Bourbaki

Corollary 3

Tag 0DL2, permanent. Cite it as [Bourbaki, Tag 0DL2].

Algebra, chapter IX, § 4. Sous-espaces totalement isotropes. Théorème de Witt

Let Q be a non-degenerate quadratic form on a finite-dimensional vector space E over an algebraically closed field A. The group of metric automorphisms of E permutes transitively the non-isotropic subspaces of given dimension of E.

This follows immediately from Theorem 1 and Corollary 2 of Proposition 3.

Exercises. — 1) a) Let K be a field of characteristic 2, J : ξ → ¯ξ an involutory antiautomorphism of K, Z the center of K. Show that if the restriction of J to Z is not the identity, every element μ of K such that ¯μ = μ is of the form λ + ¯λ (observe that there is an element ρ ≠ 0 in Z which can be written in the form ζ + ¯ζ, with ζ ∈ Z); every hermitian form on a vector space over K then satisfies condition (T).

b) Give examples of fields of characteristic 2, admitting an involutory antiautomorphism ξ → ¯ξ distinct from the identity map, and for which there are elements μ=μˉ\mu = \bar{\mu} which are not of the form λ+λˉ\lambda + \bar{\lambda} (cf. Chapter VIII, § 11, Exercise 4).

2) Let A be a field, E a vector space over A, Φ\Phi (resp. Q) a non-degenerate ε\varepsilon-hermitian form on E, satisfying condition (T) (resp. a non-degenerate quadratic form on E, Φ then denoting the symmetric bilinear form associated with Q).

a) Show that in order that a plane P \subset E be isotropic (resp. singular) and not totally isotropic (resp. not totally singular), it is necessary and sufficient that it contain only one isotropic (resp. singular) right (cf. Exercise 14 e)).

b) Suppose that dim E⩾3E \geqslant 3, and that there exist in E isotropic vectors ≠0\neq 0. Show that if P is a non-totally isotropic plane in E, there exists a non-isotropic vector subspace V \subset E, of dimension 3, containing isotropic vectors ≠0\neq 0, and such that P⊂VP \subset V.

3) The hypotheses being the same as in Exercise 2, show that if dim E⩾3E \geqslant 3, every isotropic right in E is the intersection of two non-isotropic planes.

4) The hypotheses are those of Exercise 2, and suppose in addition that E is finite-dimensional.

a) If the index ν\nu of Φ\Phi (resp. Q) is ⩾1\geqslant 1, show that for every isotropic (resp. singular) vector a≠0a \neq 0 in E, there exists a basis (ei)(e_i) of E consisting of isotropic (resp. singular) vectors, such that e1=ae_1 = a (cf. Exercise 14 e)).

b) Let V, W be two totally isotropic (resp. totally singular) subspaces of the same dimension r⩽νr \leqslant \nu; show that there exist two maximal totally isotropic (resp. totally singular) subspaces V_1, W_1, such that V \subset V_1, W \subset W_1, and V1∩W1=V∩WV_1 \cap W_1 = V \cap W. (If U=V∩WU = V \cap W, reason in U0/UU^0/U).

c) Let V, W, V1,W1V_1, W_1 be four totally isotropic (resp. totally singular) subspaces of the same dimension, such that V+WV + W and V1+W1V_1 + W_1 are non-isotropic. Show that there exists a metric automorphism uu of E such that u(V)=V1u(V) = V_1 and u(W)=W1u(W) = W_1.

d) Let ff be a linear form on E, α\alpha an element of A of the form λ+ελˉ\lambda + \varepsilon \bar{\lambda} (resp. an element of Λ\Lambda). Consider the sesquilinear form on E

(x,y)→Φ1(x,y)=Φ(x,y)+f(x)αf(y)‾ (x, y) \to \Phi_1(x, y) = \Phi(x, y) + f(x)\overline{\alpha f(y)}

(resp. the quadratic form

x→Q1(x)=Q(x)+α(f(x))2). x \to Q_1(x) = Q(x) + \alpha(f(x))^2).

Show that if Φ1\Phi_1 (resp. Q1Q_1) is non-degenerate and if ν1\nu_1 denotes its index, one has ∣ν1−ν∣⩽1|\nu_1 - \nu| \leqslant 1.

