This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
if o r a l l q E Q . b ) I f T p = { l } t h e n Noq= f o r a l l q € K e r ( z l i d + F o ) ,
Noq=
i
otherwise. c ) I f T 3 = C 1 1 t h e n Noq=f o r a l l q E K e r ( z l l i d t F o ) u K e r ( z 2 1 i d t F o ) , Noq =
t
otherwise.
Pmod. Consider t h e map hq:No
+
( n o t e t h a t Q i s a l s o a N,-module). q E K e r no. C l e a r l y Noq =I f TI
t
Noq, no
+ noq. h i s an No-epimorphism q Furthermore n o € Ann q i f and o n l y i f
.
Now we use m a i n l y b ) o f Theorem 1.
= { 1 I then No i s a f i e l d and a ) i s proved.
If T7 = 11) then < z l i d + Fo> i s t h e unique proper i d e a l i n No, hence N,
is a
l o c a l r i n g . Thus e x a c t l y t h e elements o f < z l i d t F o > a r e n o t b i j e c t i v e ; moreover, t h e y have t h e same k e r n e l ( z ( z l i d t F o ) ( q ) = O where ~ € 7 2 , ~ $ 0 , i m p l i e s ( z l i d + F o ) ( q ) = O s i n c e 2-l e x i s t s i n i s trivial
1.
If q E Ker(zlid
t
(Z ,.), t h e converse d i r e c t i o n
PI Fo) then Fo(q) = -zlq, hence Noq_c. B u t
=Z
q and p1 i s prime, t h u s Noq=. i f q $ K e r ( z l i d + F o ) P1 {O} and b ) i s proved. I f T3 = 111 then < z l l i d t
Fos,
The isomorphic images hl(),
< ( ~ ~ ~ , l )Since > . z l l* z 2 1 y we g e t
then Ann q i s
Fo> a r e t h e unique p r o p e r i d e a l s i n No. h 1 ( < z Z 1 i d t F o > ) a r e g i v e n by <(zll,l)>,
Z X Z =<(zll,l)> p1
p1
i <(z21,1)>, hence
129
Ideals and reachability in machines
N =
. The i d e a l generated by an element
0
* 0 of
No i s
o ro r No i t s e l f . Since No i s a c o m u t a t i v e r i n g
e i t h e r
w i t h i d e n t i t y , exactly the elements o fU a r e n o t i n v e r t i b l e , t h a t means n o t b i j e c t i v e . The r e s t f o l l o w s s i m i l a r t o b ) . Remark 3. With t h e same assumption and n o t a t i o n as i n Theorem 3, t h e f o l l o w i n g i s a l s o v a l i d . I f T 3 = l l I , then No= < z Z 1 i d t F o > hence = 0. Thus Im(zl,id+Fo) is (zllid+Fo) 0 ( z 2 1 i d + F o ) = (zZlid+Fo) o(zllid+F0) a submodule o f Ker(z2,id+Fo) and Im(zZlid+Fo)
i s one o f Ker(zllid+Fo).
Since Ann q * N o f o r a l l q * O , t h e i n t e r s e c t i o n o f Ker(zllid+Fo) Ker(zZlid+Fo)
with
i s t r i v i a l . I f Q i s f i n i t e o r a finite-dimensional
vector space,
Fo) = Im(zl1id+F0),respectively. The l a s t a s s e r t i o n needs an explanation. I f IQ( = n F: M, then = Im(z21id+Fo), Ker(zZlid+
then Ker(zllid+Fo)
Ker(zZlid+ Fo) = Im(zllid+ Fo) f o l l o w s by n = IKer(zllid+Fo)[
IIm(zllid+Fo)[
5 [Ker(zllid+Fo)(
( K e r ( z 2 1 i d + F o ) (< n .
The e q u a l i t y holds because of t h e homomorphism theorem, the f i r s t i n e q u a l i t y because of Im(zllid + Fo) 5 Ker(z2.id + Fo), the second because the sum o f both kernels i s d i r e c t , together w i t h t h e homomorphism theorem. I f Q i s a f i n i t e dimensional vector space then t h e a s s e r t i o n f o l l o w s s i m i l a r l y by dim Q = d i m Ker(zllid+Fo)+dim +
Im(zllid+Fo)
< d i m Ker(zllid+Fo)
+
dim Ker(zZlid + Fo) < dim Q. Now we present two methods f o r determining r e a c h a b i l i t y i n A i f char No(A)
i s a s.r. w i t h
i s n o t prime. L e t A=(Q,A,F)R be a MSA where No := N,(A) generator F,
s.tr.
(a,B;n),
and l e t q € Q * :
F i r s t method. Determine a := o r d q and b : = minCm6 IN1 mFo(q) E). Then the following i s easy t o see (cf. [41): Za i d + Zb Fo i s a system o f representatives and, by considering hq : No o f No,Ann "smallest" representation o f Noq. If n =
n
-+
Noq, Noq = Z a q
t
Zb Fo(q) i s t h e
pi i s squarefree, then we can use a l s o the f o l l o w i n g
i€A Second method. 1) Determine a l l i d e a l s i n No ( t h a t means determine T1, T2, T3) 2) Determine Ann q
3 ) There are some
TcA,
SsT, z € Z (because of Th. 2 and i t s Cor.) such t h a t
p.(z+r)>. T h e n Z a i d + Z b F o , w h e r e a : = n pi, jEs J iET i s a system o f representatives o f No,Ann (see [41). Hence
A n n q = < n pi>;<" b :=
n jcS
iET
pj,
Noq = Za q + Zb Fo(q) i s the "smallest" representation o f Noq.
130
G. Hofer
The f i r s t method works more g e n e r a l l y . B u t t h e second method works sometimes f a s t e r and o f f e r s t h e o p p o r t u n i t y t o determine Noq f o r a l o t o f s t a t e s q
s imul t a neous 1y : P r o p o s i t i o n 1. L e t A=(Q,A,FlR be a MSA where No : = No(A) i s a s.r. generator Fo and s . t r .
(a,B;n)
where n =
n
pi i s squarefree and T1 = A .
a := o r d
i€A then t h e f o l l o w i n g holds:
a ) Ann
i
4, ;EQ, i =
with If
b ) Noq = Za; + Z a F o ( { )
Mb de-
c ) I f B i s the s e t o f a l l proper d i v i s o r s o f a, Mb : = K e r ( b i d ) f o r b E B ,
notes the complement of Mb, then we get: Noq = Z a q
+
Za F o ( q ) f o r a l l q E n M, n itb.
bEB
Pmod. n q = n i d ( 4 )
= 0 f o r a l l q € Q because No i s a s.r.,
hence o r d q
d i v i d e s n f o r a l l q € Q . Note t h a t f o r t q E Z n q we get:
.
t q = 0 o q E Ker t i d
Q t i d E Annq The i d e a l generated by t i d i s g i v e n byand a l s o c o n t a i n e d i n Ann q. Since T 1 = A , Ann 6 must be o f t h e
did>
form d i d >
i(see b ) o f t h e Corol!ary). Now a = o r d
4 yields
the desired
minimal t. T h i s proves a ) , and b ) f o l l o w s . q E r l M a n M b = q € K e r a i d h A q 4 K e r b i d o o r d q = a . This proves c ) . bEB b€B Hence, f o r T, = A , a l l q E f l ManRb have t h e same o r d e r , consequently t h e bEB same a n n i h i l a t o r , which i n t u r n g i v e s i n s t a n t knowledge about Noq f o r a l l these q. The s i t u a t i o n i n t h e case
T2 = A i s s i m i l a r .
P r o p o s i t i o n 2. L e t A = (Q,A,F)R be a MSA where No : = No(A) i s a s.r. generator Fo and s . t r . Let z E Z with z=zi
i :=
(a,B;n) (mod pi)
where n =
n
with
pi i s s q u a r e f r e e and T2 = A .
iEA f o r i € A , q € Q and a := o r d q, b := o r d
4, where
( z i d + F o ) ( q ) , then t h e f o l l o w i n g hold:
a ) Ann q = b ) Noq Pk006.
Ha q
+
Z b Fo(q)
On t h e one hand we have Ann q =
by c ) o f t h e
C o r o l l a r y f o r some c,d 6 n. Hence cq = 0 and dq = 0, t h a t means a \ c and b l d . But then
;
i d + F o ) > z < a id>
and bq = 0, so
;
i d + F o ) > . On t h e o t h e r hand, a q = O
FAnn q, and we a r e done.
F o r "mixed" cases (i.e.
T,
A, T2 9 A) and f o r more i n f o r m a t i o n about
p r o p e r t i e s o f i d e a l s i n s y n t a c t i c r i n g s , t h e Jacobson-radical o f a s y n t a c t i c r i n g and so on. see [41.
Ideals and reachability in machines
131
In private communications, G. Betsch (Tubingen) suggested the following point of view. I f a s t a t e q E Q i s reachable from another s t a t e q ' E Q ( i . e . i f qENq' o r qENoq'), we can say t h a t q ' d i u i d a q ( q ' l q ) . The relation I i s then a preorder on Q.The corresponding equivalence classes consist of s t a t e s which can mutually be reached from another. These and related subjects will be studied in a subsequent paper.
REFERENCES
101 [ll] [121 [131
Blyth, T.S., Module theory, Clarendon Press, Oxford, 1977. Eilenberg, S., Automata, languages and machines, Academic Press, New York, 1974. Hlawka, E. and Schoibengeier, J., Zahlentheorie, Manzsche Verlags- und Universitatsbuchhandlung, Wien, 1979. Hofer, G., Near-rings and group-automata, Doctoral d i s s e r t a t i o n , Univ. Linz, Austria, 1986. Hofer, G. and P i l z , G., Group-automata and near-rings, Contrib. Gen. Algebra 2 , Klagenfurt, Austria, 1983. Holcombe, W.M.L., Algebraic automata theory, Cambridge University Press, Cambridge, 1982. Holcombe, W.M.L., The s y n t a c t i c near-ring o f a l i n e a r sequential machine, Proc. Edinba. Math. SOC. 26 (1983). 15-24. Kalman, R.E:, Algebraic theory o f . i i n e a r systems, i n : Topics i n math. systems theory, McGraw-Hill, New York, 1969. Lang, S., Algebra, Addison-Wesley Publishing Company, Reading, Mass., 1984. Lidl, R. and P i l z , G . , Applied a b s t r a c t algebra, Springer-Verlag, New YorkHeidelberg-Berlin, 1984. P i l z , G . , Near-rings, 2nd e d . , North-Holland/American Elsevier, AmsterdamNew York, 1983. P i l z , G., Algebra, Universitatsverlag R. Trauner, Linz, Austrla, 1984. Pilz, G . , S t r i c t l y connected group automata, submitted.
This Page Intentionally Left Blank
Ne~-ringsand Near-fElds,G. Barh(editor) 0 Elsevier Science Publbhm B.V.(North-Holland), 1987
133
COUPLINGS AND DERIVED STRUCTURES Helmut KARZEL Mathematisches Institut, Technische Universitat Miinchen, Arcisstr. 21, D - 8 0 0 0 Miinchen 2 Near vector spaces which can be derived from usual vector spaces by the method of couplings will be discussed. The main results are stated in the theorems ( 8 ) and ('16). INTRODUCTION A structure group (P,c,*) (cf.[21, §2) is a group (P,t) where the set P is provided with a structure such that for any a EP the map at: P P; x 4 ax is an automorphism of the structure (P,C) If Aut(P,L) denotes the group of all automorphisms of the is structure (P,c) then a map rp : P + Aut(F,C, - ) ; a + a CP called a coupling if for all a,b EP the functional equation
'I.
.
(C)
a b = (a-a (b))rp cprp
c
+
cp
holds.
Then we have the statement: Let (P,C,.) be a structure group, let cp : P +Aut(P,C) be a coupling and let a o b : = aaa ( b ) , then ( P , c , o ) = : ( P , c , * ) e p is cp again a structure group called the rp-derivation of (P,c,*) For a class & of structure groups we denote by D(B) the class of all structure groups which are cp-derivations of structure groups of the class 6 . Since the constant mapping rp: x +id is a coupling for any structure group we have Q c a ( & ) . A l l members o f a(&) are called Dickson structure groups with respect to ('I)
.
c. Now we can formulate the following Dickson problems: PI' Let S I be a class of structure groups. Find couplings and determine XI(&) P2 Let E be a class of structure groups and 6 a subclass of Q characterized by additional properties. When do we have D(6)=B?
.
Remarks. .I' A n y (left) nearfield ( N , + , * ) can be a structure group if we set P :=N = N \ ( O ) , (P,.): = if c denotes the additive structure + on N = P ~ ( of the left distribution law for any a EP the map is an automorphism of (P,C).
*
considered as (N , * ) and 0 ) Because a&: x a.x
* .
+-
H. Kanel
134
2. If 'J1 denotes the class of all left nearfields and 8 the subclass of all fields, then the members of the class %(a) are called Dickson nearfields. For the classes 'J1f of finitenearfields and 2f of finite fields we know by the theorem of Zassenhaus that 91f\'S)(3f) consists of exactly seven nearfields. Till now all known examples of infinite nearfields belong to D ( 3 ) (cf. [ I l l ) 3 . Let I be the class of all affine incidence p r o u p s , i.e.the class of all structure groups (P,L!,*) where (P,i?) is either an or a translation plane. Then 2l conaffine space of dim(P,Q)23 tains the subclass Z of all translation geometries (€',Q,+) defined by the property, that for each a EP the map a+: P P;x a + x is a translation. (Here (P,+) is always a commutative group.) By [2] we have here the result U=S(Z),which tells us that any affine incidence group is Dicksonian withrespect to a translation geometry. Any translation geometry (P,Q,+)has the nice property that the set R of all lines through 0 obeys the followingconditions: (F'l) For each A € R , As(P,+) and A {O],P,
.
+
+
+
(F2) U R = P . (F3) For each A,B € I , A = B or A n B = (O}. Now let (G,+) be any group (not necessarily commutative) and R a collection of subgroups of G satisfying (F'l,2,3). Then (G,R) is called a fibered g r o u p and R a fibration (or partition). R is called a kinematic fibration if further
(FK) For each a E G , for each A E R , a + A - a € R is satisfied. For a fibered group (G,R) let EndG be the set of all endomorphisms of the group ( G , + ) and E(G,L) : = (f €End G I v X €R,f(X) cX] Then E(G,R) is a semigroup with identity. But if (G,+) is not abelian in general the sum of two endomorphisms is not an endomorphism and the sets d.g.(EndG) : = [CbifiI b i t ( l , - 7 } , fi EEndG) and
.
d.g.(E(G,R)) : = {c bifil 6iE ('l,-'l], fie E(G,R)) are near-rings distributively generated by EndG and E(G,R) resp. using pointwise addition and composition (cf.[8] Part 111, B6). F o r g E d.g.(E(G,R)) we have still g(X) c X if X E R At the Harrisonburg Conference on near-rings and near-fields in 7983 it was suggested to study near-rings of the type d.g.(E(G,R)). By [51,[71 we have the result: (2) If ( G , R ) is a kinematic fibration with IRI >I and E(G,R)$(o,id) then G is abelian, E(G,R) an integral domain and ( G , R ) can be embedded into
.
Couplings and derived structures
135
h
a vector space ( t , L ) where G = G4 and LZZ(E(G,R))~ with A =Z(E(G,K))\{O) are the quotient structures. Therefore if ( G , R ) is kinematic and (G,+) not commutative then E(G,R) = [o,id) and hence d.g.(E(G,R)) is isomorphic to one of the rings ( Z , + , * ) or (z,,+,.) with nEn\l. To avoid this trivial case and also the case where d.g.(E(G,R)) =E(G,R) , we have to consider non abeliangroups (G,+) with a fibration R which is not kinematic such that
.
E(G,R) 4 l0,idI Here we will discuss the problem how to obtain examples of fibeyed groups using the method of couplings. As the simplest c l a s s 8 of fibered groups we consider all examples (V,R) where (V,+) is over a skewfield the additive group of a left vector space ( V , K ) K and where R is a fibration consisting of vector subspaces of the same dimension. F o r instance the set s l : = [ K a l a E V \ [ o ) of all one dimensional vector subspaces is such a fibration. If ( V , R ) is any fibered group of the class 9) then L :-E(V,R) is a skew field containing K and any X E R is a L-vector subspace over L. Therefore we will always assume K=L=E(V,R). R contains a subset 8 such that V = c {BI B € B is the direct sum of the subspaces of B ; we call 181 thedimension of the fibered group ( V , R ) If we set !2 : = [a+Xl a € V , X E R ) then it is easy to prove that (V,D) is an affine space of dimension 181 where V is the set of points and Q the set of lines, and ( V , ! 2 , + ) is a translation geometry. If ) % I 2 3 , then ( V , D ) is a desarguesian affine space (cf.p.e.[6] §‘lo) and hence R=(Lxl xEV ). If 181 = 2 ,this result is true if and only if(V,Q) is desarguesian; in all other cases ( V , O ) is a proper translation plane. In C4lp.207 and [ J ] the so called near vector spaces (G,@,P,*) are studied, i.e. (G,@) is a group, (I?,+,’) is a near field and : F xG G is a operation such that (V’I) V X , y t F , V x t G : ( X + p ) * x = X * x e p * x (V2) t r X , b E F , V X E G : (X*p)*x=X*(p*x) (V3) V x E G : ’I *x = x T o any near vector space there belongs also a fibration R : = (F *a] a t G\[O)) such that each fiber F *a is isomorphic to the additive group (F,+) of the near field F . So we have the questions:
.
*
*
P3
+
h
If 3 denotes the class of all additive groups ( V , + ) belongA ings to a vector space ( V , K ) does then P ( 8 ) contain examples which can be made to near vector spaces?
H. Karzel
136
Do t h e examples of n e a r v e c t o r s p a c e s g i v e n i n [ 3 ] and [ 4 ] belong t o ~ ( 8? )
P4
2.
CONSTRUCTION OF NEAR VECTOR SPACES, SENILINEAR CASE To d e a l with t h e problems P3 and P4 l e t (V,K) be a ( l e f t ) v e c t o r space. By M ( V ) we denote t h e n e a r - r i n g o f a l l maps from V i n t o V. and $ : V M ( V ) ; a -+a Let cp : V M(V); a a rp 0 : = acp- id Then g f u l f i l l s t h e c o u p l i n g f u n c t i o n a l e q u a t i o n (C)
Va,btV:
i f and only i f va,btV:
(C')
+
+
+
a b = ( a + a (b)) cprp
T
c
-
p
fulfills
$
( a + b + a (b)) %
( 3 ) For t h e o p e r a t i o n
a @b
% :=
=a +b t a b
w
00'
UI
a + a (b) cp
=
a + b + a (b) Ib
we have
a)
xbo =x
0
~ ~ (= 0 O )
b)
O@x=x
0
0 -0
c)
(V,@)
d)
Let a E V , t h e n f o r each b E V t h e e q u a t i o n a Q x = b h a s a s o l u t i o n i f and only i f a = a + i d i s a permutation.
4-
(zero map)
i s associative
0
( C ' ) and
$ ( V ) c End ( V , + )
.
c p u
From now on we assume f o r t h e map 0
(C'),
4
=
group with (4)
Let
0 and 0
S ( V ) c End(V,t)
.
$ : V +M(V)
Then by (3)
the conditions (V,&) i s aserni-
a s n e u t r a l element. By i n d u c t i o n we have
n,mEbI,
a E V ,
...
n * a :='a@aQ @ a , then n-t ime s
c)
n * a @ m * a = ( n + m )* a
e)
I f there i s a respect t o Q
f)
a h a s a n e g a t i v e ea with r e s p e c t t o d l e t Then c ) and d ) a r e v a l i d for a l l (-n) * a : = n * ( e a ) and hence z * a i s a c y c l i c subgroup of ( V , e t ) If
nEhJ with n * a = O t h e n a h a s a n e g a t i v e w i t h and M*a i s a f i n i t e c y c l i c subgroup of ( V , a ) ,
.
.
m,nEz
137
Couplings and derived structures
g)
If
a$
(5) I f
q
i s n i l p o t e n t then
has t h e additional property t h a t for each a Q i s semilinear then we have e i t h e r
: V 4 M(V)
t h e map
a EV
e a e x i s t s and
i s l i n e a r , or
a)
for each
b)
t h e r e i s an aEV such t h a t a i s not l i n e a r , then u v = 0 B 9 Qf o r a l l u , v E V and t h e r e i s an automorphism : K K; h + A such t h a t b (Ax) = x b (x) f o r a l l b,xEV and X E K . yl B
a E V , a$
-+
Proof. For
a,bEV
let
a
b 40
0' V
and l e t
a,p
phisms corresponding t o t h e semilinear maps By ( C ' ) we have f o r
A EX,
xEV
a V ,b
:
(a@b)o(lix)= a ( > ) * a(x) + p(X).bJl(x)+ap(A)a b (x)
vul
$
i s semilinear we have (6) (V'I)
a=P
and i f
be the automorrespectively
. Since
a = p $ i d then
JlJl
*
( A * a > (cl * a > = (P * a I p ( h * a >
b)
( h * a ) 0 (P * d 9 (cl =*a)*(X*a)$.
(a@b))
a b =O.
If t h e r e i s an e x t e r i o r operation : KxV j i s v a l i d then f o r a l l h , E~K and a l l a E V :
a>
.
V such t h a t
ul
Proof. a) i s a consequence of (V'l).
By (C') we have
( a Ceb)$ =
= a + b + a b hence "a b = b a o a 8 b = b 8 a f ' . Now by (V'I) 0 'p UlJI V Q vul obtain X * a @ p * a = ( A + p ) * a = p * a @ A * a which gives us b ) .
we
Next we w i l l show t h a t t h e examples of near vector spacesgiven i n [ 4 ] belong t o 9 ( B ) L e t (V,K) be a 3-dimensional vector space over a commutative f i e l d K with char K S 2 , l e t : K -3 K be an automorphism of the f i e l d K and l e t e,l,e2,e3 be a b a s i s
.
of
(V,K)
. For
xEV
let
X , ~ , X ~ , X € K~
be defined by
, .
x=x,e + x e + x e and f o r aEV l e t a (x) : = a x e Then I 1 2 2 3 3 UI -a y E Eud(V,+) and since a (Ax) = : a3Ax,,c2 = A a3x,r e2 = X a (x) f o r 9 Ib A E K , t h e map a i s semilinear. I f a 4 0 then a ( V ) =Ke and JI 3 V 2 a -1 (0) =Ke2 +Ke and i f a3 = O then a i s t h e zero map. This yl 3' Yl shows u s a b = O and ( a ( b ) ) = O for a l l a , b E V . Since
--
ulyl
- Q-
Q
(a+b) (x) = (a3 + b3)x,le2 = a3x,,e2 +F3%,e2, (a+b)* = a yl
+b4
and s o ( C ' ) :
= a + b + ( a (b)) = a + b = a + b + a b u l u l y l yl B V S Y I J l V I y l ' (V,@) with a 8 b : = a + b + a(b) i s a group which i s not 0
( a + b + a (b)) yl
By ( 3 )
$
138
H. Karzel
el @ e 3 = e,l + e3 f e,, + e3 + e2
commutative because a (b) =a3E,le2,
0 t h e formula ( 4 ) a ) i s reduced t o
=
-
= n . a + -I' n ( n - l ) a 3 a l e 2 .
