Cosets against a subgroup of least prime index
ProvedBFFPyramidal.lemma_IILemma II. Let be a finite group with a subgroup whose index is the least prime factor of , and let be any subgroup and . Then exactly one of two things happens: either the coset lies entirely inside the single coset , or meets every coset of , that is for all .
This dichotomy is what allows an induction along a chain of subgroups of least prime index: a coset of the partition either descends to the quotient or spreads over all of it, with nothing in between.
Formalization note. The conclusion is stated as a disjunction of the inclusion and the assertion that is nonempty for every . The hypothesis is G₁.index = Nat.minFac (Nat.card G). The two alternatives are not asserted to be mutually exclusive; only that at least one holds. Cosets are pointwise multiples of sets (open Pointwise).
import Mathlib open Pointwise
namespace BFFPyramidal
theorem lemma_II {G : Type*} [Group G] [Finite G] (G₁ : Subgroup G)
(hG₁ : G₁.index = Nat.minFac (Nat.card G)) (K : Subgroup G) (a : G) :
(a • (K : Set G)) ⊆ (a • (G₁ : Set G)) ∨
∀ b : G, ((a • (K : Set G)) ∩ (b • (G₁ : Set G))).Nonempty := by
sorry
end BFFPyramidal
Read-back
What the Lean code literally says, in plain math · claude-opus-5
Read-back of BFFPyramidal.lemma_II
Ambient data and how it is bound. The declaration fixes a type (an implicit argument, at arbitrary universe level), together with two instance-implicit typeclass assumptions: carries a group structure, and is finite. It then takes three explicit arguments, universally quantified in this order: a subgroup ; a hypothesis about (below); a subgroup ; and an element . No relation whatsoever is assumed between and , and no relation between and either subgroup. In particular is an arbitrary subgroup — it is not assumed proper, nontrivial, normal, of prime index, or comparable to .
The hypothesis on . The single hypothesis is the equation
where is the index of in (as a natural number), is the cardinality of as a natural number, and is the smallest prime factor function on , which is a total function with the conventions , , and for , is the least prime dividing . Since is a finite group, , so the branch is unreachable. Note that primality of the index is not stated as a hypothesis; it is only implied, and only when . In the boundary case where is the trivial group, and , so the hypothesis reads , i.e. . Nothing forces such a to exist: for groups having no subgroup whose index equals the least prime divisor of (for example a nonabelian simple group of even order, where but no index- subgroup exists), the hypothesis is unsatisfiable and the statement is vacuously true for that . For each such the assertion carries no content, though the statement is not vacuous across all .
The conclusion. The conclusion is an inclusive disjunction of two claims about the left translate , where and are first coerced to their underlying subsets of and the translation is the pointwise action of a group element on a set:
Here the inner universal quantifier over ranges over all of , and it sits inside the right-hand disjunct (so the statement is ", or else meets every left translate of ", not "for all , [disjunction]"). The second disjunct asserts nonemptiness of the intersection, i.e. existence of an element lying in both and ; it is an existence claim, not an equality or an inclusion. Set inclusion in the first disjunct is the (non-strict) subset relation.
What the two disjuncts amount to. Because left translation by is a bijection of , the first disjunct is equivalent to , so the element plays no role in it. Likewise, as ranges over all of , the sets range over exactly the left cosets of , and is equivalent to ; so the second disjunct says that meets every left coset of , equivalently that the product set is all of . Hence the whole conclusion is equivalent to the -free statement
and the explicitly quantified element is a logically inert binder: the truth value of the conclusion does not depend on it.
Degenerate and boundary cases included by the quantifiers. The trivial subgroup always satisfies the first disjunct, and always satisfies the second (taking the witness -coset representative in ), so those instances are automatic. Taking in the second disjunct is likewise automatic, since itself lies in ; the content of that disjunct comes only from cosets . When is trivial, both disjuncts hold trivially. No natural-number subtraction or division appears anywhere in the statement; the only -valued operations are the index, the cardinality, and , and the only arithmetic relation is an equality, not an inequality. Nothing is asserted about existence or uniqueness of , and no positivity or nontriviality side condition on , , or is imposed beyond finiteness of .
Confirmed by the mission captain (proposal self-audit).