ising_inverse_square_order_two_formula_limit
For every real y, the finite-ring rational expression from the second-order coefficient formula tends to (1-y)(1+9y)/180 as the natural ring size tends to infinity.
Statement
theorem ising_inverse_square_order_two_formula_limit (y : ℝ) : Filter.Tendsto (fun L : ℕ => let u : ℝ := (L : ℝ)⁻¹ let a := (1-u^2)/3 let b := (1+10*u^2-11*u^4)/45 (1-y^2)/4*b + y*(1-y)/2*(a^2-b)) Filter.atTop (nhds ((1-y)*(1+9*y)/180))
- Source
- Helper for the second-order coefficient from arXiv:2610.39815, equations (3.15)–(3.16), and OP-806.
- Verified
- 02 Nov 2026
- Axioms
Classical.choiceQuot.soundpropext- Used by 1
- ising_inverse_square_bulk_two_of_finite_formulaIf the displayed second-order formula holds for every ring size L at least 2 and every real field, then the polynomial (1-y)(1+9y)/180…
The proof, explained
The limit is found by viewing the sequence as a polynomial expression in u = 1/L. Because 1/L tends to 0 as L → ∞ and polynomials are continuous, the limit as L → ∞ is simply the expression evaluated at u = 0. Substituting u = 0 and simplifying the resulting quadratic in y yields the desired limit (1 - y)(1 + 9y)/180.
As L → ∞ in ℕ, the reciprocal (L : ℝ)⁻¹ tends to 0.
The natural cast L ↦ (L : ℝ) tends to ∞ as L → ∞, and the real inversion map x ↦ x⁻¹ tends to 0 as x → ∞. Composing these two limits shows that (L : ℝ)⁻¹ → 0.
uses
tendsto_inv_atTop_zero,tendsto_natCast_atTop_atTophave hu : Filter.Tendsto (fun L : ℕ => (L : ℝ)⁻¹) Filter.atTop (nhds 0) := tendsto_inv_atTop_zero.comp tendsto_natCast_atTop_atTop
The rational expression is a continuous function of u on ℝ.
For any fixed real parameter y, the expression (1 - y²)/4 · b(u) + y(1 - y)/2 · (a(u)² - b(u)) is a polynomial in u, as a(u) = (1 - u²)/3 and b(u) = (1 + 10u² - 11u⁴)/45 are polynomials. It is therefore continuous everywhere on ℝ.
have hc : Continuous (fun u : ℝ => let a := (1-u^2)/3 let b := (1+10*u^2-11*u^4)/45 (1-y^2)/4*b + y*(1-y)/2*(a^2-b)) := by fun_prop
The sequence of expressions evaluated at u = (L : ℝ)⁻¹ tends to the value of the expression at u = 0 as L → ∞.
By continuity, the function tends to its value at 0 as u → 0. Composing this limit with the convergence (L : ℝ)⁻¹ → 0 as L → ∞ yields the convergence of the sequence to the expression evaluated at u = 0.
have hh := (hc.tendsto 0).comp huEvaluating the expression at u = 0 yields (1 - y)(1 + 9y)/180, concluding the proof.
Setting u = 0 gives a(0) = 1/3 and b(0) = 1/45, so a(0)² - b(0) = 1/9 - 1/45 = 4/45. The expression at u = 0 evaluates to (1 - y²)/180 + 2y(1 - y)/45 = (1 + 8y - 9y²)/180, which factors as (1 - y)(1 + 9y)/180. Ring arithmetic confirms this algebraic equality.
convert hh using 1 <;> first | rfl | (congr 1; norm_num; ring)
tendsto_inv_atTop_zero- The reciprocal function x ↦ x⁻¹ tends to 0 as x tends to infinity.
tendsto_natCast_atTop_atTop- The natural cast from ℕ to an Archimedean ordered semiring tends to infinity as n tends to infinity.
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.
theorem ising_inverse_square_order_two_formula_limit (y : ℝ) : Filter.Tendsto (fun L : ℕ => let u : ℝ := (L : ℝ)⁻¹ let a := (1-u^2)/3 let b := (1+10*u^2-11*u^4)/45 (1-y^2)/4*b + y*(1-y)/2*(a^2-b)) Filter.atTop (nhds ((1-y)*(1+9*y)/180)) := by have hu : Filter.Tendsto (fun L : ℕ => (L : ℝ)⁻¹) Filter.atTop (nhds 0) := tendsto_inv_atTop_zero.comp tendsto_natCast_atTop_atTop have hc : Continuous (fun u : ℝ => let a := (1-u^2)/3 let b := (1+10*u^2-11*u^4)/45 (1-y^2)/4*b + y*(1-y)/2*(a^2-b)) := by fun_prop have hh := (hc.tendsto 0).comp hu convert hh using 1 <;> first | rfl | (congr 1; norm_num; ring)