• 検索結果がありません。

NOTES ON INVARIANT RADICALS OF LOCALLY FINITE DERIVATIONS

N/A
N/A
Protected

Academic year: 2021

シェア "NOTES ON INVARIANT RADICALS OF LOCALLY FINITE DERIVATIONS"

Copied!
4
0
0

読み込み中.... (全文を見る)

全文

(1)

Bull. Kyushu Inst. Tech.

(Math. Natur. Sci,) No. 34, 1987, pp. 5-8

NOTES ON INVARIANT RADICALS OF LOCALLY

FINITE DERIVATIONS

By Fujio KuBo

(Received November 29, 1986)

Introduction

For a Lie algebra L over an algebraically closed field f of characteristic zero, Krempa shows in [5] that every subspace invariant under all automorphisms of L is also invariant under all locally finite derivations of L. From this result we can easily derive the propo- sition that for a Lie algeie.ra L given above, A which is one of the relations si, asc, desc, lsi, ser and for any class X, every subalgebra generated by all zdsc-subalgebras of L is invariant under all locally finite derivations of L. In this paper we present some classes se for which the above proposition holds for a locally fini'te Lie algebra L without the assumption that f is algebraically closed.

1. Classes and radicals

A class Xf is a collection of Lie algebras over a field f together with their isomorphic copies and the O-dimensional Lie algebra. A class sc is said to be linear if LE eet implies L(g)f S2Ealg for any extension field S;2 of f, and if L(g)f S2Eaig for some extension field 9 offimplies LE Xf (cf. Jacobson [3; p. 146]).

Let f be a field of characteristic zero, 2 be an extqnsion field of { and G(91b be the group of all f-automorphisms of 9. We call ec G(91b-invariant if for a Lie algebra L over f and a subalgebra H of L(g)f 9 with He ecQ, (1(g)g)(H) belongs to ceg for any gE G(S2/b.

We shall give some classes which are linear and G(9/b-invariant, so called the varie- ties: Let F. be the free Lie algebra over a field e of rational numbers on an infinite set {xi, x2,...}, Fi be the free Lie algebra over e on {xi,•••,xi}• Let s=Z.,.a.,.[x.(i),•••, x.(.)] be an element of F., where o(i) is a natural number Iarger than or equal to 1 and a.,.EC. An element s is called multilinear if the numbers of times that xi occurs in [x.(i),•••, x.(.)] are same and they are O or 1. Let M be the set of all multilinear elements of F.. Let S be a subset of M, The variety 8s corresponding to S is a class of all Lie algebras L such that s(ei,.,., e.)=O for any sES with sEF. and for any ei,..., e.GL. For instance, if S={[xi,..., x,+i]}, then 8s is the class of all nilpotent Lie algebras of class

5c (Amayo and Stewart [1; Chap, 14]).

(2)

Let se=8s be a variety. Since every element of S is multilinear, X is obviously linear.

Let L be a Lie algebra over f and H be a subalgebra of L(g)f S2 with HE2Eg. Take sES with s= s(xi,..., x.) E F.. Since (1 (g) g)( [u, v]) = [(1 (g) g)(u), (1 (g) g)(v)] for any u, ve

L (g) , S2 and g E G( S2/b, s((1 (8) g)(h,),.,., (1 (g) g)(h.)) == O for any h,,..., h. E H. Therefo re we have (1(g)g)(H)ece., which shows that sc is G(9!b-invariant. It is clear that the union of ljnear classes invariant of G(9/{) is also linear and G(9!b-invariant. Hence the class ER of all nilpotent Lie algebras holds these two conditions, similarly so does the class ES2I of all solvable Lie algebras. We notice that the class 89 of all finite-dimensional Lie algebras satisfies these two conditions.

Every Lie algebra considered in this paper will be over a field of characteristic zero, and we shall use the notations and terminology of [1]. In particular `Åq ', `si', `asc',

`desc', `lsi', and `ser' denote the relations `ideal', `subideal', `ascendant subalgebra',

`descendant subalgebra', `local subideal', and `serial subalgebra' respectively.

