Alphabeta Math
TheoremStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedprecheck passaudited 2026-10-02
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

Sturm comparison for scalar jacobi equations

Statement

Let L>0 and let a,b:[0,L]→R be continuous functions with a(t)≤b(t) for every t∈[0,L]. Let u,v:[0,L]→R be twice continuously differentiable solutions of u′′+a u=0,v′′+b v=0on [0,L], with the common initial data u(0)=v(0)=0,u′(0)=v′(0)=c>0. Let τ≥0 be the first positive zero of v when v has one in (0,L] and put τ:=+∞ when it does not; thus v(t)>0 for every t∈(0,L] with t<τ and, when τ≤L, v(τ)=0. Then:

  1. v(t)>0 and u(t)≥v(t)>0 for every t∈(0,L] with t<τ;
  2. u has no zero in (0,L] before τ; when τ≤L one has u(τ)≥v(τ)=0, so the first positive zero of u is no earlier than τ.

The result concerns two scalar ordinary differential equations on a closed interval. It uses no manifold, no completeness hypothesis and no choice principle: only the pointwise ordering a≤b of the coefficients, the common positive initial slope and the two equations enter. The positivity c>0 is essential for the direction stated; the degenerate initial slope c=0 admits the solution u≡v≡0, for which no inequality of the statement is asserted.

Facts & Assumptions

Given: A length L>0, continuous functions a,b:[0,L]→R with a≤b pointwise, twice continuously differentiable u,v solving u′′+au=0 and v′′+bv=0 on [0,L], the common initial data u(0)=v(0)=0, u′(0)=v′(0)=c>0, and the first positive zero τ of v as in the Statement.

[F1]

Differentiation is linear and the product rule holds: for differentiable f,g one has (fg)′=f′g+fg′ and (f±g)′=f′±g′, and these rules apply on an interval (Sums, scalar multiples, products and quotients: (f+g)′(c)=f′(c)+g′(c), (αf)′(c)=αf′(c), (fg)′(c)=f′(c)g(c)+f(c)g′(c), and (f/g)′(c)=(f′(c)g(c)−f(c)g′(c))/g(c)2 when g(c)≠0). In particular h:=u′v−uv′ is differentiable on [0,L] with h′=u′′v−uv′′.

[F2]

Sign preservation and the mean value theorem: a function w continuous at 0 with w(0)=c>0 satisfies w≥c/2>0 on some interval [0,δ], and if w is continuous on [0,t] and differentiable on (0,t) then w(t)−w(0)=w′(ξ) t for some ξ∈(0,t); consequently u(t)=u′(ξ)t>0 and v(t)=v′(η)t>0 for all sufficiently small t>0 (If lim⁡x→cf(x)=L≠0 then ∣f∣>∣L∣/2 on a punctured neighbourhood of c; in particular if L>0 then f>L/2>0 there, The mean value theorem, as the case g(x)=x of Cauchy's: for f continuous on [a,b] with a<b and differentiable on (a,b) there is c∈(a,b) with f(b)−f(a)=f′(c)(b−a)).

[F3]

The real numbers have the least-upper-bound property: every nonempty set of reals that is bounded above has a least upper bound (Complete ordered field (least-upper-bound property), Dedekind completeness: the least-upper-bound property).

[F4]

Model data for a positive constant coefficient k: sn⁡k′′+k sn⁡k=0 on R with sn⁡k(0)=0 and sn⁡k′(0)=1 (Model functions solve the constant curvature jacobi equation), while sn⁡k(t)>0 for 0<t<π/k and sn⁡k(π/k)=0 (Comparison sine, cosine and cotangent functions).

Proof

1.1F1given

The reduced Wronskian identity. [F1, given] Define h(t):=u′(t)v(t)−u(t)v′(t) on [0,L]. By [F1] the product rule applies to both terms and the two differential equations give h′=u′′v+u′v′−u′v′−uv′′=(−au)v−u(−bv)=(b−a)uv. At the left endpoint, h(0)=u′(0)v(0)−u(0)v′(0)=c⋅0−0⋅c=0.

1.2F2given

Both solutions are positive immediately to the right of 0, and the quotient has limit 1. [F2, given] The functions u′ and v′ are continuous at 0 with u′(0)=v′(0)=c>0, so [F2] yields δ>0 with u′≥c/2>0 and v′≥c/2>0 on [0,δ]. For 0<t<δ the mean value theorem in [F2] applied to u and to v on [0,t] produces ξ,η∈(0,t) with u(t)=u′(ξ)t>0 and v(t)=v′(η)t>0. Hence the quotient f:=u/v is defined on (0,δ), and since u(t)/t→u′(0) and v(t)/t→v′(0) as t↓0 by the definition of the one-sided derivative at the endpoint, f(t)=u(t)/tv(t)/t⟶u′(0)v′(0)=1(t↓0).

