International Mathematical Forum, Vol. 12, 2017, no. 19, 915 - 927 HIKARI Ltd, www.m-hikari.com https://doi.org/10.12988/imf.2017.7980 Fully Annihilator Small Stable Modules Mehdi Sadiq Abbas and Hiba Ali Salman Department of Mathematics, College of Science Mustansiriyah University, Baghdad, Iraq Copyright Β© 2017 Mehdi Sadiq Abbas and Hiba Ali Salman. This article is distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Abstract Let R be an associative ring with non-zero identity and M be a left R-module. A submodule N of M is called annihilator small (briefly a-small), if for every submodule L of M with N+L=M, then π (L)=π (M). The properties of a-small π π submodules have been studied and characterizations of a-small cyclic submodules have been investigated. The sum of a-small submodules is studied. Moreover, we shall introduce fully annihilator small stable module (briefly FASS module) where M is called a FASS module if every annihilator small submodule of M is stable. Characterizations of FASS modules are proven. Keywords: Annihilator small submodules, Fully stable modules, Annihilator small regular modules 1. Introduction Throughout this work R will denote an associative ring with non-zero identity, M a left R-module. A submodule N of M is called small, if for every submodule K of M with N+K=M, then K=M [5]. Recently, many authors have been interested in studying different kinds of a-small submodules as in [3] and [4], where the authors in [3] introduced the concept of R-annihilator small submodules, that is; a submodule N of an R-module M is called R-annihilator small, if whenever N+K=M, where K a submodule of M; then π (K)=0. This has motivated us in turn to introduce π the concept of annihilator small submodules, in way that a submodule N of M is called annihilator small (briefly a-small) in case π (K)=π (M), where K is a π π submodule of M; whenever N+K=M. It is clear that every small submodule is a- small, but the converse is not true generally as examples can show next, while the two definitions become equal if M is faithful, recalling that M is called faithful in case π (π) = 0. Remember that singular submodule of an R-module M denoted by π Z(M)={mβM | π (π) is essential in R} [5], We shall study the properties of a- small π 916 Mehdi Sadiq Abbas and Hiba Ali Salman submodules, and define a subset of M that consists of all annihilator small elements (πππππ‘ππ ππ¦ π΄π ), as well as; we shall denote the sum of all annihilator small π submodules of M by π½ (π), and study its properties and the relation between it and π the Jacobson radical. Finally, we shall introduce the concept of fully annihilator small stable modules as a generalization of fully stable modules [1]. Recall that a submodule N of an R-module M is called stable in case for every R-homomorphism πΌ:π βΆ π we have πΌ(π) β π and M is called fully stable if every submodule of M is stable. Characterizations and properties of this concept is studied involving the satisfaction of Baerβs criterion on a-small cyclic submodules and its effect on M being a FASS module. Recall that, a submodule N of M is said to satisfy