Since
2
t h e same o r d e r with r e s p e c t t o
(7)
: V
The map
e5 8 e,,
and s i n c e
i s t h e a d d i t i o n d e f i n e d i n ['I.], 8ec.V. S i n c e
@
v
=
n *a = n . a
+ (z)aI(a)
(Xe2)y = 0 ,
each element
+ and
So we have:
6.
=
a EV
has
Ehd(V,+) d e f i n e d above has t h e f o l l o w i n g
p r o p e r t i e s a ) , b ) , c ) and (C') i s a consequence of a ) , b ) , c ) . a EV
=O,
b) a b
vul
,
i s s e m i l i n e a r and i f 0 (a$(b)) = O
a ) F o r each
C K
a
,
t h e n (Aa)
=
ya
u I J 1
UI
c ) ( a + b ) $ = a $+ b
0
Now l e t u s assume t h a t t h e r e i s an
a E V such t h a t t h e semii s not l i n e a r . Then by ( 5 ) b ) t h e r e i s an automor4 of ( K , + , * ) d i f f e r e n t from t h e i d e n t i t y , such t h a t phism u (Av) = T u ( v ) and u v = 0 f o r a l l u , v f V I f x (x) = 0 f o r Q 0 0 ) v a l l xEV t h e n b y ( 4 ) a ) n * x = n . x for n t N and X a & p a = l i n e a r map
a -B
.
=
( h + p ) a + ( a a ) (pa) $
x(pa) i s an
(Xp)a
=
with
1
( h + p ) a + pX
that
SO
aCV
=
(V,@,K,.)
a (a) + O
I
then
Trying t o extend t h e o p e r a t i o n
(Aa)#(Aa) = ( A
+p)a
and
i s a n e a r v e c t o r space. I f t h e r e a , a ( a ) a r e l i n e a r l y independent.
*
uI
of ( 4 ) a ) onto K x V s u c h t h a t
(V,CH,K,*) becomes a n e a r v e c t o r space we make t h e f o l l o w i n g approach: Let f , g , h : K +K be f u n c t i o n s such t h a t for each xEV
we have
Then
0
h('l)
=
Since
=
0 *a
f('l) ='I
X*x : = h ( h ) - x + g ( X ) x b ( x ) and
.
and
a ='I + a
We s e t
implies
a
h(0)
=
.
(x*~)~=f(X)x$ g(0)
and r e c a l l t h a t
=
g('l)
=
and
X * a @ p * a = @ ( h ) + h ( p ) )a + ( g ( X ) + g ( p ) ) a ( a ) + f ( X ) a ( h ( p ) a +
+ g(l-l)a(a)) = (h(A) + h ( ~ l ) ) +a ( g ( h ) +g(l-l) + f ( A > h ( c ) ) a ( a )
,
( A +I)
+ a =h(A + v ) a + g ( X + p ) a ( a > (v") i s e q u i v a l e n t with
( a ) h(X +>I
=h(X) + h ( p )
( p ) g(X+Ci) = g ( X ) + g b ) +f(X).h(P) and s i n c e h + ( p * a ) =h(A).p + a + g ( A ) . ( p * a ) Q ( k * a ) = h ( A ) h ( p ) a + h(h)g(l-l)a(a) + g ( A ) e f ( p ) a ( h h ) a + g ( p ) a ( a ) ) = h ( h ) h ( p ) a + ( h ( h ) g ( p ) + g(X>* f ( u ) h ( p ) ) a ( a )
=
( b p ) * a = h ( X p ) a +g(A p > a ( a >
(V2) i s e q u i v a l e n t with
(y) (6)
0
a2 = O .
~ ( X * V=)h ( X ) * h ( F ) g(Xep) = h ( X ) * g ( p ) + g ( A ) * f ( p ) E ( P )
=
Couplings and derived structures
139
+K
Since h('l) ='I the equations (a) and ( y ) tell us that h : K is a monomorphism, and from ( p ) and f ( l ) ='I we obtain f(h) =g(h+'l) - g ( X ) -g('l) =h(X) so that ( 5 ) gets the form ( 5 ' ) g(X*p) =h(X).g(P) +g(X)E(P)+E(P) By g(l) = 0 , h('1) = f(l) ='I , (a) and ( p ) we have h(n) = nh(1) for n E N , g(2) ='I and ( 5 ' ) gives us g ( 2 . h ) =h(X) +g(X).4= =
-
2.g( a ) + H( X) -E(a )
h(1-p)
(Y) -
= h(p).h(i)
of
and
Vr
c(p>
=
(PI =
$ (m - h( A ) ) . g ( h +I)
n
Furthermore
-g(X) - g ( p ) =f(k)
x(X)
=
= h ( p rA ) shows that K has to be commutative.
A s in ]I'[
and
=f(X)
h(X).E(p)
=
thus g( X)
=
:=
and [2] we consider the s u b g r o u p s U := (x E VI x = 0 ) 9 (x EVI V (x) = 01 of (V,+) Then V r is also a subgroup
.
9
and if V V = O
(V,@),
ulu x,yE V ) c V r
I
(x,(y)
.
then also U
is a subgroup of ( V , @ )
(u + v ) = ~ u +v
In any case
o
or v E U u V r and "I, is a vector subspace of (V,K)
v
if u E U U v r
because VV (h( X)x) +
consists of semilinear maps. Therefore (A +x) = o and ( 6 ) implies (h( A h ) ,(x> + (g( A ) . x,(x)) ,(x> = + (g( X>x)(x)) 1-2 = h(A)X9(x) , hence CI x,(x> = (px)$(x) + ( ~ ( c r - ~)xv(x))~(x) if
.
Eh(K) On the other hand one shows by calculation that also the part 2. of c) of the following theorem is valid:
P
be a vector space, let c p : V +Aut(V,+) be a coupling such that f o r each x € V the map x : = x -id is u r p semilinear and there is an a E V where a is not linear, and let V - : K jK ; a + T denote the automorphism of a Further let 0' x & t y = x + x (y) = x + y + x (y). Then ( 8 ) Theorem. Let ( V , K )
V
cp
a) For all u,v E V means that
c ) Let
(V,e,K,*)
a EV
X EK , u
v ) = x u (v) and
V( A
I
is a stronq couplinq i.e.
cp
b) If x (x) = O Q space. f,g,h : K
,
f o r all
with
4K
with
x EV
a (a) + O .
u
then 'I.
u v
= 0 which O Q ( U + V ) ~ =up vv.
(V,@,K,,)
is a near vector
If there are functions
+ g ( X ) . x (x)) f(A).xg such that 0 yl= 1 *x : = h( X),x + g( A ) -x,(x) becomes a near vector
with
(h(X).x
space; then K is commutative, Char K + 2 , h is a monomorphism and of the field K , f(h) =h(h),g(h) = $(m'-h(X)),
(N) For each x E V
,
p
E h(K)
,
-p x,(x)
=
(px) ,(x) +(&i -p)xy(x)&(x).
H Karzel
140
2. If the field K is commutative with Char K 4 2 and if there is a monomorphism h such that (N) is valid then (V,@) becomes a 'I 2 - h(ok))x,(x), near vector space (V,@,K,*) for X *x : = h( k)x + ~ ( h 4. I f
Remark.
(px) = E x
o
and
v
(x (x)) = O for each x E V
v 1 u
,
EK then (N) is valid. In this case h =id determines exactly the exterior multiplication for the examples given in [4l p.206.
p
3.
CONSTRUCTION OF NEAR VECTOR SPACES, LINEAR CASE The near vector spaces considered in [ 3 ] one obtains in this way: (9) Let ( A , K ) be an associative nilpotent algebra and let m E N
such that x m = 0 for all x E A but ym-"f 0 f o r at least one y E A. If char K = O or m ( p :=char K then (A,O,K,*) with m-7
xely :=x+y+xty and A*x= ( I t x ) x -'I space.
=
For an associative nilpotent algebra with
a,(x)
(ay(b))v
:=
=a b
9u1
a €A
each
a.x has the properties hence
the map
( a + b + a (b)) 0
a$
$
.C,(h)x
i
1=I 1
is a near vector
(A,K) the map $ (a+b)$ = a + b and
e
C
$
M(A)
*
= a + b +a b
$
:A 3
$
and f o r
is K-linear and nilpotent. Therefore by
(3)~) and (4)g) (A,$) with a&b = a + b + a JI (b) = a+b + a - b is a Hence group and so the examples of [ 3 ] are also members o f a($ P4 is answered positively. We remark that (A,CB,K,*) is not a near vector space if in (9) m > p : = char K ; then there is an a € A with ap f 0 but aP+''= 0 in and hence p *a = C(Pi)ai = ap f 0. Further if x E A then (x,) = O and there is an y E A with (yq)m-2 f 0.
.
The result of (9) can be reversed. ('10) If : V + M(V) has the following properties is linear, a) for each a E V , a b)
c)
there is a m € k t , there is an y E V for all
then
(V,@>
a,b EV with
,
0 m>l
with (a+b)*
with (X,)~-"(X) = O f o r all (~$)~-~(y) 4 0 if m > 2 , =
a$ + b $ and
a@b : = a + b + a (b) 0
(ag(b))
x EV
and
c = a9b J I '
is a group and (V,+,r) with xmf"= 0 for all
is an associative nilpotent ring with
a - b= a (b) JI x E V and there are y , z E V with ym-"z$O. (V,+,*) is a K-algebra if and only if (ha) = Xa for all X € K and all a E V . P
O
141
Couplings and derived structures
Finally we djscuss the problem P3 for the case of a coupling cp : V +Aut(V,+) where the corresponding map 0 : a - % avi := aV - id has the properties a) and b) of ('lo), If m = 2 , i.e. x (x) = O for all x E V then ( V , @ , K , * ) is a
v
near vector space (cf.(7)b)). Therefore let m > 2 . Let a E V am-2(a)) such that (a0)m-2(a) f 0 , a : = a and A : = (a,aa Then dim A
g
: KxN
-+K
=
m-l
and
a(A) c A
such that for
.
,...,
0
be
.
Again we ask if there is a function m-2 h *x : = 1 = 0 g(l,i)(x$)i(x) , (V,w,K,*)
.c
m- 7
is a near vector space. Then by (6) (x*a) (a) =igog(h,i)ai+l(a). B If we further follow (4)b) and make the assumption that the linear i map (X*a)* is a linear combination of the a then
Now (V'l)
gives us g(a +p,i) al(a> =C(g(x,i) + g(p ,i>)ai(a> j+i+'l + CCg(h,j) g(p,i> a (a) and by comparing the coefficients we obtain for all i with al(a> + O :
The part 2. of b) of the following theorem one obtains bycalculation.
H. Karzel
142
('16) Theorem. Let rp : V --tAut(V,+) be a coupling such that for each x f V the map x v := xcp- id is linear, there is a m E N , m>'l with (x~)~-'' (x) = O at least one y E V
for all x E V
.
a)
(V,&,K,.)
If m = 2 then
and
40
(~,)~-'(y)
for
is a near vector space,
b) Let m > 2 . 'I. If there is a function g- : K x M + M with m- 2 iGog(l,i)(xk)i(x))v E (X ,xw,...,x 2 m-2 ) for all x E V such that
(
w
with
(V,@,K,*)
space then K
igo g(A,i)(x
i ) (x) becomes a near vector I is a commutative field with char K = O or 2 m , A *x : =
m-j
-
T=g(k,O) is a monomorphism of the field K , g(X,i> = (i:,,) for all i E {O,'l, m-2) and 0 fulfills (N'). 2. If K is a commutative field with char K = O or z m , h + y a monomorphism of K such that (N') holds then (V,@,K,*) with m-2 A *x : = .C ( (x >i(x> is a near vector space. 1=0 ' V I
-+
...,
i2,,)
Remark. 5. If the map 0 : V +End(V,+) is linear (i.e. for A €K) then (N') is valid (x+y), = x I + y o and (Ax), = Ax
I
(C'
hence
'
1
a + b +(a (b))yf=(a+b+a (b)) = a + b + a b P v v 0 0 1 i+l (av(b))v=a b in particular ((x,) (x))$= x)
because then
v v
.
v v
Remark. 6. By [J1,(4.'lO) the fibrations belonging to the nearvector spaces o f theorem (9) are kinematic so that by (2) we have either d.g.(E(V,R)) = E(V,R) or d.g.(E(V,R)) is trivial. It remains the question whether there are other couplings in the sense of theorem ('16) leading to near vector spaces whose fibrations are not kinematic. In the theorems (8)b) and ('16)a) we considered couplings with x (x) = O for all x € V . F o r these maps we have: 0
('17) Let u, : V +Aut(V,+) be a coupling such that for all x,y E V , xI :=xcg- id is semilinear, x (x) = 0 and x y = 0. Then V v v a) x,(Y) = -Y,(x) , x @ y O x = y + 2x (y) I b)
( X + Y ) , = X ~ + Y ~ (Ax), , =Ix
i.e.
$ :
(V,K) +Ehd(V,K)
is a
semilinear map. c)
vr
d)
The fibration R : = {K.x I xEV\{O] of the near vector space (V,fb,K,*) is kinematic if and only if char K = 2 o r x is I linear for each x € V
=
u
.
143
Couplings and derived structures
If U)V
e)
then E((V,@),R)
Proof. a) By =
x,(x)
=
0
(xJI+ y,>(x + y + xv(y))
= ( A CK
IT'= A )
, xhyY= 0 and (C'), =
x,(y)
0 = (x(i~)~(x&y) =
x ui (Y) = -ye(x)
+ yv(x> hence
,
-
(X+Y)~(Z) =-zY (x+y) = - z (x) -z,(y)
b) By a),
by (5) (XX>,(Y)
=
-yyl(XX)= -XYui(x)=kui(Y)
.
=x,(z) ty u (z) and
.
. .
c) Let u € U and x € V then 0 = u0 (x) = - xe ( u ) hence u E V r Now v E V r and x C V implies 0 = xIk (v) =-vUI (x) thus v € U d) is a consequence of a). e) Since U + V there are x,yEV\U with xy(y) 40. Then h(x&y)=Xx+Xy+Ax 0 (y) and by b) X x @ X y = X x + h y + X 2 x 0 (y) hence XEE((V,@),R) if A2=A. (Examples for (17). Let K be a commutative field and K 4K an automorphism of K then the map 0 : K3 4 End(K3,K) defined by (a,l,a2,a3)$(xl ,x2,x3) = (0,0,a,,x2 - a2x,,) has the prop('18)
-
:
- -
erties of ('17). If K = z (T) is the quadratic extension field of 5 z 5 with t 2 = 2 then E((V,e),ST)=(O,'l,2+2t,2+37J. Remark. 7. If (V,m,K,*) is a near vector space of the type of theorem (8)c) and if ('I,@) is not commutative then E((V,Q),R) = {a € K I A ) (cf.[41, p. 206).
=
x2=
REFERENCES ]I'[ Karzel, H., Unendliche Dicksonsche Fastkorper. Arch.Math. ('1965),
[21
[3] C41
C51 [6]
[7] [8]
[9] ['lo]
'16
247-256
-, Affine
incidence groups. Rend. Sem. Mat. Brescia Vol. 2 ('1984), 409-425 -, Fastvektorraume, unvollstandige Fastkarper und ihre Abgeleiteten Strukturen. Mitt. Math. Sem. Giessen, Coxeter-Festschrift , Teil IVI ' ( 984) , I' 27-7 79 -, and C.J. Maxson, Fibered groups with non-trivial centers. Res. Math. 2 ('l984), '192-208 -, and C.J. Maxson, Kinematic spaces with dilatations. J. Geometry 2 ('1984), '196-20'1 -, K. Sorensen and D. Windelberg, Einfiihrung in die Geometrie Gottingen I' 973 Marchi, M. and C. Perelli Cippo, Su m a erticolare classe di S-spazi. Rend. Sem. Mat. Brescia 4 (1979p, 3-42 $5.12, G,, Near-Rings. North-Holland Math. Studies 23, '1977.
'1983.
Pokropp, F., Gekoppelte Abbildungen auf Gruppen. Abh. Math. Sem. Univ. Hamburg 2 ('l968), '147-159 Waling, H., Bericht uber Fastkorper. Jber. Deutsch. Math. Verein. 76 ('1974). 'l-'l05 - _ . _ ,4
[ 11 1 Wahling
-H .
,
I
Theorie d e r Fastkiirper , Thales Verlag
E s s e n 1987.
This Page Intentionally Left Blank
Near-rings and Near-fields, C. Betsch (editor) 0 Elsevier Science Publishers B.V.(North-Holland), 1987
145
MAXIMAL IDEALS I N N E A R - R I N G S H e r m a n n KAUTSCHITSCH I n s t i t u t f u r Mathematik Un iv e r s it a t K 1 a g e n f u r t 9022 K1 a g e n f u r t The i n t e n t i o n o f t h i s p a p e r i s , t o r e l a t e t h e i d e a l s t r u c t u r e o f a s p e c i a l n e a r - r i n g N t o t h a t o f some h o m o m o r p h i c In particular image N ' o f N o r t o i t s c o n s t a n t p a r t N one g e t s i n f o r m a t i o n a b o u t maximal i d e s l s . Such a r e l a t i o n s h i p i s w e l l known i n p o l y n o m i a l - o r f o r m a l p o w e r s e r i e s r i n g s a n d i t t u r n s o u t , t h a t one i s s u c c e s s f u l i n f i n d i n a such a r e l a t i o n s h i p f o r those n e a r - r i n g s , which have a 1i t t l e b i t o f a "polynomial-" o r "power s e r i e s - s t r u c t u r e " .
.
1.
IDEALS AND H O M O M O R P H I S M S T h r o u g h o u t t h i s p a p e r we assume t h a t (N,+,,)
i s a r i g h t near-
r i n g w i t h i d e n t i t y 1 and R i s a c o m m u t a t i v e r i n o w i t h i d e n t i t y . R[[x]]
denotes t h e s e t o f a l l f o r m a l power s e r i e s o v e r R ( i n one
indeterminate x),
R o [ [ x ] ] t h e s e t o f a l l f o r m a l power s e r i e s o f
positive order. F i r s t we s h a l l t r y t o d e s c r i b e m a x i m a l i d e a l s b y c e r t a i n homomorphisms. We s t u d y t h e c a s e ,
i s a near r i n s w i t h i d e n t i t y 1,
t h a t (N,+,O)
w h i c h h a s a n e p i m o r p h i s m h : N-+N', o f i n v e r t i b l e elements i n (N,o).
such t h a t h - ' ( l ' )
consists only
h-'If')s 1").
i.e.,
ExampLes: 1 . Any f i n i t e d g n r N w i t h i d e n t i t y : By P i l z [ 4 1 , 6 . 3 1 we h a v e h ( I ( N ) ) = I ( h ( N ) ) .
... ) : = P o ;
we s e t h ( p .+ p 1 x + p 2 x Z +
2. F o r N = ( R [ [ x ] l , + , . , )
i s a n e p i m o r p h i s m f r o m N o n t o (R,.),
then h
such t h a t h-1(l)=(l+plx+p2xz...)
c o n s i s t s o n l y o f i n v e r t i b l e elements o f (N,.). 3. F o r N = ( R o [ [ x ] l + , ~ ) , where
series,we
set h(r,x+r2xz+
o n t o (R,.),
... ) = r l ;
such t h a t h - ' ( l ) + + r 2 x
t i b l e elements o f (N,o) THEOREM I :
d e n o t e s t h e c o m p o s i t i o n o f power t h e n h i s an e p i m o r p h i s m f r o m
'...
f
(see [ I ] ) .
I f ( M o i } i s t h e s e t o f a l l maxima
e p i m o r p h i c image N ' ,
N
consists only o f inver-
then fh-'(MIi)f
ideals of the
i s the set o f
aLL m a x i m a l
i d e a l s o f N.
P ~ u o ~1 .: h - ' ( M I i ) w i s e 1~ h - ' ( M U i )
otheri s a maximal i d e a l o f N h-'(MIi)+N, a n d t h e n l ' : = h ( l ) E M I i a n d M I i N'. I f 8,4N w i t h
H Kautschitsch
146
,
B 3 h-' (MIi
then h ( B ) 3 MIi,
hence h ( B ) = N ' and l ' e h ( B ) .
Since
' ) i s i n v e r t i b l e i n ( N , o ) , we g e t l = b o v e B a n d B=N. ( T h e 1 n o t a t i o n b = h - (1') means, t h a t we c h o o s e a n e l e m e n t b e N w i t h
b:=h-'(l
h ( b ) = l ' .) 2.
I f M i s any maximal i d e a l o f N,then h(M) i s a maximal i d e a l o t h e r w i s e 1 ' E h ( M ) and M c o n t a i n s t h e i n v e r t i b l e
i n N':h(M)+N',
h e n c e 1 = m o u s M a n d M=N ( a g a i n t h e n o t a t i o n
element m:=h-'(l'), m:=h-'(t')
means, t h a t we c h o o s e a n e l e m e n t m r N w i t h h ( m ) = l ' ) .
T h e r e f o r e t h e r e e x i s t s an element a e N ' \ h(M). e x i s t s an element b r N \ M w i t h h ( b ) = a .
Hence t h e r e
By t h e m a x i m a l i t y o f M we
g e t M+(b)=N. By [ 5 1 , p. 111, t h e r e a r e mEM and p € P o ( N ) := ( P ( N ) ) o w i t h l=m+p(b). Here P(N) i s t h e s e t o f polynomial f u n c t i o n s on N ( i . e .
the near-ring
generated by i d N and t h e c o n s t a n t maps on N). Since l ' = h ( l ) = h ( m ) + p ' ( h ( b ) ) = h(m)+p'(a) w i t h p ' € P o ( " ) ,
we g e t 1 ' E h(M)+(a) and t h e r e f o r e h(M)+(a)=N', so
h(M) i s maximal i n N.
COROLLARY I:
( i ) A l l maximal i d e a l s o f t h e f o r m a l power s e r i e s o v e r a commutative r i n ? R w i t h 1 a r e g i v e n by
r i n g (R[[xIl,+,.) the ideals
Mi+(x)
,
w h e r e Mi
i s any maximal i d e a l o f R.
( i i ) A l l maximal i d e a l s o f t h e z e r o symmetric n e a r - r i n g o f f o r m a l power s e r i e s ( R o [ [ x f l , + , o )
over a commutative r i n g R w i t h 1
a r e g i v e n b y t h e i d e a l s {mlx+m2xP+
... I ml
E Mi]
,
w h e r e Mi
i s any
maximal i d e a l o f R.
P t l u o , j : ( i ) By T h e o r e m 1 we d e s c r i b e a l l m a x i m a l i d e a l s M o f (R[[xll,+,.)
w i t h t h e e p i m o r p h i s m o f E x a m p l e 2,
...
x+m x 2 + 1 mi 6 M i 1 2 (R,+,.), SO M=M.+(x). \m,+m
] ,
where Mi
hence M=h-'(Mi
)=
i s any maximal i d e a l o f
1
( i i ) S i m i l a r l y we d e s c r i b e a l l maximal i d e a l s M of (R,"~ll,+,o) w i t h t h e e p i m o r p h i s m o f E x a m p l e 3, h e n c e M = h - ' ( M i ) = { m x+m2xz+ 1 (R,+, . ) .
... I m l c M i 1 ,
where Mi
i s any maximal i d e a l o f
COROLLARY 2: ( i ) A non-simple f i n i t e dgnr N w i t h 1 i s never a subdirect
product o f simple near-rings.
( i i ) The r i n g R [ [ x l l o f f o r m a l p o w e r s e r i e s o v e r a c o m m u t a t i v e r i n g R w i t h 1 i s never a subdirect
product of f i e l d s ,
even i f R
i s such a r i n g . ( i i i ) The z e r o s y m m e t r i c n e a r - r i n a R , [ [ x ] ]
o f f o r m a l power
s e r i e s over a commutative R w i t h 1 i s never a s u b d i r e c t o f simple near-rings,
even i f R i s a s u b d i r e c t
P t l u o d : L e t Rad N ( o r R a d N ' ,
product
product o f f i e l d s .
r e s p e c t i v e l y ) denote t h e i n t e r -
s e c t i o n o f a l l maximal i d e a l s o f N ( o r N ' ) and J ( R ) t h e Jacobson
Maximal ideals in nearrings
147
r a d i c a l o f R.
N ( o r R, r e s p e c t i v e l y ) i s a s u b d i r e c t product o f s i m p l e n e a r - r i n g s ( f i e l d s ) i f f Rad N=O ( J ( R ) = O ) . By T h e o r e m 1 we get: = A h - l ( M I ) = h - ' (AM ' ) = h - ' (RadN ) Rad N = n M M maximal M ' maximal in N in N' I
( i ) I f N i s n o t simple,
then there exists a non-injective
N and h - l ( R a d ( N / I ) ) + O , even i f Rad(N/I)=O. ( i i ) By C o r o l l a r y l ( i ) R a d ( R [ [ x ] ] , + , . ) = J ( R ) + ( x ) ~ O , even i f
e p i m o r p h i s m h:N-+N/I,
14
J (R)=O. ( i i i ) By C o r o l l a r y l ( i i ) R a d ( R o [ [ x l l , + , ~ ) = \mlx+m2x2+
.../ml
E J ( R ) ] PO,
even i f J(R)=O.
To g e t some i n f o r m a t i o n a b o u t m a x i m a l l e f t - i d e a l s o f N , we i n t r o d u c e an a d d i t i o n a l a s s u m p t i o n on N: I ( N ) i s c o m m u t a t i v e and e,~e2=mo(n+el)-m,n Vel,e2E
w i t h s u i t a b l e m,n E N ,
I(N).
Exampbeb: 1. Any n e a r - r i n g ,
i n which t h e i n v e r t i b l e elements
a r e z e r o symmetric and p e r m u t a b l e : elne2=e20(0+el)-e200. 2 . The z e r o s y m m e t r i c p a r t R o [ x I o f t h e p o l y n o m i a l n e a r - r i n g (R[x],+,o).
I n R o [ x 1 , + , ~ ) a l l i n v e r t i b l e elements a r e g i v e n by t h e
p o l y n o m i a l s ex,
e a u n i t i n (R,.),
which are permutable w i t h
respect t o o . THEOREM 2 :
I f { L a i } i s t h e s e t o f a l l maximal l e f t i d e a l s o f
t h e e p i m o r p h i c image N ' o f N,
then {h-'(Lti)]
i s the set o f a l l
maximal l e f t i d e a l s o f N. P4ua6: 1 . N:h-l(L'i)PN,
I f LIi
i s maximal i n N ' ,
otherwise 1 E h-'(LIi)
then h-'(Loi)
and l ' = h ( l ) € LIi.
i s maximal i n By t h e a b o v e
a s s u m p t i o n we g e t :
n ' = l ' ~ n ' = p ' ~ ( q ~ + l ' ) - p ' ~ qe 'L I i v n ' E N ' , I f B slN
and B = h - l ( L ' i ) y
so LIi=N'.
t h e n h ( B ) = N ' and
1 ' 6
h(B).
b & B , so B c o n t a i n s t h e i n v e r t i b l e e l e m e n t b : = h - ' ( l ' ) .
i.e.,
l'=h(b),
Then
l = b o u = m o ( n + b ) - m o n E B a n d B=N. 2.
I f L i s any maximal l e f t i d e a l o f N, t h e n h ( L ) i s a maximal
l e f t ideal o f N': h ( L ) + N ' , o t h e r w i s e l ' ~ h ( L ) ,h e n c e l ' : = h ( l ) l:=h-'(l') a:=h-l(a')c
i s i n v e r t i b l e i n L,
So l = l + p ( a ) w i t h some z e r o
Then
w i t h p ' c P,(N'), 1 ' e h ( L ) , hence ' h (~L ) + ( a ' ) a n d t h e n h ( L ) + ( a ' ) = N ' , s o h ( L ) i s m a x i m a l i n N ' .
l'=h(l)=h(l)+p'(h(a))=l'+p'(a') 1
s o L=N. Now l e t a ' € N ' \ h ( L ) , t h e n
N \ L. B u t then L+(a)=N.
s y m m e t r i c p o l y n o m i a l p~ P,(N).
f o r some l c L a n d
H. Kautschitsch
148
I f N i s a f i n i t e and d i s t r i b u t i v e l y g e n e r a t e d n e a r -
C U R O L L A R Y 3:
r i n g w i t h commutative I ( N ) , t h e n a l l maximal l e f t i d e a l s a r e g i v e n by t h e i d e a l s
{
,
h-'(LIi)}
where L I i
i s any maximal l e f t i d e a l o f
an e p i m o r p h i c image h ( N ) o f N .
2.
IDEALS A N D THE C O N S T A N T P A R T N c Now we c o n s i d e r t h e c a s e t h a t N d o e s n o t h a v e a n e p i m o r p h i s m
o f t h e type given i n s e c t i o n 1.
I n order t o describe the ideal
s t r u c t u r e o f N by t h a t o f i t s c o n s t a n t p a r t N c , N an a d d i t i o n a l m u l t i p l i c a t i o n - i s d e f i n e d , ( i ) (N,+,.,o)
we a s s u m e ,
that i n
such t h a t
i s a composition-ring
aN 3 s e N : n o ( m + i ) = n o m + i . s
(ii)vn,m,i T h i s means,
i n particular,
Exampeen: ( R [ X ] , + , . , ~ )
t h a t (N,+,.)
i s a ring.
and ( R N [ [ x l l + , . , o ) ,
where R " [ x l l
de-
n o t e s t h e s e t o f f o r m a l power s e r i e s w i t h n i l p o t e n t i n i t i a l c o e f f i c i e n t (see [2] ) . A nonempty s e t I i s an i d e a l o f a c o m p o s i t i o n - r i n a N w i t h property ( i i ) i f f ( i ) i - j€ 1 ( i i ) i,neI ( i i i ) i.n C I
and n . i C I
vi,j
€1, neN.
F i r s t we i n t r o d u c e t w o s p e c i a l i d e a l s ( s e e a l s o [ 3 ] ) : 1.
If I g(N,+,.,o),then
IC~(Nc,+,.,o,): V i , j e I c , c eNc. 2.
i - j E I n N c ,
If 1 ' s (Nc+,.),
we s e t I c : = I n N c 9 ( N c , + , . ) i.cEInNc,
a n d we g e t
c . i ~ I n C'N i o c r I n N c ,
then ( 1 ' ) denotes t h e i d e a l o f N generated
[>
pioj'ilpi e P O ( N ) , j l i t I ' > , where P o ( N ) finite denotes t h e zero symmetric polynomials over N (see i51). b y 1': ( I 1 ) =
R e m a r k : The r i n g - i d e a l s i d e a l s o f Nc,
3.
I' o f
Nc c o i n c i d e w i t h t h e c o m p o s i t i o n
because i c c = i f o r a l l i E I' and c € N C .
I f I'<(Nc,+,.),
t h e n [ I l l : =n{ ENlnoc 61' vctNc]QN:
F o r a l l i , j t [ I ' ] ,n t N we a e t : ( i - j ) o c = i o c - j o c E I , h e n c e i - j f [ I ' l , ( i ~ n ) o c = i o ( n o c ) = i o Ec ' I ' w i t h c ' E N ~ , h e n c e i o n € [ I 1 ] , (i.n)oc=(ioc).(noc)=(io ).c'E I ' ,
-
h e n c e i . n E [ I 'a n]d s i m i l a r l y
n . i E [ I l l . We l i s t some p r o p e r t i e s o f t h e s e t w o i d e a l s : L e t I ' , J ' d e n o t e i d e a l s o f Nc: 4. [ I ' I c = I ' I f i E. [ I l l c ,t h e n i t [ I 1 n N c y S O i = i o c e 1 ' . I f i 6 1 ' 3 Nc, t h e n i o c = € 1 ' f o r a l l c e N c , hence i e [ I ' ] .
Maximal ideals in near-rings
149
5 . I f I ' c J ' , t h e n [ I ' ] c [ J ' l and ( I ' ) C ( J ' ) a n d v i c e v e r s a . L e t i e [ I ' I , t h e n i o c t I ' C J ' Y c6NC, h e n c e i t ' [ J ' l . I f i e ( ~ l ) t, h e n i = x p i o j i w i t h jisO mod I ' c J ' , S O jizO mod J ' and p i o j ~ p i o O = O mod J ' and i e ( J ' ) . 6. [ I ' n J ' l = [ I ' l n [ J ' l and ( I ' n J ' ) = ( I ' ) n ( J ' ) . T H E O R E M 3 : For a l l i d e a l s I SN t h e r e e x i s t s a n u n i q u e l y d e t e r minded i d e a l I , $ ( N C , + , . ) , such t h a t ( I , ) S I C [ I c ] . I c i s c a l l e d t h e e n c e o b i n g i d e a L of I . PI LOO,^: We s e t I c = I nNc. If i € ( I c ) , then i = 7 p i o j i w i t h j i E I c and p i t . P o ( N ) . finite we g e t from jizO mod I a l s o p i o j i ~ p i o O = O mod I , Since Ic 5 1 hence i € 1 . I f i € 1 , t h e n i , C C I n N c = I c f o r a l l c eNC, h e n c e i & [ I c ] . I f K i s a n o t h e r s u c h an i d e a l w i t h ( K ) 5 I 5 [ K ] , t h e n (K) 2 [Ic: a n d eNc, h e n c e K G I , . f o r a l l k 6 K we g e t k = k o c c I C ,
v
From (I,)c[K]
we get I c 5 K in a similar way, hence K=Ic.
C O R U L L A R Y 4 : The i d e a l l a t t i c e of N i s a homomorphic image o f t h e ideal l a t t i c e of N . P/zo06: We s e t : h ( I ) = I c f o r I 5 N . a ) h ( 1 ) i s u n i q u e l y d e t e r m i n e d by Theorem 3 . b ) h ( I n J ) = (I r\ J ) c = (I r \ J ) ' \ N c = ( I oNc) A ( J o N c ) = I c A J c = h ( I ) n h ( J ) c ) h ( I t J ) = (I t J ) n N c = ( I n N c ) t ( Jr l N c ) = I c + J c = h ( I ) t h ( J ) .
To g e t a l l maximal i d e a l s , we assume f o r N : ( i ) ( N , . ) h a s an i d e n t i t y l c " N c . or ( i i ) N, g e n e r a t e s N , i . e . ( N c ) = N . T H E O R E M 4 : All maximal i d e a l s M of ( N , t , . , o ) a r e o i v e n by t h e i d e a l s M = [ M ' I , where M ' i s a maximal i d e a l o f ( N c , + , . ) . Phoo6: a ) I f M ' i s maximal i n N c , t h e n [ M ' ] i s maximal i n N : F i r s t [ M ' ] + N , otherwise [M'],=M'=N,. I f t h e r e e x i s t s a proper ideal B B N with N 3 B 7 [ M ' l , then B c 3 M ' , o t h e r w i s e B C [ B c l C [ M ' l , so Bc=N. By p r o p e r t y ( i ) : l c € B c & B and n = l c . n E B V n E N , s o B = N . By p r o p e r t y ( i i ) : I f B c = N c , t h e n N = ( B c ) C B , s o B = N . otherb ) I f M i s maximal i n N , t h e n M c i s maximal i n N c : M c 9 N , w i s e we g e t b y s i m i l a r a r g u m e n t s a s i n a ) : M = N . I f t h e r e e x i s t s a proper i d e a l A 4 N C with McC A CNc, then M E [McI c [ A I c [ N c I = N i n
H. Kautschitsch
150
c o n t r a d i c t i o n t o t h e m a x i m a l i t y o f M i n N.
C O R O L L A R Y 5: A l l m a x i m a l i d e a l s M o f t h e p o l y n o m i a l compos i t i o n - r i n g (R[X],+,.,~) o v e r a commutative r i n g w i t h i d e n t i t y a r e g i v e n by t h e i d e a l s M=(M':R), whereM' i s a maximal r i n g i d e a l o f R.
C O R O L L A R Y 6: to
I f N has o n l y t h e t r i v i a l a n n i h i l a t o r w i t h r e s p e c t
and g e n e r a t e s N o r possesses an i d e n t i t y w i t h r e s p e c t t o
m u l t i p l i c a t i o n , then t h e composition r i n g N i s a s u b d i r e c t product o f simple composition-rings
i f f N,
i s a subdirect
product o f
simple rings. Ph006:
Rad N = [ R a d ( N c ) ] b y Theorem 4 a n d P r o p e r t y 6, s o Rad N = O o t h e r w i s e t h e r e e x i s t s a n nlO w i t h noc=O v c eNC.
i f f Rad(Nc)=O,
C O R O L L A R Y 7: The p o l y n o m i a l c o m p o s i t i o n - r i n g ( R [ x ] , + , . , o ) an i n f i n i t e i n t e g r a l domain i s a s u b d i r e c t composition-rings
over
product o f simple
i f f R i s a subdirect product o f f i e l d s .
REFERENCES
[ I ] J e n n i n g s , A.S., [2] [31
[41 [51
S u b s t i t u t i o n g r o u p s o f f o r m a l power s e r i e s , 6(1954), 325-340. Canad.J.Math. K a u t s c h i t s c h , H. a n d M u l l e r , W.B., Ideale i n Kompositionsringen formaler Potenzreihen m i t nilpotenten Anfangskoeffi34(1980), 517-525. z i e n t e n , Arch.d.Math. L a u s c h , H. and N o b a u e r , W . , A l g e b r a o f P o l y n o m i a l s ( N o r t h 1973). H o l l a n d , Amsterdam, P i l z , G., N e a r - r i n g s ( N o r t h - H o l l a n d , Amsterdam, 1977). P i l z , G., N e a r - r i n g s , w h a t t h e y a r e a n d w h a t t h e y a r e g o o d f o r , C o n t e m p o r a r y M a t h . , 9(1982), 97-119.
Near-rings and Near-fields, G . Betsch (editor) 0 Elsevier Science Publishers B.V. (North-Holland), 1987
151
D. G. NEAR-RINGS ON THE INFINITE DIHEDRAL GROUP S.J. MAHMOOD and J.D.P. MELDRUM Department of Mathematics, University of Edinburgh, King’s Buildings, Mayfield Road, Edinburgh, EH9 3JZ. Scotland.
Department of Mathematics, University Studies Centre for Girls, King Saud University, P. 0. Box 22452, Riyadh, Saudi Arabia.
In this paper, we look at the behaviour of a l l the d. g. near-rings that can be defined on the infinite dihedral group. We are mainly interested in determining which d. g. near-rings are faithful and in characterizing the lower and upper faithful d. g. near-rings for those which are not faithful. The upper and lower defects are also found in all these cases and in three cases they turn out to be not equal. 1. INTRODUCTION.
This paper is concerned with the d. g. near-rings on the infinite dihedral group. These have been determined by J. J. Malone 121. Using the techniques of J. D. P. Meldrum
[Sl, we determine the upper and lower faithful d. g. near-rings for
[51 and
near-rings and two sets of generators for each d. g. near-ring.
a l l the d. g.
Following [Sl.we also
determine the upper and lower defects and defects in a l l these cases. Before giving more details, we give a summary of the definitions.
We write maps on the right and hence we use left near-rings. As a general source for definitions and results we refer to J. D. P. Meldrum [71.especially chapter 13. See also G. Pilz [8l for a more complete survey of near-rings.
(R,+,.) is a left zero-symmetric
near-ring if (R,+) is a group with neutral element 0, (R,.) for a l l x,y,z
E
R and Ox = x0 = 0 for a l l x
Mo(G) := {f: G If s
E
+
G; 0Gf =
E
i s a semigroup, x(y+z) = xy + xz
R. The standard example of such a near-ring i s
OG}, and all near-rings are embeddable in Mo(G) for a suitable G.
R satisfies (x+y)s = xs + ys for all x,y E R then s is called distributive.
R is
distributively generated (d. 9.) if there exists a semigroup (S,.) contained in R such that (R,+) = Gp<S>.
End G, the semigroup of all endomorphisms of G, forms a distributive
semigroup in Mo(G) and generates the d. g. near-ring (E(G),End G). We always write d. g. near-rings
in the form (R,S)
as which semigroup S
is taken t o be the distributive
generating set for R is very important. Let
8: (R,S)
+
(RS),
(T,U) be two
(T,U) such that
d.
g.
near-rings.
a
Then
near-ring
homomorphism
SO 5 U is called a d. g. (near-ring) homomorphism. We say that
(R,S) is faithful if there is a d. g. rnonomorphism 8: (R,S)
+
(E(G),End G) for some group
G. Not a l l d. g. near-rings are faithful, but to each d. g. near-ring (RS) are associated two canonical faithful d. g. near-rings. near-ring
(&S) and a
The lower faithful d. g. near-ring for (R,S) i s a d. g.
d. g. epimorphism 8: (RS)
+
(R.2) such
that S8 =
2,and
for any d. g.
152
S.J. Mahmood and J.D.P. Meldrum
4: (R,S)
homomorphism
+
(T,U)
where
(T,U)
is faithful, there exists a unique d. g
6: (R.3)+ (T,U) such that 41 = 0$. The upper faithful' d. g. near-ring for * * * * ( R S ) is a d. g. near-ring (R ,S ) and a d. g. epimorphism 0: (R ,S ) + (R,S) such that 01s is an isomorphism of semigroups, and for any d. g. homomorphism 4: (T,U) + (R,S) where (T,U) is faithful, there exists a unique d. 9. homomorphism 9: (T,U) + (R*,S*) such that homomorphism
4
=
$0. For these results see Meldrum [31 for the prehistory and Mahmood [11 for the
details. There is one minor change of notation from the earlier papers, involving the upper To end t h e definitions we define g(R,S) := Ker((R,S)+(R,S)) and
faithful d. g. near-ring. D*(R,S) := Ker((R
* * ,S )+(R,S)),
the
lower
and
upper
defects
for
(R,S)
respectively.
D(R,S) := Ker((R*,S*)+(R,S)) is the defect for (R,S).
We need the idea of presentations from group theory. For each set X, we can define Fr(X), the free group o n X. Let R b e a set of elements in Fr(X), N the normal subgroup of Fr(X) generated G = Gp<X;R>.
by
R. Then
G := Fr(X)/N
We write <X;R>
is
for Gp<X:R>.
said
to
be given
by the
presentation
Presentations are far from unique: for a
given group G , different generating sets X can be chosen, and for each X different sets R of defining relations can be used. If we start with a presentation R = c S ; Y > for tho d. g near-ring ( R S ) we wish to obtain presentations of R * = <S*;Y*>, and lower faithful d. g. near-rings. multiplication in S
*
and
2
group we start with is D,
Multiplication in R
*
and
5
=
<s;y> for the upper
i s defined completely by the
(12.7, [71),which in turn is given by that in S. In this paper the the infinite dihedral group. J. J. Malone [2] has determined all
the d. g. near-rings (f3.S) such that (R,+) is an infinite dihedral group.
For those that are
not faithful, w e determine the upper and lower faithful d. g. near-ring for t w o sets S of generators, and the lower faithful d. g. near-ring in a number of other cases. We cover the lower faithful d. g. near-ring first, in section 2, then the upper faithful d. g. near-ring, in section 3. Finally w e consider defects in section 4. Before w e leave this section we quote the results we need from Malone 121. We write (R.+) as the infinite dihedral group, with what is probably the standard presentation
From [2] we know that there are six d. g. near-rings defined on (R,+): (i) Ro, the zero d. g. near-ring on (R,+).
This i s a distributive near-ring,
i. e. all
elements are distributive. (ii) R1, with multiplication determined by the following table
-y b
b
b
This near-ring is also distributive. (iii) R2, with
multiplication determined
distributive near-ring.
by the following table,
again
giving a
153
D.g.near-rings on the infinite dihedral group
-
a
b
a
b
O
b
0
0
(iv) R3. with multiplication determined by the following table, another distributive near-ring.
(v) R4, with multiplication given by
xy = y if x t G p < a > ,
= O i f x Gp ~ The distributive elements are precisely the elements of finite order (vi) R5, with multiplication given by
xy = y if x tGp, = 0 if x E Gp
The distributive elements are again precisely the elements of finite order. Note that both R4 and R 5 have left identities, namely any element not in G p i a > for R4. any element not in Gp
for R5. So they are both faithful (13.12, 171). For the
rest of the paper, therefore, w e will only be interested in Ro, R1, R2 and R3.
