Alphabeta Math
LemmaStatement: AI-adaptedProof: AI-adaptedprecheck passjudge pass (z-ai/glm-5.2)audited 2026-07-27
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.

A function on an interval satisfying f(x)≤f(y) whenever x≤y, whose image is order-convex, is continuous

Statement

Let I⊆R be order-convex (Intervals of R: the nine order-convex forms, nondegeneracy, and length) and let f:I→R satisfy

f(x)≤f(y)whenever x,y∈I and x≤y.

If the image f[I] is order-convex, then f is continuous on I (Continuity of f:A→R at a point of A and on A: the ε-δ condition, its agreement with lim⁡x→cf(x)=f(c) at a limit point, and continuity at an isolated point).

No definition of a monotone function is used, and none is available at this point in the reading order. The hypothesis is written out as the displayed inequality; the classification of monotone functions and their discontinuities comes later in the library. Equivalently, by A subset of R is connected if and only if it is order-convex, that is, an interval, the hypothesis on the image is that f[I] is connected (Separated sets, disconnection, and connected subset of R).

The hypothesis on the image cannot be dropped. Define f on I=[0,1] by f(x):=x for x<1 and f(1):=2. It satisfies the displayed inequality, its image is [0,1)∪{2}, which is not order-convex, and it is not continuous at 1: no δ works for ε=1/2, since points of [0,1) arbitrarily close to 1 have values close to 1, at distance close to 1 from f(1)=2.

This is a genuine converse to the intermediate value property, in the presence of the inequality. It does not need one-sided limits of monotone functions, which are not available at this point in the reading order; the entire proof is the two paragraphs below, which read the required δ off the image.

Facts & Assumptions

Given: An order-convex set I⊆R and a function f:I→R with f(x)≤f(y) whenever x,y∈I and x≤y, such that f[I] is order-convex; and a point c∈I together with a real ε>0.

[L1]

Continuity of f at c: for every real ε>0 there is a real δ>0 with ∣f(x)−f(c)∣<ε for every x∈I satisfying ∣x−c∣<δ (Continuity of f:A→R at a point of A and on A: the ε-δ condition, its agreement with lim⁡x→cf(x)=f(c) at a limit point, and continuity at an isolated point, The ε-neighbourhood and the punctured ε-neighbourhood of a point of R).

[L2]

Order-convexity of f[I]: if u,v∈f[I] and u≤w≤v then w∈f[I] (Intervals of R: the nine order-convex forms, nondegeneracy, and length); equivalently f[I] is connected (A subset of R is connected if and only if it is order-convex, that is, an interval, Separated sets, disconnection, and connected subset of R).

[L3]

Order and field arithmetic in R: trichotomy and totality of the order, so any two reals are comparable and exactly one of r<c, r=c, r>c holds; ε>0 gives ε/2>0 and f(c)−ε<f(c)−ε/2<f(c) (Ordered field).

[L4]

The minimum of a two-element set of reals exists and is one of the two elements (Every nonempty finite set of reals has a maximum and a minimum, Maximum and minimum of a set).

[L5]

Absolute value: for s>0, ∣t∣<s holds exactly when −s<t<s (Basic properties of the absolute value).

Proof

technique · direct
1.1

A point of I below c with a value close to f(c), when one exists at all. Suppose some p∈I has p<c. We claim there is q∈I with q<c and f(q)>f(c)−ε. If there were none, then f(q)≤f(c)−ε for every q∈I with q<c, and in particular f(p)≤f(c)−ε. Put w:=f(c)−ε/2, so f(p)≤f(c)−ε<w<f(c) by [L3]. Since f(p),f(c)∈f[I], [L2] gives r∈I with f(r)=w. By [L3] exactly one of r<c, r=c, r>c holds: the first gives w=f(r)≤f(c)−ε<w, the second gives w=f(c), and the third gives w=f(r)≥f(c) by the monotonicity hypothesis; each contradicts f(c)−ε<w<f(c). So the claimed q exists.

L2L3
2.1

The left radius. If some p∈I has p<c, fix q as in step 1.1 and put δ1:=c−q>0; then every x∈I with c−δ1<x≤c satisfies q<x≤c, hence f(q)≤f(x)≤f(c) by monotonicity, hence f(c)−ε<f(x)≤f(c). If no point of I lies below c, put δ1:=1; then the only x∈I with c−δ1<x≤c is x=c, for which f(c)−ε<f(c)=f(x)≤f(c) holds as well. In both cases δ1>0 and every x∈I with c−δ1<x≤c satisfies f(c)−ε<f(x)≤f(c).

step 1.1L3choose
2.2

The right radius, symmetrically. Suppose some p∈I has p>c. If every s∈I with s>c had f(s)≥f(c)+ε, then with w′:=f(c)+ε/2 we would have f(c)<w′<f(c)+ε≤f(p), so [L2] would give r∈I with f(r)=w′; but by [L3] exactly one of r>c, r=c, r<c holds, and the first gives w′=f(r)≥f(c)+ε>w′, the second gives w′=f(c), and the third gives w′=f(r)≤f(c) by the monotonicity hypothesis, each contradicting f(c)<w′<f(c)+ε. So there is s∈I with s>c and f(s)<f(c)+ε; put δ2:=s−c>0. If no point of I lies above c, put δ2:=1. In both cases δ2>0 and every x∈I with c≤x<c+δ2 satisfies f(c)≤f(x)<f(c)+ε.

step 1.1L2L3choose
3.1

Combining. Put δ:=min⁡{δ1,δ2}, which is a positive real by [L4]. Let x∈I with ∣x−c∣<δ, so c−δ<x<c+δ by [L5]. By totality either x≤c, and then c−δ1≤c−δ<x≤c, so step 2.1 gives f(c)−ε<f(x)≤f(c); or x≥c, and then c≤x<c+δ≤c+δ2, so step 2.2 gives f(c)≤f(x)<f(c)+ε. In either case −ε<f(x)−f(c)<ε, that is ∣f(x)−f(c)∣<ε by [L5].

step 2.1step 2.2L3L4L5
4.1

The point c∈I and the real ε>0 were arbitrary, so by [L1] the function f is continuous at every point of I, that is, continuous on I.

step 3.1L1∎

Remarks

  • Where order-convexity of the image is used, and where it is not. It is used exactly twice, in steps 1.1 and 2.2, each time to convert a value strictly between two attained values into an attained value. Nothing else in the argument looks at the image. In particular, no continuity of f is assumed anywhere, which is what makes the lemma a converse rather than a reformulation.

  • The endpoint cases are not a technicality. If c is the left endpoint of I there is no point of I below it, and the left half of the estimate is vacuous; the same at the right. Handling them by the fixed radius 1 keeps the proof free of any hypothesis that I be open or nondegenerate.

  • What this lemma is for. It is the standard route to continuity of a function defined by a monotone construction whose image is known independently — the Cantor function is the classical instance, its image being all of [0,1] — and it is stated here as a standalone lemma so that a later page may cite it rather than repeat the argument.

Depends on

Used by

Dependency tree · two levels

28 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