Baerβs criterion if for each π½:π βΆ π there exists an element π β π such that π½(π) = ππ for each π β π [1].In this paper, we are also interested to study the relation between M being a FASS module and πΈππ (π) being commutative. π 2. Annihilator small submodules Definition 2.1: A submodule N of an R-module M is called annihilator small (briefly a-small) in M, and denoted by N aβͺ M; if whenever N+K=M for each submodule K of M, then π (K)=π (M). Where π denotes the left annihilators in R. π π π A left ideal I of R is annihilator small if for each left ideal J of R with I+J=R, implies that π (J)=0. π Examples and remarks 2.2: 1. It is clear that every small submodule is annihilator small, but the converse is not true generally. For example, in the β€-module β€, (0) is the only small submodule while for every n>1, there exists m such that nβ€+mβ€=β€ and π (πβ€)=0= π (β€). π π 2. If M is a faithful R-module then the concepts of annihilator small submodules and R-annihilator small submodules are equivalent. 3. There are annihilator small submodules that are direct summands as in the β€ - 2 module M=β€ β¨β€ , where it is clear that A=β€ β¨(0) is a direct summand of M, 2 2 2 π = π΄β¨β€ = π΄β< (1Μ ,1Μ ) > and π (M)=0=π (< (1Μ ,1Μ ) >). 2 β€2 β€2 Recall that, M is called prime if π (N)=π (M) for every non-zero submodule N π π of M[5]. M is called quasi-Dedekind if Hom(M/N, M)=0 for every proper submodule N of M[6], it is mentioned in [6] that every quasi-Dedekind module is prime. The proof of the following proposition is obvious. Proposition 2.3: Let M be a prime R-module. Then every proper submodule of M is annihilator small. In particular, every proper submodule of a quasi-Dedekind R- module is annihilator small. It is mentioned in [6, p.25] that β ππ β€-module is quasi-Dedekind, and hence by the use of proposition (2.3) we get that every proper submodule of β is annihilator small, but only finitely generated submodules of β are small. Fully annihilator small stable modules 917 Proposition 2.4: Let M be an R-module with submodules Aβ π. If N aβͺ π then A aβͺ π. Proof: Let X be a submodule of M such that A+X=M, since Aβ π hence N+X=M. By N being a-small in M then π (π) = π (π) and hence A aβͺ M. β π π Proposition 2.5: Let M be an R-module with submodules Aβ π, if A aβͺ π and π (π) = π (π) then A aβͺ π. π π Proof: Let X be any submodule of M such that A+X=M, now Nβ©M=Nβ© (A+X) implies that N=A+(Nβ©X) by the modular law. Since A aβͺ π, thus π (πβ©π) = π π (π). But π (π) β π (πβ©π) = π (π) = π (π) implies that π (π) β π (π) and π π π π π π π then π (π) = π (π), hence X aβͺ π. β π π Proposition 2.6: Let M and N be R-modules and πΌ:π βΆ π an R-monomorphism if W aβͺ M then πΌ(π) aβͺ πΌ(π). Proof: Let U be a submodule of N such that πΌ(π)+U=πΌ(π), now Uβ π implies πΌβ1(π) β πΌβ1(π) = π and πΌ(πΌβ1(π))=Uβ©πΌπ(πΌ) = πβ©πΌ(π) = π. Now, πΌβ1(πΌ(π))+πΌβ1(π) = πΌβ1(πΌ(π)) and then W+πΌβ1(π) = π this implies that π (πΌβ1(π))=π (π) since W aβͺ M. Let X=πΌβ1(π) then π (π)=π (π). Let rβ π π π π π (π)=π (πΌ(π)), thus rπΌ(π)=0 βΉ πΌ(ππ)=0 βΉ ππ = 0 βΉ π β π (π) βΉ π π π π (π) β π (π) = π (π) βΉ π (π) = π (π) β π (πΌ(π)) βΉ π (π) = π (πΌ(π)). π π π π π π π π Hence, πΌ(π) aβͺ πΌ(π). β Corollary 2.7: Let M and N be R-modules and πΌ:π β π an R-monomorphism such that π (πΌ(π)) = π (π), if W aβͺ M then πΌ(π) aβͺ π. π π In the same manner of the definition of Jacobson radical related to small submodules, we will state a definition related to annihilator small submodules in the following. But first we need this definition. Definition 2.8: Let M be an R-module and aβ π. We say that an element a in M is annihilator small if Ra is annihilator small submodule of M. let π΄π = {π β π π|π π aβͺ π}. Note that π΄π is not a submodule of M. In fact, it is not closed under π addition, for example in the β€βππππ’ππ β€ we have that 3,-2 β π΄π but 3-2=1β β€ π΄π β€. We can see by the use of proposition (2.4) that if M is an R-module and aβ π΄π , then Ra β π΄π . Moreover, if A aβͺ M then Aβ π΄π . π π π Definition 2.9: Let M be an R-module. Denote π½ (π) for the sum of all annihilator π small submodules of M. 918 Mehdi Sadiq Abbas and Hiba Ali Salman It is clear that π΄π β π½ (π) for every R-module M. The β€βππππ’ππ β€ is π π an example of this inclusion being proper, where πβ€ is a-small for each nβ 1,β1 in β€, hence π½ (β€) = β πβ€ = β€, but π΄π = {π β β€|πβ€ π βͺ β€} = π πβ€ πβͺβ€ β€ {πβ€|π β 1,β1}. Recall that, if T is an arbitrary proper submodule of a right R-module M and N a submodule of M, then N is called T-essential provided that N β T and for each submodule K of M, Nβ©KβT implies that KβT [8]. We introduce the following singularity of modules. Definition 2.10: Let M be an R-module and J be an arbitrary left ideal of R. define the subset Z(J,M)of M by Z(J,M)={xβM| π (x) is J-essential in R}, it is easy to π show that Z(J,M)={xβM| Ix=0 for some J-essential left ideal I of R}. It is clear that Z(0,M)=Z(M) for any R-module M. Proposition 2.11: Let M be an R-module and J an arbitrary proper left ideal of R. Then Z(J,M) is a submodule of M, and it is called the singular submodule of M relative to J. Proof: It is clear that Z(J,M) is non-empty. Let x,y β Z(J,M), then there exist two J-essential left ideals A and B of R with Ax=0 and By=0. Now, Aβ©B is J-essential and (Aβ©B)(x-y)=0 [7] and thus x-y β Z(J,M). For each rβR, since π (π₯) β π (ππ₯) π π and π (π₯) is J-essential in R hence rxβ Z(J,M). β π Lemma 2.12: Let M be a non-zero R-module and N a submodule of M. If π (π) is π π (π)-essential in R, then π (π (π)) is a-small in M; in particular, N is a-small in π π π M. Proof: Let X be a submodule of M with X+π (π (π))=M. Then π (π)β© π π π π (π (π (π))) = π (π)β©π (π) = π (π), since π (π) is π (π)-essential in R π π π π π π π π then π (π) β π (π) and hence π (π (π)) is a- small in M. The last assertion π π π π follows from proposition (2.4). β Corollary 2.13: Let M be a non-zero R-module. If m β Z(π (π),M), then Rm is π a-small in M. Proof: Let mβ π(π (π),π). Then π (π) ππ π (π)-essential in R, and by lemma π π π (2.12) we have Rm is a-small in M. β Note that the converse of lemma(2.12) is true if π (π΄β©π΅) = π (π΄)+ π π π (π΅) for each left ideals A and B of R. For this, let T be a left ideal of R with π π (π)β©π