Alphabeta Math
TheoremStatement: AI-adaptedProof: AI-adaptedSession-authored (Fable 5 assisted)precheck passverified 2026-07-26 (claude-opus-5)
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.

Transfinite recursion

Statement

Let (W,<)(W, <) be a well-order (Well-order and well-ordered set) and let GG be a class function: a rule, given by a formula in the language of set theory, that assigns a set G(h)G(h) to every function hh whose domain is a proper initial segment of WW (Initial segment of a well-order). Then there is exactly one function FF with domain WW such that

F(a)=G(FW<a)for every aW.F(a) = G(F \restriction W_{<a}) \qquad \text{for every } a \in W.

Here FW<aF \restriction W_{<a} is the restriction of FF to the initial segment determined by aa, so the value of FF at aa is prescribed in terms of all its earlier values at once.

Because GG is a class function rather than a set, this is a theorem schema of ZF: one theorem for each formula defining GG. It uses Replacement, and it uses no form of the Axiom of Choice.

Facts & Assumptions

Given: A well-order (W,<)(W, <), a class function GG as in the statement, and the axioms of ZF. Call a function hh an attempt when dom(h)\mathrm{dom}(h) is an initial segment of WW and h(b)=G(hW<b)h(b) = G(h \restriction W_{<b}) for every bdom(h)b \in \mathrm{dom}(h). The empty function is an attempt, with domain the initial segment \emptyset.

[A1]

G(h)G(h) is a well-determined set for every function hh whose domain is a proper initial segment of WW.

[A2]

The axioms of ZF are available, in particular Separation, Union and Replacement. No choice principle is assumed.

[L1]

Transfinite induction on (W,<)(W, <): if W<aSW_{<a} \subseteq S implies aSa \in S for every aWa \in W, then S=WS = W (Transfinite induction).

[L2]

W<a={xW:x<a}W_{<a} = \{x \in W : x < a\} and Wa=W<a{a}W_{\le a} = W_{<a} \cup \{a\} are initial segments, an intersection of initial segments is an initial segment, and (W<a)<b=W<b(W_{<a})_{<b} = W_{<b} for b<ab < a (Initial segment of a well-order).

[L3]

The order of WW is total and every nonempty subset of WW has a least element (Well-order and well-ordered set).

Proof

technique · direct
1.1

Let hh and kk be attempts and put D=dom(h)dom(k)D = \mathrm{dom}(h) \cap \mathrm{dom}(k); then DD is an initial segment of WW, being an intersection of two downward closed sets.

L2construct
2.1

Put S={aW:aD or h(a)=k(a)}S = \{a \in W : a \notin D \text{ or } h(a) = k(a)\} and suppose W<aSW_{<a} \subseteq S for some aDa \in D; since DD is downward closed, W<aDW_{<a} \subseteq D, so hh and kk agree at every x<ax < a, that is hW<a=kW<ah \restriction W_{<a} = k \restriction W_{<a}, and therefore h(a)=G(hW<a)=G(kW<a)=k(a)h(a) = G(h \restriction W_{<a}) = G(k \restriction W_{<a}) = k(a), giving aSa \in S.

step 1.1A1L2
3.1

If instead aDa \notin D then aSa \in S outright, so for every aWa \in W the implication "W<aSW_{<a} \subseteq S implies aSa \in S" holds, and transfinite induction gives S=WS = W: any two attempts agree on the intersection of their domains.

step 2.1L1
4.1

In particular, for each aWa \in W there is at most one attempt with domain W<aW_{<a}, and at most one attempt with domain WW, since two of them would have that same set as the intersection of their domains.

step 3.1
5.1

Let T={aW:there is an attempt with domain W<a}T = \{a \in W : \text{there is an attempt with domain } W_{<a}\}, which is a set by Separation applied to WW, and note that the attempt in question is unique when it exists.

step 4.1A2construct
6.1

Suppose W<aTW_{<a} \subseteq T; for each b<ab < a let hbh_b be the unique attempt with domain W<bW_{<b} and put hb+=hb{(b,G(hb))}h_b^{+} = h_b \cup \{(b, G(h_b))\}, a function with domain WbW_{\le b} which is again an attempt, because it agrees with hbh_b below bb and hb+(b)=G(hb)=G(hb+W<b)h_b^{+}(b) = G(h_b) = G(h_b^{+} \restriction W_{<b}).

