Alphabeta Math
TheoremStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedjudge pass (gpt-5.6-terra)audited 2026-09-22
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.

DMC implies Urysohn's lemma

Statement

ZF+DMC proves Urysohn's lemma: in a normal space (Normal spaces and T4 spaces, with the source disagreement over whether normality includes T1 stated explicitly) any two disjoint closed sets F,G admit a continuous f:X→[0,1] (Continuity of a map of topological spaces at a point and globally, Intervals of R: the nine order-convex forms, nondegeneracy, and length) with F⊆f−1({0}) and G⊆f−1({1}).

DMC is used once, in its successor-menu form of Dependent multiple choice in finite-level tree form: it supplies the finite menus of dyadic nodes, and the finitely many open sets inside each menu are intersected coordinatewise to obtain a single dyadic scale.

Facts & Assumptions

Given: A normal space X; disjoint closed sets F,G⊆X; the principle DMC.

[F1]

Normality via shrinking: if A is closed, U is open and A⊆U, then there is open V with A⊆V⊆V‾⊆U (A space is normal if and only if every closed A inside an open U admits an open V with A⊆V⊆V‾⊆U).

[F2]

The dyadic rationals D⊆[0,1] are an increasing union of finite levels Dn, the level Dn+1 inserts one new point strictly between each pair of Dn-consecutive elements, every two elements of D lie in a common Dn, and the positive dyadics have infimum zero (the displayed dyadic growth bound gives 2−n→0) (The dyadic rationals of [0,1], their finite levels Dn, and their density in [0,1]).

[F3]

Dyadic scale lemma: if (Ur)r∈D are open subsets of X with Ur‾⊆Us whenever r<s and U1=X, then f(x):=inf⁡({r∈D:x∈Ur}∪{1}) is a continuous map X→[0,1] (If (Ur)r∈D are open with Ur‾⊆Us whenever r<s and U1=X, then x↦inf⁡{r∈D:x∈Ur} is a continuous map X→[0,1], and no choice principle is used).

[F4]

Finite choice: a finite list indexed by a natural number, all of whose entries are nonempty sets admits a choice function for its family of values (Every natural-number-indexed list of nonempty sets has a choice function on its family of values).

[F5]

DMC: every serial relation on a nonempty set admits nonempty finite successor menus (Dependent multiple choice in finite-level tree form).

[L2]

A nonempty subset of N has a least element, and a specified set self-map with an initial state admits recursion on N (The well-ordering principle, The recursion theorem).

Proof

technique · direct
1.1givenF5

Assume DMC, let X be normal and let F,G be disjoint closed subsets of X.

2.1step 1.1F1L1

Under step 1.1, X∖G is open and F⊆X∖G; by [F1] fix open V0 with F⊆V0⊆V0‾⊆X∖G, and put V1:=X∖G.

3.1step 2.1L1

Under step 1.1, define a node of level n to be a tuple ⟨U1,…,U2n⟩ of open subsets of X such that Ui‾⊆Ui+1 for 1≤i<2n, F⊆U1, and U2n=X∖G. Let T be the set of such nodes over all n∈N, and let the relation S on T be: aSb when b is a node of level n+1 whose even entries are the entries of a, that is b2i=ai for 1≤i≤2n.

4.1step 2.1step 3.1F1F4L1

Under step 3.1, ⟨V1⟩=⟨X∖G⟩ is a node of level 0 in T: it has the single required entry, F⊆X∖G and its last entry is X∖G, so T≠∅. And S is serial on T: given a node a=⟨U1,…,U2n⟩, apply [F1] to the closed set F inside the open set U1 to obtain an open W0 with F⊆W0⊆W0‾⊆U1, for each 1≤i<2n apply [F1] to the closed set Ui‾ inside the open set Ui+1 to obtain an open Wi with Ui‾⊆Wi⊆Wi‾⊆Ui+1, and use [F4] to choose all of W0,…,W2n−1 at once (when n=0 one may take W0 to be the V0 of step 2.1, which has exactly the required inclusions); then b:=⟨W0,U1,W1,U2,…,W2n−1,U2n⟩ has 2n+1 entries, its even entries are b2i=Ui for 1≤i≤2n, and F⊆b1, bj‾⊆bj+1 for 1≤j<2n+1, b2n+1=X∖G by step 3.1, so b is a node of level n+1 with aSb.