Let A be any relation between a Lie algebra and its subalgebra, for example, si, asc, lsi, ser. A class 2E is A-coalescent if A, BE sc, AAL and BAL then ÅqA, BÅr E se and ÅqA, BÅrA L. For instance, itg, IYnEn and IYnEQ( are A-coalescent for A==si, asc, lsi (Hartley [2], Amayo & Stewart [1; pp. 257, 263]).

For zl given above, a class a; and a Lie algebra L, .IAor(L) is defined as J, re(L) = ÅqHAL: H E XÅr ,

We notice that J,iee(L)gJ.,.ee(L)gJ,.,es(L) and J,ies(L)gJi,ire(L)•

PRoposmoN 1. Let XgEg and L be a locally finite Lie algebra. Then Ji,ift-(L)=

Jserce(L)•

PRooF. It is enough to show that for a finite-dimensional subalgebra H of L, H lsi L ifand only ifHserL. Kashiwagi shows in [4] that HlsiLimplies HserL. Conversely assume that HserL. Then by [1;Proposition13.2.4], HnKsiK for any finite- dimensional subalgebra K of L. Now let X be any finite subset of L. Since ÅqH, XÅr is finite-dimensional,H =Hn ÅqH, XÅr sj ÅqH, XÅr. This shows thatH lsi L. Q. E. D.

ExAMpLE 1 : Let f be a field ofcharacteristic zero and P=f[t] be a polynomial algebra.

Considered as an abelian Lie algebra P has derivations x: pHtp and y: pHdpldt. Then [x, y] ==z is the identity on P. Let L be the split extension P+Åqx, y, zÅr (the Hartley algebra [1; Example 6.3.6]). For this Lie algebra L, we have the following relations of inclusions :

Jsig(L)

ll E Jascg(L) EIii Jserrv(L){

Jisig(L)

'

(3)

Notes on lnvariant Radicals of Locally Finite Derivations 7

We first show that J.,,g(L)==ÅqP, yÅr. Take an ascendant subalgebra H including z.

Then by transcendental induction we have H n P=P. Hence every &scendant subalgebra of L of finite dimension does not contain z. On the other hand ÅqP, yÅrgJ.,.g(L).

Therefore J.,.rv(L)=ÅqP, yÅr as we claimed.

Next we show that J,ig(L)==Ji,ig(L)=P. Since L is finitely generated, HsiL if and only if HIsiL. Hence J,ig(L)=Ji,ig(L). ObviouslyPsJ,ig(L)gJ.,,g(L)=ÅqP, yÅr. For a subideal H ofL including y, we have Pg[L, .y]g[L, .H]gH for some n. Therefore every finite-dimensional subideal of L can not contain y, and we have J,ig(L) =P.

It is obviOus that J,.,g(L)=L since ÅqP, yÅr -C J,,,g(L) and ÅqxÅr desc L.

2. Invariantradicals

For a Lie algebra L we denote by d(L), di(L), a(L) the set of all derivations, all locally finite derivations, all automorphisms of L respectively, and denote by D(L), Dt(L), A(L) the lattice of subspaces of L invariant under all elements of d(L), di(L), a(L) respectively.

LEMMA 2 (Krempa [5]). For a Lie algebra L over an algebraically closed field of characteristic zero, we have A(L)gDi(L).

ExAMpLE 2: For the Lie algebra L given in Example 1, we have J.,.g(L) EA(L)XD(L), observing that J.,.g(L) =ÅqP, yÅr•sdL. Let S be a finite-diemsnional non-abelian simple Lie algebra and put L'=SÅ~S. Then a subspace SÅ~{O} is a characteristic ideal of L' but does not belong to the lattice A(L').

The purpose of this section is to prove the following

THEoREM. Let L be a tocally finite Lie algebra over a .field f of charaeteTistic zero, and A be one of the relations si, asc, desc, lsi, ser. Let a class EE have the following conditions

(1) ee is linear subclass of g,

(2) ee is G(flb-invariantfor anyfield f ofcharacteristic zero and its algebraic closure l, (3) ec is A-coalescent.