step 5.1step 4.1A1L2construct
7.1

The assignment bhb+b \mapsto h_b^{+} is given by a formula, so Replacement makes {hb+:b<a}\{h_b^{+} : b < a\} a set; let uu be the union of that set.

step 6.1A2construct
8.1

uu is a function: any two of the hb+h_b^{+} are attempts and therefore agree on the intersection of their domains, so no argument receives two different values.

step 7.1step 3.1
8.2

dom(u)\mathrm{dom}(u) is the union of the sets WbW_{\le b} for b<ab < a, which equals W<aW_{<a}: each such WbW_{\le b} is contained in W<aW_{<a} by transitivity of the order, and each x<ax < a lies in WxW_{\le x}.

step 7.1step 6.1L2L3
9.1

uu is an attempt with domain W<aW_{<a}: for c<ac < a we have uW<c=hc+W<c=hcu \restriction W_{<c} = h_c^{+} \restriction W_{<c} = h_c and u(c)=hc+(c)=G(hc)u(c) = h_c^{+}(c) = G(h_c), so u(c)=G(uW<c)u(c) = G(u \restriction W_{<c}).

step 8.1step 8.2step 6.1
10.1

Hence aTa \in T whenever W<aTW_{<a} \subseteq T, so T=WT = W by transfinite induction: for every aWa \in W there is a unique attempt hah_a with domain W<aW_{<a}.

step 9.1step 5.1step 4.1L1
11.1

Replacement applied to the formula aha+=ha{(a,G(ha))}a \mapsto h_a^{+} = h_a \cup \{(a, G(h_a))\} makes {ha+:aW}\{h_a^{+} : a \in W\} a set; let FF be its union, which by the arguments of steps 8.1 and 8.2 is a function whose domain is the union of the sets WaW_{\le a} for aWa \in W, namely WW itself.

step 10.1step 8.1step 8.2A2construct
12.1

FF satisfies the recursion: for aWa \in W we get FW<a=haF \restriction W_{<a} = h_a and F(a)=ha+(a)=G(ha)F(a) = h_a^{+}(a) = G(h_a), so F(a)=G(FW<a)F(a) = G(F \restriction W_{<a}), and in particular FF is an attempt with domain WW.

step 11.1step 9.1
13.1

Uniqueness: any function FF' with domain WW satisfying F(a)=G(FW<a)F'(a) = G(F' \restriction W_{<a}) for all aa is by definition an attempt with domain WW, hence equals FF.

step 12.1step 4.1
14.1

There is therefore exactly one function FF with domain WW and F(a)=G(FW<a)F(a) = G(F \restriction W_{<a}) for every aWa \in W.

step 12.1step 13.1

Remarks

Which axioms are spent. Separation produces TT in step 5.1; Replacement produces the sets of attempts in steps 7.1 and 11.1; Union assembles them. Replacement is genuinely needed and cannot be traded away: the attempts hah_a are not subsets of any single set given in advance, so Separation alone cannot collect them. This is the standard reason transfinite recursion is unavailable in Zermelo set theory without Replacement.

No choice is used, and the reason is uniqueness. At every stage the object to be used is the unique attempt with a given domain, singled out by step 4.1, never one of many candidates. A construction phrased as "at stage aa pick some element not yet used" is a different theorem and does cost a choice principle; the difference is exactly the content of FALSE: transfinite induction and recursion need the Axiom of Choice.

Class function, not a set function. GG has to be allowed to take values anywhere in the universe, since F(a)F(a) may be an arbitrary set, so GG cannot be assumed to be a set of ordered pairs. That is why the result is a schema. When GG happens to be an honest set function, defined on the set of all functions from proper initial segments of WW into a fixed set AA and taking values in AA, the statement is an ordinary theorem of ZF and the same proof applies verbatim.

Relation to ordinary recursion. Taking W=NW = \mathbb{N} recovers the familiar recursion on the natural numbers, but not literally the published The recursion theorem: that theorem prescribes g(σ(n))g(\sigma(n)) in terms of g(n)g(n) alone, whereas GG here sees the whole earlier history FW<aF \restriction W_{<a}. Course-of-values recursion on N\mathbb{N} is the special case of the present theorem, and it is what later pages use whenever a construction at stage nn consults all earlier stages.

Depends on

Used by

Dependency tree · next 3 levels

Direct dependencies and their dependencies through the next three levels: 20 results over 9 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