riemann

The Lean code claims a proof of the Riemann Hypothesis. This site has not checked the claim yet. See what is checked.

Algebra

The algebra that the rest of the proof assumes: congruences, groups and their characters, modules and their length, fields, finite fields and Galois theory, matrices and minors, Hermitian forms, tensor products and coinvariants.

The other chapters assume this algebra. The window bound VWin measures modules over ℤℓ\mathbb{Z}_\ell by their length and cuts fields out of ℚ(ζN)\mathbb{Q}(\zeta_N) by congruences, and the theta branch works with Hermitian and skew-Hermitian forms and their Gram matrices. Here a ring is commutative with 11, except a ring of matrices.

Integers and congruences

Definition. Let a,b∈ℤa, b \in \mathbb{Z}. Then aa divides bb, written a∣ba \mid b, if b=acb = ac for some c∈ℤc \in \mathbb{Z}. A prime is an integer p≥2p \ge 2 whose only positive divisors are 11 and pp.

Definition. Let a,b∈ℤa, b \in \mathbb{Z}, not both 00. The greatest common divisor gcd⁡(a,b)\gcd(a, b) is the largest integer dd with d∣ad \mid a and d∣bd \mid b.

Theorem (Bézout identity). Let a,b∈ℤa, b \in \mathbb{Z}, not both 00. There are x,y∈ℤx, y \in \mathbb{Z} with ax+by=gcd⁡(a,b)ax + by = \gcd(a, b).

Definition. Fix N≥1N \ge 1. Integers a,ba, b are congruent modulo NN, a≡b(modN)a \equiv b \pmod N, if N∣a−bN \mid a - b; the classes form the ring ℤ/Nℤ\mathbb{Z}/N\mathbb{Z}. Its unit group (ℤ/Nℤ)×(\mathbb{Z}/N\mathbb{Z})^\times is the set of classes aa with gcd⁡(a,N)=1\gcd(a, N) = 1, which by Bézout are the invertible classes. Its order is Euler's function φ(N)\varphi(N).

VWin takes primes q≡3(mod8)q \equiv 3 \pmod 8 and d≡7(mod16)d \equiv 7 \pmod{16} with q∣d+1q \mid d + 1, and cuts its fields out by classes a≡±1(mod2e+2)a \equiv \pm 1 \pmod{2^{e+2}} in (ℤ/Nℤ)×(\mathbb{Z}/N\mathbb{Z})^\times.

Groups

Definition. A group homomorphism is a map f:G→Hf : G \to H of groups with f(xy)=f(x)f(y)f(xy) = f(x) f(y). Its kernel ker⁡f={x∈G:f(x)=1}\ker f = \{x \in G : f(x) = 1\} is a normal subgroup of GG.

Theorem (first isomorphism theorem). Let f:G→Hf : G \to H be a group homomorphism. Then xker⁡f↦f(x)x \ker f \mapsto f(x) is an isomorphism G/ker⁡f→f(G)G / \ker f \to f(G).

Definition. Let XX be a set with a structure (a group, a field, a vector space with a form). The automorphism group Aut(X)\mathrm{Aut}(X) is the group, under composition, of bijections X→XX \to X that preserve the structure and whose inverses preserve it.

Definition. A group GG is cyclic if G={gk:k∈ℤ}G = \{g^k : k \in \mathbb{Z}\} for some g∈Gg \in G, a generator. Each cyclic group is isomorphic to ℤ\mathbb{Z} or to ℤ/nℤ\mathbb{Z}/n\mathbb{Z} for one n≥1n \ge 1.

Definition. Let AA be a finite abelian group. A character of an abelian group AA is a homomorphism χ:A→ℂ×\chi : A \to \mathbb{C}^\times. The characters form a group A^\widehat{A} under pointwise product, with unit the trivial character 11.

Proposition. Let AA be a finite abelian group. Then A^≅A\widehat{A} \cong A (not canonically), and for χ∈A^\chi \in \widehat{A}, ∑a∈Aχ(a)={∣A∣if χ=1,0otherwise.\sum_{a \in A} \chi(a) = \begin{cases} |A| & \text{if } \chi = 1, \\ 0 & \text{otherwise.} \end{cases}

Example. A character of ℤ/nℤ\mathbb{Z}/n\mathbb{Z} is determined by χ(1)\chi(1), which can be any nn-th root of unity in ℂ\mathbb{C}. A Dirichlet character modulo NN is a character of (ℤ/Nℤ)×(\mathbb{Z}/N\mathbb{Z})^\times (Chapter 6).

