Alphabeta Math
TheoremStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedjudge pass (gpt-5.6-terra)audited 2026-09-14
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.

Generic factorization and ccc preservation for two-step iterations

Statement

In ZFC, a P∗Q˙-generic K factors as a P-generic G and Q˙G-generic H with M[K]=M[G][H], and conversely G∗H is generic. If P is ccc and P⊩Q˙ is ccc, then P∗Q˙ is ccc.

Facts & Assumptions

Given: AC, a two-step iteration in a transitive ground model, and the stated chain hypotheses for the last clause.

[F1]

Two-step forcing iterations fixes the set-sized order and, under AC, supplies a bounded-carrier name forced equal below any condition to an arbitrary local name for a member of Q˙.

[F2]

Forcing theorem supplies name evaluation and the truth lemma; Forcing relation for all formulas supplies the existential and negation clauses, Atomic forcing relation supplies atomic membership, and Monotonicity, density, and decision for forcing supplies dense decisions and density closure.

[F4]

Forcing preserves ordinals says each ordinal of a generic extension is a ground ordinal.

Proof

1.1F1F2

From K put G={p:∃q˙∈R (p,q˙)∈K} and H={q˙G:∃p∈G (p,q˙)∈K}. Preimages of ground dense subsets of P are dense in the iteration, so G is generic. If D˙G is dense in Q˙G, the forcing theorem produces below each pair a name for a second-coordinate extension in D˙; [F1] replaces that local name by a forced-equal name in the set carrier R. The resulting pairs are dense, and meeting them proves H generic over M[G].

1.2F1F2

Conversely, for generics G,H take the filter generated by pairs (p,q˙) in the restricted iteration with p∈G and q˙G∈H. Every quotient condition has a name occurring in Q˙, or a locally forced-equal representative in R by [F1], so the same dense-set translation meets every ground dense subset of P∗Q˙. Recursive evaluation of names first by G and then by H proves M[K]=M[G][H].

1.3F1F2

Suppose A={(pα,q˙α):α<ω1M} were an antichain and put I˙={(αˇ,pα):α<ω1M}. We prove syntactically that 1P forces the map α↦q˙α from I˙ into Q˙ to have pairwise incompatible, hence distinct, values. Otherwise some s forces distinct α,β∈I˙ and a common Q˙-extension of their values. The forcing membership clause lets us strengthen below pα and pβ; the existential clause of [F2] then supplies a further condition t and a name r˙ with t⊩r˙≤Q˙q˙α,q˙β. By [F1], replace r˙ below t by a forced-equal ρ∈R. Then (t,ρ) belongs to the restricted iteration and extends both members of A, a contradiction. Thus 1P⊩I˙ injects into a Q˙-antichain. Since 1P⊩Q˙ is ccc, 1P⊩I˙ is countable.

Because P is ccc, [F3] preserves the ground regular ω1M. Hence 1P⊩I˙ is bounded below ωˇ1M. The forcing existential clause supplies, densely below any condition, a name for such a bound; F4 says that bound is a ground ordinal, and the check-name membership clause then densely decides the name equal to some ground βˇ with β<ω1M. Thus D={p∈P:∃β<ω1M p⊩I˙⊆βˇ} is dense. Choose a maximal antichain C⊆D; it is countable by ccc of P. For each c∈C choose a ground bound βc, and put β∗=sup⁡c∈C(βc+1)<ω1M. Predensity of C and the forcing negation clause give 1P⊩I˙⊆βˇ∗. But the definition of I˙ gives pα⊩αˇ∈I˙ for every α<ω1M, contradicting the bound when α≥β∗. Therefore P∗Q˙ is ccc. AC supplies the antichain indexing, the maximal antichain C, and its bound choices; no external generic is used. [F1, F2, F3, F4] ∎

Depends on

Used by

Dependency tree · two levels

23 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