Alphabeta Math
Pipeline-generated
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.

Decomposition Inertia and Frobenius — Examples

1 · Prerequisites

2 · Summary

Quadratic fields display all three (e,f,g) possibilities. Gaussian and Eisenstein integers make the exceptional ramified primes explicit; the fifth-root field and its real quadratic subfield give a concrete tower. The splitting field of T3−2 shows why a nonabelian base prime determines a conjugacy class.

Two counterexamples isolate the hypotheses: ramified residue data admits multiple Frobenius lifts, and the bad generator 5 at 2 has repeated reduction even though the field is unramified and inert. The integral generator (1+5)/2 recovers the correct cycle data.

3 · Logical flowchart

4 · Definitions, theorems and proofs

None yet.

5 · Examples, counterexamples and false statements

ExampleConstruction: AI-adaptedVerification: AI-adaptedjudge pass (gpt-5.6-terra)audited 2026-09-07Open item page →

Decomposition inertia in a quadratic field

Example

In a quadratic Galois extension of number fields, let G=C2. For any nonzero base prime the three possibilities are: split: (e,f,g)=(1,1,2), D=I=1, Frobenius identity; inert: (1,2,1), D=C2, I=1, Frobenius the nonidentity element; ramified: (2,1,1), D=I=C2, arithmetic Frobenius coset identity in D/I.

Facts & Assumptions

Given: The data and hypotheses of the statement.

[F1]

Galois prime decomposition efg: For a finite Galois extension L/K and nonzero prime p, every P above p has the same ramification index e and residue degree f. If there are g such primes, then efg=[L:K].

[F2]

Orders of decomposition and inertia groups: For finite Galois L/K and nonzero P∣p, writing e and f for its ramification index and residue degree, ∣D(P/p)∣=ef,∣I(P/p)∣=e,∣D(P/p)/I(P/p)∣=f. The prime P is unramified over p if and only if its inertia group is trivial.

[F3]

Frobenius order is residue degree: For finite Galois L/K and nonzero P∣p, the arithmetic Frobenius coset has order f(P/p) in D/I. If P is unramified, Frob⁡P has the same order in D.

Verification

1.1F1

Positive integers e,f,g with efg=2 have exactly the three displayed triples: the single factor 2 occurs in exactly one coordinate. These correspond respectively to split, inert, and ramified ideal factorizations.

2.1F2F3step 1.1∎

The formulas ∣D∣=ef and ∣I∣=e determine the subgroups, since C2 has only its identity subgroup and itself. For e=1 the Frobenius order is f, giving identity in the split case and the unique element of order two in the inert case. In the ramified case D/I has order f=1, so the residue coset is identity; there is no assertion of a unique lift.

ExampleConstruction: AI-adaptedVerification: AI-adaptedjudge pass (gpt-5.6-terra)audited 2026-09-07Open item page →

Gaussian and eisenstein frobenius

Example

In Q(i), an odd prime p splits if p≡1(mod4) and is inert if p≡3(mod4); arithmetic Frobenius sends i↦ip. The prime 2 ramifies. In Q(ζ3), a prime p≠3 splits if p≡1(mod3) and is inert if p≡2(mod3); arithmetic Frobenius sends ζ3↦ζ3p. The prime 3 ramifies.

Facts & Assumptions

Given: The data and hypotheses of the statement.

[F1]

Frobenius cycle type and prime splitting: Let F∈Z[T] be monic separable with splitting field L, and let p be a rational prime not dividing Disc⁡(F). Then p is unramified in L and Fˉ∈Fp[T] is squarefree. The degrees of its monic irreducible factors, with each distinct factor counted once, are exactly the cycle lengths of arithmetic Frobenius on the roots of F.

[F2]

Decomposition inertia in a quadratic field: In a quadratic Galois extension of number fields, let G=C2. For any nonzero base prime the three possibilities are: split: (e,f,g)=(1,1,2), D=I=1, Frobenius identity; inert: (1,2,1), D=C2, I=1, Frobenius the nonidentity element; ramified: (2,1,1), D=I=C2, arithmetic Frobenius coset identity in D/I.

