Alphabeta Math
False statementConstruction: AI-adaptedVerification: AI-adaptedPipeline-generatedjudge pass (gpt-5.6-terra)audited 2026-09-10
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.

Riemannian distance is defined by the length of a unique shortest curve

Statement

Riemannian distance is the length of a unique shortest curve. In fact both attainment and uniqueness can fail.

Facts & Assumptions

Given: First use the Euclidean metric on P=R2{0} with endpoints (1,0),(1,0). Then use the induced metric on S1 with endpoints (1,0),(1,0).

[F1]

Riemannian distance on a connected manifold: On a connected Riemannian manifold define dg(p,q)=inf{Lg(γ):γ is piecewise C1 from p to q}. Lengths are those of def-riemannian-speed-and-length. For each pair p,q, lem-any-two-points-in-a-connected-smooth-manifold-can-be-joined-by-a-piecewise-c-one-curve supplies a curve, so the set of lengths is nonempty, contains a finite real number and is bounded below by zero. Applying the least-upper-bound property cor-cauchy-reals-lub-complete to the negatives gives a finite nonnegative infimum. On the empty connected manifold this defines the empty distance function; there are no pairs to evaluate. No minimizing curve is part of this definition.

[F3]

Length is additive under concatenation and invariant under reversal: Length adds under finite concatenation and is unchanged by reversal.

[F4]

Newton–Leibniz needs only continuity on [a,b], differentiability on (a,b), and a Riemann-integrable extension of the interior derivative: Let a<b. Suppose G:[a,b]R is continuous on [a,b] and differentiable on (a,b). If f:[a,b]R is Riemann integrable and f(x)=G(x)(a<x<b), then abf=G(b)G(a). No derivative of G at either endpoint is assumed, and the two endpoint values assigned to the integrable extension f do not enter the conclusion.

Refutation

technique · direct
1.1

For a piecewise C1 path γ=(x,y) in P between the prescribed points, its Euclidean speed is x2+y2. On each smooth piece this is at least x. Integrating and telescoping the endpoint differences gives L(γ)x(b)x(a)=2; Newton–Leibniz applies to the continuously differentiable coordinates on every closed piece.

F4given
1.2

For any piecewise C1 path in S1, subdivide its parameter interval so that each piece lies in one open arc admitting a smooth angle coordinate. Such a finite subdivision exists: the inverse images of these arcs cover the compact interval; a finite subcover has a positive Lebesgue number, so sufficiently short equal subintervals refine it. Refine further at the original smooth-piece endpoints. Choose an angle value at the initial point and successively add integer multiples of 2π to each local angle so adjacent values agree at the joining parameter. This gives a continuous piecewise C1 angle θ with γ=(cosθ,sinθ). Differentiation gives speed θ.

given
2.1

For 0<ε<1, travel along the horizontal axis from (1,0) to (ε,0), along the upper semicircle of radius ε, and along the axis from (ε,0) to (1,0). All three pieces avoid the origin. Their lengths are 1ε, πε, and 1ε: the semicircle parametrization (εcost,εsint) has speed ε for 0tπ, with reversal giving the required direction. Additivity and reversal yield 2+(π2)ε. Consequently the infimum is 2.

F1F3step 1.1
3.1

If an admissible path had length 2, then the integral of x2+y2x would be zero. This function is continuous and nonnegative on each smooth piece, so it vanishes on each piece (a positive value would give a positive integral on a small interval). Thus y=0 and x0 there. Newton–Leibniz and continuity at the subdivision points imply y0. The intermediate value theorem gives a parameter with x=0, contradicting avoidance of the origin. Hence the infimum on P is not attained.

F4step 1.1step 2.1
4.1

For the antipodal endpoints, take θ(a)=0. Then θ(b)=(2k+1)π for some integer k. Integrating θ and applying Newton–Leibniz on the pieces gives Lθ(b)θ(a)π. The paths t(cost,sint) and t(cost,sint) for 0tπ have speed 1, length π, and distinct images. Both attain the distance. This proves failure of uniqueness as well as the failure of existence in step 3.1.

F1F4step 3.1step 1.2

Source locator

Lee, pp. 337–338, Riemannian length and distance. The nonattainment and antipodal calculations, including the finite angle-lift construction, are supplied above; no geodesic existence theorem is used.

Depends on

Used by

Nothing in the library uses this result yet.

Dependency tree · two levels

18 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