ebook img

Strong $3$-Commutativity Preserving Maps on Standard Operator Algebras PDF

0.14 MB·
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Strong $3$-Commutativity Preserving Maps on Standard Operator Algebras

STRONG 3-COMMUTATIVITY PRESERVING MAPS ON STANDARD OPERATOR ALGEBRAS MEIYUN LIU,JINCHUANHOU 6 1 0 Abstract. LetX beaBanachspaceofdimension≥2overtherealorcomplexfieldFandA 2 astandardoperatoralgebrainB(X). AmapΦ:A→Aissaidtobestrong3-commutativity n a preservingif [Φ(A),Φ(B)]3 =[A,B]3 for all A,B∈A, where [A,B]3 is the 3-commutator of J A,B defined by [A,B]3 =[[[A,B],B],B]. The main result in this paper is shown that, if Φ 4 2 is a surjective map on A, then Φ is strong 3-commutativity preserving if and only if there exist a functional h:A→F and a scalar λ∈F with λ4 =1 such that Φ(A)=λA+h(A)I ] A for all A∈A. F . h t a m [ 1. Introduction 1 Let R be a ring (or an algebra over a field F). Then R is a Lie ring (or Lie algebra) v 6 under the Lie product [A,B] = AB −BA. Recall that a map Φ from A into itself is called a 3 3 commutativity preservingmap if [Φ(A),Φ(B)] = 0 whenever [A,B] = 0for all A, B ∈ R. The 6 0 problemofcharacterizing commutativity preservingmapshadbeenstudiedintensively (see[2, . 1 0 3,9,13]andthereferencestherein). BellandDailgavetheconceptionofstrongcommutativity 6 1 preserving maps in [1]. A map Φ : R → R is said to be strong commutativity preserving : v if [Φ(A),Φ(B)] = [A,B] for any A,B ∈ R. Clearly, strong commutativity preserving maps i X must be commutativity preserving maps, but the inverse is not true generally. The structure r a of linear (or additive) strong commutativity preserving maps has been investigated in [1, 4, 5, 8]. For nonlinear strong commutativity preserving maps, with R being a prime unital ring containing a nontrivial idempotent element, Qi and Hou in [12] proved that every nonlinear surjective strong commutativity preserving map Φ :R → R has the form Φ(A)= λA+f(A) forallA∈ R,whereλ ∈ {−1,1}andf isamapfromRintoZR,thecenterofR. Particularly, this result is true for maps on factor von Neumann algebras. In [11] the nonlinear surjective strong commutativity preserving maps on triangular algebras are studied. Recently, Liu in [7] obtained that a surjective strong commutativity preserving map Φ :A → A on von Neumann 2010 Mathematics Subject Classification. 47B49; 47B47. Key words and phrases. 3-commutators, standard operator algebras, Banach spaces, preservers. This work is partially supported by Natural Science Foundation of China (11171249, 11271217). 1 2 MEIYUNLIU,JINCHUANHOU algebras A without central summands of type I has the form Φ(A) = ZA + f(A) for all 1 A∈ A, where Z ∈ ZA satisfies Z2 = I and f is a map from A into ZA. For a ring R and a positive integer k, recall that the k-commutator of elements A,B ∈ R is defined by [A,B]k = [[A,B]k−1,B] with [A,B]0 = A and [A,B]1 = [A,B] = AB − BA; a map Φ : R → R is said to be strong k-commutativity preserving if [Φ(A),Φ(B)] = k [A,B] for all A,B ∈ R. Obviously, strongk-commutativity preservingmaps are usualstrong k commutativity preserving maps if k = 1. It seems the study of the problem of