5) a) Let B be a ring, ξ→ξˉ\xi \to \bar{\xi} an involutory antiautomorphism of B, ε\varepsilon an element of the center of B such that εεˉ=1\varepsilon \bar{\varepsilon} = 1. Show that if β\beta is an invertible element of B such that β+εβˉ≠0\beta + \varepsilon \bar{\beta} \neq 0, there exists an invertible element μ≠1\mu \neq 1 in B, such that μ(β+εβˉ)μ=β+εβˉ\mu(\beta + \varepsilon \bar{\beta})\mu = \beta + \varepsilon \bar{\beta}. (Show that one can take μ\mu such that μβμˉ=β\mu \beta \bar{\mu} = \beta).

b) Let A be a field, E a vector space over A, Φ\Phi a non-degenerate ε\varepsilon-hermitian sesquilinear form on E, satisfying (T). Show that if Φ\Phi is not alternating, for every non-isotropic hyperplane HH of E, there exists a metric automorphism of E, distinct from the identity, leaving every element of HH invariant (use a).

6) The hypotheses being those of Exercise 2, let aa be an isotropic vector ≠0\neq 0 in EE (resp. a nonsingular isotropic vector (note that such vectors exist only if AA is of characteristic 2)). Let λ∈A\lambda \in A such that λ+ελˉ=0\lambda + \varepsilon \bar{\lambda} = 0 (resp. λ=(Q(a))−1\lambda = (Q(a))^{-1}); show that the transvection x→x+Φ(x,a)λax \to x + \Phi(x, a)\lambda a (Chapter II, § 6, Exercise 7) is a metric automorphism of EE; converse.

7) The hypotheses being those of Exercise 2, let GG be the group of metric automorphisms of EE. Show that the only semilinear bijections of EE onto itself which commute with all the elements of GG are the homotheties of EE, except in the following three cases: dim E=2E = 2, GG is the group of metric automorphisms corresponding to a quadratic form of index 1 on EE, and AA is one of the three fields F2,F3\mathbf{F}_2, \mathbf{F}_3 or F4\mathbf{F}_4. (Use Exercises 5, 6 and 3; examine separately the case of a quadratic form on a vector space of dimension 2).

*8) Let AA be a field, EE a finite-dimensional vector space >0> 0 over AA, Φ\Phi a non-degenerate ε\varepsilon-hermitian sesquilinear form on EE, satisfying (T). Let M(Φ)M(\Phi) be the group of multipliers of the similarities of EE for Φ\Phi (§6,no 5§ 6, \mathrm{n}^o\ 5).

a) Let V1,V2V_1, V_2 be two vector subspaces of EE, of the same dimension, and let Φ1,Φ2\Phi_1, \Phi_2 be the restrictions of Φ\Phi to V1,V2V_1, V_2 respectively. For there to exist a similarity uu such that u(V1)=V2u(V_1) = V_2, it is necessary and sufficient that there exist α∈M(Φ)\alpha \in M(\Phi) such that Φ2\Phi_2 is equivalent to αΦ1\alpha \Phi_1 (use Witt's theorem).

b) Let (F,F′,G)(F, F', G) be a Witt decomposition of EE (no 2\mathrm{n}^o\ 2), and let Φ0\Phi_0 be the restriction of Φ\Phi to the non-isotropic subspace GG. Show that M(Φ)=M(Φ0)M(\Phi) = M(\Phi_0) if G≠{0}G \neq \{0\}. (Use Witt's theorem and Proposition 2 of no 2\mathrm{n}^o\ 2).

c) Now prove that if the index vv of Φ\Phi is such that dim E=2vE = 2v, M(Φ)M(\Phi) is the group of elements ζ≠0\zeta \neq 0 of the center of AA such that ζˉ=ζ\bar{\zeta} = \zeta. If dim E=2v+1E = 2v + 1, M(Φ)M(\Phi) is the group of elements of the form ρρˉ\rho \bar{\rho}, where ρ\rho runs through the multiplicative group of the elements ≠0\neq 0 of the center of AA (use Witt's theorem). (Cf. § 10, exerc. 18).*

9) Let AA be a commutative field, EE a vector space over AA, QQ a non-degenerate quadratic form on EE. A similarity for QQ is any automorphism uu of EE such that there exists an element α≠0\alpha \neq 0 of AA for which Q(u(x))=αQ(x)Q(u(x)) = \alpha Q(x) for all x∈Ex \in E; uu is then also a similarity for the bilinear form associated with QQ. Assuming the dimension of EE to be finite and >0> 0, state and exercise prove, for the similarities relative to QQ, the analogues of the results of exerc. 8.