[F3]

The multiplicative group Fq× of a finite field is cyclic: The multiplicative group F×=F∖{0} of every finite field F is cyclic.

[F4]

Integers in a quadratic field: For squarefree d≠1, OQ(d)=Z[(1+d)/2] if d≡1(mod4), and Z[d] otherwise.

Verification

1.1F4

The quadratic integral-basis theorem gives OQ(i)=Z[i] and OQ(ζ3)=Z[ζ3], since ζ3=(−1+−3)/2. The polynomials are T2+1 and T2+T+1, with discriminants -4 and -3. At the stated nonexceptional primes they have good reduction.

2.1F1F2F3step 1.1

For odd p, roots of T2+1 are elements of order four in Fp×. Cyclicity says they exist exactly when 4∣p−1. For p≠3, roots of T2+T+1 are elements of order three, since (T−1)(T2+T+1)=T3−1 and T=1 is not a root unless p=3. Cyclicity gives roots exactly when 3∣p−1. A quadratic without a root is irreducible. Thus the cycle types are two fixed points or a transposition, giving split or inert cases in the quadratic table.

2.2F1step 1.1

In good reduction the two roots are distinct. The Frobenius congruence sends each chosen root to its p-th power modulo P; that power is itself a root, so injectivity on the two root reductions forces equality in the number field. This gives both claimed formulas.

3.1step 1.1algebra∎

In the Gaussian ring, (1+i)2=2i and Z[i]/(1+i)=F2 by substituting i=-1. Thus (1+i) is prime and (2)=(1+i)2 as ideals. In the Eisenstein ring, (1−ζ3)2=−3ζ3 and the quotient by (1−ζ3) is F3 by substituting ζ3=1. Hence (3)=(1−ζ3)2 as ideals. Both exceptional primes ramify.

ExampleConstruction: AI-adaptedVerification: AI-adaptedjudge pass (gpt-5.6-terra)audited 2026-09-07Open item page →

Frobenius in a small cyclotomic field

Example

Let ζ=ζ5 be a primitive fifth root of unity and L=Q(ζ). Then Gal⁡(L/Q)≅(Z/5Z)×=C4 by σa(ζ)=ζa. For every prime p≠5, p is unramified and its arithmetic Frobenius is σp mod 5. In particular 2 is inert, with (e,f,g)=(1,4,1).

Facts & Assumptions

Given: The data and hypotheses of the statement.

[F1]

Frobenius cycle type and prime splitting: Let F∈Z[T] be monic separable with splitting field L, and let p be a rational prime not dividing Disc⁡(F). Then p is unramified in L and Fˉ∈Fp[T] is squarefree. The degrees of its monic irreducible factors, with each distinct factor counted once, are exactly the cycle lengths of arithmetic Frobenius on the roots of F.

[F2]

Frobenius order is residue degree: For finite Galois L/K and nonzero P∣p, the arithmetic Frobenius coset has order f(P/p) in D/I. If P is unramified, Frob⁡P has the same order in D.

[F3]

Galois prime decomposition efg: For a finite Galois extension L/K and nonzero prime p, every P above p has the same ramification index e and residue degree f. If there are g such primes, then efg=[L:K].

[F4]

Eisenstein criterion over the integers: Let f=anxn+⋯+a0∈Z[x] be primitive with n≥1. If there is a prime p such that p∤an,p∣ai for every i<n,p2∤a0, then f is irreducible in Q[x].

[F5]

The discriminant of a monic polynomial as the coefficient expression of Δn2: By prop-vandermonde-square-is-symmetric and thm-fundamental-theorem-of-symmetric-polynomials, there is a unique polynomial Dn∈Z[T1,…,Tn] such that Δn(x1,…,xn)2=Dn(e1,…,en). For a monic polynomial f(t)=tn+a1tn−1+⋯+an over a commutative ring, its discriminant is Disc⁡(f):=Dn(−a1,a2,…,(−1)nan). Equivalently, in any algebra in which f splits with roots α1,…,αn, this coefficient expression evaluates to Δn(α1,…,αn)2. The definition therefore depends only on the coefficients and not on a choice or ordering of roots. For a monic constant polynomial, Disc⁡(1)=1.