GeoOne composes a surjection J→ℤ/ℓmJ \to \mathbb{Z}/\ell^m with a nontrivial character of ℤ/ℓm\mathbb{Z}/\ell^m.

Multisets

Definition. A multiset on a set SS is a function m:S→ℤ≥0m : S \to \mathbb{Z}_{\ge 0} with finite support. The value m(s)m(s) is the multiplicity of ss. For example, {0,0,1}\{0, 0, 1\} is the multiset with m(0)=2m(0) = 2, m(1)=1m(1) = 1.

In Row0b the Hodge-Tate weights HT(τ)HT(\tau) at each embedding τ\tau form a multiset of integers.

Rings, ideals and modules

Definition. An ideal of a ring RR is an additive subgroup I⊆RI \subseteq R with rI⊆IrI \subseteq I for each r∈Rr \in R. The quotient ring R/IR/I is the set of cosets r+Ir + I with the induced operations. The ideal II is prime if R/IR/I is an integral domain, and maximal if R/IR/I is a field.

Definition. A module over RR is an abelian group MM with a bilinear map R×M→MR \times M \to M, (r,m)↦rm(r, m) \mapsto rm, such that 1m=m1m = m and (rs)m=r(sm)(rs)m = r(sm). A ℤ\mathbb{Z}-module is an abelian group.

Definition. Let RR be an integral domain and MM an RR-module. An element m∈Mm \in M is torsion if rm=0rm = 0 for some r≠0r \ne 0. The torsion elements form the torsion subgroup MtorsM_{\mathrm{tors}}, and MM is torsion free if Mtors=0M_{\mathrm{tors}} = 0. A group GG has no torsion if its only element of finite order is 11.

Definition. A module MM is simple if M≠0M \ne 0 and its only submodules are 00 and MM. A composition series of MM is a chain 0=M0⊊M1⊊⋯⊊Mn=M0 = M_0 \subsetneq M_1 \subsetneq \dots \subsetneq M_n = M of submodules with each Mi/Mi−1M_i / M_{i-1} simple.

Theorem (Jordan-Hölder). Any two composition series of MM have the same number nn of steps, and the same quotients up to isomorphism and order.

Definition. The length of a module MM, written len⁡RM\operatorname{len}_R M, is the number nn of steps of a composition series, and ∞\infty if MM has none. For a submodule N⊆MN \subseteq M, len⁡M=len⁡N+len⁡M/N\operatorname{len} M = \operatorname{len} N + \operatorname{len} M/N.

Fix a prime ℓ\ell. Chapter 4 defines the ℓ\ell-adic integers ℤℓ\mathbb{Z}_\ell and their fraction field ℚℓ\mathbb{Q}_\ell; here it is enough that ℤℓ\mathbb{Z}_\ell is a principal ideal domain with nonzero ideals ℓkℤℓ\ell^k \mathbb{Z}_\ell and ℤℓ/ℓℤℓ=𝔽ℓ\mathbb{Z}_\ell / \ell \mathbb{Z}_\ell = \mathbb{F}_\ell.

Proposition (ℤℓ\mathbb{Z}_\ell-modules). The only simple ℤℓ\mathbb{Z}_\ell-module is 𝔽ℓ\mathbb{F}_\ell. So a ℤℓ\mathbb{Z}_\ell-module MM of finite length is finite, and #M=ℓlen⁡M\#M = \ell^{\operatorname{len} M}.

Example. len⁡ℤℓ/ℓ3=3\operatorname{len} \mathbb{Z}_\ell/\ell^3 = 3. The module ℚℓ/ℤℓ\mathbb{Q}_\ell / \mathbb{Z}_\ell has infinite length, because it contains the strictly increasing chain ℓ−kℤℓ/ℤℓ\ell^{-k}\mathbb{Z}_\ell / \mathbb{Z}_\ell, k≥0k \ge 0.

VWin bounds ∑ℓlog⁡ℓ⋅len⁡ℤℓШ\sum_\ell \log \ell \cdot \operatorname{len}_{\mathbb{Z}_\ell} \text{Ш} for Tate-Shafarevich groups with coefficients in twists of ℚℓ/ℤℓ\mathbb{Q}_\ell/\mathbb{Z}_\ell, a sum equal to log⁡#Ш\log \# \text{Ш} when Ш is finite. GeoOne asks for an open compact subgroup with no torsion.