β π (π). Then π π Mβ π (π (π)) β π (π (π)β©π) = π (π (π))+π (π). π π π π π π π Since π (π (π)) is a-small in M, then π β π (π (π)) β π (π). This shows that π π π π π π (π) ππ π (π)-essential in R. π π Fully annihilator small stable modules 919 Proposition 2.14: Let M be a non-zero finitely generated R-module and K a submodule of M. If K is a-small in M, then so is K+J(M)+Z(J,M) where J=π (π). π Proof: Let X be a submodule of M such that K+J(M)+Z(J,M)+X=M. Since M is finitely generated, then {π }π is a set of generators of M and M= βπ π π , and π π=1 π=1 π J(M) is small in M; that is, K+Z(J,M)+X=M. Now, for each π β M we have π = π π π +π§ +π₯ where π β K, π§ β Z(J,M) and π₯ β X for each i=1,β¦,n. Thus M= π π π π π π K+βπ π π§ +X and since K is a-small in M by our assumption. π=1 π Thus π (π)=π (βπ π π§ +X)=π (βπ π π§ )β©π (π)=(β©π π (π π§ ))β© π π π=1 π π π=1 π π π=1 π π (π (π)). But π§ β Z(J,M), thus π (π§ ) is π (π)-essential in R for each i=1,β¦,n, and π π π π π hence β©π π (π π§ ) is π (π)-essential in R [2]. Thus π (π) β π (π), and hence π=1 π π π π π K+J(M)+Z(J,M) is a-small submodule of M. β Corollary 2.15: Let M be a finitely generated R-module. Then J(M)+Z(J,M) is a- small in M where J=π (π). π The proof of the following proposition is as that in lemma (2.12). Proposition 2.16: let M be an R-module such that Z(J,M) is finitely generated. If K is an a-small submodule of M, then so is K+Z(J,M). In the following we give a characterization of cyclic annihilator small submodules. Theorem 2.17: Let M be an R-module and mβM. Then the following statements are equivalent: 1. Rm aβͺ M. 2. β© π (π βππ) = π (π) πππ πππβ π β π . πβπΌ π π π π π 3. There exists jβI such that ππ β π ππ for all π β π (π). π π Proof: (1)βΉ (2) For each iβI, π = π βππ+ππ and hence M=β π (π β π π π π πβπΌ π ππ)+π π. By (1) we have π (π) = π (β π (π βππ)) =β© π (π βππ). π π π πβπΌ π π πβπΌ π π π (2)βΉ (1) Let X be a submodule of M with X+Rm=M. Then for each iβI π = π π₯ +ππ, π β π πππ π₯ β π. Let tβ π (π), then π‘π = π‘ππ+π‘π₯ = π (M). π π π π π π π π π (2)βΉ (3) Let π β π (π) and assume that ππ β π ππ for all iβ I. Then ππ = π π π πππ = πππ for all iβI, so by (1) π ββ© π (π βππ) = π (π) which is a π π πβπΌ π π π π contradiction. (3)βΉ (2) Let π ββ© π (π βππ) and hence π β π (π βππ) for all π β πΌ. πβπΌ π π π π π π Thus ππ πππ = πππ for all π β πΌ, so ππ β π ππ. By (2) π β π (π) and hence π= π π π π β© π (π βππ) β π (π) and β© π (π βππ) = π (π) for all π β π . β πβπΌ π π π π πβπΌ π π π π π Theorem 2.18: Let R be a commutative ring, M=β π π and K a submodule of πβπΌ π M. Then the following statements are equivalent: 1. K aβͺ M. 2. β© π π (π βπ ) = π (π) for all π β πΎ. πβπΌ π π π π π 920 Mehdi Sadiq Abbas and Hiba Ali Salman Proof: (1)βΉ (2) For each π β πΌ, let π β πΎ. Then π = π βπ +π for each π β π π π π π πΌ. Then π = β π (π βπ )+πΎ, by (1) we obtain π (π) = πβπΌ π π π π (β π (π π )) =β© π (π (π βπ )). π πβπΌ πβ π πβπΌ π π π (2)βΉ (1) Let A be a submodule of M with M=A+K. Then for each π β πΌ π = π π +π where π β π΄ πππ π β πΎ. Hence π = π βπ for each π β πΌ and π = π π π π π π π β π (π βπ )+πΎ. Now, let π‘ β π (π΄) then π‘π = π‘(π βπ ) for each π β πΌ and πβπΌ π π π π π π hence π‘ β π (π (π βπ )) = π (π) by (2), so π (π΄) β π (π). Thus K aβͺ M. β π π π π π π Next, properties and characterization of π½ (π) are given. π Proposition 2.19: Let M be an R-module such that π΄π β π, then we have the π following: 1. π½ (π) is a submodule of M and contains every annihilator small submodule π of M. 2. π½ (π) = {π +π +β―+π ;π β π΄π πππ πππβ π,π β₯ 1}. π 1 2 π π π 3. π½ (π) is generated by π΄π . π π 4. If M is finitely generated, then J(M) β π½ (π). π Proof: 1. Let {π |π β Ξ} be the set of all annihilator small submodules of M, thus π π½ (π) = β π . Let x,yβ π½ (π), this means that π₯ = β π₯ πππ π¦ = π πβΞ π π πβΞ π β π¦ π€βπππ π₯ ,π¦ β π πππ πππβ π β Ξ πππ π₯ ,π¦ β 0 for at most a finite πβΞ π π π π π π number of π β Ξ. Then x+y=β (π₯ +π¦ ) such that π₯ +π¦ β π for each πβΞ π π π π π π β Ξ, x+yβ π½ (π). Now, let rβ π and π₯ β π½ (π) it is an easy matter to see π π that ππ₯ β π½ (π). Hence, π½ (π) is a submodule of M. it is clear from the π π definition of π½ (π)that it contains every a-small submodule of M. π 2. Follows from (1) and π΄π β π½ (π). π π 3. Since π΄π β π½ (π), then < π΄π >β π½ (π). Clearly, π½ (π) β< π΄π >. π π π π π π Hence, π½ (π)is generated by π΄π . π π 4. Since M is finitely generated then J(M) βͺ M, hence J(M) aβͺ M and by (1) J(M) β π½ (π). β π Proposition 2.20: Let M be an R-module such that π΄π β π. Then the following π statements are equivalent: 1. π΄π is closed under addition; that is, a finite sum of a-small elements is a- π small. 2. π½ (π) = π΄π . π π Proof: (1)βΉ(2) Let π +π +β―+π β π½ (π), π β π΄ i=1,β¦,n, π΄ is a-small in M for 1 2 π π π π π each i=1,β¦,n. then π π aβͺ M by proposition (2.4). Hence π β π΄π for each π π π i=1,β¦,n, by the assumption in (1) we get that π +β―+π β π΄π . thus π½ (π) β 1 π π π π΄π and hence π½ (π) = π΄π . π π π (2)βΉ (1) Let x,y β π΄π , since π΄π β π½ (π) then π₯,π¦ β π½ (π) and by using π π π π proposition (2.19) we have π₯+π¦ β π½ (π). Hence, π₯+π¦ β π΄π (by our assump- π π Fully annihilator small stable modules 921 tion); that is, π΄π is closed under addition. We can prove that a finite sum of π annihilator small elements is annihilator small by the use of induction. β Proposition 2.21: Let M be an R-module such that π΄π β π. If considering the π following statements: 1. π½ (π) is an annihilator small submodule of M. π 2. If K and L are annihilator small submodules of M, then K+L is an annihilator small submodule of M. 3. π΄π is closed under addition; that is, sum of annihilator small elements of π M is annihilator small. 4. π½ (π) = π΄π . π π Then (1) βΉ (2) βΉ (3) βΊ (4) . If M is finitely generated, then (1) βΊ (2). Proof: (1)βΉ (2) Let K,L be a-small in M, then K+Lβ π½ (π) which is a-small by π assumption. Thus by using proposition (2.4) we get K+L aβͺ M. (2)βΉ (3) Let x, y β π΄π , then Rx, Ry are a-small in M, and hence by (2) Rx+Ry is π annihilator small in M. But R(x+y)βRx+Ry and by using proposition (2.4) we get R(x+y) aβͺ M. Hence, x+y β π΄π . π (3)βΊ (4) By proposition (2.20). Now, let M be finitely generated to prove (2)βΉ (1). Consider {π ,π ,β¦,π } to be the set of generators of M. Let X be a submodule of M such 1 2 π that π½ (π)+X=M, then π = π +π₯ such that π β π½ (π) and π₯ β X for each π π π π π π π i=1,β¦,n. Thus βπ π π = βπ π π +βπ π π₯ and hence M=βπ π π +π. π=1 π π=1 π π=1 π π=1 π Now, since π β π½ (π) and since (2) βΉ (3) βΊ (4) we get π½ (π) = π΄π ; that is, π π π π π β π΄π and hence π π aβͺ M thus π (π) = π (π) implies that π½ (π) aβͺ M. β π π π π π π Proposition 2.22: Let M be a finitely generated R-module and π½ (π) aβͺ M. Then π we have the following statements: 1. π½ (π) is the largest annihilator small submodule of M. π 2. π½ (π) = β{π|π is a maximal submodule of M with π½ (π) β π}. π π Proof: 1. Clear by the definition of π½ (π). π 2. Let a β β{π|π is a maximal submodule of M with π½ (π) β π}. Claim π that Ra aβͺ M, if not then M=Ra+X where X is a submodule of M and π (π) = π (π). Since π½ (π) aβͺ M then π½ (π)+Xβ M. But M is finitely π π π π generated, thus there exist a maximal submodule B of M such that π½ (π)+ π π β B. Now, if a β B then B=M a contradiction! But a β β{π|π is a maximal submodule of M with π½ (π) β π} a contradiction! Thus Ra aβͺM π and hence aβ π½ (π). Hence, π½ (π) = β{π|π is a maximal submodule of π π M with π½ (π) β π}. β π 922 Mehdi Sadiq Abbas and Hiba Ali Salman 3. Fully annihilator small stable modules Definition 3.1: An R-module M is called fully annihilator small stable; (briefly FASS-module), if every annihilator small submodule of it is stable. Characterizations of FASS-modules are given in the following. Proposition 3.2: Let M be an R-module. Then the following statements are equivalent: 1- M is a FASS-module. 2- Each a-small cyclic submodule of M is stable. 3- For each xβ π΄π , yβ M if π (π₯) β π (π¦) then π π¦ β π π₯. π π π 4- M satisfies Baerβs criterion on a-small cyclic submodules. 5- π (π (π π₯)) = π π₯ for each π₯ β π΄π . π π π Proof: (1)βΉ (2) Obvious (2)βΉ (3) Let π₯ β π΄π , π¦ β π such that π (π₯) β π (π¦). Define π:π π₯ βΆ π by π π π π(ππ₯) = ππ¦ if rx=0 then rβ π (π₯), hence π β π (π¦) and ry=0, this shows that π is π π well-defined which is clear a homo. Now, since π₯ β π΄π then Rx aβͺ M by π definition of π΄π . Thus π(π π₯) β π π₯ implies that π π¦ β π π₯. π (3)βΉ (1) Let N be an a-small submodule of M and let πΌ:π βΆ π be an R- homomorphism. Now, let π¦ = πΌ(π₯) β πΌ(π) then π₯ β π and hence π π₯ β π implies that π π₯ is a-small by proposition (2.4) and π₯ β π΄π . Now, let π β π (π₯) βΉ ππ₯ = π π 0 βΉ πΌ(ππ₯) = 0 βΉ π(πΌ(π₯)) = 0 βΉ ππ¦ = 0 βΉ π β π (π¦) βΉ π (π₯) β π (π¦) βΉ π π π π π¦ β π π₯ β π and since in particular π¦ = 1.π¦ β π π¦ β π π₯ β π then πΌ(π) β π. (2)βΉ (4) Let π π₯ be a-small cyclic in M and let πΌ:π π₯ βΆ π be an R-homo. Then by (2) πΌ(π π₯) β π π₯ βΉ β π β π π₯,πΌ(π) β π π₯ βΉ β π β π π₯ β π β π π π’πβ π‘βππ‘ πΌ(π) = ππ. (4)βΉ (5) Let π¦ β π (π (π π₯)), define π:π π₯ βΆ π by π(ππ₯) = ππ¦ if π π₯ = π π₯ βΉ π π 1 2 (π βπ )π₯ = 0 βΉ (π βπ ) β π (π₯) βΉ (π βπ )π¦ = 0 βΉ π π¦ = π π¦ βΉ π is 1 2 1 2 π 1 2 1 2 well-defined and clearly a homo. Now, by assumption there exists π‘ β π such that π(π€) = π‘π€ β π€ β π π₯ since π π₯ aβͺ M. In