characterizing strong k-commutativity preserving maps was started by [10], in where it is shown that a nonlinear surjective map on a unital prime ring containing a nontrivial idempotent is strong 2-commutativity preserving if and only if it has the form A 7→ λA + f(A), where λ is an element in the extended centroid of the ring satisfying λ3 = 1 and f is a map from the ring into its center. With k increasing, the problem of characterizing strong k-commutativity preserving maps becomes much more difficult. Let R be a unital prime ring containing a nontrivial idempotent. It is reasonable to conjecture that a surjective map Φ : R → R is strong k-commutativity preserving if and only if there exist an element λ in the extended centroid of R with λk+1 = 1 and a map f : R → ZR such that Φ(A) = λA+f(A) for all A ∈ R. However, we even do not know whether or not the conjecture is true for k = 3. The purpose of this paper is to answer the above conjecture affirmatively for the case when k = 3 and the maps act on standard operator algebras on Banach spaces. Let F be the real field R or the complex field C and X be a Banach space over F. As usual, denote by B(X) the Banach algebra of all bounded linear operators acting on X. Recall that, a subalgebra A ⊆ B(X) is called a standard operator algebra if it contains the identity I and all finite rank operators. The following is our main results. Theorem 1.1. Let X be a Banach space over the real or complex field F with dimX ≥ 2 and A ⊆ B(X) be a standard operator algebra. Assume that Φ : A → A is a surjective map. Then Φ is strong 3-commutativity preserving if and only if there exist a functional h :A → F and a scalar λ ∈ F with λ4 = 1 such that Φ(A)= λA+h(A)I for all A∈ A. 2. Proof of the main result Before provingTheorem 1.1, wegive several lemmas. In this section we always assumethat A is a standard operator algebra on a real or complex Banach space X and F stands for the realfieldRorthecomplex fieldC,dependingonX is realorcomplex. For x ∈ X andf ∈ X∗, STRONG 3-COMMUTATIVITY PRESERVING MAPS ON STANDARD OPERATOR ALGEBRAS 3 we denote x⊗f for the rank one operator defined by z 7→ hz,fix, where hz,fi = f(z). Note that, every operator of rank ≤ 1 can be written in this form. The first lemma is obvious by the main result in [6]. Lemma 2.1. Let A be a standard operator algebra, and let A , B , C , D ∈ A, i = i i j j 1,2,...,n, such that n A TB = n C TD for all T ∈ A. If A ,...,A are linearly i=1 i i i=1 j j 1 n P P independent, then each B is a linear combination of D ,...,D . Similarly, if B ,...,B i 1 m 1 m are linearly independent, then each A is a linear combination of C ,...,C . In particular, i 1 m if ATB = BTA for all T ∈ A, then A and B are linearly dependent. Lemma 2.2. Let A be a standard operator algebra and A∈ A, if [A,P] = 0 for any rank 3 one idempotent P ∈ A, then A ∈ Z(A)= {λI : λ ∈ F}. Proof. For any nonzero vector x ∈ X, there exists f ∈ X∗ such that hx,fi = 1. Let P = x⊗f; then P is a rank one idempotent. By the assumption, ∗ 0 = [A,P] = [A,x⊗f] = [A,x⊗f]= Ax⊗f −x⊗A f, 3 3 which implies that Ax is linearly dependent of x. Thus, for any x ∈ X, there