Polynomials and power series

Definition. The polynomial ring R[X]R[X] is the set of finite sums ∑iaiXi\sum_i a_i X^i with ai∈Ra_i \in R. The degree of f≠0f \ne 0 is the largest ii with ai≠0a_i \ne 0, and ff is monic if this aia_i is 11. Over a field kk, a nonconstant polynomial is irreducible if it is not a product of two polynomials of smaller degree.

Definition. The ring of formal power series R[[X1,…,Xd]]R[[X_1, \dots, X_d]] is the set of all formal sums ∑αaαXα\sum_\alpha a_\alpha X^\alpha over multi-indices α∈ℤ≥0d\alpha \in \mathbb{Z}_{\ge 0}^d, multiplied as polynomials. No convergence is asked.

The prime labels of Green are monic irreducible polynomials over ℤ/g\mathbb{Z}/g with nonzero constant term, and a formal group law in CM divisibility is a dd-tuple of power series in 2d2d variables.

Fields and extensions

Definition. A field extension L/KL/K is a field LL that contains the field KK. Its degree [L:K][L : K] is the dimension of LL as a KK-vector space. For K⊆L⊆MK \subseteq L \subseteq M, the tower law is [M:K]=[M:L][L:K][M : K] = [M : L][L : K].

Definition. Let L/KL/K be an extension. An element α∈L\alpha \in L is algebraic over KK if f(α)=0f(\alpha) = 0 for some nonzero f∈K[X]f \in K[X]. Its minimal polynomial is the monic f∈K[X]f \in K[X] of least degree with f(α)=0f(\alpha) = 0. It is irreducible, it divides each g∈K[X]g \in K[X] with g(α)=0g(\alpha) = 0, and [K(α):K]=deg⁡f[K(\alpha) : K] = \deg f.

Definition. A field Ω\Omega is algebraically closed if each nonconstant polynomial in Ω[X]\Omega[X] has a root in Ω\Omega. An algebraic closure of KK is an algebraically closed extension Ω/K\Omega/K in which each element is algebraic over KK.

Theorem (Steinitz). Each field KK has an algebraic closure K‾\overline{K}, unique up to an isomorphism that is the identity on KK.

Finite fields

Theorem. Let pp be a prime and q=pdq = p^d with d≥1d \ge 1. There is a field 𝔽q\mathbb{F}_q with qq elements, unique up to isomorphism: the set of roots of Xq−XX^q - X in 𝔽‾p\overline{\mathbb{F}}_p. Each finite field is one of these. The subfields of 𝔽pn\mathbb{F}_{p^n} are the 𝔽pd\mathbb{F}_{p^d} with d∣nd \mid n, and 𝔽q×\mathbb{F}_q^\times is cyclic of order q−1q - 1.

Definition. Let qq be a power of pp and d≥1d \ge 1. The Frobenius automorphism of 𝔽qd\mathbb{F}_{q^d} over 𝔽q\mathbb{F}_q is ϕ(x)=xq\phi(x) = x^q. It is a field automorphism that fixes exactly 𝔽q\mathbb{F}_q.

Proposition. The automorphisms of 𝔽qd\mathbb{F}_{q^d} that fix 𝔽q\mathbb{F}_q form a cyclic group of order dd, generated by ϕ\phi. For x∈𝔽qdx \in \mathbb{F}_{q^d}, the Frobenius conjugates x,xq,…,xqd−1x, x^q, \dots, x^{q^{d-1}} are the roots of its minimal polynomial over 𝔽q\mathbb{F}_q, each repeated d/deg⁡d / \deg times. They are distinct if and only if 𝔽q(x)=𝔽qd\mathbb{F}_q(x) = \mathbb{F}_{q^d}.

Green labels the cuspidal characters of GLd(𝔽g)\mathrm{GL}_d(\mathbb{F}_g) by these polynomials of degree dd, and its value formula sums over the Frobenius conjugates zgiz^{g^i}; GreenOfRegular takes the minimal polynomial of an xx with xgd=xx^{g^d} = x.

Roots of unity

Definition. Let KK be a field and n≥1n \ge 1. An nn-th root of unity in KK is an element ζ\zeta with ζn=1\zeta^n = 1. They form a finite cyclic group μn(K)\mu_n(K). A primitive root of unity of order nn is a ζ\zeta whose order in K×K^\times is exactly nn.