Verification

1.1F4

The polynomial Φ(T)=T4+T3+T2+T+1 satisfies Φ(T+1)=T4+5T3+10T2+10T+5, Eisenstein at 5. Translation preserves reducibility, so Phi is irreducible. Its four roots ζa for a=1,2,3,4 already lie in L. They give four automorphisms, with composition multiplying exponents modulo 5. The element 2 has successive powers 2,4,3,1, hence generates this group.

2.1F5step 1.1

At a root r of Phi, differentiating (T−1)Φ(T)=T5−1 gives Φ′(r)=5r4/(r−1). The product over its four roots is 54/5=53: the root product is 1, and ∏(r−1)=Φ(1)=5. Pairing opposite root differences shows ∏rΦ′(r)=(−1)6Disc⁡(Φ)=Disc⁡(Φ). Hence its discriminant is 53.

3.1F1F2F3step 1.1step 2.1∎

For p≠5 the good-reduction theorem gives unramifiedness and distinct root reductions. Frobenius sends the residue of zeta to its p-th power; since ζp is another root, distinctness forces Frob⁡(ζ)=ζp. At p=2 that automorphism has order four, so f=4; e=1 and efg=4 then give g=1, namely inertness.

ExampleConstruction: AI-adaptedVerification: AI-adaptedjudge pass (gpt-5.6-terra)audited 2026-09-07Open item page →

Nonabelian frobenius conjugacy class

Example

The splitting field of T3−2 over Q has Galois group S3. At p=5 its Frobenius conjugacy class consists of all three transpositions: distinct choices of prime can give distinct elements.

Facts & Assumptions

Given: The data and hypotheses of the statement.

[F1]

Frobenius cycle type and prime splitting: Let F∈Z[T] be monic separable with splitting field L, and let p be a rational prime not dividing Disc⁡(F). Then p is unramified in L and Fˉ∈Fp[T] is squarefree. The degrees of its monic irreducible factors, with each distinct factor counted once, are exactly the cycle lengths of arithmetic Frobenius on the roots of F.

[F2]

Frobenius elements above a prime are conjugate: In a finite Galois extension L/K let the nonzero prime p be unramified. If σP=P′ above p, then Frob⁡P′=σFrob⁡Pσ−1. Thus p determines one conjugacy class. If the Galois group is abelian, the element is independent of P.

[F3]

Eisenstein criterion over the integers: Let f=anxn+⋯+a0∈Z[x] be primitive with n≥1. If there is a prime p such that p∤an,p∣ai for every i<n,p2∤a0, then f is irreducible in Q[x].

Verification

1.1F3

Eisenstein at 2 proves T3−2 irreducible. Let a be its positive real root. The real cubic field Q(a) does not contain the nonreal cube root of unity ζ3. Its quadratic polynomial remains irreducible over that real field, so L=Q(a,ζ3) has degree six and contains all roots. The faithful permutation action on its three roots embeds its order-six Galois group into S3, hence is an isomorphism.

1.2algebra

For roots r of T3−2, the derivative is 3r2. Their product is 2, so ∏r3r2=27⋅4=108. Pairing differences contributes (−1)3, giving discriminant -108. Thus 5 is a good prime. Modulo 5, direct multiplication gives (T+2)(T2+3T+4)=T3−2; the quadratic discriminant is 9−16=3 modulo 5, which is not among the squares 0,1,4. Its degrees are therefore 1 and 2.

2.1F1F2step 1.1step 1.2∎

The Frobenius cycle theorem gives cycle type (1,2), a transposition. The Frobenius elements above 5 constitute its entire conjugacy class. Conjugating a transposition in S3 yields each of the three transpositions and no other permutation. Thus the class contains three distinct elements.

ExampleConstruction: AI-generatedVerification: AI-adaptedjudge pass (gpt-5.6-terra)audited 2026-09-07Open item page →

Decomposition groups in a tower

Example

Let M=Q(ζ5), L=Q(5), and K=Q. At p=2 there is a unique prime in each field above p. All ramification indices are one, and the residue degrees are four in M/K and two in each step. The decomposition sequence is 1⟶C2⟶C4⟶C2⟶1, and all inertia groups are trivial. The relative Frobenius for M/L is the square of the Frobenius for M/K.