5.1step 3.1step 4.1F5L1L2

Under step 4.1, apply [F5] to obtain finite nonempty menus Fn⊆T. They need not be level-aligned. Let k be the least level represented in F0 by [L2] and let H0 consist of its level-k members. Define Hn+1={b∈Fn+1:(∃a∈Hn) aSb}. This is a definable recursion (encode the index and subset in a set state, with a default for other states), licensed by [L2]. Induction shows that Hn is nonempty finite, every member has level k+n, and the Hn have both successor and predecessor properties: each retained a has a successor in the original next menu and that successor is retained. To recover levels below k, for a level-k node a and 0≤m≤k put πm(a)i=ai2k−m for 1≤i≤2m. Each projection is a level-m node: all entries contain F, the last is X∖G, and the closure inclusions follow by skipping along the original chain. Moreover πm+1(a)2i=πm(a)i and πk(a)=a. Now put Fm′={πm(a):a∈H0} for m<k, and Fm′=Hm−k for m≥k. These are nonempty finite menus of exactly level m, with both predecessor and successor properties, including the transition into level k. In particular F0′={⟨X∖G⟩}. No enumeration of infinitely many finite menus or branch choice is made.

6.1step 5.1L1

Under step 5.1, for n∈N and 1≤i≤2n put Un,i:=⋂{ ai:a∈Fn′ }, the intersection of the i-th entries of the finitely many menu elements; each Un,i is open, F⊆Un,1, Un,2n=X∖G, and Un,i‾⊆Un,i+1, because the intersection is contained in every entry, its closure is contained in every entry closure by monotonicity, and each such closure is contained in the corresponding next entry. Monotonicity follows directly from the smallest-closed-superset characterization in [L1].

7.1step 5.1step 6.1

Under step 6.1, Un+1,2i=Un,i for all n and 1≤i≤2n: every b∈Fn+1′ has b2i=ai for the predecessor a∈Fn′ with aSb, and every a∈Fn′ occurs as the predecessor of some b∈Fn+1′ by the successor property, so the two intersections have the same entries.

8.1step 6.1step 7.1F2F3

Under step 7.1 define a family on all dyadics by U0:=∅, U1:=X, and, for 0<r<1, Ur:=Un,i whenever r=i/2n with 1≤i<2n. This is well defined by step 7.1 and [F2]. If 0<r<s<1, choose N with r=i/2N, s=j/2N and i<j by [F2]; then Ur=UN,i, Us=UN,j, and Ur‾⊆Us by step 6.1. The same inclusion is automatic when r=0 or s=1. Thus (Ur)r∈D satisfies the hypotheses of [F3] literally.

9.1step 8.1F3

Under step 8.1, [F3] applies to the scale and gives the continuous f(x)=inf⁡({r∈D:x∈Ur}∪{1}):X→[0,1].

10.1step 6.1step 8.1step 9.1F2

Under step 9.1, F⊆f−1({0}): if a∈F and 0<r=i/2n<1, then a∈F⊆Un,1⊆Un,i=Ur, and also a∈U1=X. Hence {r∈D:a∈Ur}=D∖{0}, whose infimum is 0, so f(a)=0.

10.2step 6.1step 8.1step 9.1

Under step 9.1, G⊆f−1({1}): for b∈G, first b∉U0=∅. For r=i/2n∈D with 0<r<1 we have i≥1, i<2n and Un,i⊆Un,2n=X∖G, so b∉Ur; also b∈X=U1; hence {r∈D:b∈Ur}∪{1}={1} and f(b)=1.

11.1step 9.1step 10.1step 10.2F5∎

Under steps 9.1, 10.1 and 10.2 the map f is continuous with F⊆f−1({0}) and G⊆f−1({1}), which is Urysohn's lemma for the arbitrary normal space X and disjoint closed sets F,G; the only choice principle used was DMC.

Remarks

Depends on

Used by

Dependency tree · two levels

66 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