exists some scalar λ ∈ F such that Ax = λ x. It follows that there exists a scalar λ ∈ F such that x x A= λI. Hence A∈ Z(A) = {λI :λ ∈F}. (cid:3) It was proved in [10] that, for k = 2, if an element S satisfies [A,S] = 0 for all A in a k prime ring, then S is a central element. But, for k ≥ 3, this result is not true anymore. For k = 3, we have the following lemma, which reveals one of the main difficulties to solve the problem of characterizing the maps preserving strong 3-commutativity. Lemma 2.3. Let A be a standard operator algebra, N (A) = {N | N2 = 0,N ∈ A} and 2 S ∈ A. Then [A,S] = 0 holds for all A ∈ A if and only if there exist a scalar λ ∈ F and an 3 element N ∈ N (A) such that S = λI +N. 2 Proof. To check the “if” part, assume that S = λI+N with N2 =0. It is easily seen that [A,S] = [A,λI +N] = [A,N] = AN3−3NAN2+3N2AN −N3A= 0 3 3 3 for any A∈ A. Next we check the “ only if ” part. Assume that [A,S] = AS3−3SAS2+3S2AS −S3A = 0 (2.1) 3 for all A∈ A. By Lemma 2.1, we see that I, S, S2 and S3 are linearly dependent. If I and S are linearly dependent, then S = λI for some λ ∈ F and the proof is done. So we may suppose that I and S are linearly independent. We assert that I, S and S2 are linearly dependent. In fact, if I, S and S2 are linearly independent, then, by Lemma 2.1, we 4 MEIYUNLIU,JINCHUANHOU get S = λI, a contradiction. So there exist two scalars α,β ∈ F such that S2 = αI +βS. Then S3 = S(αI +βS) = αS +β(αI +βS) = αβI +(α+β2)S and it follows from Eq.(2.1) that 0 = AS3−3SAS2 +3S2AS −S3A = A(αβI +(α+β2)S)−3SA(αI +βS)+3(αI +βS)AS −(αβI +(α+β2)S)A = αβA+(α+β2)AS −3αSA−3βSAS +3αI +3βSAS −αβA−(α+β2)SA = (4α+β2)(AS −SA). Since there always exists A∈ A such that AS −SA6= 0, we must have 4α+β2 = 0. Let λ = β, N = S −λI, then λ2 = −α, and 2 N2 = S2−2λS +λ2I = S2−βS −αI = 0. So S = λI +N with N ∈ N (A). The proof of the Lemma 2.3 is completed. (cid:3) 2 Now we are at the position to give our proof of the main theorem. Proof of Theorem 1.1. The “ if ” part of the theorem is obvious. In the sequel, we always assume that Φ : A → A is a surjective strong 3-commutativity preserving map. we will check the “ only if ” part by several steps. Step 1. For any A,B ∈ A, there exists a scalar λ ∈ F such that Φ(A+B) = Φ(A)+ A,B Φ(B)+λ I. A,B For any A,B and T ∈ A, we have [Φ(A+B)−Φ(A)−Φ(B),Φ(T)] 3 = [Φ(A+B),Φ(T)] −[Φ(A),Φ(T)] −[Φ(B),Φ(T)] 3 3 3 = [A+B,T] −[A,T] −[B,T] = 0. 3 3 3 By the surjectivity of Φ and Lemma 2.2, above equation implies Φ(A+B)−Φ(A)−Φ(B)∈ {λI : λ ∈ F}. So the assertion in Step 1 is true. Step 2. Φ(FI)= FI and Φ(FI +N (A))= FI +N (A). 2 2 Φ(FI) = FI is obvious by Lemma 2.2 and the surjectivity of Φ. If N ∈ N (A) and λ ∈ F, then, for any A ∈A, we have 2 [Φ(A),Φ(λI +N)] = [A,λI +N] = [A,N] = 0. 3 3 3 By the surjectivity of Φ and Lemma 2.3, we see that Φ(λI +N)∈ FI +N (A). 2 On the other hand, if Φ(B)= λI +N for some scalar λ and N ∈ N (A), then 2 [A,B] = [Φ(A),Φ(B)] = [Φ(A),λI +N] = [Φ(A),N] = 0 3 3 3 3 STRONG 3-COMMUTATIVITY PRESERVING MAPS ON STANDARD OPERATOR ALGEBRAS 5 for all A∈ A. By Lemma 2.3 again, we see that B ∈ FI +N (A). 2 Hence Φ(FI +N (A))= FI +N (A), completing the proof of the Step 2. 