Alphabeta Math
Pipeline-generated
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.

Quasilinear Characteristics and Cauchy Kovalevskaya — Examples

1 · Prerequisites

2 · Summary

These computations show the two distinct characteristic boundaries: smooth lifting can outlive an inverse-projection graph, and a fully nonlinear Cauchy problem can lose uniqueness when its rank condition fails. The final examples also distinguish analytic normal-form hypotheses from merely smooth transport solutions.

3 · Logical flowchart

4 · Definitions, theorems and proofs

None yet.

5 · Examples, counterexamples and false statements

ExampleConstruction: Literature-sourcedVerification: AI-generatedaudited 2026-09-06Open item page →

Semilinear characteristics with logistic growth

Example

Let g:R→R be C1. For ut+ux=u(1−u) and u(0,x)=g(x), the characteristic formula is

u(t,x)=g(x−t)et1+g(x−t)(et−1),

on the open set where 1+g(x−t)(et−1)>0. For each fixed label ξ=x−t, this is the time interval containing zero before any pole.

Facts & Assumptions

Given: A C1 datum g:R→R and a point (t,x) where the displayed denominator is positive.

Verification

technique · direct
1.1givenalgebra

The projected characteristics in the independent-variable plane are X(t,ξ)=(t,ξ+t). Thus their spatial component is ξ+t, so ξ=x−t, and the Jacobian of (t,ξ)↦(t,ξ+t) is 1.

2.1step 1.1algebra

Along one such curve, z′=z(1−z) and z(0)=g(ξ); put q=g(ξ) and D(t)=1+q(et−1). On D(t)>0, direct differentiation gives z=qet/D and z′=qet(1−q)/D2=z(1−z), with z(0)=q. This includes q=0 and q=1 without division by either. Since D(0)=1 and D′(t)=qet has constant sign, {t:D(t)>0} is exactly the interval containing zero on which the denominator is nonzero.

3.1step 2.1given∎

Substitute ξ=x−t to obtain the formula, and characteristic reconstruction verifies the PDE on its stated domain.

ExampleConstruction: Literature-sourcedVerification: AI-generatedjudge pass (gpt-5.6-terra)audited 2026-09-06Open item page →

Rarefying inviscid Burgers data

Example

For u0(ξ)=ξ, X=(1+t)ξ, so for t≥0 the solution is u(t,x)=x/(1+t). The fan expands and has no forward caustic.

Facts & Assumptions

Given: The Burgers characteristic formula with u0(ξ)=ξ and t≥0.

Verification

technique · direct
1.1givenalgebra

The formula gives X=ξ+tξ=(1+t)ξ and Xξ=1+t>0.

2.1step 1.1algebra∎

Inverting gives ξ=x/(1+t), while u(t,X)=u0(ξ)=ξ; hence u(t,x)=x/(1+t).

ExampleConstruction: Literature-sourcedVerification: AI-generatedjudge pass (gpt-5.6-terra)audited 2026-09-06Open item page →

Inviscid Burgers gradient catastrophe

Example

For u0(ξ)=−tanh⁡ξ, the map is X=ξ−ttanh⁡ξ. Its first crossing is t=1 at ξ=0, and the slope there tends to −∞.

Facts & Assumptions

Given: The Burgers formula and u0′(ξ)=−sech⁡2ξ.

Verification

technique · direct
1.1givenalgebra

Xξ=1−tsech⁡2ξ, which first vanishes at t=1, ξ=0, since 0<sech⁡2ξ≤1.

2.1step 1.1givenalgebra∎

The Riccati formula gives ux(t,X(t,0))=−1/(1−t), which tends to −∞ as t↑1.

CounterexampleConstruction: Literature-sourcedVerification: AI-generatedjudge pass (gpt-5.6-terra)audited 2026-09-06Open item page →

Quasilinear characteristics can cross before the lifted ODE blows up

Statement refuted

If the lifted characteristic ODE continues, then the quasilinear PDE remains a single-valued classical graph.

Counterexample

Given: Burgers data u0(ξ)=−tanh⁡ξ.

Proof technique: direct.

1.1givenalgebra

The lifted characteristic is Z(t,ξ)=−tanh⁡ξ and X(t,ξ)=ξ−ttanh⁡ξ, both defined for every t≥0.

2.1step 1.1algebra

Yet Xξ(1,0)=0, so the projected map has a caustic at t=1.

3.1step 2.1given∎

Thus the lifted ODE persists while inverse projection to a graph fails, refuting the statement.

ExampleConstruction: Literature-sourcedVerification: AI-generatedjudge pass (gpt-5.6-terra)audited 2026-09-06Open item page →

Clairaut complete integral and its nondegenerate stationary envelope

Example

For u=xux+(ux)2, S(x;a)=ax+a2 is a complete integral. Its stationary envelope is u=−x2/4.

Facts & Assumptions

Given: The equation and the family S(x;a)=ax+a2.

