Alphabeta Math
TheoremStatement: Literature-sourcedProof: AI-generatedSession-authored (Fable 5 assisted)precheck passjudge pass (z-ai/glm-5.2)audited 2026-07-26
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 subset of R\mathbb{R} is connected if and only if it is order-convex, that is, an interval

Statement

Let ERE \subseteq \mathbb{R}. Then EE is connected (Separated sets, disconnection, and connected subset of R\mathbb{R}) if and only if EE is order-convex (Intervals of R\mathbb{R}: the nine order-convex forms, nondegeneracy, and length), that is, if and only if

x,zE and xwz    wE.x, z \in E \text{ and } x \le w \le z \;\Longrightarrow\; w \in E .

On the word "interval". Order-convexity is exactly the defining property that Intervals of R\mathbb{R}: the nine order-convex forms, nondegeneracy, and length proves for each of its nine forms, and in that sense the theorem says that the connected subsets of R\mathbb{R} are the intervals. The converse classification, that every order-convex subset of R\mathbb{R} is empty or one of the nine forms, is true and is explicitly not proved anywhere in this library; Intervals of R\mathbb{R}: the nine order-convex forms, nondegeneracy, and length records that omission in its own remarks. So the statement proved below is the equivalence with order-convexity, and the phrase "is an interval" is to be read as "is order-convex" throughout this page.

Facts & Assumptions

Given: A subset ERE \subseteq \mathbb{R}.

[L1]

Separated sets, disconnection, connectedness; separated sets are disjoint (Separated sets, disconnection, and connected subset of R\mathbb{R}).

[L2]

A\overline{A} is the smallest closed superset of AA, so ABA \subseteq B gives AB\overline{A} \subseteq \overline{B} and AF\overline{A} \subseteq F for every closed FAF \supseteq A; and A\overline{A} is exactly the set of points every neighbourhood of which meets AA (The closure equals the set together with its limit points, equals the set of points every neighbourhood of which meets it, and is the smallest closed superset; a set is closed iff it contains its limit points, Interior, closure, boundary and exterior of a subset of R\mathbb{R}).

[L3]

Order-convexity, and the interval forms: (,w](-\infty,w] and [w,)[w,\infty) are closed sets, (,w)(-\infty,w) and (w,)(w,\infty) are open sets, and the order is total and transitive (Intervals of R\mathbb{R}: the nine order-convex forms, nondegeneracy, and length, Open subset of R\mathbb{R} (every point has a neighbourhood inside it), closed subset (complement open), and clopen, Ordered field, Complete ordered field (least-upper-bound property)).

[L4]

Least-upper-bound property: a nonempty subset of R\mathbb{R} bounded above has a unique least upper bound (Complete ordered field (least-upper-bound property), Suprema and infima are unique, Lower bound, bounded below, bounded set).

[L5]

Epsilon characterisation: for nonempty SS bounded above and c=supSc = \sup S, every ε>0\varepsilon > 0 admits sSs \in S with cε<sc - \varepsilon < s (Epsilon characterisation of the supremum).

[L6]

Nε(x)={y:yx<ε}N_\varepsilon(x) = \{\, y : |y - x| < \varepsilon \,\} (The ε\varepsilon-neighbourhood and the punctured ε\varepsilon-neighbourhood of a point of R\mathbb{R}).

[L7]

Every nonempty finite set of reals has a minimum, which is one of its members (Every nonempty finite set of reals has a maximum and a minimum, Maximum and minimum of a set).

[L8]

Ordered-field arithmetic: 0<10 < 1, so 2:=1+1>02 := 1+1 > 0 and 21>02^{-1} > 0; for d>0d > 0 one has 0<d21<d0 < d \cdot 2^{-1} < d; adding a constant preserves an inequality (The multiplicative identity is positive, Order is preserved by adding a constant and by adding inequalities, Sign rules for products and monotonicity of multiplication, Ordered field). These order-arithmetic facts are stated by their sources for the strict order only; the nonstrict forms used below follow by adjoining the equality case, in which the two sides coincide.

Proof

technique · direct
1.1

Suppose EE is not order-convex: there are x,zEx, z \in E and wRw \in \mathbb{R} with xwzx \le w \le z and wEw \notin E; then wxw \ne x and wzw \ne z, so x<w<zx < w < z. Put A:=E(,w)A := E \cap (-\infty,w) and B:=E(w,)B := E \cap (w,\infty); then xAx \in A and zBz \in B, so both are nonempty, and AB=EA \cup B = E because no element of EE equals ww.

assume-hypL3
1.2

Suppose instead that EE is order-convex and that (A,B)(A,B) is a disconnection of EE; fix pAp \in A and qBq \in B. Separated sets are disjoint by [L1], so pqp \ne q, and interchanging the names AA and BB if necessary, which is legitimate because the hypotheses on the pair are symmetric, we may assume p<qp < q.

assume-hypL1choose
1.3

For a nonempty SRS \subseteq \mathbb{R} bounded above, supSS\sup S \in \overline{S}: for every real ε>0\varepsilon > 0 the fact [L5] supplies sSs \in S with supSε<ssupS\sup S - \varepsilon < s \le \sup S, so ssupS<ε|s - \sup S| < \varepsilon and sNε(supS)Ss \in N_\varepsilon(\sup S) \cap S; thus every neighbourhood of supS\sup S meets SS, and [L2] gives supSS\sup S \in \overline{S}.