2 2 Step 3. For any nontrivial idempotent P ∈ A, there exist three scalars λP, µP,µI−P ∈ F and an element N ∈ N (A) such that Φ(P) = λ P +µ I +N and Φ(I −P) = λ (I − P 2 P P P P P)+µI−PI −NP. Let P ∈ A be a nontrivial idempotent and write P = P and P = I − P. Then A 1 2 can be written as A = A + A + A + A , where A = P AP (i,j ∈ {1,2}). Write 11 12 21 22 ij i j Φ(P)= S +S +S +S with S ∈ A . 11 12 21 22 ij ij We check Step 3 by several claims. Claim 3.1. For any A∈ A, we have [A,Φ(P)] ∈A +A . 3 12 21 Note that, for any A∈ A, we have [A,P] = PA(I −P)+(I −P)AP ∈ A +A . 2 12 21 Also notice that [A,Q] = [A,Q] holds for any A,Q ∈ A with Q an idempotent. So we have 3 [A,P] = [A,P] = [[A,P] ,P] , which implies that 3 5 3 2 [Φ(A),Φ(P)] = [[Φ(A),Φ(P)] ,P] 3 3 2 holds for any A ∈ A. By the surjectivity of Φ one gets [A,Φ(P)] = [[A,Φ(P)] ,P] for all A∈ A. 3 3 2 Therefore, for any A we have [A,Φ(P)] ∈ A +A . 3 12 21 The next claim is one of the key steps for our proof. Claim 3.2. S = S = 0. 12 21 Taking A = A ∈ A and applying Claim 3.1, we have 11 11 0 = (I −P)[A ,Φ(P)] (I −P) 11 3 = 3S S A S +3S S A S −3S A S S −3S A S S , 21 11 11 12 22 21 11 12 21 11 11 12 21 11 12 22 which implies that (S S +S S )A S = S A (S S +S S ) 21 11 22 21 11 12 21 11 11 12 12 22 holds for all A ∈ A , that is, 11 11 (S S +S S )PAPS = S PAP(S S +S S ) (2.2) 21 11 22 21 12 21 11 12 12 22 6 MEIYUNLIU,JINCHUANHOU holdsforallA∈ A. Thus,byLemma2.1, S S +S S andS mustbelinearlydependent. 21 11 22 21 21 It follows that S S +S S = λ S (2.3) 21 11 22 21 1 21 for some λ ∈ F. Again by Eq.(2.2), we get 1 S S +S S = λ S . (2.4) 11 12 12 22 1 12 Taking A= A ∈ A and applying Claim 3.1, one gets 12 12 0 =P[A ,Φ(P)] P 12 3 =A S S2 +A S S S +A S S S +A S S S 12 21 11 12 21 12 21 12 22 21 11 12 22 21 12 −3S A S S −3S A S S +3S2 A S +3S S A S 11 12 21 11 11 12 22 21 11 12 21 12 21 12 21 and 0 =(I −P)[A ,Φ(P)] (I −P) 12 3 =−S S2 A −S S21S A −S S S A −S2 S A 21 11 12 22 11 12 21 12 21 12 22 21 12 −3S A S S −3S A S2 +3S S 1A S +3S S A S . 21 12 21 12 21 12 22 21 1 12 22 22 21 12 22 Thus A [S (S2 +S S )+S (S S +S S )] 12 21 11 12 21 22 21 11 22 21 (2.5) = 3S A (S S +S S )−3(S2 +S S )A S 11 12 21 11 22 21 11 12 21 12 21 and [S (S2 +S S )+S (S S +S S )]A 21 11 12 21 22 21 11 22 21 12 (2.6) = 3S A (S2 +S S )−3(S S +S S )A S 21 12 22 21 12 21 11 22 21 12 22 hold for all A ∈ A . 12 12 Similarly, by taking A = A and applying Claim 3.1, one gets 21 0 =P[A ,Φ(P )] P 21 1 3 =S2 S A +S S S A +S S222A +S S S A 11 12 21 11 12 22 21 12 21 12 21 12 21 −3S S A S −3S S A S +3S A S2 +3S A S S 11 12 21 11 12 22 21 11 12 21 11 12 21 12 21 and 0=(I −P)[A ,Φ(P)] (I −P) 21 3 =A S2 S +A S S S +A S S2 +A S S S 21 11 12 21 11 12 22 21 12 22 21 12 21 12 −3S A S S −3S A S S +3S2 A S +3S S )A S , 22 21 11 12 22 21 12 22 22 21 12 21 12 21 12 STRONG 3-COMMUTATIVITY PRESERVING MAPS ON STANDARD OPERATOR ALGEBRAS 7 which imply that [S (S S +S S )+S (S2 +S S )]A 11 11 12 12 22 12 22 21 12 21 (2.7) = 3(S S +S S )A S −3S A (S2 +S S ) 11 12 12 22 21 11 12 21 11 12 21 and A [S (S S +S S )+S (S2 +S S )] 21 11 11 12 12 22 12 22 21 12 (2.8) = 3S A (S S +S S )−3(S2 +S S )A S 22 21 11 12 12 22 22 21 12 21 12 hold for all A ∈ A . Obviously, Eqs.