particular, π(π₯) = π¦ = π‘π₯ β π π₯ βΉ π¦ β π π₯ βΉ π (π (π π₯)) = π π₯. π π (5)βΉ (1) Let N be an a-small submodule of M and πΌ:π βΆ π be an R-homo. Suppose π¦ = πΌ(π₯) β πΌ(π) βΉ π₯ β π βΉ π π₯ β π βΉ π π₯ π βͺ π ππ¦ (2.4) βΉ π₯ β π΄π βΉ πππ‘ π β π (π π₯) βΉ π πΌ(π₯) = πΌ(π π₯) = πΌ(0) = 0 βΉ πΌ(π₯) β π π π (π (π π₯)) βΉ πΌ(π₯) β π π₯ ππ¦ ππ π π’πππ‘πππ βΉ π¦ = πΌ(π₯) β π π ππππ π π₯ β π, π π which implies that M is a FASS module. β Proposition 3.3: Let M be an R-module such that π (πβ©πΎ) = π (π)+π (πΎ) for π π π every finitely generated a-small submodules N and K of M. Then M is a FASS module if and only if M satisfies baerβs criterion on finitely generated a-small submodules of M. Fully annihilator small stable modules 923 Proof: βΉ) Let N be a finitely generated a-small submodule of M and let π:π βΆ π be an R-homomorphism. Now, π = π π₯ +π π₯ +β―+π π₯ for some 1 2 π π₯ ,β¦,π₯ ππ π. Now, the proof goes by induction if n=1 then it is the same as for 1 π proposition (3.2). Assume that Baerβs criterion holds for all a-small submodules generated by m elements for m β€ n-1, there exists two elements r, s in R such that f(x)=rx for each π₯ β π π₯ +π π₯ +β―+π π₯ and π(π₯β) = π π₯β for each π₯β β π π₯ . 1 2 πβ1 π Now, for each π¦ β ((π π₯ +π π₯ +β―+π π₯ )β©π π₯ ) we have ry=sy and hence 1 2 πβ1 π r-sβ π ((π π₯ +π π₯ +β―+π π₯ )β©π π₯ ), thus by hypothesis there exists π’+ π 1 2 πβ1 π π£ β π (π π₯ +β―+π π₯ )+π (π π₯ ) such that r-s=u+v and then r-u=v+s=t. For π 1 π π π each π§ β π, π§ = βπ ππ₯ for some π β π , i=1,β¦,n and π(π§) = π(βπ ππ₯ ) = π=1 π π π π=1 π π π(βπβ1ππ₯ )+π(π π₯ ) = π(βπβ1ππ₯ )+π (π π₯ ) = π(βπβ1ππ₯ )β π=1 π π π π π=1 π π π π π=1 π π π’(βπβ1ππ₯ )+π (π π₯ )+π£(π π₯ ) = (πβπ’)(βπβ1ππ₯ )+ (π +π£)(π π₯ ) = π=1 π π π π π π π=1 π π π π π‘(βπβ1ππ₯ )+π‘(π π₯ ) = π‘(βπ ππ₯ ) = π‘π§. π=1 π π π π π=1 π π βΈ) If Baerβs criterion holds for a-small finitely generated submodules then it holds for a-small cyclic submodules and proposition (3.2) ends the discussion. β Proposition 3.4: Let M be a FASS R-module such that for each x in π΄π and left π ideal I of R, every R-homo π:πΌπ₯ βΆ π can be extended to an R-homomorphism πΌ:π π₯ βΆ π. If any a-small submodule N of M satisfies the double annihilator condition; that is, π (π (π)) = π then so does N+Rx. π π Proof: Denote π (π) and π (π π₯) by A and B respectively. Then by our assumption π π π (π΄) = π, and since M is a FASS module then π (π΅) = π π₯. The proof of π+ π π π π₯ β π (π (π+π π₯)) is obvious, since π (π+π π₯) = π (π)β©π (π π₯) = π΄β©π΅. π π π π π It is enough to show that π (π (π+π π₯) β π+π π₯. Now, let π¦ β π (π΄β©π΅) and π π π define π:π΄π₯ βΆ π by π(ππ₯) = ππ¦ for each π β π΄, if ax=0 then π β π (π₯) = π΅ π hence π β π΄β©π΅ and since π¦ β π (π΄β©π΅) then ay=0. Therefore, π is a well- π defined clearly a homo. The use of our assumption implies that there exists an extension πΌ:π π₯ βΆ π of π, and πΌ(π π₯) β π π₯ since M is a FASS module implies that ππΌ(π₯) = πΌ(ππ₯) = ππ¦ for each a in A. Then π(πΌ(π₯)βπ¦) = 0 implies that πΌ(π₯)βπ¦ β π (π΄) = π; that is, there exists π β π such that πΌ(π₯)βπ¦ = π ππ π¦ = π π+πΌ(π₯) β π+π π₯. Thus