2 . THE LOWER FAITHFUL D. G. NEAR-RING. We first have t o establish the method w e will use. Essentially we are using a form of the method used in section 3 of [51, based on theorems 2.5 and 2.6 of that paper, and on lemma 4.7 of [31. See also theorem 13.32 of [71. We adapt these results as follows. Let ( R S ) be a d. 9. near-ring without identity. Without loss of generality, assume that
(R,S)is not faithful, and has lower faithful d. g. near-ring
(R.2). Adjoin an identity t o (5,S) { S } U (1). Then (T,+) is
as described in theorem 13.32 of [71 t o obtain (T,U) say, where U = the free product of (+ ,!)
(T,+) = + (,!)
and a free group on the single element 1. say
< 1>
In fact the free product, denoted by *, can be taken t o be in any variety t o which (R,+) belongs. Which variety we choose does not affect the argument. From the construction of
(R,?),
it follows that (T,+) is the free (R,S)-module on the element 1 (lemma 13.20 and
theorem
13.21
of
[7]).
Let
N = Ker 8,
where
8 : (R,S)
+
(5,s).
is
the
canonical
S.J. Mahmood and J. D. P. Meldrurn
154
. . .,
homomorphism. Then w(s1, lw(s1, . . ., sn) = 0 i n (T,+). w(s1,. .
., sn)
such
Sn), a word in the elements of S, is in N precisely when
Using theorems 2.5 and 2.6 of 151, we find those words
that
wfsl, . . ., sn) E (T’,+)r,
where
r E X
for
a
presentation
(R,+) = <S;X>, and (T’,+) = (R,+)
* < 1>.
We add these words to the relations in the presentation of (R,+).
We then repeat the
process starting now with the new presentation. Eventually we reach a presentation for This is the method described in theorem 2.6 of [51. In practice, by judicious choice
(!,+).
of the relations t o add on in the first stage, the process takes only two steps, not a possibly infinite number. Familiarity with the details of the structure of the infinite dihedral group is assumed. For those who are not familiar with it, any reasonable text-book or tame group theorist colleague will supply the necessary information. We will start by listing some properties of the four d. 9. near-rings (Ri,S), 0
5 i 5 3,
which are easily deducible, and are some help in
subsequent work, though not all of them are completely necessary for what we do later. We use the presentation given in (1.1).
To show that Rj, for j
= 0.1.2.3, is not faithful, we use (R
* *
,S ) The definition of
R* i s
given by theorem 2.2 of [61 or theorem 13.26 of [71, namely for each s E S, define sd as an endomorphism of (T*,+) := (R,+) *
< 1 > by ls* =
s
E
R, rs* = rs, and extending, as we may
by the properties of free products. Then the d. g. near-ring generated by S
*
*
in Mo(T ) is
(R*,S*). Lemma 2.1.
* *
Using the presentation ( l . l ) , we have that a
+
b
+
a
+
b i s not a relation
in(Rj ,S ) f o r j = 0,1,2,3.
Proof.
We just need to find g E T* such that g(a+b+a+b) # 0 in (R,*,+)
For Ro, let g = 1 + 1. Then
* * * *
g(a +b +a +b ) = 2a + 2b + 2a + 2b = 4a.
For R1, let g = 1 + a
-
1. Then
g(a*+b*+a*+b*) = a
+
b - a
+
b
+
b
-b
+
a + b
-a+b
=a+b-a-b+a+b-a-b=4a.
For R2, let g = 1 + a. Then
g(a*+b*+a*+b*) = a
+
b
For R3, let g = 1 + b. Then
+
b + 0 + a + b + b + 0 = 2a.
+ b
- b
155
D.g.nearrings on the infinite dihedral group * * * *
g(a +b +a +b ) = a + 0 + b
+
b + a + 0 + b + b = 2a.
Corollary 2.2. (R,,S) is not faifhful forj = Proof. This follows directly from lemma
*
0.1.2.3. 2.1, since a + b
+
a + b is a relation in (Rj,+),
but not in (R, .+). The normal subgroups of D,
are well-known.
This enables us t o determine all the
ideals of Rj, for j = 0.1.2.3. First we list the normal subgroups of Dm. This list can be found in [21 or in [41.
Lemma 2.3. The normal subgroups of@,+) given by (1.1) are R, (0) and N1 := <2a,a+b>, N2 := i 2 a , b > ,
hA :=, f o r h
2.4.
Corollary h
2
1.
The complere set of ideals of (R0,S) is given b y {Ro, N1, N2, hA;
2 0). Proof. This follows immediately from lemma 2.3 since in Ro, a l l products are 0. Theorem 2.5. (i) The complete ser of ideals of (R1.S) is given by (R1, N1, N2, hA; h an
even inregel). (ii) The complete set of ideals of (R2.S) is given by {Rg, N2, A, hA; h an even integelj. (iii) The complete set of ideals o f ( R 3 , S ) is given by(R3, N2, hA; h
2 0).
Proof. All three near-rings are distributive. So the ideals are normal subgroups H such that RjH
5 H.
HR,
5H
for each R,. The results are then found by a simple checking
process, using the l i s t i n lemma 2.3. Before w e start o n the determination of lower faithful near-rings, we remark that R/c2a> is an abelian group of order 4. Theorem 2.6. R,/
f o r j = 0,1,2,3, is always faithful, being a ring
Proof. This follows since a d. g. near-ring with an abelian group structure is a ring (Corollary 9.26 of [71). Another remark that needs t o be made is that since Rj, j = 0,1,2,3, is distributive, the semigroup S of distributive generators can be taken t o be any multiplicative semigroup that generates R. For each near-ring, we find the lower faithful d. g. near-ring for a number of different generating semigroups. Theorem 2.7. (i) lf S
5
Consider (R0.S). the zero d. g. near-ring on the infinite dihedral group
the set of elements of order 2, then (R0,S) is faithful
(ii) /f S = {a,b). then (I3o.S) is (Ro/<4a>,S),
the zero d. g. near-ring on the dihedral
group of order 8. (iii) l f S > {a,b) end S contains only one element of order 2, namely b, then (R0.S) is the same as in (ii). (iv) l f S
2
{a.b,a+b}, then
(50,s)is (R0/<2a >,S), the zero
ring on the group of order 4,
exponent 2. Proof. We use lemma 3.3 of [51. presentation
(R,+) = <S,X>
is
This states that (R,S) is faithful if and only if the
integrally
closed,
i.
e.
if
w(s1,
. . ., Sn)
E X
then
S.J. Mahmood and J.D.I! Meldrum
156
w(ms1,
. . ., ms,)
is a relation in R for a l l integers m. If S consists entirely of elements of
order 2, this is automatically satisfied, as w(ms1, w(s1,
. . ., sn)
. . .,
ms,)
= 0 if
m is even, and is
5 S.
The second example
if m is odd. This proves (i).
To proceed, w e note that (l+l)(a+b+a+b) = a + a + b + b + a + a + b + b = 4a So 4a must b e a relation in (@)
f o r j = 1,2,3, as long as {a.b}
on p. 55 of 151 gives us (ii). In (iii), S = (a,nja,b} for i E I. nja + b + nia + b, nia
- a -...- a
The relations are a
+
b
+
a
+
b.
o r nia + a + . . . + a, with an appropriate number of
summands, depending on whether n i is positive or negative. As for the second example on p. 55 of [51, we get again that the lower faithful d. g. near-ring is the same in case (iii)
as in case (ii). Finally we deal with case (iv). We are now in the position described in lemma 13.7 of [71. So in
(RO,~)w e
b = b + 5.
must have g +
This is sufficient t o prove the result.
We now turn t o R1. From theorem 2.5, we know that the ideals of R 1 are R1, N1, N2 and hA for all even integers h. The following result simplifies some calculations. Lemma 2.8. Ann(R1) = N1. -This result is immediate from the multiplication table.
Now a bit of notation which
will be used consistently in the rest of the section. Notation. An arbitrary word in (T',+) is of the form w := n i l + r1 + n21
+
, , ,
where n1 may be 0, rs may be 0, ni
+ n s l + rs E
2, ri E R1, for 1 5 i 5 s. Write
n : = n 1 + . . . + ns
Theorem 2.9. (R1.S) is faithful, where S = {b,a+b}. Proof. x := Q, y :=
A presentation for (R1,+) is given by R1 =
a+b.
b+b, a+b+a+b>.
Write
Consider w(x+x). Then
w(x+x) = n l b + r l b + . .
,
+nsb+rsb+nlb+rlb+.
. . +n,b+rsb
= 2(nl+m1+ , . , +ns+ms)b = 0
where rib = mib, m i E 2, f o r 1
i
5 s, from
the multiplication table. Now consider w(y+y).
Then
w(y+y) = nl(a+b)+O+ . . . +ns(a+b)+O+nl(a+b)+O+. . . +ns(a+b)+O = 2n(a+b) =
using lemma 2.8. Thus required.
0,
no new relations are introduced and so (Rl,{b,a+b}) is faithful, as
157
D.g.near-rings o n fhe infinite dihedral group Lemma 2.10.
Let (R1.S) be a d. g. near-ring. where 2ra + b
S. Then 4rg is a relation
6
in (Fl1.2).
Proof. Let
x := 2ra + b. Then x + x is a relation in (R1.f).
Consider (l+b)(x+x). We
obtain (l+b)(x+x) = 2ra +
& + b + 2ra + b + 5 = 4ra.
This suffices t o prove the result. We can use this in several cases. First we look a t the situation in which S consists of two elements of order 2.
Theorem 2.11.
Lets = {2ra+b.(Zrrl)a+b}.
Then (El,?) = Rl/<4ra>.
-
Proof. Lemma 2.10 shows that 4ra is a relation in (R1,+). Let x := 2ra
( 2 r i l ) a + b. Then wx = nl(2ra+5) 4r5 = 0,
it
m = mi +
follows
...
that
+ m,.
+
mlb
+
2ra commutes
.
,
. + ns(2ra+b)
with
b.
+
+ b, y :=
m&, where rix = mib. Since
Hence wx = n2ra + (n+m)b where
Also wy = nl((Zr*l)a+b) + 0 + . . . + ns((2ril)a+b) = n((Zr+l)a+b). We
now have w(x+x) = 2(nZrg+(n+m)b) = 4nra + 2(n+m)b = 0, w(y+y) = 2n((2rtl)g+b) = 0. Thus we have a new relation. namely 4r(x+y). Then
w(4r(x+y)) = 4r(wx+wy)
But in R1/<4ra>
all elements have order dividing 4r.
Hence w[4r(x+y)) = 0.
Thus
R1/<4ra> is faithful. Next we consider the case in which
S
= {b.a+b.Zra+b}.
Denote b by x, a+b by y and
Zra+b by z.
~Theorem 2.12. Let S = {b,a+b,2ra+b}.
Then (51.2) = Rl/.
Proof. Again lemma 2.10 shows that 4ra i s a relation in (El,+), and so 2ra commutes with
b.
By the above work, we know that the relations x + x, y + y and L + z do not
introduce new relations into (el,+). But we have also a relation in (R1,+) given by 2r(y+x)
+
x + z. Consider w(Zr(y+x)+x+z).
w(Zr(y+x)+x+z) = 2r(wy+wx) + wx + wz
= 2r(n(g+k)+(n+m)b) + (n+m)b + n2ra + (n+m)b where n = n1 + .
..+
ns, mb = ( r l +
. . . +rs)b
= (rl+
,
. . +rs)(2ra+b), since
We use a similar argument to that used in the proof of theorem 2.11.
(2.13) 2ra E Ann(R1). We divide the
consideration of this element into four cases, depending on whether n or m are even or Odd. (i) n even, m even. Then (2.13) is 0 as 4rg = 0. (ii)n even, m odd. Then (2.13) becomes 2rb +
+
b = 0.
S.J. Mahmood and J. D.P.Meldrum
158
(iii) n odd, m even. Then (2.13) becomes 2r(a+b+b) +
b+
2rg +
b=
2rg +
b+
2rg +
b=
0.
(iv) n odd, m odd. Then (2.13) becomes 2r(a+b) + 2rg = 2ra. Hence 2ra is a relation in 51. Having added this relation w e can repeat the process. The new relation is 2r(y+x). So w(2r(y+x)) = Zr(wy+wx). But in R1/<2ra> order dividing 2r. Hence w(2(y+x)) = 0 always. Thus R1/<2ra> Theorem 2.14.
L e t s = {b,a+b,(2r+l)a+b}.
all elements have
is faithful.
Then (51.2) = R1/<2ra>.
Proof. From the proof of theorem 2.9 we know that b + b and a + b + a + b do not
introduce new relations. As for a
+
b
does not introduce a new relation.
+
a
+
b, we can check that (2r+l)a + b
So we consider (2r+l)(y+x)
+
x
+
(2r+l)a + b
z, where x := b,
+
y := a + b and z := (2r+l)a + b. With n and m defined as in theorem 2.12, w e obtain, since
z
E
Ann(R1),
(2.15)
We consider separately four cases as in theorem 2.12. (i) n even,
m even. Then (2.15)
i s 0.
b=
(ii) n even, m odd. Then (2.15) becomes ( 2 r + l ) b +
0.
(iii) n odd, m even. Then (2.15) becomes (2r+l)(a+b+b)
= a
b=
+ (2r+l)g +
+
(iv) n odd, m odd. Then (2.15) becomes (2r+l)(g+b) + (2r+l)g
+
b=g
+
0.
Q + (2r+l)g + b
- (2r+l)g = -2ra. Hence 2ra is a relation in 51. Having added this relation w e can repeat the process.
The new relation is 2r(y+x). R1/<2ra>
As in theorem 2.12, this introduces no new relations.
Thus
is faithful.
There are a number of other results giving the lower faithful d. g. near-ring for (R1,S) where S consists entirely of elements o f order 2. They are somewhat more complicated than the previous cases. So w e list them without proof.
-Result 2.16. /f(r-r')/t
1. 1.
(i) Lets = {Zra+b,(2rfl)a+b,Zr'a+b}, with r # r', r,r'
is even, then(fl1,Z) = R1/<4ta>.
/f(r-r')/t
L e t t = H.C.F.(r,r').
is odd, then (FIl,S) = R1/<2ta>.
(ii) Let S = {2ra+b,(2r+l)a+b,(2r'+l)a+b}, with r # r', r,r'
Let t = H.C.F.(4r,2r-2r').
1.
Then (fi1,S) = R l / < t a > . (iii) Let S = {(2r-l)a+b,2ra+b,(2r'+l)a+b},
with r # r', r,r'
2
1. L e t t = H.C.F.(4r,2(r-r'-l)).
Th8n (51.5) = R l / < t a > . We now turn t o generating sets involving an element o f infinite order.
= R/<4a>.
Theorem 2.17. L s t S = (a,ra+b}, where r E. Z. Then (F11.3)
Proof. Write x :=a, y := ra + b. Then the relations in a presentation of R1 are y + y,
x + y + x + y. By considering (l+l)(x+y+x+y) = 2a + 2(ra+b) + 2a + 2(ra+b) = 4a, w e see that 4a is a relation i n !41. We f i r s t find wx and wy: wx = n l a ny, if y E N1, wy = n l y
+
mlb + . .
,
+ ny,
+
msb if y
Since 4g = 0, it follows that 25 commutes with
b.
4 N1.
+
mlb
+
-
. . . + nsg + msb, wy $ N1, then r is even.
Now if y
So wy = n y +
mb
if y
4
N1. Consider
159
D.g.nearrings on the infinite dihedral group
w(y+y). If r is odd, then y E N1, so w(y+y) = 2ny = 0 since y has order 2. If r is even, then w(2y) = 2(ny+mb). The possible values of ny + mb are ra. ra +
k, b,
0. Since r is even
Z(ny+mb) i s zero in a l l cases as 45 = 0.
look
Now
2(nla+m@+
at
w(x+y+x+y).
If
. , . +nsg+msb+n(ra+b)).
y
N1,
E
Suppose
r
then
n
is
is
odd,
even.
and
Then
w(x+y+x+y)
this
=
reduces
to
2(nla+mlb+ . . . +nsg+msb) and n = n1 + .
. . + ns. If m i + . . , + m, is odd, this is of the form Z(qZ+b) which i s 0. If m i + , . . + m s is even, this is of the form . , +Esnsg) = 2(n1+~2n2+., . +Esns)a where c i = f l for 2 i1 .s. But 2(nla+c2n2g+, n l + . , . + ns even forces n1 + c2n2 + . . . + cSnS even. So w(x+y+x+y) reduces to a multiple of
43
hence
is 0.
Now
suppose
. . . +nsa+msb+rg+b). Since n1 + 2(nla+mlb+ -
..
n is
that
odd.
Then
can repeat the above process to show that this is zero. Now l e t y and
w(x+y+x+y) =
- 2(nla+mlb+
.
2(nlg+rnlb+.
.
. +nsa+msb+ny+mb).
If
n
is
The
4 N1.
even, this
coefficient
m l + . . . + rns + 1 + m = 2m + 1. So this element i s of the form 2(qa + This covers a l l cases for the relation x + y The final relation is 4 3 i. e. 4x.
+
x
+
Then r is even becomes
If n i s odd, then this
. . +nsa+(ms+m)tj and can be treated as above.
2(nlg+m1Q+ . . . +nsa+msb+ra+b+mb).
becomes
w(x+y+x+y) =
. + ns + r is even (as n and r are odd), we
b
of
b)
is
which i s 0.
y.
But w(x+x+x+x) = 4wx = 0, since in R1/<4a>
all
elements have order divisible by 4. Thus no new relations are introduced and we have our result. Our final look at d. g . near-rings on R1 i s a t the case where Theorem 2.18.
Let
S = [a,b,ra), for r > 1.
(r-1)/2 is also odd, o r r is even andr/2 is odd
S
= {a,b,ra}, for r
m e n (Ft1.S) is R1/<2a>
>
1.
if r is odd and
Otherwise (El,?) isR1/c4a>.
Proof. Let x := a, y := b, z := ra. As at the beginning of the proof of theorem 2.17 we see that 45 is a relation in y + y, x + y + x + y, rx - z.
nlg+mlb+. nlz+m$+.
El.
The relations in a presentation for R 1 can be taken as
We first find what effect x, y and
. . +nsg+msE, wy = (n+m)k, as in . . +nS;+m& if L 4N1, i. e. r is odd,
z have on w: wx
=
the proof of theorem 2.17, and wz = and
WL
= nz if z E N1, i. 8. r is even. We
consider the innocuous relations first. We have w(y+y) = 2(n+m)F = 0 in a l l cases. Also w(x+y+x+y) = 2(nlg+mlh+. . . +nsa+msE+(n+m)b). This is of the form 2(qa+b) = 0 unless n i s even since m l
+
. ..+
ms = m. But then an argument similar t o that in the proof of
theorem 2.17 shows that we get 0.
Now we deal with rx
- z.
We have w(rx-z) = r(nig+rnib+
. . . +nsa+msb) -
wz. If r is
even then we have r(nla+mlb+ . . . +nsg+msb) - nre. If 4(r, then this i s 0 as R1/<4a>
has
exponent 4. If 2)r but r/2 is odd, then take n l = 1, rnl = 1, s = 1. We get r(a+b) - rg = -ra as r is even. But 4a = 0, -ra = 0 implies tg = 0, where t = H.C.F.(4,r) = 2, by hypothesis. So R1/<2a>
is (El,?) as it i s a ring (theorem 2.6).
r(nlg+rnlb+ . . . +nsg+mse)
-
Now let r be odd.
Then w(rx-z) =
. +nsra+msb). But nlg+mlh+ . . . +nsa+msb = n l + ~ 2 n p+ . . .. +ESnS for some choice of e i = fl,
(nirg+mib+. .
b, where m l + . . . + m, = m, q = 5 i 5 s. Then n1ratrnlb+ . . +nsrg+rnsb = qra + mb. So w(rx-z) = r(qi+h) - (qra+mb). If m is odd this is r(qe+b) + qra + b = 95 + b + qrg + b = q(1-r)g. By taking n1 = 1 = m l , qa +
2
,
160
S.J.Mahmood and J.D.P. Meldrum
s = 1, we get (1-r)a
0. If (l-r)/Z is odd, as before this leads t o R1/<2a>
=
= (I31.S).
If
41(r-l), then this is always zero. If m is even w(rx-z) becomes rqe - qra = 0 for all choices of q. The final check is whether 42 introduces new relations in the t w o cases 41r, 4\(r-l). has exponent 4, no new relations are introduced.
But as before, since R1/<4a>
This finishes our look at the lower faithful d. g. near-rings on R1. We turn t o R 2 next. Fortunately the work i s much easier here. The first result parallels lemma 2.8. Lemma 2.19. --
Ann(R2) = N2.
Again this is obvious from the multiplication table. The following result is useful.
-Lemma 2.20.
Ler (R2.S) be a d. g. near-ring
/f r is odd and ra + b E S, rhen 2rg = 0
in R2.
Proof. Put x b + rb + ra + b
.= r a + b. Then x + x is a relation.
So (l+a)(x+x) = x + ax + x
+
ax = r a +
rb = 2ra since b + rb = 0.
+
~Theorem 2.21. / f a + b
E
S , then(R2.S) = R2/<2a>.
Proof. Use lemma 2.20 and theorem 2.6. If a E S. rhen (t3.5) = R2/€2a
Theorem 2.22.
>.
Proof. Note that S must contain at least one element of order 2, say r a + b. Put x := a, y := r a
+
b. Consider (l+l)(x+y+x+y). As before this forces 42 = 0. If r is odd, then
lemma 2.20 forces 2 r a
0, which together with 42 = 0 gives 22 = 0. Another appeal t o
=
theorem 2.6 gives the result.
x_
+
12 + y + 5 + b + y.
So assume that r is even.
since (l+a)x = 5
2(1-r)2 = 0. Together with 4 :
+
5,
(l+a)y = y.
= 0 we again get
25
=
Consider (l+a)(x+y+x+y) =
So (l+a)(x+y+x+y) = Z(g+b+ra+Q) =
0 and the result follows from theorem
2.6.
Finally we look at the case when S = {2ra+b,(Zrfl)a+b} and r > 0. This gives the only non-ring lower faithful d. g. near-ring for R2. Theorem 2.23. Let S = {2ra+b,(Zrfl)a+b}.
Proof. Put
t = 2rfl.