Then JAee(L) is invariant under all locally finite derivations of L.

LEMMA 3. Let L, f, A, EE be as in the theorem andi be an algebraic closure of f.

Then for HAL(g)ff with Hescr, there exists a subalgebra U of L such that HgU (g)J, UAL and UGXf.

PRooF. Since L is locally finite and H is finite-dimensional, there exists a finite-

dimensional subalgebra VofL such that HgV(g)fl. Let G be the Galois group G(ilf)

off over f. Since Åq(1(g)g)(H):geGÅr is invariant under 1(g)g for any gEG, there exists

a subspace U ofL such that

(4)

Åq(1(g)g)(H): gEGÅr -U (g),fgV(g), f.

Since V(gÅri f is finite-dimensional, there exist g,,..., g. in G such that Åq(1 (g)g)(H): g E GÅr = Åq(1(lig)gi)(H):i=1,...,nÅr. It is obvious that (1(Ei)gi)(H)AL(El)ff. By our assumptions (2), (3) for X we have (1 (g)gi)(H)EXi, U (g)ff =Åq(1(Eg)g,)(H): i=1,..., nÅreXi and U (g), fA

L(Ei)ff. Therefore UEecf and UAL by the assumption (1) for se. Q.E.D.

PRooF oF THEoREM. Let i be an algebraic closure of f. By Lemma 3 we immediately have ÅqHAL(g)fi: HEecTÅrgÅqU(g)fi: UAL, UEXfÅr. From the linearity ofX we have

ÅqHAL(Ei)ff: HEXrÅr=ÅqU Qff: UAL, UEXfÅr =ÅqUAL: UE2EiÅr (iDff.

This subalgebra belongs to A(L(g)fi), so to Di(L(g)ff) by Lemma 2. Therefore ÅqUAL:

UG XfÅr belongs to the lattice D,(L) by the linearity of se. Q, E. D.

References

[1] R.K. Amayo and I. N. Stewart, Infinite-dimensional Lie algebras, Noodhoff, Leyden, 1974.

[2] B. Hartley, Locally nilpotent ideals of a Lie algebra, Proc, Cambridge Philos. Soc. 63 (1967), 257-272.

[3] N. Jacobson, Lie algebras, Interscience, New York, 1962.

[4] Y. Kashiwagi, Lie algebras which have an ascending series with simple factors, Hiroshima Math.

J. 11 (1981), 215-227.

[5] J. Krempa, On invariant subspaces of locally finite derivations, Bull. London Math. Soc. 12 (1980), 374-376.

Depart of Mathematics

Kyushu Institute of Technology

参照

関連したドキュメント

σ(L, O) is a continuous function on the space of compact convex bodies with specified interior point, and it is also invariant under affine transformations.. The set R of regular

In this context, the Fundamental Theorem of the Invariant Theory is proved, a notion of basis of the rings of invariants is introduced, and a generalization of Hilbert’s

In my earlier paper [H07] and in my talk at the workshop on “Arithmetic Algebraic Geometry” at RIMS in September 2006, we made explicit a conjec- tural formula of the L -invariant

The algebra of noncommutative symmetric functions Sym, introduced in [2], is the free associative algebra (over some field of characteristic zero) generated by an infinite sequence (

Kartsatos, The existence of bounded solutions on the real line of perturbed non- linear evolution equations in general Banach spaces, Nonlinear Anal.. Kreulich, Eberlein weak

Another technique we use to find identities is the repre- sentation theory of the symmetric group. The process of studying identities through group representations is indi- rect

It is shown that the space of invariant trilinear forms on smooth representations of a semisimple Lie group is finite dimensional if the group is a product of hyperbolic

This applies to the case where the induced action 1 ϕ acts transitively on the base manifold and states that each point in the bundle gives rise to a bijection between the set