Verification

technique · direct
1.1givenalgebra

Sx=a, so xSx+(Sx)2=ax+a2=S; thus S is a complete integral.

1.2givenalgebra

Sa=x+2a=0 gives a=−x/2, and Saa=2 is invertible.

2.1step 1.1step 1.2algebra∎

Substitution gives u=S(x,−x/2)=−x2/4; the nondegenerate-envelope lemma makes it a classical solution.

ExampleConstruction: Literature-sourcedVerification: AI-generatedjudge pass (gpt-5.6-terra)audited 2026-09-06Open item page →

Eikonal cones are not classical at the vertex

Example

For n≥1, on Rn the function u(x)=∣x∣ solves ∣Du∣=1 away from x=0, but is not a classical solution at its vertex.

Facts & Assumptions

Given: An integer n≥1 and the Euclidean norm function u(x)=∣x∣.

Verification

technique · direct
1.1givenalgebra

For x≠0, Du(x)=x/∣x∣, so ∣Du(x)∣=1.

1.2givenalgebra

Along a unit vector e, the directional quotients at zero are u(te)/t=1 for t>0 and u(te)/t=−1 for t<0.

2.1step 1.2given∎

They have no common limit, so u is not differentiable at zero and cannot be classical there.

CounterexampleConstruction: Literature-sourcedVerification: AI-generatedjudge pass (gpt-5.6-terra)audited 2026-09-06Open item page →

Characteristic initial data need not determine a fully nonlinear solution

Statement refuted

Characteristic initial data always determine a unique local fully nonlinear classical solution.

Counterexample

Given: The equation F(x,y,z,px,py)=px2=0 and data u(x,0)=0.

Proof technique: direct.

1.1givenalgebra

Here Fp=(2px,0)=0 on F=0, so [Fp,Dγ] has rank 1<2 for γ(y)=(y,0).

1.2givenalgebra

Every u(x,y)=f(y) with f(0)=0 has ux=0, hence solves F=0 and attains the data.

2.1step 1.1step 1.2construct∎

Choosing distinct such f gives distinct local solutions, and step 1.1 identifies the failed rank hypothesis.

ExampleConstruction: Literature-sourcedVerification: AI-generatedjudge pass (gpt-5.6-terra)audited 2026-09-06Open item page →

Cauchy–Kovalevskaya normal form with analytic data

Example

For analytic g0,g1, the wave equation utt=uxx with u(0,x)=g0(x) and ut(0,x)=g1(x) is a second-order analytic Cauchy problem in normal form. This checks its hypotheses only; it does not invoke a recorded theorem.

Facts & Assumptions

Given: Analytic functions g0,g1 and the wave equation utt=uxx.

Verification

technique · direct
1.1givenalgebra

The equation is solved for the second normal derivative: utt=uxx, whose right side is analytic in the relevant jet variables.

2.1step 1.1given

The normal line is t, and the data supply exactly ∂t0u∣t=0=g0 and ∂tu∣t=0=g1, the orders 0 and 1 required for order 2.

3.1step 1.1algebra∎

The coefficient of utt is 1, so t=0 is noncharacteristic for this solved normal form.

ExampleConstruction: Literature-sourcedVerification: AI-generatedjudge pass (gpt-5.6-terra)audited 2026-09-06Open item page →

Quadratic Hamilton–Jacobi data produce an explicit caustic time

Example

For ut+(ux)2/2=0 with u0(ξ)=−ξ2/2, the characteristic map is X(t,ξ)=(1−t)ξ. Its first caustic time is 1.

Facts & Assumptions

Given: The Hamiltonian H(p)=p2/2 and initial momentum p0(ξ)=−ξ.

Verification

technique · direct
1.1givenalgebra

Since Hx=0, the momentum equation gives p˙=0, hence p(t,ξ)=−ξ.

2.1step 1.1givenalgebra

As X˙=Hp(p)=p, integration from X(0,ξ)=ξ gives X=(1−t)ξ.

3.1step 2.1given∎

Xξ=1−t is nonzero for t<1 and zero at t=1, which is the first caustic by the stated convention.

ExampleConstruction: Literature-sourcedVerification: AI-generatedjudge pass (gpt-5.6-terra)audited 2026-09-06Open item page →

Smooth nonanalytic transport data give a smooth nonanalytic solution

Example

Let g(x)=e−1/x2 for x≠0 and g(0)=0. The transport solution u(t,x)=g(x−t) is smooth but not analytic at each point x=t.

Facts & Assumptions

Given: The displayed flat function g and u(t,x)=g(x−t).

Verification

technique · direct
1.1givenalgebra

Every derivative of g at 0 is 0, while g(x)>0 for x≠0; thus g is C∞ but cannot equal its Taylor series near 0.

1.2givenalgebra

The chain rule gives ut=−g′(x−t) and ux=g′(x−t), hence ut+ux=0 and u(0,x)=g(x).

2.1step 1.1step 1.2given∎

Translation carries the flat nonanalytic point from 0 to x=t, so u is nonanalytic there despite being smooth.

Sources