Then (R2.S) = R2/c2(2r*l)a>.
That 2ta = 0 follows from lemma 2.20. We just need t o show
that no more relations come in. Let x := 2ra + b, y := ta + b. The relations w e do have are x + x, y
+
y, 2t(x+y). Also wx = nx, wy = nl(ta+lJ
0. ta commutes with
b.
+
rnlh + . . . +
ns(ta+tJ + m&.
As 2tg =
So wy = ny + (n+m)Q. Further w(x+x) = 2n5 = 0 as 2x = 0. Also
w(y+v) = Z(ny+(n+m)b). If n + m is even, this is 2ny which is 0. If n + m is odd, we get either 2b (m odd, n even) or 2(y+p) = 2ta (m even, n odd). Finally w(2t(x+y)) = Ztw(x+y). R2/<2ta>
In either case w(y+y) = 0.
has exponent 2t. So w(2t(x+y)) = 0. This finishes
the proof. This brings us t o the final d. g. near-ring t o consider, namely R3. We first note the annihilator of R3. Lemma 2.24. --
Ann(R3) = A.
This is again obtained directly from the multiplication table. Lemma 2.25. --
/f r a + b E S, rhen 2ra is a relation in R3.
-
Proof. Consider (l+b)(ra+b+ra+b) = ra + b + b + ra + b + b 2ra, -Lemma 2.26. /f {a,ra+b) 5 S, then 2(r+a)a is a relation in R3.
since b(ra+b) = b.
161
D.g. near-rings on the infinite dihedral group Proof. Consider -
(l+b)(a+ra+b+a+ra+b) =
5 + 0 + rg
+
+
21
+
5
0 + rg +
+
+
=
2(r+l)g. This leads t o our final result in this section. Theorem 2.27.
/f either a
E
S or a + b
E
S or {ra+b,(r+l)a+b} 5 S, then (Fl3.5) =
R3/< 2a >. Proof. If a -
E
S, then S must contain an element of the form ra
+
b as well.
By
lemmas 2.25 and 2.26 it follows that 2rg and 2(r+l)g are relations, hence so is 2g. since H.C.F.(r,r+l) = 1. If a + b
E
S, then lemma 2.25 gives the result. Finally if {ra+b,(r+l)a+b}
S, then 2rg and 2(r+l)g are relations, hence so is
25
as before.
5
Just apply theorem 2.6
now.
3. THE UPPER FAITHFUL 0. G. NEAR-RINGS. We will now determine the upper faithful d. g. near-rings for Ro, R1, R 2 and R3 for two sets of generators in each case, namely S = (a.b} and S = (b,a+b}
This i s done as the
determination of upper faithful d. g. near-rings is harder than that of lower faithful d. g. near-rings. For (Ro,(b,a+b}) and (Rl,(b,a+b}), the upper faithful d. g. near-rings are RO and R 1 respectively.
So we only deal with the remaining cases.
The method we use for
finding upper faithful d. g. near-rings is that of theorem 2.2 of [61. A useful result in calculating upper faithful d. g. near-rings is theorem 13.12 of [71,
which says that a free group in any variety is faithful. Hence it follows that if (R,+) lies in a variety, then so does (R',+). We quote this as a lemma.
-Lemma 3.1. Let (R,S) be a d g. near-ring. N(R,+) lies
in a variety, then so does (R ,+).
An application of this to our case is as follows. Since (Ri,+) is metabelian, i. e. has
*
abelian derived group, then so i s (Ri ,+). cases we consider.
*
Also Ri is generated by two elements in all the
So (Ri ,+) will be a homomorphic image of the free metabelian group
on two generators, a group which is easy to handle.
*
In practice this means that to
determine (Ri ,+) we need only determine the orders of the two generators, x and y say, find out what [x,yl := -x
-y+
x + y, is. and what the conjugates of [x.yl are. We write
conjugates in the form xy := -y + x + y. We establish the following notation for the rest of this section.
Let T := (Dm,+)
*
<1>, be the free product of an infinite dihedral group given by (1.1) and an infinite cyclic group generated by 1. Since we will generally be identifying Dar with (Rj,+) for some j, we write a general element of T as w = n i l + r1 + where ni
E
Z, ri
E
. .. +
n s l + rs,
D-, for 1 5 i
5s
and the first and last elements may be zero, s
We remind the reader that theorem 2.2 of endomorphisms of T by 1 . s = s,
ri.s = ris
(3.2)
2
1.
[Sl shows that if we define the elements of S as
S.J. Mahmood and J. D.P. Meldrum
162
S generates
then
R,* in Mo(T). We start with (Ro,{a,b}).
Note that in Ro, we may assume
without loss of generality that w = n l since ris = 0 in a l l cases with Ro. Then (Ro*,S) is given by
Theorem 3.3. Let S = (a,b}.
Ro* =
*
Proof. We
first check that 2b is a relation in RO . But n l .(b+b) = 2nb = 0.
(Ro*,+) =, metabelian.
Hence
a has infinite order and b has order 2, and by lemma 3.1, Ro
Consider
m l a + b + mga + b + .
an ,
,
arbitrary
element
(Ro*,+).
of
2
+ mta + b + mt+la, where mi E Z, t
is
It
of
the
*
is
form
1, and m i and mt+1 may
be zero. Then it i s a relation if
nl(m1a + b +
... +
mta + b + mt+la) = mlna + nb + .
for a l l n. If n is even, this means n(m1 + has
infinite
mlna + b +
order,
...+
this
forces
. ..+
mi +
- m2na
..-
+ .
m l - m2 + . . . - mZq + mZq+l = 0.
have
., .+
+
mtna
nb
+
+
mt+lna = 0.
mt+l)a = 0 for a l l even values of n. As a . + m t + l = 0.
If
n
is
odd,
then
mtna + b + mt+lna = 0. This means that there must be an even number
of b's, i. e. t = 2q and m l n a mla + b +
, ,
..
mZqa + b + r n z q + l a
m2qna + mzq+ina = 0.
Conversely
satisfies
annihilates n l for all n. This is equivalent to
*
C$, m2i
easy
is
it
C i mi = 0
to
check
X i (-1)'mi = 0,
and
= 0,
As before we must
,C,:
that
if
then
it
m2i+l = 0. So the set of
relations for Ro is
We now check what happens to the commutators. All commutators are words of the + m t + l a in which El::
form Cl,,(mia+b)
mi = 0. A standard technique from group theory
shows that words in commutators are precisely words of this form. So a l l the relations in
Ro*,
other than
We now consider [a,bla - [a,bl
are in the derived group.
Zb,
-2a+b+a+b+a+b-a+b+a,i.e.m1 =-2,m2=m3=1,m4=-1.m5=
- [b.al
makes it easy to check that this is a relation. Also we look a t [a,blb
=
1. This
= 0, simply
using 2b = 0. This finally shows t h a t the presentation in the theorem is that of (Roil,+). W e now move o n t o R1. Here again we have only one case, namely S = {a,b}. Theorem 3.4. Let S = (a,b).
Then (Rl*,S) i s given by
R1* = . Proof. We first shows
that
mla + b +
nla
the
is
Let w E T as in (3.2). Consider
a
relation
in
R1
*
.
An
arbitrary
As 2b = 0, this
element
Of
R1
*
is
mta + b + mt+la = u as in theorem 3.3. Consider wu. First we point out multiplication
. . . + nsa + + (n+k)b + . . . +
klb +
ml(wa)
2b
...+
from +
.
Then w.2b = 2fb for some integer f. as all products lie in .
w.2b.
that
*
check that 2b is a relation in R1
table,
we
have
=
ria
rib
=
kib
ksb, wb = (n+k)b, where n = C i ni and k = mt(wa) + (n+k)b + rnt+l(wa).
If
n
+
k
is
say.
xi
ki.
even
So
wa
=
Hence wu = this
gives
D.g.near-rings on the infinite dihedral group (ml+.
. . +mt+l)(wa)
163
= 0. Take n1 = 3, n2 = 1, k l = 1 = k2. Then wa = 3a + b + a + b = 2a.
Hence we must have ( m l + . . . +rnt+l)2a = 0 and so C i mi = 0. Then wu = 0 for all cases in which n + k i s even. If n + k i s odd, this gives ml(wa) + b
...+
f
mt(wa)
b + mt+l(wa) = 0. Take "1 = 1,
f
k1 = 2. Then wa = a and we must have m l a + b +
. ..+
proof
t = 2q.
of
theorem
.
m2 + m4 + . wa
By
ml(Za+b) + b ml
. + mzq = 0.
.
E
+
this
gives
us
Then wu = 0 for n1 = 2,
taking
a l l cases
k l = 1,
we
ml
+
+
As in the
m t + l = 0.
m3 + .
in which
have
. . . + mzq(2a+b) + b + mZq+i(2a+b).
. . . + rnzq+l
f
3.3,
mta + b
..
+ mZqtl = 0,
n + k is odd and
wa = 2a
+
The coefficients
So
b. of
wu
=
b add up to
+ 2q = 2q. Hence wu will always lie in c a > in this case. We rewrite
wu as ml(2a+b) + m2(-2a+b) + . . . + mgq(-2a+b)
+
mgq+l(2a+b). Note that mi(i2a+b) = 0
if mi i s even and is i 2 a + b if mi i s odd. So wu reduces to CiEJ((-1)i-12a+b), where i f J
I:?,+'
if and only if mi i s odd. Since even, it follows that E, =
IJ( is
even, say 2f.
_+1. So wu = ~ 1 2 a- E22a + .
need ~1 - € 2
+
da(C,EJE,(-l)i-')
. .
.-
mi = 0, and we have left out a l l mi such that mi is
..
f
This can be rewritten as CjEJ(E,2a+b), where
~ 2 f - 1 2 a- ~ 2 f 2 a = 2 a ( E 1 - ~ 2 + . . - ~ 2 f ) . So we
~ 2 =f 0. But if wa
4
,
we have wa = da + b, say. Then wu =
= 0. We conclude that the full set of relations i s
where J i s defined by i f J if and only if mi is odd, and j is a dummy variable which enumerates the elements of J. We now check what happens to the commutators. subset of those in (Ro*,+).
relations are in the derived group. Let c := [a,bl. power of d is equal to a power of c. nd - n'c, -a
f
or
The relations in this case are a
Hence (Ro*,+) i s a homomorphic image of (R1*,+).
So all
Consider d := [a,bla. We show that no
If not then nd = n'c say, i. e. we have a relation
n(-Za-b+a+b+a) - n'(-a-b+a+b)
n(-2a+b+a+b+a) + n'(b-a+b+a)
=
n(-a+b+a+b) + a + n'(b-a+b+a). Without loss of generality we may assume that
=
n >
0.
We have to consider separately n' > 0 and n' < 0. First let n' > 0. Then we have a relation m l a + b which q = n + n', m1 = -2, mi = (-l)i for 2 as
can easily
So
be checked.
i
5 2n,
J = (2, . . .,2q+l},
relation m l a + b +
5 i 5 2n-1,
. ..
f
mzqa
+
b in which
mzn = 2, mi = (-l)i for 2n+l
q = n
...
f
mzqa
f
b
Now let n'
-
5 i 5 2q.
n ' - 1, m l
< 0. =
f
mZq+ia in
< i 5 2(n+n')+l,
for 2n
and (-l)i+J = (-l)i+i-l
Z i E ~ ( - l ) i +=j -2q # 0 and this cannot be a relation.
2
f
mi = (-l)i+'
= -1.
Thus
Then we have a
-2,
mi = (-1)'
for
But then all mi for i odd are
negative. Hence C Z , m2i-1 # 0, and this cannot be a relation. So c and d are independent elements. Consider cb = -b
-a-b+
a + b
+
b = -b
-
a + b + a = [b,al = -c. Also, since R1* is
metabelian. it follows that db = ca+b = cbfa = ( - c ) ~= -(ca) = -d. -2a
-
-c
The final consideration is
i s a relation to finish the proof. But da - c = ca+a - c = We a - b + a + b + 2a - ( - a - b + a + b) = -3a - b + a + b f 2a + b - a + b + a.
da, We wish t o show that da
S.J. Mahmood and J.D.P. Meldrum
164
get, using the general form for u, m l = -3, -3+2;1=0,
1-1=0, J = (1,2,4,5}.
1+ 1- 1
-
1 = 0. Hence da
m2 = 1, m3 = 2, m4 = -1,
So ,Z,c~(-l)i+i = (-l)’+’
- c is
So
m5 = 1.
+ (-1)2+2 + (-1)4+3 +
(-lp4
=
a relation.
The next near-ring to consider is (R2.S) with S = {a.b} and {b,a+b}.
We consider
(R2*,{a,b}) first. Theorem 3.5.
Ler
Then (R2*,S) is given b y
S = {a,b}.
R2* = ca,b; 2b. c := [a,bl, d := [a,bla, da = c, cb = -c, db = -d>.
Proof. We first
check that 2b i s a relation in R2*. Let w
E
T as in (3.2). Consider
w(b+b). From the multiplication table wb = nb. Hence w(b+b) E 2 = {0}, for all w We now consider wa.
1
i
1.s.
We can write wa = n l a + k l b + .
Then wa = ml(wa) + nb
+
. ..
..+
E
T.
nsa + ksb, where ria = kib,
+ mt(wa) + nb + mt+l(wa).
From now on the
proof follows a line very similar to that of theorem 3.4 and leads to the same additive structure.
*
Now we turn t o (R2 ,{a+b,b}).
We, perhaps surprisingly, obtain the same additive
structure as before. Theorem 3.6.
Lei S = {a+b,b}.
Wriiex := a + b, y := b.
Then (R2*,S) is given b y
R2* = <x.y; 2y, c := [x,~], d := [x,yIx. dX = c, cy = -c, dv = -d>. Proof. First we show that x has infinite order. that
wx = n l x + k l b + . . . + nSx + k,b.
nl(a+b) + k l b + . . . + ns(a+b) + ksb.
where
We start by calculating wx. rix = kib.
15i
1.s.
So
We see wx
=
Take n1 = 1, k l = 1 to obtain wx = a + b + b = a.
Hence w(fx) = f(wx) = fa is never 0 unless f = 0. Thus x has infinite order. Next we show that y has order 2.
But wy = nb as in theorem 3.5. Hence 2y = 0. We now see that
wu = ml(wx) + nb + . .
.+
mt(wx)
+
nb + mt+l(wx). The rest follows as in the proof of the
two previous theorems. The last d. g. near-ring that needs t o be considered is (R3.S). We look a t two cases, namely S = {a.b} and S = {a+b,b}. as we have mentioned earlier. To maintain continuity, we consider S = {a+b,b} first. Ler S = {a+b,b}.
Theorem 3.7.
Write x := a + b, y := b. Then (R3*,S) is given by
R3* = <x,y; 2y. c := [x,y], d := [x,yIX. dX = c, cy = -c, dy = -d>.
Proof.
We start by looking a t wx and wy: wx = n l x + k l b +
rjx = kib, 1 5 i 5 s. Also
WY
...
+ nsx + k,b,
where
As in the proof of theorem 3.6, this forces x to have infinite order.
= nib + klb + .
. . + nsb +
ksb E .
As in the proof of theorem 3.4, we get 2y
as a relation. The rest of the pattern should be familiar from the last three theorems. This brings us to the last of the upper faithful d. g. near-rings, namely (R3*,{a,b}). Theorem 3.8. Ler S = {a,b}.
men
R3* = ca,b; 2b, [a,bla = [a,b], [a,blb = [b,al>.
Proof. We can see that 2b i s a relation as in the proof of the previous theorem.
In
this case though, we have wa = na from the multiplication table. Again wb = (n+m)b, in the same way.
So wu = mlna
+
(n+m)b +
...
+ mtna + (n+m)b + mt+lna.
needs a slight mOdi?iC8tjOn of the proof of theorem 3.3 to finish this proof.
It now only
165
D.g. near-rings on the infinite dihedral group 4. DEFECTS.
In this last section we find the upper defect, lower defect and defect in the six cases in which the upper faithful d. g. near-ring has been found. Theorem 4.1. Lets = {a,b}.
Then
(i) Q(R0.S) is a zero ring on an infinite cyclic group,
*
(ii) D (R0.S) i s a zero ring on an infinite cyclic group, (iii) D(R0.S) is the direct sum of g(R0,S) and D*(Ro,S).
Proof.
From theorem 2.7, D(R0.S) i s c4a>. i. e. an infinite cyclic group.
From
theorem 3.3 we see that D*(Ro,S) = <[a,b] + 2az. again an infinite cvclic group. Finally we see, again from theorem 3.3, that 4a and [a.b] commute additively.
Hence the additive
structures of the three defects are as stated. The rest follows since all products are zero in Ro. Theorem 4.2.
Lets
= (a.b).
Then
(i)D(R1.S) is a zero ring on an infinite cyclic group,
*
(ii) D (R1.S) i s a zero ring on the direct sum of two infinite cyclic groups, (iii) D(R1.S) is a zero ring on the direct sum of three infinite cyclic groups. Proof. -
From
theorem
4na.4ma = 16nmb = 0.
2.17,
From
(Q(Ri,S).+)
theorem
3.4,
=
14a>.
(D*(Rl,S),+)
Products =
are
zero
<[a,bl + 2a. [a.bla
since +
2a>.
Since a l l products of a and b are equal to b. we see
Consider (-a+b+a+b)(-2a+b+a+b+a).
that the product above i s 24b = 0.
Similarly all other products are zero.
theorem 3.4, (D(RI,S),+) = c[a,b], [a,bIa, 4a>.
Again from
Since 2a commutes with [a,bl, it follows that
(D(Rl,S),+) is the direct sum of three infinite cyclic groups.
As above any product in
D(R1.S) results in an element of the form 2nb = 0. This finishes the proof. Theorem 4.3.
Lets = {a,b}.
Then
(i) D(R2.S) is a zero ring on an infinire cyclic group,
*
(ii) D (R2.S) i s a zero ring on the direct sum of two infinite cyclic groups, (iii) D(R2.S) is a zero ring on the direct sum of three infinite c y c k groups.
*
Proof. From theorem 2.22, (g(Rz,S).+) = C z a r and from theorem 3.5, (D (Rz,S).+) = <[a,b] + 2a, [a,bIa + 2a>. The rest follows as in the proof of theorem 4.2. Theorem 4.4.
Lets = {b,a+b}.
Then
(i) D(R2.S) is a zero ring on an infinite cyclic group,
*
(ii) D (R2,S) i s a zero nng on the direct sum of two infinite cyclic groups, (iii) D(R2.S) is
a zero ring on the direct sum of three infinite cyclic groups.
Proof. From theorem 2.21, (g(Rz,S),+)= <2a>. -
= 4nmb = 0.
All products are zero, since 2na.2ma
From theorem 3.6 we can see that, in the notation of that theorem,
(Dm(R2,S).+)= <2x, c+d>, as can be checked by writing these elements in terms of a and b and using the known structure of Da. theorem 3.6. In (R**,S)
2x commutes with both c and d as we see from
the products are given by x2 = y, y2 = xy = yx = 0. As 2y = 0, it
*
follows that, as before, D (R2.S) i s a zero ring. Finally, again from theorem 3.6, and in the notation of that theorem, (D(Rz,S),+) = <2x, c, d > and as above 2x. c, d a l l commute, and all products are zero.
S.J. Mahmood and J. D.P. Meldrum
166
Theorem 4.5. Lets = {a.b).
Then
(i) Q(R3,S) is a zero ring on an infinite cyclic group, (ii) D*(R3,S) is a zero ring on an infinite cyclic group, (iii) D(R3.S) i s a zero ring on the direct sum of two infinite cyclic groups.
ploof. From theorem 2.27, (lp3,S),+) = < 2 a r . The multiplication table shows that this i s a zero ring.
From theorem 3.8 we get that (DX(R3,S).+) = <[a.bl + 2a>
(D(R3,S),+) = <[a,b], 2a>.
and
Again the multiplication table shows that a l l products are zero,
since 2b = 0 and a l l elements involving b have an even number of b's in them.
Theorem 4.6.
Let
S
= {b,a+b).
Then
(i) D(R3,S) is a zero ring on an infinite cyclic group,
*
(ii) D (R3.S) is a zero ring on the direct sum of two infinite cyclic groups, (iii) D(R3.S) is a zero ring on the direct sum of three infinite cyclic groups.
Proof.
From theorem 2.27 we see that Q(R3.S) is the same as in theorem 4.5. We
now consider theorem 3.7. From this we again see that (D*(R3,S),+)= <2x, c+d> as in the last theorem, and the remainder of the theorem follows as before. In [S] all defects were calculated for some d. g. near-rings.
*
*
In all cases the defects
were zero rings, Q(R,S) = D (R,S) and D(R.S) = p(R,S) 63 D (R,S). Here we have examples in which Q(R.S) is not isomorphic to D*(R,S).
In [51,theorem 3.5, we have an example in
which (Q(R,S),+) is not a n abelian group, hence Q(R,S)
i s not a ring.
But the other
statement still holds. This statement is most unlikely t o be generally true. So a next stage would be t o seek counter examples to this statement.
REFERENCES
S. J. Mahmood. Limits and colimits in categories of d. g. near-rings. Proc. Edinburgh Math. SOC. 23 (1980). 1-7. J. J. Malone. D. g. near-rings on the infinite dihedral group. Proc Royal SOC. Edinburgh 78A (1977), 67-70. J. D. P. Meldrum. The representation of d. g. near-rings. J. Austral. Math. SOC. 16 (1973). 467-480. J. D. P. Meldrum. The endomorphisrn near-ring of an infinite dihedral group. Proc. Royal SOC.Edinburgh 76A (1977). 311-321. J. D. P. Meldrum. Presentations of faithful d. g. near-rings. Proc. Edinburgh Math. SOC.23 (1980), 49-56. J. D. P. Meldrum. Upper faithful d. g. near-rings. Proc. Edinburgh Math. SOC. 26 (1983). 36 1-370. J. D. P. Meldrum. Near-rings and their links with groups. Pitman Research Notes No. 134, London 1985. G. Pilz. Near-rings. 2nd. Edition. North-Holland. Amsterdam 1983.
Near-ringsand Near-fieldsG.Betsch (editor) 0 Elsevier SciencePublishers B.V. (North-Holland), 1987
167
NEAR-RINGS ASSOCIATED WITH COVERED GROUPS C.J.
MAXSON
Mathematics Department Texas A6M University College Station, TX 77843 From the time of Descartes, early in the 17th century, mathematicians have been interested in associating algebraic structures with geometric structures and investigating the interplay. In this paper we continue this line of investigation by associating near-rings to generalized translation spaces and noting how the geometry influences the algebraic structure.