Example. In ℂ\mathbb{C}, the primitive nn-th roots of unity are e2πik/ne^{2\pi i k/n} with gcd⁡(k,n)=1\gcd(k, n) = 1.

Lemma K of the window bound needs a field that contains a primitive root of unity of order ll.

Galois theory

Definition. Let L/KL/K be a finite extension, and let Aut(L/K)\mathrm{Aut}(L/K) be the group of field automorphisms of LL that fix KK. Then ∣Aut(L/K)∣≤[L:K]|\mathrm{Aut}(L/K)| \le [L : K], and L/KL/K is a Galois extension if equality holds; then Gal(L/K)=Aut(L/K)\mathrm{Gal}(L/K) = \mathrm{Aut}(L/K) is its Galois group. For a subgroup HH, the fixed field is LH={x∈L:h(x)=x for each h∈H}L^H = \{x \in L : h(x) = x \text{ for each } h \in H\}.

Theorem (fundamental theorem of Galois theory). Let L/KL/K be a finite Galois extension with group GG. The maps H↦LHH \mapsto L^H and M↦Gal(L/M)M \mapsto \mathrm{Gal}(L/M) are inverse bijections between the subgroups of GG and the fields MM with K⊆M⊆LK \subseteq M \subseteq L. The extension LH/KL^H/K is Galois if and only if HH is normal in GG, and then Gal(LH/K)≅G/H\mathrm{Gal}(L^H/K) \cong G/H.

By Chapter 4, Gal(ℚ(ζN)/ℚ)≅(ℤ/Nℤ)×\mathrm{Gal}(\mathbb{Q}(\zeta_N)/\mathbb{Q}) \cong (\mathbb{Z}/N\mathbb{Z})^\times, with aa acting by ζ↦ζa\zeta \mapsto \zeta^a. The field EE of IsChiField in VWin is the fixed field of the classes aa with a≡±1(mod2e+2)a \equiv \pm 1 \pmod{2^{e+2}}, ψ(a)=1\psi(a) = 1 and two Jacobi symbols equal to 11.

Matrices

Let RR be a ring. Mm×n(R)M_{m \times n}(R) is the set of m×nm \times n matrices, and Mn(R)=Mn×n(R)M_n(R) = M_{n \times n}(R). The transpose of A=(aij)A = (a_{ij}) is AT=(aji)A^T = (a_{ji}), and (AB)T=BTAT(AB)^T = B^T A^T.

Definition. The determinant of A∈Mn(R)A \in M_n(R) is det⁡A=∑σ∈Snsgn⁡(σ)∏i=1nai,σ(i).\det A = \sum_{\sigma \in S_n} \operatorname{sgn}(\sigma) \prod_{i=1}^n a_{i, \sigma(i)}. The trace of a matrix is tr⁡A=∑iaii\operatorname{tr} A = \sum_i a_{ii}.

Proposition. det⁡(AB)=det⁡Adet⁡B\det(AB) = \det A \det B, det⁡AT=det⁡A\det A^T = \det A, and tr⁡(AB)=tr⁡(BA)\operatorname{tr}(AB) = \operatorname{tr}(BA). A matrix AA is invertible in Mn(R)M_n(R) if and only if det⁡A∈R×\det A \in R^\times. The invertible matrices form the group GLn(R)\mathrm{GL}_n(R).

Definition. The characteristic polynomial of A∈Mn(R)A \in M_n(R) is χA(X)=det⁡(X⋅1−A)\chi_A(X) = \det(X \cdot 1 - A). It is monic of degree nn, with coefficient −tr⁡A-\operatorname{tr} A at Xn−1X^{n-1} and constant term (−1)ndet⁡A(-1)^n \det A.

Definition. Let A∈Mm×n(R)A \in M_{m \times n}(R) and s≤min⁡(m,n)s \le \min(m, n). A minor of size ss is the determinant of an s×ss \times s submatrix. The leading principal minor Δs(A)\Delta_s(A) is the one on the first ss rows and columns.

Proposition. Let x∈Md×n(R)x \in M_{d \times n}(R) with n≤dn \le d, and let b∈Mn(R)b \in M_n(R) be upper triangular with diagonal b11,…,bnnb_{11}, \dots, b_{nn}. Then Δs(xb)=Δs(x) b11⋯bss\Delta_s(xb) = \Delta_s(x) \, b_{11} \cdots b_{ss} for s≤ns \le n. For each g∈GLn(R)g \in \mathrm{GL}_n(R), Δn(xg)=Δn(x)det⁡g\Delta_n(xg) = \Delta_n(x) \det g.