*10) Let AA be a field, EE a vector space over AA of dimension >2> 2, Φ1\Phi_1 (resp. Φ2\Phi_2) a non-degenerate sesquilinear form ε1\varepsilon_1-hermitian (resp. ε2\varepsilon_2-hermitian) on EE for an involutory antiautomorphism J1J_1 (resp. J2J_2) of AA, satisfying condition (T). Now prove that if the group of metric automorphisms of EE for Φ1\Phi_1 is a subgroup of the group of similarities for Φ2\Phi_2, there exists α∈A\alpha \in A such that Φ2=Φ1α\Phi_2 = \Phi_1 \alpha (use exerc. 5 b) and 6).

Exercise prove the analogous property when A is assumed commutative, and when Φ1\Phi_1 and Φ2\Phi_2 are replaced in the statement by two non-degenerate quadratic forms Q1,Q2Q_1, Q_2 on E.*

11) Let A be a ring, J an involutory antiautomorphism of A, E an A-module admitting a finite basis (ei)(e_i), Φ\Phi a non-degenerate ε\varepsilon-hermitian form on E, R the matrix of Φ\Phi with respect to (ei)(e_i); the group of metric automorphisms of Φ\Phi is identified with the group G of invertible matrices U such that tU.R.Uj=R^tU . R . U^j = R.

a) Suppose that there exists a matrix P such that R=tP+εPjR = ^tP + \varepsilon P^j. Now prove that for every matrix S such that tS+εSj=0^tS + \varepsilon S^j = 0, and such that P+SP + S is invertible, U=(tPj−εjS)−1(P+S)U = (^tP^j - \varepsilon^j S)^{-1}(P + S) belongs to G, and εI+U\varepsilon I + U is invertible. Conversely (show that for every matrix U∈GU \in G such that εI+U\varepsilon I + U is invertible, we have

ε(εI+tU)−1R+εjR(εjI+Uj)−1=R. \varepsilon(\varepsilon I + ^tU)^{-1}R + \varepsilon^j R (\varepsilon^j I + U^j)^{-1} = R.

b) Show that the condition in a) is satisfied when Φ\Phi satisfies condition (T). Case where in A the equation 2ξ=α2\xi = \alpha has one and only one solution for every α∈A\alpha \in A.

¶ 12) Let A be a commutative field, E a finite-dimensional vector space of dimension n over A, Φ\Phi a nondegenerate ε\varepsilon-Hermitian sesquilinear form on E.

a) Let u be an endomorphism of E; show that if the ri(1≤i≤m)r_i (1 \leq i \leq m) are the similarity invariants of u (Chapter VII, § 5, No. 1, Def. 1), the similarity invariants of the adjoint u∗u^* of u with respect to Φ\Phi are the polynomials rˉi(1≤i≤m)\bar{r}_i (1 \leq i \leq m), where rˉi\bar{r}_i is obtained from rir_i by applying the automorphism J to each coefficient (cf. Chapter VII, § 5, exerc. 2). For every monic irreducible polynomial p∈A[X]p \in A[X] dividing the minimal polynomial of u, let Fk(u,p)F_k(u, p) be the kernel of (p(u))k(p(u))^k in E, and let F(u,p)F(u, p) be the union of the Fk(u,p)F_k(u, p) for all integers k>0k > 0. Show that if p and q are distinct monic irreducible polynomials dividing the minimal polynomial of u, the subspaces F(u,p)F(u, p) and F(u∗,qˉ)F(u^*, \bar{q}) are orthogonal (use Bezout's identity). Finally, if G is a vector subspace of E such that u(G)⊂Gu(G) \subset G, we have u∗(G0)⊂G0u^*(G^0) \subset G^0.