I.
GEOMETRY We start with some geometric background and some definitions. is a translation plane if its group
that an affine plane A
In this case T is
tions operates transitively on the set of points of A. an abelian group with congruence partition
(Ti),
Recall
T of transla-
i.e., each Ti is a sub-
group of T, lTil = ITj[ T = u T Ti n T = ( 0 1 , i # j and T T = T, i i’ j i j i f j. On the other hand let C = {Gal be a non-trivial congruence partition (at least two components) consisting of subgroups Ga this congruence partition of G
a translation plane
of a group G.
taking the elements of G as points, the cosets of the components of lines and setting XG
parallel to
yG
B
if and only if a
known that every translation plane has the form C(G) G
with a suitable partitlon.
From
is obtained by
C(G)
=
p.
C
as
It is well
for some abelian group
More will be said about the equivalence of cer-
tain geometries and groups with special properties later. translation planes see [l], 141
For material on
or 1101. = ( P , L, 1) P called lines,
We now define generalized translation structures. Let Z where P
is a set of points, 1 a collection of subsets of
with an incidence relation defined in P x L x
L satisfying the following axioms:
(Al) Every two p o i n t s
(A2)
L and a parallelism defined in
ILI
22
a, b in P are incident with at least one line;
and for each A
E
L,
IAI
2;
(A3)
Parallelism is an equivalence relation;
(A4)
For each x
E
P.
A
E
1 there exists a unique line containing x
and parallel to A. Further let @
be a one-one map of
incidence structure C
such that O(P)
?’into the collineations of the is a point transitive group of fixed
168
C.J. Maxson We say
point free collineations which map lines onto parallel lines. is a generalized translation structure.
For a generalized translation structure , let 0 ment in P a
in P
denote the ele-
is the identity map in the group @(PI.
such that @ ( O )
0,
there exists a unique map
in @ ( P I such that $,(O)
=
For each a.
With-
out l o s s of generality we take Q(a) = Qa. Note that if every two points determine a unique line then we have a For other work on translation
translation structure studied by Andre [ 2 ] .
structures see the paper by Biliotti and Herzer [ 3 ] and the references
If further we require ( A ( = IBI for each A, B E L then we have a Sperner translation space ((111). We now extend the relationship between translation planes and congruence be a group partitions to generalized translation structures. Letwritten additively, but not necessarily abelian, with identity 0 . A collecgiven there.
tion
When
C = {H,K,. . . I (C1)
U C
(C2)
VH 6 C ,
H
(G,C)
IH6
is a
if
c c,
C,
and VK C C ,
H f G
the pair
G
H
=
the cover.
or H r i K
K
(OK
=
<
K
then H E K .
=
P(G),
{OI.
By taking P( G ) L(G),
ther. by associating to each a C G
The
If further,
and setting x + H II y + K
tine to verify that E ( G )
H f
partition) and the pair
be a covered group. x C G}
if
(G.C) is called a covered group.
cells of
is called a fibration
group_. Let
+
H # {O},
are called
VH, K
(c3) Then C
C
G
G
=
is a cover of
C
elements of
{x
of subgroups of
II
>
= C,
iff H
is a fibered
(G,C)
L(G) =
satisfies (All
=
K, it is rou-
-
(A4).
Fur-
the left translation La, denoted by
@(G):
a + ha we find that we have a generalized translation structure
(C(G),
Q(G)).
Conversely every generalized translation structure arises in this manner (see ( 1 2 1 ) .
From this we see that a generalized translation structure may be
and we henceforth do s o . considered as a covered group
(HI =
(“1
VH,
K
C
C, and as noted above each translation
plane arises from a group with a congruence fibration.
Thus fibrations, equal
fibrations and congruence fibrations “tighten” the structure.
We now tighten
the structure in an alternate fashion namely by requiring that there be a semigroup of operators. Let
<E,O>
domorphism of C
be a generalized translation structure and recall that an enis an incidence preserving function u: P
+
P,
that is,
169
Near-rings associated with covered groups
for A
E
L,
5B
u(A)
E
L.
The collection end I: of endomorphisms of
a semigroup with identity under the operation of composition. a subsemigroup of (01)
The
end C
identity id
For each u E S
Is
is
with the following properties: and the constant map w(a)
previously identified) for each (02)
I:
Suppose S
a
p
E
=
0 (where 0 has been
are both in S ;
the following diagram commutes
1
1
end E
end C Then we say T =
is a generalized translation space with operators,
henceforth denoted by GTSO. If
is the group with cover giving rise to the GTSO
u
then property (02) ensures that each
E
is an endomorphism of G.
S
then since ~ ( 0 =) 0 , each line through 0 is mapped by through 0. Hence, S
But u into a line
is a semigroup of endomorphisms satisfying
(01)'
The identity endomorphism and zero endomorphism are in S ;
(02)'
For each a U(Gi)
5 Gj
.
E
for each Gi
S,
Conversely if G
T = <E,@,S>
E
C, there exists G
j
is a group with a cover
C
E
C with
and a semigroup of endo-
morphisms satisfying (01)' and (02)' then we obtain a GTSO. GTSO arises in this manner, thus we identify the GTSO
Furthermore every
with the asso-
ciated group with operators and cover.
We now show how to associate near-rings with GSTO's,. First, we consider the
E(G.C)
set
=
{f E End G
I f(H)
that the semigroup of operators plays no role here.) functions E(G,C) tations of E(G,C),
5 H,
VH € C}.
(Note
Under composition of
is a semigroup with identity, called the semigroup of
. We consider the near-ring distributively generated by
denoted by
sider MS(G,C)
= {
dg E ( G , C ) .
f E MO(G)
For our second associated near-ring we con-
I
f(H)
5 H, VH
C
C, f u = u f, Vo € S )
which
is a zero-symmetric near-ring with identity under the operations of function
addltion and composition.
Unfortunately the word "kernel" has been used in
the literature for both of these near-rings. the kernel of
and
M (G,C) S
In this paper we call dg E ( G , C ) In the next
the centralizer of.
section we discuss the centralizer and in Section 111 the kernel. 11.
THE CENTRALIZER OFW e begin this section with the result that centralizers of GTSO's are
very general, Indeed every zero-symmetric near-ring with identity arises in
C.J. M a s o n
170
this manner. The next result, whose proof is due to Peter Fuchs extends the corresponding result in [12] where it was established only in the finite case. Theorem 11.1. [12]. Let N be a zero-symmetric near-ring with identity in which ab
=
1 implies ba
1 , a,b E N. Then there exists a GTSO, T
=
=
such that
M (G,C).
N
S
Proof.
Let G
{(l,O),
(0,l))
=
N x N.
Clearly G
N x N
H C
#
is contained in a maximal one.
denote the set of all maximal N-subgroups of G.
{H )
%
N x N
Hence maximal N-subgroups of
is a generating set).
exist and every N-subgroup C =
is a finitely generated N-group (e.g.,
C
Let
is a cover of
H and if (a,b) C G, then (a,b) C N(a.b) 5 Ha B for some maximal N-subgroup Ha. Therefore 11 Ha = G. For each a C N, let p : G + G be the function defined by p ( u , v ) = (ua,O) and let t: G + G be the function defined by t(u,v) = (v,O). Then pa, t are N-endomorphisms of G. We let S be the semigroup of endomorphisms
G
since Ha 2
implies Ha
a C N) U It).
so
is an N-subgroup of G semigroup of operators. As
1
{id} U {pa
generated by
=
a(Ha)
a
For each for some H
HB
B
S
C
and
Ha E C,
Thus
C C.
u(H,)
is a
S I
Proceeding as in [ 1 2 ] one then shows N
=
Ms(G,C).
with most representation results the above leads to several interest-
ing avenues of investigation.
Here we consider only the problem of character-
izing structural properties of M (G,C)
in terms of the geometry
S
.
(Further representation results are considered in r121.1 For the remainder of this section, all structures are finite and
Convention:
all near-rings are zero-symmetric with identity.
We first characterize when MS(G,C) some further concepts.
Let T
is a generating set if G
=
Syi and
=
finite, generating sets exist.
is a near-field.
be a GTSO. S y i F Syj
For each yi
C
....
if
ifj.
Y, define
I
For this we need
A set
Y
=
{yl,...,y t)
Since G
Ii = n {Ha
yi C Ha) and let I = {sequences (xl, xt) xi c If and F(Yi, F(x. x.)) where F(u,v) = { ( a , B ) E S x S au = Bv}. Further let J’ J i-th projection map on I. We sayis kernel dependent if, pi I Ker a P { O } implies pi I Ker a ( s e e [ll] 1. Recall also
I
c
x,
y C G*
a l , . . . , un,
= G \{O)
are connected if there exist w~,...,w~-~E
...,t
tl, a1
x
= tl
a2 w 1
C S
G
I
is Ha C C,
yj) c pi be the for a 6 S, that and
such that
w1 f 0
= t2
w2 # O
u n wn-1 = t n y f O This relation is an equivalence relation on
G
*
and the equivalence classes
171
Neatvings associated with covered groups
are called the connected components.
G is said to be connected if G"
is a
connected component. Theorem 11.2.
Let T
[12].
=
. Then MS(G,C)
is a near-field iff T
is connected and kernel dependent. To obtain definitive structural results one places some restrictions on the semigroup
o€
operators. One considers the situation In which
group of automorphisms (with 0 adjoined).
As
S
is a
one might suspect (from the pre-
vious work on centralizer near-rings ( 1 7 1 ) the orbits of the action and stabilizers of elements in G"
play an important role.
For results in the case of
fibrations see 1141 and for the general results see [51. Next, one considers the situation in which S = U
{O, id}.
We write M,(G,C)
ble and not nilpotent, Ma(G,C)
S
for MS(G,C).
i s a cyclic semigroup,
When
a is not inverti-
is not a simple near-ring.
Therefore one
restricts to the case of a nilpotent operator. Theorem 11.3. 151, 1141.
Let a be a nilpotent operator on
IKer a I - I
(G,C).
connected components.
(11.3.1)
There are
(11.3.2)
There is a unique generating set, Y.
(11.3.3)
If
C is a fibration, Ker a
is contained in a unique cell,
say €Io. When
G
is connected much can be s a i d .
Theorem 11.4. 151, [141. Let a lowing are equivalent.
be a nilpotent operator on
(11.4.1)
Ma(G,C)
is a near-field.
(11.4.2)
Ma(G,C)
is a simple near-ring.
(11.4.3)
M (G,C) <(G,C)
(11.4.4)
(G,C).
The f o l -
is a 2-semisimple near-ring.
2
Z2.
Further, if C (11.4.5)
Ma(G,C)
(11.4.6)
Y I1 Ho = 0 .
is a fibration the above are equivalent to
is a local near-ring.
If G is not connected and a
is a nilpotent operator, necessary and
sufficient conditions in terms of the geometry which characterize when Ma(G,C)
have been determined
is simple. Again €or the case in which
C
is a fibration see [14] and for the general case see 151. We conclude this section with two suggestions for further research, first a rather specific problem and then a more general program. Sug 1:
Characterize 2-semisimplicity in terms of
S = U { O , id).
for
172
C J. Maxson
Sug 2:
For other classes of operators, (e.g.,
inverse semigroups,completely
0-simple semigroups) characterize the structure of MS(G,C) of 111.
THE KERNEL OF
.
We recall from Section 1 that the kernel of
is the near-ring
dg E distributively generated by the dilatations E
=
E(G,C).
When
is a
C
is well-known [ 3 ] , [ 6 1 , [ 7 ] ,
fibration, the structure of the semigroup E
[lo],
in terms
.
[Is].
Theorem 111.1.
Ifis a finite fibered group then E*
group of fixed point free automorphisms of
G.
is a
= E\{O)
Further, E* is a cyclic group.
A now classical result states that in the caseis a translation plane, G
is an abelian group.
finite translation plane, dg E
Thus from the above theorem, if is a finite field [ 3 ] , [ l o ] .
is a
Therefore to
generalize to arbitrary fibrations one is lead to determining the structure of Such a study has been carried out in [ 7 1 and [ B ] .
fibered groups.
For our
particular case we specialize one of the structural results (see also [6]). Theorem 111.2.
If
is a finite €ibered group with
E* f {id)
then G
is a finite p-group for some prime p, of exponent p and of nilpotency class at m s t
2.
Using the above result concerning E and G along with some work of Herzer [6] (recently generalized in [ 1 3 ] ) the following rather surprising result has been obtained. Theorem I I G . [ 1 5 ] . commutative ring. If
E* = {id)
Examples ( [ 7 ] )
If
is a finite Eibered group then dg E
If, further, E*
f
then dg E
{id}
then clearly dg E = Zn
show that in this case n
where
n
is a
is a finite field. is the exponent of G.
need not be a prime.
Theorem 111.3,
together with the classical work of Andre ( [ l ] ) , shows that whether or not G is abelian, whenever
E*
f
in a natural manner.
{id)
there is a field associated wtth the geometry
We note that in the non-abelian case, although
the elements of the associated field are sums of automorphisms, they need not themselves be morphisms. We also mention that i n the abelian case the field dg E applications.
The significance of the field dg E
still unknown. Recently (191, [16]) a study of the near-ring groups
has been initiated.
i n the above studies
G
has geometric
in the non-abelian case is dg E
for finite covered
In general, very little is known.
In fact,
is restricted to be a finite elementary abelian
173
Near-rings associated with covered groups
p-group and the structure of the ring dg E
is investigated.
In [91 several
basic results, several examples and a complete solution to the case G are presented.
In [I61 necessary and
3
P sufficient conditions in terms of an
associated lattice are given for the kernel dg E ring or a semisimple ring.
= (2 )
to be a field or a simple
The surface has just been scratched, there is
still much to be investigated here. REFERENCES Andr6. J., Uber nicht-Desarguessche Ebenen mtt tansitiver Translationsgruppe, Math. Zeitschr., 0 (1954), 156-186. , Uber Parallelstrukturen, 11: Translationsstrukturen, Math. Zeitschr., -/a (1961 ), 155-163. BILIOTTI. M. and Herzer, A . , Zur Geometrie der Translationsstrukturen mit eigentlichen Dilatationen. Abh. Math Sem. Univ. Hamburg 2 (1983), 1-27. Dernbowski, H. P., Finite Geometries, Springer-Verlag, New YorkHeidelberg-Berlin, 1968. Fuchs, P. and Maxson, C. J., Kernels of Covered Groups with Operators (in preparation). Herzer, A., Endliche nichtkornmutative Gruppen mit Partition Il und ftxpunktfreiem n-Automorphismus. Arch. Math. 34 (1980), 385-392. Karzel, H. J. and Maxson, C. J., Fibered groups with non-trivial centers, Result. der Math., 7 (1984). 192-208. , Fibered p-groups. Abh. Math. Sem. Univ. Hamburg (to appear). Karzel, H., Maxson, C. J. and Pil-z,G. F., Kernels of Covered Groups. Result. der Math. (to appear). Luneburg, H., Translation Planes, Springer-Verlag, New YorkHeidelberg-Berlin, 1980. Maxson, C. J., Near-rings Associated with Sperner Spaces, JOUK. of Geom., 20 (1983). 128-145. , Near-rings associated with Generalized Translation Structures. JOUK. of Geom. 24 (19851, 175-193. Maxson, C. J. and Meldrum, J. D. P., D G Near-rings and Rings (submitted). Maxson, C. J. and Oswald, A., Kernels of Fibered Groups with Operators, (submitted). Maxson, C. J . and Pilz, G. F., Near-rings determined by fibered groups, Arch. Math. 44 (1985), 311-318. Maxson, C. J. and Pilz, G. F., Kernels of Covered Groups 11, (submitted). Pilz. G. F., Near-rings, Revised Edition, North-Holland, New York, 1983.
This Page Intentionally Left Blank
Near-ringsand Near-fclds,C. Betxh (editor) 0 Elsevier Science Publishers B.V. (North-Holland), 1987
175
KRULL DIMENSION AND TAME NEAR-RINGS J.D.P. MELDRUM and A.P.J. VAN DER WALT Department of Mathematics, University of Edinburgh, Mayfield Road, Edinburgh EH9 3JZ. Scotland.
Department of Mathematics, University of Stellenbosch, Stellenbosch 7600, South Africa.
A tame near-ring R is one which has a faithful module in which all R-modules are R-ideals. Such a near-ring is very ring-like. In this paper we extend some results on rings with Krull dimension to tame near-rings. Krull dimension i s a generalization of the minimal condition. The main results are as follows. (i) A tame right noetherian near-ring has a possibly transfinite power of its J2 radical zero. (ii) The Jg radical of a tame near-ring does not contain a non-zero idempotent ideal. (iii) The nil radical of a tame near-ring,is residually nilpotent.
1.
INTRODUCTION AND DEFINITIONS. Contrary to what is suggested by their name, general near-rings are in some respects
very far from rings. One such is the fact that none of the obvious generalizations of the Jacobson radical to near-rings i s necessarily nilpotent even in the finite case. This is one reason for introducing tame near-rings, a reasonably large class of near-rings which behave very like rings. They were first introduced by Scott [lo] and [ l l ] and he showed how close to rings they were in their behaviour. In particular tame near-rings with the descending chain condition have a l l three J (for Jacobson) radicals nilpotent. In this paper we generalize results about rings with Krull dimension to near-rings.
Apart from those
results that are essentially lattice theoretic in nature, the proofs of the results about near-rings are not straight generalizations of the proofs from ring theory. This is mainly due to the fact that the ring-like properties lie in a faithful module, whereas the near-ring has the Krull dimension. A (left) near-ring i s a set R with two operations, + and x,y,z
E
R.
., such that
(R,+) is a not
is a semigroup and x(y+z) = xy + xz, Ox = 0, for all
necessarily abelian group, (R,.)
If (G,+) i s a group written additively then Mo(G) = (f : G
+
G, OGf = O G }
is a
near-ring under point-wise addition and composition of functions. A homomorphism from
R into Mo(G) defines G as an R-module. of G such that HR (g+k)r - gr
E
5 H.
A submoduie H of the R-module G is a subgroup
An R-ideal of G is a normal subgroup K of G such that
K for all g E G, k
E
K, r
E
R.
under the right regular representation.
The group (R,+) is a n R-module, denoted RR,
A right invariant subnear-ring
of R i s an
R-submodule of RR and a right ideal of R is an R-ideal of RR. A left invariant subnear-ring of R i s a subgroup S of (A,+)
such that RS
5S
and a left ideal of R is a normal left
invariant subnear-ring. An ideal i s a left ideal which is also a right ideal. An R-module G
J.D.P. Meldrum and A.P.J. van der Walt
176
is monogenic if G = gR for some g
E
G. A monogenic R-module G is of type 0 if it has no
non-trivial proper R-ideals, of type 1 if it is of type 0 and g
E
g R = (0). of type 2 if it has no non-trivial proper R-submodules. J.+(R) as the intersection of all annihilators of R-modules
G implies gR = G or For v = 0,1,2,
of type v.
define
These definitions
follow closely those t o be found in Meldrum [81, or in Pilz [91, except that he uses right near-rings. and his nomenclature for the substructures of R-modules is different.
Most
standard results can be found there. Recall that a tame R-module G is an R-module in which a l l R-submodules are R-ideals.
A near-ring is tame i f it has a faithful tame R-module.
Two large families of
tame near-rings are given by firstly, the subnear-rings of Mo(G) generated by a semigroup S of endornorphisms of G containing Inn G, the inner automorphisms of G, and secondly, the zero-symmetric parts of near-rings of polynomials. We n o w come t o Krull dimension. Defined originally for commutative rings, it has been extended to arbitrary rings where it has been used extensively (Gordon and Robson [2]). In this context it is essentially a condition on the lattice of substructures generalizing the descending chain condition. Let G be an R-module.
Denote the Krull dimension of G by kd(G). Then kd(G) = - 1 i f and only
if G = (0). If a is an ordinal and kd(G) + a then kd(G) = a if there is n o infinite descending
chain
G = Go
2 G1 1.G2 2 . ..
of
R-ideals
of
G
such
that
kd(Gi/Gi+l) + a
for
i = 0.1.2,. . _. In Gi/Gi+l we define the Krull dimension relative to the R-ideals of G/Gi+l.
Note that kd(G) = 0 corresponds t o the descending chain condition on R-ideals. that R has Krull dimension if the R-module RR has Krull dimension.
need not necessarily have Krull dimension. Finally for an ideal I of the near-ring R w e define lo = R, I 8 + l= 18.1, for limit ordinals
a and 10 =
I, 16+1 =
n;=, (18)"
and l a =
We say
Any given R-module la =
n ~I8 < ~
nB
In the next section w e gather together results about Krull dimension and powers of ideals needed later. In the last section we present our results about tame near-rings with Krull dimension.
2. PRELIMINARY RESULTS.
In the applications we need a number of results about modules with Krull dimension from Gordon and Robson 121 and Lenagan 171. The difficulty in using these results is t o unravel those results that arise from the properties of the lattice of substructures from those which use results about rings. Here we list the results we need with indications of proof where necessary.
Lemma 2.1.
(121 Lemma l.l(i)). L e t G bs a tame R-module with Kfull dimension, H an
R-ideal of G . Then kd(G) = sup{kd(G/H), kd(H)} if either side exists Note that w e need tameness since R-ideals of H are not necessarily R-ideals of G. Indeed the result is no longer true if G is not tame.
Example 2.2. The wreath product G of two infinite cyclic groups satisfies the maximal condition on normal subgroups but contains a normal subgroup H which is the direct sum of an infinite number of infinite cyclic groups. Take R t o be the integers with 1 acting as
177
KruN dimension and tame near-rings
the identity map on G. Then G has Krull dimension (since it satisfies the maximal condition on R-ideals, see lemma 2.4) but H does not (see lemma 2.6). Lemma 2.3.
Let G b e an R-module with KruN dimension
Then every homomorphic
image o f G has Krull dimension less than o r equal to the Krull dimension of G. See proposition 1.2(i) in [2]. This result follows because the lattice of R-ideals of a homomorphic image of G is a sublattice of the lattice of R-ideals of G, and it does not need tameness of the R-module. Lemma 2.4.
Every R-module with the maximal condition on R-ideals has Krull
dimension This result can be proved by the same methods as proposition 1.3 of [2]. Before we generalize proposition 1.4 of [21, we need the concept of, and a result about uniform dimension. We say that the R-module G has finite uniform dimension if there is a fixed (finite) bound on the number of summands in a direct sum of non-zero summands. Lemma 2.5. Let G be a tame R- moduJe which does nor contain an infinite direct sum
of nan-zero R-ideals
Then G has finite uniform dimension
Proof. We first show that any non-zero R-ideal of G contains an R-ideal of G which is uniform, i.e. a n R-ideal with the property that any two non-zero R-ideals of G contained in it intersect non-trivially.
Let H be a non-zero R-ideal of G. If it is uniform, we are home.
If not there exist K1, L1 R-ideals of G contained in H such that K1 n L 1 = {O}.
So K1 0 L1
< H. If either K1 or L1 are uniform, we are home. If not we obtain K1 0 Kg 0 L2
5H
where K2, L2 are non-zero R-ideals of G contained in L1. Since G does not contain an infinite direct sum of non-zero R-ideals, this process must stop and it can only stop when we have a non-zero uniform R-ideal. The second step is to show that there is a finite number of uniform R-ideals of G whose sum i s direct and is an essential R-ideal of G, i.e. it has non-trivial intersection with every non-zero R-ideal of G. By above, G has a uniform R-ideal H i . If H i is not essential, then there is a non-zero R-ideal K1 such that H1
n K1 = {O}.
Then K1 contains a uniform
non-zero R-ideal H2 and H i Q H2 is direct. This process must stop after a finite number of steps as before, for the same reason. This gives us the second step. The last step is to obtain the fixed bound on the number of non-zero direct summands. The method used is a version of the Steinitz Exchange Theorem used with finite dimensional vector spaces. At first we consider direct sums consisting of uniform R-ideals. Let H i G I . . . @ Hn be a direct sum of uniform R-ideals which i s essential. To be able to use this sum for the exchange process, we need to show that if K is an R-ideal of
G which intersects each Hi non-trivially, then K is essential in G. This we now do. For i = 1,.
.@
. . , n, define
Ln is direct and L
5 K.
Li = K n H i , a l l non-zero by hypothesis. Hence L = L1 C3. We show that L is essential in G. Let g
E
G - ( 0 ) and let gR
denote the R-ideal of G generated by g, since G i s tame. Since H i Q . we have gRn(H1 Q . . . @ H ),
# {O}.
Thus there exists g1 = h i +
...
. .. +
. Q Hn
is essential,
hn where hi
E
H,,
g1 t gR, and some hi # 0, without loss of generality h i # 0. Because H i is uniform, and L1
J.D.P. Meldrum and A.P.J. van der Walt
178
So we have h l r l
is a non-zero R-ideal in HI, L1 i s essential in H i . Hence h l R n L 1 = {O}. = I1 # 0. Then g1r = h'l + h'2 +
have 0 # h'l
E
. .. +
where h'l = I1 # 0. If h'2,
h',
L c\ gR and we are home. If h'2,
. . ., h'n
..
., h,'
are all zero we
are not a l l zero, then without loss
of generality h'2 # 0. We now repeat the process to get h'2R n L 2 = (0) and there exists r2
E
...+ . . . + h",
R such that ( h ' l + h'2 +
1'1 + 12
L n g R . If h"g +
E
h',)r2
= 1'1 + 12
= 0
+
h"3 +
... +
h",.
Then 1'1
+
12 # 0 and
we are home. In any case we eventually obtain a
non-zero element of gR c)L after at most n steps. Thus L is essential. Now suppose that MI,
. . .,
Mm are non-zero R-ideals of G such that M1 8 . . . @ Mm
is a direct sum. We show that m
5 n,
which will give us our result. M2 @ . . . @ Mm i s
not essential since it intersects M i trivially. Hence (Mq I3. . . @ Mm) f~Hi = (0) for some i,
. . I3 Mm i s direct. Similarly . . @ M, i s not essential and after renumbering the Hi's if necessary, the sum @ M3 @ . . . @ M, is direct. Proceeding in this way we obtain that m 5 n.
without loss of generality i = 1. Hence the sum H i 8 M2 @ . H1 @ M3 8 3 . H i @ H2
This proof is a fairly straightforward adaptation of the proof of the corresponding result in Chatters and Hajarnavis [l]. lemma 1.9. We come now to the generalization of proposition 1.4 of Gordon and Robson [21. Lemma 2.6. --
A tame R- module with Krull dimension has finite uniform dimension
This can be proved as in [2]. We now look at critical R-modules. kd(G) =
a
We call a non-zero R-module G a-critical if
and the Krull dimension of any proper homomorphic image of G is strictly less
than a. A critical R-module is an R-module which i s a-critical for some ordinal a. Lemma 2.7. --
Any non-zero
R- module with Krull dimension has a critical R- ideal
The proof is similar to that of proposition 2.1 of [2]. We now present a technical lemma, generalizing Proposition 1 of Lenagan [7], which we will need later. Lemma
28.
Let
G
be
0 c H1 < H2 < . . . such that G R-ideals G = Go
a
module
= u"i=1
with
Hi.
an
ascending
chain
of
R-ideals
If there is an infinite descending chain of
> G1 > . . . such that for all i,j
then G does not have Krull dimension
See Proposition 1 of I71 for the proof. We now come to the question of the length of well-ordered descending chains of R-ideals in modules with Krull dimension.
This generalizes work of Gulliksen [3] and
Krause [41.
-Lemma 2.9. Let G be an R- module with Krull dimension a. Then a 1 limit (B; G has Further, if a < w, then
a well-ordered descending chain of R-ideals of length at least wB}. any R-module G with the propelty that a = limit
(8; G has a well-ordered descending
chain of R- ideals of length at least we) has Krull dimension a.
Proof.
We use transfinite induction on a. Consider the case a * 0.
Then G has
Krull-dimension 0 if and only if G has the descending chain condition on R-ideals.
That
179
KruN dimension and tame near-rings
holds if and only if every well ordered descending chain of R-ideals in G i s of finite length, i.e. 0 = limit (6; G has a well-ordered descending chain of R-ideals of length at least Now assume that the first result is true for all ordinals y and
consider
a
well-ordered
descending
u = w 6 . n l + . . , + nr, where n1, , . ., nr
5 a.
t o show that 6
E
2
of
R-ideals
to8}.
Let G have kd(G) = a in
G
of
length
Z, and the exponents of w are ordinals. We need
6 > a.
So suppose that
Let G ,
u. Since u
well-ordered descending chain of length for each m
chain
< a.
be the u " . m t h
= w B . n l , and
6 > a,
term of the
w e can find Gm
1. Then Gi/Gi+l has a well-ordered descending chain of R-ideals of G of
length wa inherited from G. By the induction hypothesis the Krull dimension of Gi/Gi+l relative to G is not less than a. for each i 2 0. By definition kd(G) f a. This finishes the proof, by induction, of the first statement. We now prove the second statement by induction. The case a = 0 has already been done. As a is finite in this case, we have limit (6; G has a well-ordered descending chain of R-ideals of length at least u6} = max
(6; G
has a well-ordered descending chain of
< a.
Suppose G
has Krull dimension a. Then kd(G) 1.a implies that G has a well-ordered
descending
R-ideals of length at least u6}. We assume that the result is true for y chain of R-ideals of length at least w", lemma, we have
a
= max
(6; G
using the induction hypothesis. Using part of the
has a well-ordered descending chain of R-ideals of length
at least uB}. Now suppose that G has the property that a = max (5; G has a well-ordered descending chain of R-ideals of length at least w6}. infinite descending chain of R-ideals.
Let G = Go
2 G1 2 G2 2 . . .
b e an
If more than a finite number of the factor modules
Gi/Gi+l are not of Krull dimension < a (relative t o G) then an infinite number of these factors have a well-ordered descending chain of R-ideals of length at least these wa.w
chains = uatl.
to
G, we
get
a
well-ordered
descending
This cannot happen by hypothesis.
factor modules Gi/Gi+l are not of Krull dimension C kd(G) + a by the induction hypothesis.
chain of
ua. By lifting
length at
least
So at most a finite number of the
a
and thus kd(G) = a. since obviously
This finishes the proof of the second part of the
lemma by induction. The final set of preliminary results concerns powers of ideals. We list them without proof because they are easy t o prove and the proofs parallel exactly those for rings in Krause and Lenagan [51, after account has been taken of the left-right switch in the definition.
Note that with near-rings, powers of ideals are just subsets o f the near-ring,
not ideals in their own right.
Lemma 2.10. Let I be an ideal of the neor-ring
R.
Then the following hold
5 lat6 for a// ordinals a, 6. (ii) (I")" 5 l a S nfor all ordinals a, positive (iii) la 5 lug for all ordinals a. (i) la.lS
integers n.
J.D.P. Meldrum and A.P.J. van der Walt
180
3. TAME NEAR-RINGS WITH KRULL DIMENSION. Here we extend some results concerning rings with Krull dimension due t o Lenagan
[S] and [7] and Krause and Lenagan [5]to tame near-rings with Krull dimension. Many of the proofs are suitable amalgams of proofs for the ring case taken from the papers mentioned above, together with one or two special features. We consider three topics: the transfinite powers of the J2 radical in tame near-rings with the ascending chain condition on right ideals, the non-existence of idempotent ideals contained in J2(R), where R i s tame with Krull dimension and finally the nil radical in tame near-rings with Krull dimension. For the first of these three topics we consider the following situation.
Let R be a
right noetherian tame near-ring, i.e. a tame near-ring with the ascending chain condition on right ideals. It will have Krull dimension by lemma 2.4, say kd(R) = a. By lemma 2.9 any well ordered descending chain of right ideals has length strictly less than ua”.
Let G be
a faithful tame R-module. Lemma &l.
With the notation given above, let g
E
G . Then the R-module gR has a
well ordered descending chain of R-ideals which reaches (0) and such that aach factor is 2- primitive By proposition 3.4
Proof.
isomorphism i s given by gr
+
of
Pilz 191. we have gR
AnnR(g) + r.
‘R
R/AnnR(g) where the
By the ascending chain condition on right
ideals, we can find a right ideal L1 of R maximal subject to containing AnnR(g). Then Ll/AnnR(g) is a maximal right ideal in R/AnnR(g) and so gL1 i s a maximal R-ideal in gR. Since gR i s tame as the submodule of a tame module, we deduce that gR/gLI i s 2-primitive. 2-primitive.
Now repeat the process with gL1 to obtain gL2 such that gL1/gL2 i s Continuing transfinitely if necessary we obtain the result that we want.
Corollary 3.2.
The conclusion of lemma 3.1 holds for any monogenic tame R-module
for a right noetherian near-ring R.
Lemma 3.3. that J2‘
With the same hypotheses as lemma 3.1,there exists an ordinal
5 AnnR(gR)
proOf. Given g
u
such
for a / / g E G. E
G , we construct the well-ordered descending chain of R-ideals
whose existence is guaranteed by lemma 3.1, say
Since each factor i s 2-primitive, we have
for all ordinals A. ordinal
The case all p C
We will prove by transfinite induction that (gR)JzU = (0) for some
u by showing that (gR)J2‘ 5 gRA for all ordinals A.
X.
If
X
= 1 has already been pointed out. So assume that the result is true for
A = p
+
1 for some p, then (gR)J2’ = (gR)JZp. J2
hypothesis and gRp. J2 gRi =
nN<xgRp
1.gRp+l
and J2’ =
5
= gRA by the remark above. If
n,
J2p.
Let gr
E
gR.,
X
gR, x E J2’.
J2 by t h e induction is a limit ordinal, then Then for all p
< A,
181
KruN dimension and tame near-rings x
E
J*p
(gr)x
E
and
by
nu<),gR,,
the
= gRA.
induction
hypothesis
This is true for all gr
E
5 gR,,.
Hence
Hence (gR)J2A
5 gRA,
(gr)x E (gR)JzP
gR, x
J2 A .
E
which is sufficient t o complete the induction argument, hence the proof of the result. Corollary
Proof.
g . Jq
w"+l
< Ann(GR).
By lemma 2.9, any well-ordered descending chain of right ideals in R has
< Ann(gR)
length strictly less than ua+'. Hence J2
for a l l g
This proves the corollary.
-Theorem 3.5.
Let R be a tame right noetherian near-ring.
E
G by the lemma.
us+i Then J2(R) = (0) i f R
has an identity and has square zero otherwisa
proOf. If R has an identity, then GR = G, and since G i s faithful, it follows from corollary 3.4 that J2
*a+%
*u+l
faithful, we have (J2
,oc+l Otherwise RJ2
= (0).
annihilates G and since G is
)2 = (0).
Theorem 3.6. Let R be a tame right noetherian near-ring.
Then ( J Z ( R ) ) ~is nilpotent if
R has an identity, and (J2(R))a+1 has square zero otherwise Proof. The second part follows from lemma 2.10 and theorem 3.5. The first pan follows from lemma 2.10 and lemma 2.9 as in the proof of theorem 3.5 once we observe that by lemma 2.9, J2u = (0) for some
u
= ua.n
+ .
. . for
some positive integer n.
We now turn to the non-existence of idempotent ideals within the J2 radical. There
are two results of this type, one where the module has Krull dimension, the other where the near-ring has Krull dimension. Let R be a tame near-ring with a faithful tame R-module G. Let G have Krull dimension.
~Definition 3.7. The critical socle series of G is defined by So(G) = (0). Sk+i(G)/Si(G) is the sum of a l l critical submodules of G/Sk(G) and
s~ =
u,,<x
S,(G)
for limit ordinals
A.
By lemma 2.7 it follows that SA+1(G) > Sl(G) for all ordinals
A
and hence that
G = Sa(G) for some ordinal a. Theorem 3.8. Let R be a tame near-ring with a faithful tame R-module H with Krull dimension Then J2(R) does not contain a non-zero idempotent idear!
Proof.
Let I
5
J2(R), I be an idempotent ideal.
8-critical submodules of H with 8
5 kd(H).
We show that I annihilates all
Note that there is no distinction between
submodules and R-ideals in H, since H is tame. The cases 6 = -1,0 are both trivial since 0-critical modules are just the 2-primitive submodules of H. Now assume that LI = (0) for all modules with kd(L) < 8. Then if M is a 6-critical submodule of H and M' is a non-zero submodule of M, we have kd(M/M')
< 6.
Hence (M/M')I = (0). i.e.
M' 2 MI. Thus MI will
generate the unique minimal submodule of M if it is not zero. Since I 5 J2(R), it follows that (0) = MI.1 = MI2 and hence ld
5
Ann M, i.e. MI = (0) since I = ld.
This
proves the result by induction. Corollary 3.9.
If R is a tame near-ring with a faithful tame R-module with Krull
dimension then, for any ideal I sequence or I is nilpotent
< Jz(R),
either I > l2
> . . . is a
strictly decreasing
J.D.P. Meldrum and A.P.J. van der Walt
182
Proof.
Otherwise J2(R) would contain an idempotent ideal contrary t o the theorem.
We now turn t o the case in which R has the Krull dimension. Theorem 3.10.
Let R be a tame near-ring with Krull dimension. Then 4 ( R ) doss not
contain a non-zero idempotent idear!
Proof. Let G be a faithful tame R-module.
Let g E G, I
J’(R), I an idempotent ideal.
Then by proposition 3.4 of Pilz [9] we have gR = R R/AnnR(g) and as R has Krull dimension, so does R/AnnR(g) by lemma 2.3.
theorem 3.8 it follows that gR.1 = (0). holds for all g
E
5 G, hence gR is Thus RI 5 AnnR(g) and
Also gR
5 AnnR(G).
G, and so I*
a tame R-module.
so I’
By
5 AnnR(g).
This
As G is faithful, it follows that l2 = (0). This
proves the result. Corollary
3.11.
contained in J2(R).
Let R be a tame near-ring with Krull dimension. Let I be an ideal Then either the sequence I
> I’ > . . . is strictly decreasing or
I is
nilpotent
Proof. Otherwise J2(R) would contain an idempotent ideal contrary t o the theorem. We come finally t o the nil radical of a tame near-ring
with
Krull dimension.
Unfortunately we cannot obtain the nilpotence of the nil radical as happens for t h e ring case. But we do obtain something.
~Lemma 3.12. Let R be a tame near-ring with Krull dimension and let G be a faithful tame R-module, with g E G . andC is nil then C2
Write A for AnnR(g).
If (C+A)/A is a critical right ideal o f R/A
5 A.
Proof. Let (3 : R
+
5 G by proposition 3.4 of Pilz [91. gR 5 G and G is tame. Let C be a
RIA where as before R/A =R gR
By lemma 2.3, RIA has Krull dimension, and is tame as
nil right ideal of R, such that CO is critical. We can find such an ideal unless all nil right
So without loss of generality w e may
ideals of R are contained i n A, by lemma 2.7. assume that CO # (0). Let cO But c” = 0 f o r some n
tame.
follows that (ce)m-’ CB.CB
E
>
CO, cO # 0. Then (cC)O is an R-ideal of R B since RO 0. So there exists a least m
f 0 and (cO)m-’
>
annihilates cO. Thus AnnC(cO)
Ce/AnnCe(CB) = C + A/AnnC+A(ce)
and
AnnC+A(Ce)
IS
1 such that (ce)m = 0. It
1.(A.cm-’}.
2 {A,cm-’}
Hence
> A.
Thus
kd(cO .CO) < kd(CO) since C B is critical. Hence as CO is critical, we have cO .CO = 0, i.e. (CB)2 = (0) and C2
5 A.
The fact that a non-zero submodule of a critical tame module
with Krull dimension has the same Krull dimension follows from lemma 2.1.
_ Lemma _ _ _ 3.13. _ Let R be a tame near-ring with Krull dimension. Let N be the nil radical of R. If N is not nilpotent, then there exists a chain of right ideals 0 = A0 such that each AiB is nilpotent but
A = AnnR(g), g
E
url AiO
< A1 <
is not nilpotent, where O : R
-c
R/A for
G, a faithful tame R-module
Proof. By hypothesis and using lemma 3.12. we deduce that there is a non-zero right ideal CB with square (0).
By Zorn‘s Lemma w e may choose a maximal right ideal
contained in N whose square is zero in R/A. Call this right ideal A l . Assume that w e have chosen
4,
then choose Ai+l so that Ai+lO/A,e is a maximal right ideal contained in
NB/AiB whose square i s zero.
Obviously each AiO is nilpotent.
If XO =
u
AiB is
nilpotent. then XO = AnB for some n, which is impossible by hypothesis and lemma 3.12.
183
Krull dimension and tame near-rings
Lemma 3.14.
Let R be a tame near-ring with Krull dimension and faithful tame
module G. Let N ‘be the nil radical of R. f o r every g E G, there exists n = n(g) such rhat
Nn(g) C AnnR(g).
Proof. Assume the result does not hold. Then by contained i n N
,
lemma 3.13, there is a right ideal X
where X0 i s not nilpotent and 0 i s defined as in lemma 3.13. We also
have X = u z 1 Ai and AiB is nilpotent for each i 1 1 .
Since NB and X8 have Krull
dimension, using lemmas 2.1 and 2.4, we can apply lemma 2.8 and deduce that for some n. j we have (XB)”
5 (XB)”+l
+
Aj0.
Hence (X0)”
1.A,B
since the powers of X0 form a
strictly decreasing sequence (corollary 3.1 1). Thus XB is nilpotent, a contradiction. Theorem 3.15.
Let R be a tame near-ring with Krull dimension, N the nil radical of
R. ThenNW= {O}.
Proof.
N~
Let
5 fl gEG AnnR(g)
G =
be
a
faithful
(01 since G i s
tame
R-module.
By
lemma
3.14,
faithful.
ACKNOWLEDGMENT. The work on this paper was done while the first named author was visiting the Department of Mathematics of the University of Stellenbosch. Financial assistance of both the Council for Scientific and Industrial Research and the University of Stellenbosch is gratefully acknowledged.
REFERENCES
Chatters, A.W. and Hajarnavis, C. R. Rings with chain conditions (Pitman Research Notes in Mathematics, 44. London. 1980). Gordon, R. and Robson. J.C. Krull dimension Mem. Amer. Math. SOC. 133 (1973). Gulliksen. T.H. A theory of length for noetherian modules J. Pure and Appl. Algebra 3 (1973). 159-170. Krause, G. Descending chains of submodules and the Krull dimension of noetherian modules J. Pure and Appl. Algebra 3 (1973), 385-397. Krause, G. and Lenagan, T.H. Transfinite powers of the Jacobson radical Comm. Alg. 7 (1979). 1-8. Lenagan, T.H. The nil radical of a ring with Krull dimension Bull. London Math. SOC.5 (1973). 307-31 1. Lenagan, T.H. Reduced rank in rings with Krull dimension Ring theory (Proc. Antwerp Conf.), 123-131. (Lecture Notes in Pure and Appl. Math. 51, Dekker. New York. 1979). Meldrum. J.D.P. Near-rings and their links with groups (Pitman Research Notes No. 134, London, 1985) Pilz. G. Near-rings (North-Holland, Amsterdam, 1977 (2nd Ed. 1983)). Scott, S.D. Near-rings and near-ring modules (Doctoral Dissertation), Australian National University) 1970. Scott, S.D. Tame near-rings and N-groups Proc. Edinburgh Math. SOC. 23 (1980). 275-296.
This Page Intentionally Left Blank
Near-rings and Near-fields, G.Betsch (editor) Q Elsevier Science PubMers B.V.(North-Holland), 1987
185
SOLUTION OF AN OPEN PROBLEM CONCERNING 2-PRIMITIVE NEAR-RINGS
J H MEYER and ANDRIES P J VAN DER WALT
Department of Mathematics, University of Stellenbosch, Stellenbosch 7600. South Africa*.
ABSTRACT We construct a 2-primitive non-ring with identity which will clarify an open problem posed in 1971. In this note we solve an open problem about 2-primitive near-rings raised by Betsch [ l ] in connection with his theorem 4.8, which reads as follows:
Let R be a r i g h t near-ring with i d e n t i t y 1 which i s not a ring and which i s 2-primitive on the R-module G. I f R contains an idempotent e of rank 1, then R contains a minimal l e f t ideal or we have e# 1
(+I
(By the rank of an element
and
r E R
eG) = 0 .
Ann ( G R
we mean the cardinality of the set of non-
zero orbits of Aut G on rG.) R The problem is to find an example to show that the exceptional case can indeed occur.
(+I
We shall show that matrix near-rings over some near-fields
furnish examples of this phenomenon. Matrix near-rings were introduced in Meldrum and Van der Walt [ 2 1 .
For the
convenience of the reader we provide the pertinent definitions: For a natural number over the near-ring {fr,:Rn+Rnl
R
n,
we define the n
as the subnear-ring
1
where
X
n m a t ~ xnear-ring
of M(Rn)
Mn.ln(R)
generated by
fr.