π+π π₯ = π (π (π+π π₯). β π π Proposition 3.5: Let M be an R-module such that for each π₯ β π΄π and left ideal π I of R, every R-homomorphism π:πΌπ₯ βΆ π can be extended to an R- homomorphism πΌ:π π₯ βΆ π. Then M is a FASS module if and only if each finitely generated a-small submodule of M satisfies the double annihilator condition. Proof: The proof goes by induction as for n=1 it implies from proposition (3.2), and for n=m+1 it implies from proposition (3.4). β The following proposition gives properties of FASS modules. Proposition 3.6: Let M be an R-module. consider the following statements: 1. M is a FASS module. 2. Every submodule N of M with π (π) = π (π) is a FASS module. π π 924 Mehdi Sadiq Abbas and Hiba Ali Salman 3. Every 2-generated a-small submodule B of M with π (π΅) = π (π) is a π π FASS module. 4. If N,K β M, K aβͺ M and N is an epimorphic image of K then N β K. Then (1)βΊ (2) βΉ (3), and (1)βΊ (4) Proof: (1)βΊ (2) Necessity, Let N be a submodule of M such that π (π) = π (π), π π let K aβͺ N and πΌ:πΎ βΆ π be an R-homo. Proposition (2.6) implies that K aβͺ M and hence π βπΌ(πΎ) β πΎ by M being a FASS module, where π:π βΆ π is the inclusion homomorphism. Thus πΌ(πΎ) β πΎ and N is a FASS module. Sufficiency, clear. (2)βΉ (3) Obvious. (1)βΉ (4) Let π₯ β π and πΌ:πΎ βΆ π be an R-epimorphism. Then π βπΌ:πΎ βΆ π is an R-homomorphism, since K aβͺ M then by (1) π βπΌ(πΎ) β πΎ where π:π βΆ π is the inclusion homomorphism. Since N is an epimorphic image of K then for each x in N there exist y in K such that πΌ(y)=x and hence Nβ πΌ(K)βK implies that NβK. (4)βΉ (1) Let N aβͺ M and πΌ:π βΆ π be an R-homo. Now, πΌ:π βΆ πΌ(π) is an epimorphism and using (4) we get πΌ(π) β π. β Next, the relation between the property of an R-module M being FASS and the commutativity of its endomorphism ring is discussed. Proposition 3.7: Let R be a commutative ring and M a FASS R-module. Then πΈππ (π) is commutative over elements in π΄π . Moreover, for each π₯ β π΄π and π π π π β πΈππ (π) there exists an element r in R (depends on x) such that f(x)=rx. π Proof: Let f and g be any two elements in πΈππ (π) and let x belongs to π΄π , then π π Rx is annihilator small in M and hence stable. But every stable submodule is fully invariant [1], which leads to that there exists two elements r, s in R such that f(x)=rx and g(x)=sx [8]. Now, (fβg)(x)=f(g(x)=f(sx)=r(sx)=s(rx) =g(rx)=g(f(x)=(gβπ)(π₯); that is, πΈππ (π) is commutative over elements in π΄π . β π π A natural question to ask is whether there exist conditions under which the converse true? Such a question leads us to define the concept of annihilator small regular modules as shown below. Definition 3.8: An R-module M is called annihilator small regular if given any element in π΄π , then there exists fβ πβ = π»ππ (π,π ) such that m=f(m)m. A ring π π R is called annihilator small regular if it is annihilator small module on itself. There are bilinear functions: π:πΓπβ βΆ π π:πβ Γπ βΆ πΈππ π Where π(π,πΌ) = πΌ(π) β π β π΄π ,πΌ β πβ and π(πΌ,π) = πΌ(π)π β πΌ β π πβ,π β π΄π . π Proposition 3.9: Every commutative a-small regular ring R is a FASS ring.
Description: