, m + 1))) Filter.atTop (nhds L) L ∑' (m : ), g m

Proposition 8.8: there exists a bounded function f : ℕ × ℕ → ℝ≥0 such that (i) for every n, the series ∑_{m=1}^∞ f(n,m) converges; (ii) for every m, the limit lim_{n→∞} f(n,m) exists; (iii) nevertheless, the limit of row sums is not equal to the sum of pointwise limits.