Norm expansion N(1+γ)=1+Tr γ+Nγ+Tr δ in prime degree
Provedprod_one_add_smul_eq_one_add_finsum_add_finprod_add_finsum_smul_of_prime_cardLet be a commutative ring and a finite group acting on by ring automorphisms (a MulSemiringAction), and suppose that the cardinality of is prime. Then for every there exists an element lying in the ideal of generated by the set of all products with distinct, such that the identity
holds in , the products and sums being the finprod and finsum of the indicated families over all of (which reduce to ordinary finite products and sums since is finite). Thus the product of over equals , plus the trace of , plus the norm of , plus the trace of an element of the ideal generated by the products of two distinct conjugates of . No invertibility or division is assumed, so the statement covers both the tame and the wild case.
This is the standard expansion of the norm in a cyclic layer of prime degree, with the error term controlled by products of two distinct conjugates of ; the primality of enters because translation permutes freely the subsets of of intermediate size, so that their contributions assemble into a trace. It is used in the ramification-theoretic estimates for discrete valuation rings with a prime-order group action, namely IsDiscreteValuationRing.exists_finset_card_le_forall_exists_sub_mul_finprod_smul_mem_pow_of_prime_card, IsDiscreteValuationRing.exists_sub_one_mem_and_finprod_smul_sub_mem_of_jump_lt_of_prime_card and IsDiscreteValuationRing.finprod_smul_sub_one_mem_maximalIdeal_pow_of_sub_one_mem_pow_herbrand_of_prime_card.
import Mathlib set_option maxHeartbeats 4000000 set_option synthInstance.maxHeartbeats 400000 set_option backward.isDefEq.respectTransparency.types false set_option autoImplicit false
theorem prod_one_add_smul_eq_one_add_finsum_add_finprod_add_finsum_smul_of_prime_card
{B : Type*} [CommRing B] {G : Type*} [Group G] [Finite G] [MulSemiringAction G B]
(hℓ : (Nat.card G).Prime) (γ : B) :
∃ δ ∈ Ideal.span {x : B | ∃ σ₁ σ₂ : G, σ₁ ≠ σ₂ ∧ x = (σ₁ • γ) * (σ₂ • γ)},
∏ᶠ σ : G, (1 + σ • γ) = 1 + ∑ᶠ σ : G, σ • γ + ∏ᶠ σ : G, σ • γ + ∑ᶠ σ : G, σ • δ := by sorry