b) Suppose that uu∗=u∗uuu^* = u^*u (the case where u is said to be a normal endomorphism for Φ\Phi; cf. § 7, No. 3); show that we then have u∗(Fk(u,p))⊂Fk(u,p)u^*(F_k(u, p)) \subset F_k(u, p) for every k, and consequently u∗(F(u,p))⊂F(u,p)u^*(F(u, p)) \subset F(u, p). If we put G(p,qˉ)=F(u,p)∩F(u∗,qˉ)G(p, \bar{q}) = F(u, p) \cap F(u^*, \bar{q}), show that E is the direct sum of the subspaces G(p,qˉ)G(p, \bar{q}), and that G(p,qˉ)G(p, \bar{q}) and G(p1,qˉ1)G(p_1, \bar{q}_1) are orthogonal if p≠q1p \neq q_1 or if p1≠qp_1 \neq q; in particular G(p,qˉ)G(p, \bar{q}) is totally isotropic if p≠qp \neq q. Show that G(p,pˉ)G(p, \bar{p}) is reduced to 0 or is nonisotropic, and that if p≠qp \neq q, no nonzero vector of G(p,qˉ)G(p, \bar{q}) is orthogonal to G(q,pˉ)G(q, \bar{p}) (use the fact that Φ\Phi is nondegenerate); deduce that if p≠qp \neq q, G(p,qˉ)G(p, \bar{q}) and G(q,pˉ)G(q, \bar{p}) are totally isotropic subspaces of the same dimension, and that G(p,qˉ)+G(q,pˉ)G(p, \bar{q}) + G(q, \bar{p}) is nonisotropic.

c) Suppose that JJ is not the identity or that AA is not of characteristic 2, and that u∗=uu^* = u. Let M\mathscr{M} be the set of non-isotropic subspaces M⊂G(p,pˉ)M \subset G(p, \bar{p}), stable under uu (hence submodules of the A[X]A[X]-module EuE_u (Chapter VII, § 5, No. 1)). Now prove that if MM is a minimal element of M\mathfrak{M}, MM is an indecomposable submodule of EuE_u (Chapter VII, § 4, No. 7). (Suppose that MM is the direct sum of an indecomposable submodule M1M_1 and a submodule M2≠{0}M_2 \neq \{0\}, the minimal polynomials php^h and pkp^k of the restrictions of uu to M1M_1 and M2M_2 respectively being such that h≥kh \geq k. Observe then that M1M_1 is necessarily isotropic and that every z≠0z \neq 0 in M1M_1 such that p(u).z=0p(u).z = 0 is orthogonal to M1M_1 (use the fact that every submodule of M1M_1 is monogenic); write z=(p(u))h−1.xz = (p(u))^{h-1}.x and that zz is not orthogonal to M2M_2, and deduce that necessarily k=hk = h. Now prove that there exists an indecomposable submodule N2N_2 of M2M_2 such that php^h is the minimal polynomial of the restriction of uu to N2N_2, and that M1+N2M_1 + N_2 is non-isotropic; conclude that M2=N2M_2 = N_2. Finally, if y∈M2y \in M_2 is not orthogonal to zz, consider the submodule PP of MM generated by w=x+λyw = x + \lambda y, where λ∈A\lambda \in \mathbf{A}, and prove that λ\lambda can be chosen so that PP is non-isotropic, by proving that Φ((p(u))h−1.w,w)≠0\Phi((p(u))^{h-1}.w, w) \neq 0; which leads to a contradiction).

d) Deduce from c) that G(p,p‾)G(p, \overline{p}) is the direct sum of pairwise orthogonal indecomposable submodules HiH_i. If php^h is the minimal polynomial of the restriction of uu to HiH_i, and if dd is the degree of pp, prove that there exists in HiH_i a totally isotropic subspace of dimension d.[h/2]d.[h/2]. Case where EE contains no isotropic vector ≠0\neq 0 (cf. § 7, No. 3).

e) State and prove the analogous properties to those of c) and d) when u∗=−uu^* = -u or u∗u=1u^*u = 1.

f) Give an example where n=4n = 4, Φ\Phi is symmetric and of index 2, p=p‾=X−1p = \overline{p} = X - 1, uu is normal, E=G(p,p‾)E = G(p, \overline{p}), but EE is not the direct sum of minimal submodules of M\mathfrak{M}, and where there exists an eigenvector of uu which is not an eigenvector of u∗u^* (cf. § 7, No. 3).