(2.3) ∼ (2.8) together are equivalent to say that 21 21 PA[S (S2 +S S )+S (S S +S S )] = [3λ S −3(S2 +S S )]AS , 21 11 12 21 22 21 11 22 21 1 11 11 12 21 21    [S21(S121+S12S21)+S22(S21S11+S22S21)]A(I −P)= S21A[3λ1S22−3(S222+S21S12)],    [S11(S11S12+S12S22)+S12(S222+S21S12)]AP = S12AP[3λ1S11−3(S121+S12S21)], (I −P)A[S11(S11S12+S12S22)+S12(S222+S21S12)]      = [3λ1S22−3(S222 +S21S12)](I −P)AS12    hold for all A ∈A. Thus, by Lemma 2.1, we see that 3λ S −3(S2 +S S ) and P are linearly dependent. 1 11 11 12 21 So 3λ S −3(S2 +S S )= λ P (2.9) 1 11 11 12 21 2 for some λ ∈ F. This entails that 2 [S (S2 +S S )+S (S S +S S )]= λ S , (2.10) 21 11 12 21 22 21 11 22 21 2 21 [S (S S +S S )+S (S2 +S S )] = λ S (2.11) 11 11 12 12 22 12 22 21 12 2 12 and [3λ S −3(S2 +S S )] = λ (I −P). (2.12) 1 22 22 21 12 2 It is easily checked by Eqs.(2.9) ∼ (2.12) that 4λ 4λ (λ2− 2)S = 0 and (λ2− 2)S =0. 1 3 21 1 3 12 If λ2− 4λ2 = 0, then we have 1 3 (Φ(P)− λ1I)2 2 = (S +S +S +S − λ1I)2 11 12 21 22 2 = (S2 +S S )+(S S +S S )+(S S +S S ) 11 12 21 11 12 12 22 21 11 22 21 +(S2 +S S )−λ S −λ S −λ S −λ S + λ21I . 22 21 12 1 11 1 12 1 21 1 22 4 = (λ S − λ2P)+λ S +λ S +(λ S − λ2(I −P))−λ S 1 11 3 1 12 1 21 1 22 3 1 11 λ2 −λ S −λ S −λ S + 1I 1 12 1 21 1 22 4 = (λ21 − λ2)I = 0. 4 3 8 MEIYUNLIU,JINCHUANHOU Let N = Φ(P)− λ1I; then Φ(P) = λ1I +N with N2 = 0. But, by Step 2, this implies P 2 2 P P P = λI + N for some λ ∈ F and some N ∈ N (A), a contradiction. So we must have 2 λ2− 4λ2 6= 0. Hence S = S = 0 and Φ(P)= S +S . 1 3 12 21 11 22 Next, let us determine S and S . 11 22 Taking A = A and applying Claim 3.1, we have 11 [A ,S ] = A S3 −3S A S2 +3S2 A S −S3 A = P[A ,Φ(P)] P = 0 11 11 3 11 11 11 11 11 11 11 11 11 11 11 3 holds for all A ∈ A . As A is clearly a standard operator algebra in B(PX), by Lemma 11 11 11 2.3, S = λ P + N with N2 = 0 for some λ ∈ F and some N ∈ N (A ). A similar 11 1 1 1 1 1 2 11 argument can show that S = µ (I −P)+N for some µ ∈ F and some N ∈ N (A ). 22 P 2 P 2 2 22 Hence Φ(P) = λ P +N +µ (I −P)+N = λ P +µ I +N , 1 1 P 2 P P P where λ =λ −µ with N = N +N , N2 = 0 and N N = N N = 0. P 1 P P 1 2 P 1 2 2 1 Similarly, we can get Φ(I −P) = λI−P(I −P)+µI−PI +NI−P for some λI−P,µI−P ∈ F and NI−P ∈ N2(A). Claim 3.3. λP = λ(I−P) and NI−P = −NP. By Step 1 and Claim 3.2, there exists some scalar c∈ F, such that Φ(I) = Φ(P)+Φ(I −P)+cI = λPP +µPI +NP +λI−P(I −P)+µI−PI +NI−P +cI (2.13) = (λP −λI−P)P +(µP +µI−P +c)I +NP +NI−P. Since Φ(I) ∈ FI by Step 2, there exist two scalars α,t ∈F, with t = λP −λI−P such that tP +NP +NI−P = αI. (2.14) So A(tP +NP +NI−P)= (tP +NP +NI−P)A for all A ∈ A. In particular, taking A= N gives P tNPP +NPNI−P = tPNP +NI−PNP. (2.15) Recall that N = N +N with N ∈ N (A ), i = 1,2. So tN P = tPN = tN and hence, P 1 2 i 2 ii P P 1 by Eq.(2.15), we have NPNI−P = NI−PNP. Then, it follows from Eq.(2.14) that (tP −αI)2 = (NP +NI−P)2 = 2NPNI−P. (2.16) Note that (NPNI−P)2 = 0 as NPNI−P = NI−PNP. Thus, Eq.(2.16) entails that t = α = 0, which, by Eq.