Facts & Assumptions

Given: The data and hypotheses of the statement.

[F1]

Frobenius in a small cyclotomic field: Let ζ=ζ5 be a primitive fifth root of unity and L=Q(ζ). Then Gal⁡(L/Q)≅(Z/5Z)×=C4 by σa(ζ)=ζa. For every prime p≠5, p is unramified and its arithmetic Frobenius is σp mod 5. In particular 2 is inert, with (e,f,g)=(1,4,1).

[F2]

Decomposition and inertia in towers: Let M/L/K be a tower of number fields with M/K and L/K finite Galois, and fix nonzero primes Q∣P∣p. With H=Gal⁡(M/L), D(Q/P)=D(Q/p)∩H,I(Q/P)=I(Q/p)∩H. Restriction gives exact sequences 1→D(Q/P)→D(Q/p)→D(P/p)→1, 1→I(Q/P)→I(Q/p)→I(P/p)→1. The intersection identities also hold without L/K Galois; the displayed quotient assertions use that hypothesis.

[F3]

Frobenius compatibility in finite towers: Let M/L/K have M/K and L/K finite Galois, and let Q∣P∣p be nonzero primes with Q unramified over p. Then Frob⁡(Q/p)∣L=Frob⁡(P/p),Frob⁡(Q/P)=Frob⁡(Q/p)f(P/p). If M=L1L2 with both Li/K finite Galois and Q unramified over p, the Frobenius elements at its two contractions determine Frob⁡(Q/p) uniquely by restriction.

Verification

1.1F1algebra

Put t=ζ5+ζ5−1. Expanding and using 1+ζ5+⋯+ζ54=0 gives t2+t−1=0. This polynomial has discriminant 5, not a rational square, so Q(t)=Q(5) has degree two. Exponent 2 sends t to ζ52+ζ5−2=−1−t≠t, while exponent 4 fixes t. Thus restriction has kernel the order-two subgroup generated by exponent 4.

2.1F1F2step 1.1

At 2, M/K is unramified and inert with degree four, so there is one prime Q, D(Q/p)=C4 and I(Q/p)=1. Every prime of L above p has a prime of M above it by integral prime factorization, so it is the contraction P of Q. The tower exact sequences now give D(Q/P)=C2 and D(P/p)=C2 with all inertia groups trivial. In the unramified residue isomorphisms these group orders are the residue degrees, giving two in each step.

3.1F3step 1.1step 2.1∎

Compatibility says that the absolute exponent-2 Frobenius restricts to the nonidentity automorphism of L, and the relative Frobenius is its f(P/p)=2 power, namely exponent 4. This realizes the displayed exact sequence.

CounterexampleConstruction: AI-generatedVerification: AI-adaptedjudge pass (gpt-5.6-terra)audited 2026-09-07Open item page →

Ramified frobenius has no canonical lift

Statement refuted

The residue Frobenius congruence does not determine a unique element of D at a ramified prime. At 2 in Q(i), with P=(1+i), one has D=I=C2 and κ(P)=F2. Identity and complex conjugation are distinct lifts of the same arithmetic Frobenius coset in D/I.

Facts & Assumptions

Given: The data and hypotheses of the statement.

[F1]

Gaussian and eisenstein frobenius: In Q(i), an odd prime p splits if p≡1(mod4) and is inert if p≡3(mod4); arithmetic Frobenius sends i↦ip. The prime 2 ramifies. In Q(ζ3), a prime p≠3 splits if p≡1(mod3) and is inert if p≡2(mod3); arithmetic Frobenius sends ζ3↦ζ3p. The prime 3 ramifies.

[F2]

Arithmetic frobenius coset: For finite Galois L/K and nonzero P∣p, the arithmetic Frobenius coset is the unique element of D(P/p)/I(P/p) corresponding under the residue isomorphism to x↦xNp on κ(P), where Np=∣κ(p)∣. It is defined also when P is ramified. Its inverse is called geometric Frobenius. The quotient element is distinguished; a representative in D need not be unique.