Definition. A block matrix is a matrix written as an array of submatrices, for example (ABCD)\begin{pmatrix} A & B \\ C & D \end{pmatrix} with AA and DD square. If C=0C = 0, its determinant is det⁡Adet⁡D\det A \det D.

Definition. A matrix A∈Mn(R)A \in M_n(R) is nilpotent if Ak=0A^k = 0 for some k≥1k \ge 1. Over a field, AA is nilpotent if and only if χA=Xn\chi_A = X^n, and then An=0A^n = 0.

Example. (0100)\begin{pmatrix} 0 & 1 \\ 0 & 0 \end{pmatrix} is nilpotent. The block matrices (1a∗01d−a)\begin{pmatrix} 1_a & * \\ 0 & 1_{d-a} \end{pmatrix} form a subgroup of GLd\mathrm{GL}_d.

The test vector at infinity of block determinants is ∏sΔsνs−νs+1\prod_s \Delta_s^{\nu_s - \nu_{s+1}}, a highest weight vector by the proposition above. The cuspidal condition of Green averages over the block unipotent matrices of the example, and CM divisibility shows that a dual Frobenius matrix is nilpotent.

Hermitian and skew-Hermitian forms

Let EE be a field of characteristic ≠2\ne 2 with an automorphism σ\sigma of order 11 or 22, and let F=EσF = E^\sigma. In the proof, EE is a CM field, F=E+F = E^+ and σ\sigma is complex conjugation (Chapter 4). Let VV be an EE-vector space of dimension nn.

Definition. A sesquilinear form on VV is a map h:V×V→Eh : V \times V \to E that is linear in the first variable, additive in the second, and satisfies h(x,ay)=σ(a)h(x,y)h(x, ay) = \sigma(a) h(x, y). It is a Hermitian form if h(y,x)=σ(h(x,y))h(y, x) = \sigma(h(x, y)), and a skew-Hermitian form if h(y,x)=−σ(h(x,y))h(y, x) = -\sigma(h(x, y)). If σ=1\sigma = 1 these are the symmetric and the alternating bilinear forms.

Definition. Let v1,…,vk∈Vv_1, \dots, v_k \in V. Their Gram matrix is J=(h(vi,vj))i,j∈Mk(E)J = (h(v_i, v_j))_{i, j} \in M_k(E). If v1,…,vnv_1, \dots, v_n is a basis, then h(x,y)=xTJσ(y)h(x, y) = x^T J \sigma(y) in coordinates. The form is Hermitian if and only if σ(J)T=J\sigma(J)^T = J, and skew-Hermitian if and only if σ(J)T=−J\sigma(J)^T = -J. A change of basis by g∈GLn(E)g \in \mathrm{GL}_n(E) replaces JJ by gTJσ(g)g^T J \sigma(g).

Definition. The form hh is nondegenerate if h(x,y)=0h(x, y) = 0 for each yy implies x=0x = 0, that is, if det⁡J≠0\det J \ne 0 for the Gram matrix JJ of a basis.

Definition. Let W⊆VW \subseteq V be a subspace. Its orthogonal complement is W⊥={x∈V:h(x,w)=0 for each w∈W}W^\perp = \{x \in V : h(x, w) = 0 \text{ for each } w \in W\}. The sum V=W1⊕W2V = W_1 \oplus W_2 is an orthogonal sum if h(W1,W2)=0h(W_1, W_2) = 0.

Proposition. Let hh be nondegenerate. Then dim⁡W⊥=n−dim⁡W\dim W^\perp = n - \dim W. If hh is nondegenerate on WW, then V=W⊕W⊥V = W \oplus W^\perp is an orthogonal sum.

Definition. A vector x≠0x \ne 0 is isotropic if h(x,x)=0h(x, x) = 0. A hyperbolic plane is a space of dimension 22 with a basis x,yx, y such that h(x,x)=h(y,y)=0h(x, x) = h(y, y) = 0 and h(x,y)=1h(x, y) = 1.