(2.14), implies that λI−P = λP and NI−P = −NP, as desired. STRONG 3-COMMUTATIVITY PRESERVING MAPS ON STANDARD OPERATOR ALGEBRAS 9 Step 4. For any idempotent P ∈ A of rank ≤ 2, we have N = 0; For any rank-1 P idempotents P,Q with PQ = QP = 0, we have λ = λ . P Q Assume that P is a rank-1 idempotent. By Step 3, there exist two scalars λ ,µ ∈ F and P P anelement N ∈ N (A), suchthatΦ(P)= λ P+µ I+N . Takingthespacedecomposition P 2 P P P 1 0   X = X +˙X so that P has the matrix representation P = . Since N = N +N 1 2 P 1 2 0 0   0 0   withN ∈ PAP andN ∈ (I−P)A(I−P),wemusthaveN = 0andhenceN = 1 2 1 P 0 S   with S2 = 0. Assume S 6= 0, then there exists rank-1 idempotent P′ ∈ B((I−P)X), such that P′S 6= 0. 2 2 Let P = 0 ⊕ P′, then PP = P P and with respect to the space decomposition X = 2 2 2 2 PX+˙P X+˙(I −P)(I −P )X, we have 2 2 1 0 0 0 0 0     P =  0 0 0  and P2 =  0 1 0 .         0 0 0 0 0 0     Let 1 0 0   Q = P +P2 =  0 1 0 .     0 0 0   Obviously, Q2 = Q. By Step 1 and Step 3, there exist scalars c,λ ,λ , λ , µ , µ , µ ∈ F P P2 Q Q P P2 and elements N ,N ,N ∈ N (A) with Q P P2 2 t t 0 0 0 0 r 0 R  11 12     11 13  NQ =  t21 t22 0 , NP =  0 s22 S23  and NP2 =  0 0 0         0 0 T   0 S S   R 0 R     32 33   31 33  such that λ Q+µ I +N = Φ(Q) = Φ(P )+Φ(P )+cI Q Q Q 1 2 = λ P +µ I +N +λ P +µ I +N . P P P P2 2 P2 P2 10 MEIYUNLIU,JINCHUANHOU So we have λ +µ 0 0 t t 0  Q Q   11 12  0 λ +µ 0 + t t 0  Q Q   21 22       0 0 µ   0 0 T   Q    λ +µ +µ +c 0 0 r 0 R  P P P2   11 13  = 0 λ +µ +µ +c 0 + 0 s S ,  P2 P P2   22 23       0 0 µ +µ +c   R S S +R   P P2   31 32 33 33  whichentails thatt =t = 0, R = 0, R = 0, S = 0andS = 0. SinceN2 = 0, weget 12 21 13 31 23 32 Q also t = t = 0. Then, as N2 = 0 and N2 = 0, one gets further that s = r = 0. So N 11 22 P P2 22 11 P 0 0 0   0 0 1 0   ′   has the form NP =  0 0 0 ; that is S = . However, as P2 = ,   0 S 0 0  0 0 S   33     33  ′ ′ we get a contradiction P S = SP = 0. Hence we must have N = 0. Moreover, 2 2 P λ +µ +µ +c = λ +µ =λ +µ +µ +c. P P P2 Q Q P2 P P2 This forces that λ = λ . It follows that T = 0; that is N = 0, too. P2 P Q Note that, for the case when dimX ≥ 3, we also have µ = µ + µ + c and hence Q P P2 λ = λ . Q P If P is an idempotent of rank 2, taking rank-1 idempotent P ,P so that P = P +P , then 1 2 1 2 the above argument shows that N = 0. So, the assertion of Step 4 is true. P Step 5. Assume dimX ≥ 3. λ4 = 1 and there exists a functional h : A → F such that Φ(A)= λA+h(A)I for any A∈ A. We first show that, if dimX ≥ 3, then there exists a scalar λ ∈ F such that, for any rank-1 idempotent P ∈ A, Φ(P) = λP + µ I for some µ ∈ F. In fact, for any rank P P one idempotent operators P,Q, as dimX ≥ 3, there exists rank-1 idempotent R such that PR = RP = QR = RQ = 0. It follows from Step 4 that λ = λ = λ . Hence, there exists P R Q a scalar λ such that λ = λ for any rank-1 idempotent P. P Thus, for any rank-1 idempotents P, Q ∈A, we have [P,Q] = [Φ(P),Φ(Q)] = [λP,λQ] = λ4[P,Q] . 3 3 3 3 It follows that λ4 = 1 as there are rank-1 idempotents P,Q so that [P,Q] 6= 0. 3 Now, for any A and any rank-1 idempotent P in A, we have [A,P] = [Φ(A),Φ(P)] = [Φ(A),λP] = λ3[Φ(A),P] , 3 3 3 3

See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.