op_258_exists_finset_sum_gt_one
If f : α → ℝ is nonnegative and its ENNReal-sum of ENNReal.ofReal (f ·) is strictly greater than 1, then some finite set I of indices and some δ > 0 satisfy 1 + δ ≤ Σ_{i∈I} f i.
Statement
theorem op_258_exists_finset_sum_gt_one {α : Type*} (f : α → ℝ) (h0 : ∀ i, 0 ≤ f i) (hsum : 1 < ∑' i : α, ENNReal.ofReal (f i)) : ∃ (I : Finset α) (δ : ℝ), 0 < δ ∧ 1 + δ ≤ ∑ i ∈ I, f i
- Source
- step towards op_258_triangle_forcing_implies_decay_ramsey
- Verified
- 18 Oct 2026
- Axioms
Classical.choiceQuot.soundpropext- Built on 2
- TotalPreorderA total preorder: a preorder where
≤is total (complete). This is weaker thanLinearOrder— it does NOT require antisymmetry or… - LinearOrder.toTotalPreorderEvery
LinearOrderis aTotalPreorder.
- TotalPreorderA total preorder: a preorder where
- Used by 1
- op_258_bounded_extensionFor a positive antitone profile a with Σ a i > 1 there is n₀ such that for every n ≥ n₀, every partially defined a-bounded colouring of…
The proof, explained
The infinite sum in the extended nonnegative reals is defined as the supremum of all finite partial sums. Since this supremum strictly exceeds 1, some finite subset I ⊆ α already has its partial sum strictly greater than 1. Because f is nonnegative, this extended-real inequality translates directly to the real partial sum ∑_{i ∈ I} f i > 1, and choosing δ = (∑_{i ∈ I} f i) - 1 > 0 provides the required excess.
There exists a finite subset I ⊆ α such that 1 < ∑ i ∈ I, ENNReal.ofReal (f i).
By definition, the infinite sum of extended nonnegative reals is the supremum over all finite subsets of their partial sums (`ENNReal.tsum_eq_iSup_sum`). Since 1 is strictly less than this supremum, the supremum property (`lt_iSup_iff.mp`) yields a finite set I whose partial sum strictly exceeds 1.
uses
ENNReal.tsum_eq_iSup_sum,lt_iSup_iff.mprw [ENNReal.tsum_eq_iSup_sum] at hsum obtain ⟨I, hI⟩ := lt_iSup_iff.mp hsum
The real partial sum satisfies 1 < ∑ i ∈ I, f i.
Since f is nonnegative everywhere, the finite sum of extended reals equals `ENNReal.ofReal (∑ i ∈ I, f i)` by `ENNReal.ofReal_sum_of_nonneg`. Because 0 < 1 ≤ ENNReal.ofReal (∑ i ∈ I, f i) (using `zero_lt_one` and `lt_of_lt_of_le`), `ENNReal.ofReal_pos.mp` implies that the real sum is strictly positive. Rewriting 1 as `ENNReal.ofReal 1` via `ENNReal.ofReal_one`, the equivalence `ENNReal.ofReal_lt_ofReal_iff` reflects the strict inequality back to real numbers, yielding 1 < ∑ i ∈ I, f i.
uses
ENNReal.ofReal_sum_of_nonneg,ENNReal.ofReal_pos.mp,lt_of_lt_of_le,zero_lt_one,ENNReal.ofReal_one,ENNReal.ofReal_lt_ofReal_iffrw [← ENNReal.ofReal_sum_of_nonneg (fun i _ => h0 i)] at hI have hpos : 0 < ∑ i ∈ I, f i := ENNReal.ofReal_pos.mp (lt_of_lt_of_le (show (0 : ENNReal) < 1 from zero_lt_one) hI.le) rw [← ENNReal.ofReal_one] at hI have hgt : 1 < ∑ i ∈ I, f i := (ENNReal.ofReal_lt_ofReal_iff hpos).mp hI
The finite set I and the choice δ = (∑ i ∈ I, f i) - 1 satisfy 0 < δ and 1 + δ ≤ ∑ i ∈ I, f i.
Setting δ = (∑ i ∈ I, f i) - 1, the strict inequality 1 < ∑ i ∈ I, f i immediately gives 0 < δ, and 1 + δ ≤ ∑ i ∈ I, f i holds as an exact equality. Both conditions are verified by linear arithmetic.
exact ⟨I, (∑ i ∈ I, f i) - 1, by linarith, by linarith⟩
ENNReal.tsum_eq_iSup_sum- The infinite sum of extended nonnegative reals equals the supremum of all finite partial sums.
lt_iSup_iff.mp- If a value is strictly less than the supremum of an indexed family, then it is strictly less than some member of the family.
ENNReal.ofReal_sum_of_nonneg- For a finite set of nonnegative real numbers, the extended nonnegative real of their sum equals the sum of their extended nonnegative reals.
ENNReal.ofReal_pos.mp- If the extended nonnegative real of a real number is strictly positive, then the real number itself is strictly positive.
lt_of_lt_of_le- If a < b and b ≤ c, then a < c.
zero_lt_one- Zero is strictly less than one.
ENNReal.ofReal_one- The extended nonnegative real of 1 equals 1.
ENNReal.ofReal_lt_ofReal_iff- For real numbers p and q with 0 < q, ENNReal.ofReal p < ENNReal.ofReal q if and only if p < q.
- Lines 8–10 are a Lean-specific technicality: because `ENNReal.ofReal` maps all non-positive real numbers to 0, the lemma `ENNReal.ofReal_lt_ofReal_iff` requires the upper real number to be strictly positive. Hence the proof must explicitly verify that the real partial sum is positive and rewrite 1 as `ENNReal.ofReal 1` before stripping the extended-real coercion.
Written by a model (gemini-3.8-flash) from the Lean proof and the goal states Lean reported at each step. Lean verified the proof; this explanation is a reading aid and is not itself machine-checked.
Lean source view module on GitHub
Choose a tactic — a marked line — to see the goals Lean had before and after it, as in Lean’s infoview. Step with ← → or j k.
/-- A sum strictly greater than 1 is already exceeded by a finite partial sum: some finite set of colours carries mass `1 + δ` with `δ > 0`. -/theorem op_258_exists_finset_sum_gt_one {α : Type*} (f : α → ℝ) (h0 : ∀ i, 0 ≤ f i) (hsum : 1 < ∑' i : α, ENNReal.ofReal (f i)) : ∃ (I : Finset α) (δ : ℝ), 0 < δ ∧ 1 + δ ≤ ∑ i ∈ I, f i := by rw [ENNReal.tsum_eq_iSup_sum] at hsum obtain ⟨I, hI⟩ := lt_iSup_iff.mp hsum rw [← ENNReal.ofReal_sum_of_nonneg (fun i _ => h0 i)] at hI have hpos : 0 < ∑ i ∈ I, f i := ENNReal.ofReal_pos.mp (lt_of_lt_of_le (show (0 : ENNReal) < 1 from zero_lt_one) hI.le) rw [← ENNReal.ofReal_one] at hI have hgt : 1 < ∑ i ∈ I, f i := (ENNReal.ofReal_lt_ofReal_iff hpos).mp hI exact ⟨I, (∑ i ∈ I, f i) - 1, by linarith, by linarith⟩