L2L4L5L6
2.1

In the situation of step 1.1 the pair (A,B)(A,B) is a disconnection: (,w](-\infty,w] is a closed set containing AA, so A(,w]\overline{A} \subseteq (-\infty,w] by [L2], whence AB(,w](w,)=\overline{A} \cap B \subseteq (-\infty,w] \cap (w,\infty) = \varnothing; symmetrically B[w,)\overline{B} \subseteq [w,\infty) and AB=A \cap \overline{B} = \varnothing. So AA and BB are separated, nonempty, and their union is EE, and EE is disconnected.

step 1.1L1L2L3
2.2

In the situation of step 1.2 put S:=A[p,q]S := A \cap [p,q]; it is nonempty because pAp \in A and ppqp \le p \le q, and it is bounded above by qq, so c:=supSc := \sup S exists by [L4], and pcqp \le c \le q since pSp \in S and qq is an upper bound.

step 1.2L3L4
3.1

cAc \in A: from SAS \subseteq A and [L2] we get SA\overline{S} \subseteq \overline{A}, and cSc \in \overline{S} by step 1.3, so cAc \in \overline{A} and hence cBc \notin B because AB=\overline{A} \cap B = \varnothing; on the other hand pcqp \le c \le q with p,qEp, q \in E and EE order-convex gives cE=ABc \in E = A \cup B, so cAc \in A.

step 1.2step 1.3step 2.2L1L2
4.1

c<qc < q, since cAc \in A and qBq \in B are distinct by [L1] while cqc \le q; and every vv with c<vqc < v \le q lies in BB: such a vv satisfies pc<vqp \le c < v \le q, so vEv \in E by order-convexity, and vAv \notin A, for otherwise vA[p,q]=Sv \in A \cap [p,q] = S would force vcv \le c.

step 1.2step 2.2step 3.1L1L3
5.1

cBc \in \overline{B}, which is impossible: given a real ε>0\varepsilon > 0, put t:=min{ε21, (qc)21}t := \min\{\varepsilon \cdot 2^{-1},\ (q - c) \cdot 2^{-1}\}, a positive real by [L7] and [L8] since qc>0q - c > 0, and v:=c+tv := c + t; then c<vc < v and vc+(qc)21<qv \le c + (q-c) \cdot 2^{-1} < q, so vBv \in B by step 4.1, while vc=tε21<ε|v - c| = t \le \varepsilon \cdot 2^{-1} < \varepsilon, so vNε(c)Bv \in N_\varepsilon(c) \cap B. Hence every neighbourhood of cc meets BB and cBc \in \overline{B} by [L2]; but cAc \in A by step 3.1 and AB=A \cap \overline{B} = \varnothing by [L1]. So the assumed disconnection cannot exist and an order-convex EE is connected.

step 3.1step 4.1L1L2L6L7L8
6.1

Step 2.1 shows that a set which is not order-convex is disconnected, hence a connected set is order-convex; step 5.1 shows that an order-convex set admits no disconnection, hence is connected. The two together are the asserted equivalence.

step 2.1step 5.1

Remarks

  • Where completeness is spent. Only in step 2.2, which produces sup(A[p,q])\sup(A \cap [p,q]); no other step uses the least-upper-bound property, and the rest is the order, ordered-field arithmetic and the definition of separation. The obstruction over an incomplete ordered field is traceable to the failure of that supremum to exist, and it is visible in Q[0,2]\mathbb{Q} \cap [0,2] is bounded and disconnected, so being an interval of Q\mathbb{Q} is not enough : the set Q[0,2]\mathbb{Q} \cap [0,2] contains all the rationals between its endpoints and is nevertheless disconnected as a subset of R\mathbb{R}, split at an irrational point that Q\mathbb{Q} does not see.

  • The two directions are of different characters. "Not order-convex implies disconnected" is a construction, step 1.1, and needs nothing beyond the order. "Order-convex implies connected" is where the work sits, and the supremum cc produced in step 2.2 is the point at which the two pieces would have to meet; the contradiction is that it is adherent to both.

  • The theorem is about subsets of R\mathbb{R} and its statement is written in order vocabulary, so it cannot even be stated where no order is present; Which results on this page use the order of R\mathbb{R} and therefore have no general-topological analogue collects the results on this page with that feature.

Depends on

Used by

Dependency tree · next 3 levels

Direct dependencies and their dependencies through the next three levels: 36 results over 14 levels. An arrow runs from a result to what uses it, and this result sits at the bottom with a heavier outline. Click the chart to enlarge it.

Sources