Counterexample

1.1F1

The Gaussian calculation gives (2)=P2 and residue field F2. There is only one prime above 2, so both automorphisms stabilize it. Modulo P, i=-1=1, and conjugation also sends i to -i=1. As every integral element is a+bi, both automorphisms act identically on every residue.

2.1F2step 1.1∎

Thus D=I=C2 and D/I is trivial. The arithmetic map on F2 is x squared, which is identity on its two elements 0 and 1. Both identity and conjugation satisfy its congruence, but they differ on i in the number field. This refutes uniqueness from residue data; it does not preclude an additional external convention from selecting a representative.

CounterexampleConstruction: AI-adaptedVerification: AI-adaptedjudge pass (gpt-5.6-terra)audited 2026-09-07Open item page →

Frobenius cycle type needs good reduction

Statement refuted

Factor multiplicities from an arbitrary integral generator need not encode Frobenius cycles. For F(T)=T2−5 at p=2, Fˉ=(T+1)2, yet Q(5) is unramified and inert at 2, with Frobenius a transposition. The integral generator ω=(1+5)/2 has minimal polynomial G(T)=T2−T−1, whose reduction T2+T+1 is irreducible over F2.

Facts & Assumptions

Given: The data and hypotheses of the statement.

[F1]

Frobenius cycle type and prime splitting: Let F∈Z[T] be monic separable with splitting field L, and let p be a rational prime not dividing Disc⁡(F). Then p is unramified in L and Fˉ∈Fp[T] is squarefree. The degrees of its monic irreducible factors, with each distinct factor counted once, are exactly the cycle lengths of arithmetic Frobenius on the roots of F.

[F2]

Dedekind--Kummer prime factorisation: Let L/K be a finite extension of number fields, let α∈OL with OL=OK[α], and let F∈OK[X] be its monic minimal polynomial over K. Let p be a nonzero prime ideal of OK, not dividing the index of this power order (the index is 1 under the stated monogeneity hypothesis). If Fˉ=∏igˉiei with distinct monic irreducibles over OK/p, then pOL=∏iPiei,Pi=(p,gi(α)),f(Pi/p)=deg⁡gˉi, Here gi∈OK[X] are any monic lifts of gˉi, and the last f denotes the residue degree from def-prime-above-and-residue-degree.

[F3]

Ramification is detected by the number-field discriminant: A rational prime p ramifies in K/Q if and only if p∣dK.

[F4]

Integers in a quadratic field: For squarefree d≠1, OQ(d)=Z[(1+d)/2] if d≡1(mod4), and Z[d] otherwise.

[F5]

Power-basis and polynomial discriminants: Let K=Q(α). If f is the degree-n monic minimal polynomial of α, then disc⁡(1,α,…,αn−1)=(−1)n(n−1)/2NK/Q(f′(α))=disc⁡(f).

Counterexample

1.1F3F4F5

The quadratic integral-basis theorem gives OL=Z[ω]. Direct substitution gives G(omega)=0, and its discriminant 5 is not a rational square, so it is the minimal polynomial. The power-basis discriminant formula gives dL=Disc⁡(G)=5. Therefore 2 is unramified by the field-discriminant criterion.

2.1F1F2step 1.1

Modulo 2, G is T2+T+1, taking value 1 at both 0 and 1. It is irreducible. Dedekind-Kummer applies to the full ring OL=Z[ω] and gives a single prime of e=1 and f=2. Alternatively the good-reduction cycle theorem for G gives the transposition Frobenius.

3.1F1F2step 1.1step 2.1∎

For the other generator, 5=2ω−1, so Z[5] has index 2 in OL, as the change-of-basis matrix has determinant 2. Its polynomial discriminant is 20 and its reduction is (T+1)2. The two characteristic-zero roots reduce to the same root, so this reduction is not a bijection of root sets. It cannot supply the cycle comparison; in particular reading its multiplicity as ramification would contradict e=1. The full-ring hypothesis of the cited Dedekind-Kummer statement fails for this generator.

Sources