2.1F1F2F3step 1.1step 1.2

The inequality propagates up to the supremum of the set of times where it holds. [F3, step 1.1, step 1.2] Put A:={ t∈(0,L]:u(s)≥v(s)>0 for every s∈(0,t] }. This set is nonempty: on the initial interval of step 1.2 both solutions are positive, so step 1.1 gives h′=(b−a)uv≥0. The mean value theorem and h(0)=0 give h≥0 there. Therefore (u/v)′=h/v2≥0, and the limit u/v→1 gives u≥v throughout this initial interval. Any sufficiently small positive endpoint thus belongs to A. The set is bounded above by L, so T:=sup⁡A exists in R by [F3] and 0<T≤L. For every s∈(0,T) choose t∈A with s<t, which is possible because T is the least upper bound of A; then u(s)≥v(s)>0. Hence uv≥0 on (0,T), and step 1.1 together with a≤b gives h′=(b−a)uv≥0 on (0,T). Since h(0)=0 by step 1.1, the mean value theorem in [F2] applied on [0,s] yields h(s)=h′(ξ)s≥0 for every s∈(0,T), so h≥0 on (0,T). The denominator v is positive on (0,T), so the quotient rule in [F1] together with step 1.1 gives f′=u′v−uv′v2=hv2≥0on (0,T), that is, f is nondecreasing on (0,T). For fixed s∈(0,T) and 0<t<s this gives f(s)≥f(t), and letting t↓0 in the limit of step 1.2 yields f(s)≥1, that is, u(s)≥v(s).

3.1step 2.1

The supremum reaches the first zero of v or the domain endpoint. [step 2.1] Put τL:=min⁡{τ,L}. We claim T≥τL. Suppose instead T<τL, so that T<L and T<τ, and since τ is the first positive zero of v, one has v>0 on (0,T]; put d:=v(T)>0. From u≥v on (0,T) and continuity of u and v at T we get u(T)≥v(T)=d. Choose ε>0 with T+ε<L and with u>d/2>0 and v>d/2>0 on [T,T+ε], possible by continuity. Then uv≥0 on (0,T+ε): on (0,T) because u≥v>0, and on (T,T+ε) because both factors are positive there. By step 1.1, h′≥0 on (0,T+ε), and from h(0)=0 the mean value theorem in [F2] gives h≥0 on (0,T+ε). Since v>0 on (0,T+ε), the quotient rule and step 1.1 give f′=h/v2≥0 on (0,T+ε); monotonicity of f together with the limit f(0+)=1 of step 1.2 then gives u(s)≥v(s) for every s∈(0,T+ε). As v>0 on (T,T+ε), every t with T<t<T+ε satisfies t∈A, contradicting the definition of T as an upper bound of A. Therefore T≥τL. For every t∈(0,τL), step 2.1 gives u(t)≥v(t)>0. If τL=L<τ, continuity extends u≥v>0 to t=L; if τL=τ≤L, continuity gives u(τ)≥v(τ)=0. This proves the stated conclusions on the domain [0,L].

4.1step 3.1

The first positive zero of u is no earlier than τ. [step 3.1] If v has no zero in (0,L], then τ=+∞ and step 3.1 gives u(t)≥v(t)>0 on all of (0,L]; in particular neither u nor v has a positive zero. If v has a zero in (0,L], let τ be its first one: by step 1.2 the zero set lies in [δ,L] for some δ>0; it is a nonempty closed subset of this compact interval, so it has a minimum τ with v(τ)=0, and no zero lies in (0,τ). Since u≥v on (0,τ) by step 3.1 and u is continuous, passing to the limit gives u(τ)≥v(τ)=0. Hence u>0 on (0,τ), and its first positive zero, when it exists, is at least τ.

5.1F4step 4.1∎

Model case and boundary conventions. [F4, step 4.1] For a≡b≡k>0 and u=v=sn⁡k, [F4] shows that both equations hold with common initial slope c=1, and step 4.1 recovers sn⁡k>0 on (0,π/k). For the comparison form, let k>0 and suppose the coefficient of a given solution v satisfies b(t)≤k for all t; take the pair of the theorem to be the given v and c sn⁡k: the first member solves w′′+b w=0, the second solves w′′+k w=0 by [F4] and has initial slope c⋅1=c, and b≤k is the required pointwise ordering. Step 4.1, applied only on the common domain [0,L], shows that v has no zero on (0,min⁡{L,π/k}); if L<π/k, it is positive also at L, and if L≥π/k, its first zero on [0,L] is no earlier than π/k. The interval is one-sided at 0, where the initial data and the one-sided derivatives live; the endpoint L enters only through the convention for τ. The degenerate coefficient b≡0 admits v(t)=ct, which solves the equation and has no positive zero, so the statement is consistent in the flat case. The excluded case c=0 has the zero solution and no comparison content, and the empty interval is excluded by L>0. No manifold, no indexed family of solutions and no selection principle occur, so the argument is choice-free.

Depends on

Used by

Dependency tree · two levels

37 results within two dependency steps of this one, each drawn at its shortest distance from it. An arrow runs from a result to what uses it, so the chart reads left to right and ends at this result, which carries a heavier outline. Every node is a link to that result. Click elsewhere on the chart to enlarge it.

Sources