Proposition. Let hh be a nondegenerate Hermitian or skew-Hermitian form and x∈Vx \in V an isotropic vector. Then xx lies in a hyperbolic plane H⊆VH \subseteq V, and V=H⊕H⊥V = H \oplus H^\perp with hh nondegenerate on H⊥H^\perp.

Now let E=ℂE = \mathbb{C} with complex conjugation, or E=ℝE = \mathbb{R} with σ=1\sigma = 1, and let hh be Hermitian. The form is positive definite if h(x,x)>0h(x, x) > 0 for each x≠0x \neq 0.

Theorem (Sylvester's law of inertia). VV has a basis e1,…,ene_1, \dots, e_n with h(ei,ej)=0h(e_i, e_j) = 0 for i≠ji \ne j and h(ei,ei)∈{1,−1,0}h(e_i, e_i) \in \{1, -1, 0\}. The number pp of ii with h(ei,ei)=1h(e_i, e_i) = 1 and the number qq with h(ei,ei)=−1h(e_i, e_i) = -1 do not depend on the basis.

Definition. The pair (p,q)(p, q) is the signature of hh. The form is nondegenerate if and only if p+q=np + q = n, positive definite if and only if the signature is (n,0)(n, 0), and definite if it is (n,0)(n, 0) or (0,n)(0, n). If hh is skew-Hermitian over ℂ\mathbb{C}, then ihih is Hermitian, and the signature of hh is defined as that of ihih.

In TLiftK the Gram matrices are JX=1J_X = 1, positive definite at each infinite place, and the skew-Hermitian JW=δ⋅1J_W = \delta \cdot 1 with iJWiJ_W of signature (d,0)(d, 0). The recursion of ThetaIntegralsAgree splits off hyperbolic planes, and the Bessel seesaw uses an orthogonal sum X=T⊕L′′X = T \oplus L''.

Tensor products and coinvariants

Definition. Let M,NM, N be RR-modules. Their direct sum M⊕NM \oplus N is the set of pairs (m,n)(m, n) with componentwise operations. Their tensor product of modules is an RR-module M⊗RNM \otimes_R N with a bilinear map (m,n)↦m⊗n(m, n) \mapsto m \otimes n, such that each bilinear map β:M×N→P\beta : M \times N \to P factors as β(m,n)=βˉ(m⊗n)\beta(m, n) = \bar\beta(m \otimes n) for exactly one linear map βˉ:M⊗RN→P\bar\beta : M \otimes_R N \to P.

Definition. A pure tensor is an element of the form m⊗nm \otimes n. The pure tensors generate M⊗RNM \otimes_R N, but not every element is one: in k2⊗kk2k^2 \otimes_k k^2, the element e1⊗e1+e2⊗e2e_1 \otimes e_1 + e_2 \otimes e_2 is not a pure tensor.

Definition. Let a group GG act linearly on a vector space VV over kk. The space of coinvariants is VG=V/span⁡{gv−v:g∈G,v∈V}V_G = V / \operatorname{span}\{gv - v : g \in G, v \in V\}. A linear form λ:V→k\lambda : V \to k factors through VGV_G if and only if λ(gv)=λ(v)\lambda(gv) = \lambda(v) for all g,vg, v.

ThetaIntegralsAgree asks that the kernel of the local map pvp_v at each good place be a space of coinvariants, so that invariant linear forms factor through pvp_v, and its matching condition compares pure tensors ⨂vΦv\bigotimes_v \Phi_v.

Zorn's lemma

Theorem (Zorn's lemma). Let PP be a nonempty partially ordered set in which each chain has an upper bound in PP. Then PP has a maximal element.

It gives bases of vector spaces, maximal ideals of nonzero rings and algebraic closures of fields. In block determinants, projEx uses it to find an irreducible representation with an equivariant projection that does not kill a given nonzero automorphic form.

Further reading

  • J.-P. Serre, A Course in Arithmetic, Chapter I (finite fields) and Chapter IV (quadratic forms).
  • S. Lang, Algebra, Chapters I to VI (groups, rings, modules, polynomials, fields, Galois theory) and Chapter XVI (tensor products).
  • M. Atiyah and I. Macdonald, Introduction to Commutative Algebra, Chapters 1, 2 and 6 (ideals, modules, length).
  • W. Scharlau, Quadratic and Hermitian Forms, Chapters 1 and 7.
  • R. Lidl and H. Niederreiter, Finite Fields, Chapters 1 and 2.

The source of this page. Markdown. Something is wrong? Report it.