13) The hypotheses are those of Exercise 2, and suppose in addition that EE admits a countable basis (en)(e_n). Let FF be a totally isotropic (resp. totally singular) subspace of EE such that F00=FF^{00} = F; now prove that there exists a totally isotropic (resp. totally singular) subspace F′F' such that: 1∘ F∩F′={0}1^\circ \ F \cap F' = \{0\}; 2∘2^\circ there exists a basis (am)m∈I(a_m)_{m \in I} of FF and a basis (am′)m∈I(a'_m)_{m \in I} of F′F' (the interval of N\mathbf{N} with origin 0) such that Φ(ai,aj′)=δij\Phi(a_i, a'_j) = \delta_{ij} for every pair of indices; 3∘3^\circ (F+F′)00=F+F′(F + F')^{00} = F + F' and EE is the direct sum of F+F′F + F' and G=(F+F′)0G = (F + F')^0. (Form by induction an increasing sequence (Ln)(L_n) of non-isotropic subspaces, whose union is EE, such that dim⁡Ln+1−dim⁡Ln=2\dim L_{n+1} - \dim L_n = 2, and apply Proposition 2 of No. 2 to each of the LnL_n; to form this sequence, consider, for each nn, the smallest integer kk such that ek∉Lne_k \notin L_n, and use Exercise 9 b) of § 1).

14) Let AA be a field of characteristic 2, EE a finite-dimensional vector space of dimension nn over AA, Φ\Phi a non-degenerate hermitian form on EE, not necessarily satisfying condition (T).

a) Show that the set VV of the x∈Ex \in E such that Φ(x,x)\Phi(x, x) is of the form α+α‾\alpha + \overline{\alpha} is a vector subspace of EE.

b) Let V1=V∩V0V_1 = V \cap V^0, q=dim⁡V1q = \dim V_1, V2V_2 a supplementary space to V1V_1 with respect to VV, V3V_3 a supplementary space to V1V_1 with respect to V0V^0. Now prove that there exists a base (ei)1≤i≤2q(e_i)_{1 \leq i \leq 2q} of (V2+V3)0=V20∩V30(V_2 + V_3)^0 = V_2^0 \cap V_3^0 such that the vectors e1,…,eqe_1, \ldots, e_q form a base of V1V_1 and that Φ(ei,eq+j)=δij\Phi(e_i, e_{q+j}) = \delta_{ij} for 1≤i≤q,1≤j≤q1 \leq i \leq q, 1 \leq j \leq q.

c) Let G(Φ)G(\Phi) be the group of metric automorphisms of EE (for Φ\Phi). Now prove that for every u∈G(Φ)u \in G(\Phi), we have u(x)=xu(x) = x for every x∈V0x \in V^0.

d) For every u∈G(Φ)u \in G(\Phi), we have u(V)=Vu(V) = V; let uvu_v be the restriction of uu to VV, and let GvG_v be the group formed by the uvu_v. Now prove that: 1o1^o the kernel of the homomorphism u→uvu \to u_v from G(Φ)G(\Phi) onto GvG_v is commutative; 2o2^o if Φ2\Phi_2 is the restriction of Φ\Phi to V2V_2 and G(Φ2)G(\Phi_2) the group of metric automorphisms of V2V_2 for Φ2\Phi_2, there exists a homomorphism from GvG_v onto G(Φ2)G(\Phi_2) whose kernel is commutative (use bb and cc).

e) Suppose that AA is commutative and JJ is the identity; let EE be a vector space of dimension 3 over AA, (ei)1≤i≤3(e_i)_{1 \leq i \leq 3} a base of EE, Φ\Phi the non-degenerate symmetric form on EE whose matrix with respect to (ei)(e_i) is

(100001010). \begin{pmatrix} 1 & 0 & 0 \\ 0 & 0 & 1 \\ 0 & 1 & 0 \end{pmatrix}.

Show that all isotropic vectors in EE are contained in the hyperplane generated by e2e_2 and e3e_3 (cf. exerc. 4 a)). Give an example of a non-isotropic plane containing only one isotropic right (cf. exerc. 2 a)). Show that there is no automorphism u∈G(Φ)u \in G(\Phi) such that u(e1)=e1+e2u(e_1) = e_1 + e_2, though we have Φ(e1,e1)=Φ(e1+e2,e1+e2)\Phi(e_1, e_1) = \Phi(e_1 + e_2, e_1 + e_2).

Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.

Cited by

Nothing.

Cites

Nothing.

Translated with machine help, not yet reviewed by a person.