...,
r
> =
17 t. = rr. and tk = 0 if k # i. Here Rn * I n denotes 8 (R,+), and n-vectors are written with pointed brackets and are 1 thought of as column vectors. The elements of Mn(R(R) are called n X n 17
< tl,t2,
. ..,
tn >,
with
matrices over R and will usually be denoted by upper case letters. In particular, if R
happens to be a (right) near-field, then
zerosymmetric (right) near-ring with identity
1
1
fll + f22
+
Mn(RI
... +
f1 nn'
is a See [ZI
for more details. In matrix ring theory, the matrices with zeroes in all the rows except in a fixed row, play an important role. near-rings:
If
1
We will generalize this concept to matrix 1. m n, we call a matrix of the form fki,Uj, with J j=l
1
*Financial assistance by the Council for Scientific and Industrial Research is gratefully acknowledged.
J.H. Meyer and A.P.J. van der Walt
186
and U E Mn(R) for all 1 < j < m, a k-th row m a t r i x . 3 j tells us more about the construction of such matrices:
r. E R
Lemma 1
Lemma 1 Let 1 < k < n. If u i s any matrix, then fziU i s a k-th rotJ m a t r b for a l l r E R, 1 < i 6 n. Furthermore, there e x i s t s an expression for fr,u which c o n s i s t s e n t i r e l y of functions of the form ki
fs ( s E R), kj
apart
porn parantheses and operators. Proof From the definition it follows that fr.U is a k-th row matrix. Now, ki
where
[x; p,q]
denotes f h .
1
Then, for each
j <mi,
we treat
r.. [rij; k,k. .IV. = fkR,, Vij 11
13
in the same
11
0
way, e t cetera. The next lemma is quite useful:
Lemna 2 If
i s a minimal l e f t ideal of
L
is an invertibZe then LV := {LVJL E Ll is aZso a
-1
element of M ~ ~ R ) i.e. , v &sts, minimal l e f t ideal of Mn,(R).
mnl,(R)and
V
Proof It is routine calculation to check that LV is a left ideal. Suppose it is not minimal.
Then we have a proper subset .$
left ideal properly contained in L,
properly contained in
LV.
But then
of
L
(sV)V-' =
such that .$V is a
s
is a left ideal
a contradiction.
0
Now let F be an infinite near-field which is not a field. The example we are going to construct works for any n 2 2, but for the sake of convenience, we stick to the case n = 2.
141,
R
is 2-primitive on
Thus, let R := M (F). As was pointed out in 2 2 G := F The next proposition is fundamental in
.
this paper:
Proposition 1 AutRG G
(F
Io},.)
where the elements of
F
103 act on
by multiplication on the r i g h t .
Proof It is easy to check that < a,B > @ ( A ) = < a A , B A > morphism, suppose
A
'$: F
{Ol
is a monomorphism.
+
defined by
To see that it is an epi-
is an R-automorphism of
since, if < 1,0 > A = < 0,B >,
Then ci f 0 ,
AutRG
G, and that < l,O>A = c a , @ >. 1 then < 1,0 > A = (fll< 1,0 >)A
1
1
f (< 1,0 28) = f l l < 0,B > = < 0,O > which implies B = 0 . But only the 11 zero-element < 0 ' 0 > can be mapped onto < 0,O > by an R-automorphism. Thus,
=
if < y,6 >
is any element of
(fTl < 1,O > ) A
+
(f:l
6 + f21 )A = 6 f21(< 1,0 >A) =
G, < y,6 >A = (fTl < 1,0 >
< 1,0 > ) A = fy ( < 1,0 > A )
11
+
187
Solution of an open problem
'
fY1 < a,B > + fZ1 < a,B > = < y a , h > = < y,6 > @ ( a ) . Thus, O(a) = A ,
a E
with
{o}.
F
0
From this proposition it follows that a full set of orbit representatives of Aut
is given by {< l,h >Ih E F] U {< 0,1 >, < 0 , O > I . Now consider 1 fll. It certainly is an idempotent in R and, since 1 = {< h,O >lh E F], fll is of rank 1. Also, fil f I, the identity
G
R
on
G
the matrix f1 G 11 matrix.
At this p i n t we choose a specific F , field
arising from Q(x),
(F,+,o)
namely the infinite Dickson near-
the field of rational functions over the
rationals, by defining multiplication as follows:
where d = := degree (p(x)) - degree (q(x)). q (x) or [ 5 1 for more details. We need the following:
Lemna
See
[3],
example 8.29,
Let F be the i n f i n i t e near-field as defined above and suppose i s a f i r s t row matrix. Then u< 1,0 > = < -,R(x) O > f o r some s (XI poZynomiaZs R(x) rmd S(x) i n Qtx]. Furthemore, there i s an i n f i n i t e sequence of functions {niliEN c_ F such that 3
U E IM2(F)
P. (XI
f o r polynomiaZs
and Q .(x) where
P ,( x )
Pi(x)
i.e.
d(-)
QiW
G k (say) for a l l
such a m y that
{d(qi)}
.
i E N
Qi (XI
1 i s bounded from
Also, we may choose the
q.
above,
in
i s not bounded from below.
h o o f The infinite sequence of functions we have in mind here is t h e set 1 1 { ~ j ~ =f o~r ,m E N large enough. Note that d(-) = -i, so that X
X
1
{d(T)lY%
is not bounded from below.
We use the following induction argument
X
to justify our claim: matrices of the form
and
U< 1
1, > ~ = < 0 X
is fixed for all that
+
Since
x fll
is a first row matrix, it can be built up from
U
and
fY,(h,y E F),
a -1,0 s(x) ' xi
i s 1.
>
for all
Now suppose V
by lemma 1.
Note that d(-) r(xJ s (x) are first row matrices such
i 3 1.
and W
J.H. Meyer and A.P.J. van der Walt
188
for all
i
> m3
(say) where m
3
> m1'
Pi(x) Since d(-)6 qi (x)
N
kl and
d(g(x h(x
+
+
+
+
i))
g(x + ; i+ i)pi(x) a
kj (say), such that
d(
h(x
+
+
i, we must have
is independent of
i)
< k3
for all
i
>
1.
i)qi(x)
go Actually, we also have to consider the case '"::f
V(V f I), but since we can g (x)
assume
to be a first row matrix, by lemma 1 ,
fh(x) V = 0. If 12
V = I, then
go
go 12
V
v
= fh(x)
12
and we have considered this case already at the beginning of our
189
Solution of an open problem Taking lemma 1 into consideration, the result follows.
induction process.
0
We will use lemma 3 to prove the next proposition, from which it will follow that Ann (G R
Proposition
1
f G) = 0, showing that the exceptional case 11
Let
2
can occur.
(+)
be t h e i n f i n i t e near-field as i n lemma 3.
F
Suppose
{ B I , B 2 , ..., Bm 1 i s a f i n i t e s e t of o r b i t s of AutRG on G. Then m
U 8 ) = 0. i=l i
Ann (G R
Proof
Bi
We may assume that the
are non-zero.
Consequently, an orbit
< 1,0 >, < 0,1 > or < 1 , x >
may have representatives:
8
for a non-zero h E F.
Now suppose m ig18i) f 0 .
AnnR(G
l<j G m
and U
B
such that U B . f 0 for a certain 1 for infinitely many orbits BT. We have to consider
U E MZ(F)
Then there is a matrix = 0
three cases: (i) 8. is generated by < 1,0 >. I Suppose U< 1,0 > = < a,B > with (say) a # 0, -1 then fyl u< 1,0 > = < 1,0 > and also,
-1 fyl U< 1.1 > = < O , O
for almost
?
-1 h E F, i.e
all
U< 1,x > = < 0,O >
fyl
E F.
a finite number of
according to lemma 1.
A E F, except maybe
for all
-1 Furthermore, f y l U
is a first row matrix,
So, without loss of generality, we may suppose
is a first row matrix such that U< 1,0 > = < 1,0 >
< 0,O >
for almost all
of functions
for all
i
different
>
1.
ni,
for almost all
But since
rl..
for infinitely many
So
B.
I
Pi (x) and such that d(-)< k (say) Qi (x) belong to different orbits for
in Q[x]
< l,rli >
we must have
Pi(X)
d(rli ) = -d(rli)
By lemma 3 there is an infinite sequence Pi (x) such that U< l,n > = < 1 + erli,O > for i Qi (X)
and Qi(x)
- - -1
h E F.
-
‘Qi’iEN
polynomials Pi(x)
U
U< 1,h > =
and
u<
1,Q. > = < 1 1
+
Pi (x)
Q~(X)O‘it0
>
=
< 0,o >
This implies that
-1
-rli
(inverse with respect to
o
)
i.
This is clearly a contradiction, since Pi (x) ( ) can be made arbitrarily large, while d
cannot be generated by
Q, (x)
< 1,0 >.
k.
J.H.Meyer and A.P.J. vun der Walt
190
i s generated by < 0,l >.
(ii) 8
1
Then, in the same way as (i), we can find
a matrix V E M2(F) such that V< 0,1 > = < 1,0 > and V< 1,x > = < O , O > for almost all A E F. But then, if U = V(f 112 + f21 1 + f22), 1 we have and U< l,h > = < 0,O >
U < 1,0 > = < 1,0 >
E F, which
for almost all
is impossible by (i).
(iii) 8
i8
1
< 1,x > f o r a mn-zero
generated by
A E
F.
As before, we can
such that V< 1 , x > = < 1,0 >
find a matrix V E M2(F)
and
V< 1,C > = < 0,O > for almost all 5 E F. But then, if 1 x 1 U = V(fll + f21 + f22), we have U< 1,0 > = < 1,0 > and U<
l,C >
>
= < 0,O
5 E F, a contradiction by (i).
for almost all
Thus, we must have m
U B
Ann (G R
-
By a result in 111, 55 (pp 89 A :=
Aut G , R
i=l i
) = 0.
0
2 there is a mapping m E MA(F ) ,
go),
where
such that
rn< X , O > andrnc
A,B
>
= < X,O >
for all
X E
0,O >
for all
h,@E F ,
= <
F
B
# 0.
Hence, by proposition 2 , cur example also shows that there are functions in 2
M (F ) A
which are not matrices.
In the next result we shall see that, although there are minimal R-subgroups
A
in this near-ring, for example ideals in R.
{fll
6 + f211X,6 E
there are no minimal left
F),
This implies that not only does the exceptional case
(+)
occur
in our example, but it occurs exclusively.
Proposition
3
Let
be t h e same near-ring as before.
R
Then R
does *lot
contain a minimal left i d e a l .
Proof matrix in
Suppose
f.
f
R
is a minimal left ideal of
and that
U
is a non-zero
According to theorem 4 . 3 [ 11, any non-zero element of
rank 1 and consequently, rank finding an element X
U =
1.
f
W e shall produce a contradiction by
in a minimal left ideal such that rank X
> 2.
Since
non-zero, it must be non-zero on an infinite number of orbits of Aut
R
according to proposition 2 .
u< and
Thus, we can choose a,@ E F '
@
on
>
l
=
< y o X , & o h >,
with
A E
F
{O}.
a
U<
1,cC > = < Y,6
UT< 1,a-l0 B > = U< 1,B > = < Y O
A,&.
h >.
Now UT E LT, which is a minimal left ideal according to lemma 2 , because -1 1 a-1 T = f l l + f22 exists. Let us define the matrix W as follows: -1
w = +
f2* 2d-l
if
6 # 0
U
is
G,
such that
f l l + f22. Then
UT< 1,l > = and
IO),
G
1,a > = < y,& > # < 0 , o >
U< 1,B
Let T be the matrix
has
Solution of an open problem Then, if
5
:= a
-1
o B , we have 5
# 1
while
wuT< 1,l >
=
w< y,6 > = < l,x >
1,5 >
=
w< y o X , 6 0 X > = < X,xoX >.
and wuT<
Moreover, W T E LT. Now consider the matrix 1 + f12 1 + f21 5 v = fll The inverse of V
v-1
=
191
+
1 2. fZ
is
(1-0-1 1 fll (€11
f;; )
+
-1 f-c(l-c) 1 22 (fll + f;;)
+
+
1
f22
which can be easily verified by showing that W-l= V-lV = I. According to 1-a 2, hV is a minimal left ideal of R and furthermore WIppv< 0,l > = WUT< 1,l > = < l,x > and WIpI11< 1,0 > = WU"< 1,c > = < X,xoX > . Let X = Y ( 1 + WTV) - Y, where Y = f1l l + f22. x Then X E LTV, and
x<
0,l > = Y ( < 0 , l > + < 1,x = Y<1,1
Clearly, < 1,(x + 1)'
-
x >
and
+
x >
-
< l,(x
of Aut G on XG which implies rank R any minimal left ideals.
X
>)
-
Y< 0 , l >
< 0,x >
+
1)x > 2.
belong to different orbits
Consequently, R
does not have 0
REFERENCES Betsch, G., Some structure theorems on 2-primitive near-rings, Colloquia Mathematica Societatis Jdnas Bolyai, 6 . Rings, modules and radicals. Keszthely (Hungary), 1971. [ 2 1 Meldrum, J.D.P. and Van der Walt, A.P.J., Matrix near-rings. To appear in Archiv der Mathematik. [31 Pilz, G . , Near-rings (North Holland, 1977). Van der Walt, A.P.J., Primitivity in matrix near-rings. To appear in [4] Quaestiones Mathematicae. [5] Zemmer, J.L., Near-fields, planar and non-planar, The Math. student, 32 [ 11
(1964)
.
This Page Intentionally Left Blank
Near-ringa and Near-fsldr, C. Betrch (editor) 0Elsevier Science Publishere B.V. (North-Holland), 1987
193
ARE THE JACOBSON-RADICALS OF NEAR-RINGS M-RADICALS ? Rainer MLITZ Institut fur Angewandte und Numerische Mathematik, TU Wien Wiedner HauptstraRe 8-10 A-1040 GIien Let U be a universal class of right near-rings (i.e. a class o f near-rings satisfying the law (x+y)z = xz+yz which is closed under taking homomorphic images and ideals). It is well known that a Kurosh-Amitsur-radical on U is a map N -t fN q N ( c means "ideal of") defined on U and satisfying : (pl) (pN+I)/I c p(N/I) for all I 4 N , (P2) P(NlPN) = 0 ( ~ 3 ) P(PN) = PN (p4) pI = I , I Q N imply I E ~ N. (pl) and ( p 2 ) define the Hoehnke-radicals and a Hoehnke-radical is a KuroshAmitsur-radical i f and only if it satisfies pN = 0 o) (0 # I a N PI # I ) (p5) (see [31 and e.g. [61). The Jacobson-radicals J2 and J3 are Kurosh-Amitsurradicals (at least for zero-symmetric near-rings), Jo and J1 are not because o f the lack o f ( p 3 ) - see [41. Y
3
-
Recently, the concept of Kurosh and Amitsur has been generalised to that of M-radicals (which include considerably many interesting factorisations of various structures) see [51. For our near-ring case, M will be a binary relation
-
on the universal class U satisfying the following conditions: (MI) A M N * A # 0 is a subnearring of N , ( ~ 2 ) N M N for all N # 0 in U , (M3) A M N * A M (pN or A = 0 for every homomorphism cp defined on N , (M4) A M N , I q N and A c I imply A M I . An M-radical is then a Hoehnke-radical satisfying (p5M) pN = 0 c) (A M N * pA # A ) . Notice that ( p l ) ,(p2) and (p5M) always imply (p4M) p A = A , A M N - A c p N . Direct application of the above yields Proposition 1 : If M i s any r e l a t i o n s a t i s f y i n g (MI) t o (M4) and containing the r e l a t i o n of being a nonzero ideal, t h e following assertions are equivalent: (1)
p
i s a Kwosh-Amitsur-radical
which is M-strong ( i . e . f u Z f i l s (p4M)) ;
R. Mlitz
194
(2)
p i s both an M-radical and a Xurosh-Amitsur-radical
;
.
p i s an &radical fuZfiZZing (p3) (3) N o t i c e t h a t P r o p o s i t i o n 1 i s v a l i d f o r Q-groups
.
A r e c e n t r e s u l t by Anderson, K a a r l i and Wiegandt ( I l l ) t h e n y i e l d s
C o r o l l a r y 1 : For zero-symmetric near-rings,
J 2 and J3 are M-radicals for
every reZation M s a t i s f y i n g (MI) t o (M4) and situated between the r e l a t i o n s of being a nonzero ideal and that of being a nonzero right-invariant subgroup of ( i . e . a subgroup A of N s a t i s f y i n g AN c A I .
N
By t h i s c o r o l l a r y , P r o p o s i t i o n 2.7 i n [51 y i e l d s C o r o l l a r y 2 : F o r zero-symetric near-rings and every r e l a t i o n M with (MI) t o (M4) situated between that of being a nonzero ideal and that of being a right-
invariant subgroup (# 01, the radical classes R2 and R3 of J2 and J3 are closed under M-extensions : A M N , A E R , N / 4 > E R mp N E R
.
(where denotes the ideal of N generated by A I
As e x h i b i t e d i n [5J, f o r a s s o c i a t i v e and a l t e r n a t i v e r i n g s . t h e M - r a d i c a l s w i t h r e s p e c t t o t h e r e l a t i o n o f being a nonzero s u b r i n g a r e e x a c t l y t h e s t r o n g Kurosh-Amitsur-radicals ; thus, f o r every r e l a t i o n M w i t h ( ~ 1 )t o (M4) s i t u a t e d between t h a t o f being a nonzero s u b r i n g and t h a t o f b e i n g a nonzero i d e a l , t h e M-radicals f u l f i l (p3). T h i s means, t h a t t h e sum R o f N c o i n c i d e s w i t h pN
.
N o f a l l r a d i c a l M-subobjects
P
I n general, we o n l y know t h a t R N i s c o n t a i n e d i n p N ; P thus, i n [51, besides t h e general case, “approximable r a d i c a l s “ (pN i s generated
as an i d e a l by a r a d i c a l M-subobject) and “attainable r a d i c a l s ” (pN = R N) were studied
. All
P
t h e n a t u r a l examples t u r n e d o u t t o be a t t a i n a b l e . I n t h e f o l l o w i n g
a n e a r - r i n g example o f an approximable, b u t n o t a t t a i n a b l e M - r a d i c a l i s g i v e n Example : F o l l o w i n g
[51, P r o p o s i t i o n 4.3
,a
.
class S o f near-rings i s the
semisimple c l a s s o f an M - r a d i c a l i f and o n l y i f N E S i s e q u i v a l e n t t o t h e prop e r t y t h a t every M-subobject o f N has a nonzero homomorphic image i n S t h e r e i t i s e a s i l y seen t h a t f o r any r e l a t i o n f i e s Nc M N
M
.
From
w i t h ( ~ 1 )t o (M4) which s a t i s -
f o r a l l non zero-symmetric n e a r - r i n g s N (where Nc denotes t h e con-
s t a n t p a r t o f N), the c l a s s S o f a l l zero-symmetric n e a r - r i n g s i s t h e semisimple c l a s s o f an M-radical on t h e c l a s s o f a l l n e a r - r i n g s
.
N a t u r a l examples o f such
r e l a t i o n s M a r e those o f being a nonzero subnearring resp. a nonzero ( i n v a r i a n t )
. The corresponding r a d i c a l i s g i v e n by pN =and i s t h e r e f o r e approximable . Suppose now t h a t p i s a t t a i n a b l e ; t h e n (by P r o p o s i t i o n 1) p i s N-subgroup
a
Kurosh-Amitsur-radical whose semisimple c l a s s i s homomorphically c l o s e d and contains a l l near-rings with zero-multiplication
.
By a r e s u l t o f Betsch and K a a r l i
(121, Theorem 3.3) S has t o be t h e c l a s s o f a l l n e a r - r i n g s , a c o n t r a d i c t i o n It follows t h a t o i s not attainable
.
.
195
Are the Jacobson-radicals of near-rings M-radicals? The above example shows t h a t (pl),(p2) q u e n t l y , t h e r e may be some hope t o f i t
and ( p 5 M ) do n o t i m p l y (p3); conse-
Jo and J 1 i n t o a w e l l - d e s c r i b e d (see 151)
r a d i c a l t h e o r y by t h e use o f a s u i t a b l e r e l a t i o n M M-strongness ( i . e . dical
.
. As
we a l r e a d y know, t h e
( p 4 ~ ) )i s a necessary c o n d i t i o n f o r a r a d i c a l t o be an M-ra-
I t i s s t r a i g h t f o r w a r d t o see t h a t (p4M) i s e q u i v a l e n t t o t h e M-regulari-
t y o f t h e corresponding semisimple c l a s s S , i . e . t o :
N E S
,A
M N
*A
has a nonzero homomorphic image i n S
.
P r o p o s i t i o n 2 : The semisimple classes of J o and J 1 are M-regular f o r every reZation M s a t i s f y i n g (MI) t o (M4) and contained i n the r e l a t i o n of being a non-
.
Consequently, the radical classes of Jo and J 1 are M-strong zero l e f t ideal f o r these r e l a t i o n s M Proof :
.
L e t L be a nonzero l e f t i d e a l o f a J1-semisimple n e a r - r i n g ; t o show
t h e l e f t - i d e a l - r e g u l a r i t y o f S1 (corresponding t o J1),
i t s u f f i c e s t o show t h a t
t h e r e i s some n o n t r i v i a l L-group o f t y p e 1. To t h i s aim t a k e an a r b i t r a r y N-group G o f t y p e 1
. LG
= 0 i m p l i e s L G AnnNG
. Since
the i n t e r s e c t i o n o f the
N - a n n i h i l a t o r s o f a l l N-groups o f t y p e 1 i s J1(N) = 0, t h e r e must be some G s a t i s f y i n g LG # 0
. In
such a group G, t h e r e i s some g w i t h Lg # 0, hence Lg = G
( s i n c e Lg i s a normal N-subgroup i n t h e simple N-group G); moreover, t h i s shows t h a t G i s a s t r i c t l y c y c l i c L-group. Thus, i f I H i / i E I l
i s a c h a i n o f normal
L-subgroups o f G w i t h Hi # G f o r a l l i E I,we have LHi = 0 f o r a l l i E I,and thus LH = 0 f o r t h e u n i o n H o f t h e Hi
.
mal L-subgroup K i n G s a t i s f y i n g LK = 0 type 1
.
By Z o r n ' s Lemma, t h e r e i s a maximal nor-
.
G/K i s t h e n a n o n t r i v i a l L-group o f
F o r Jo t h e p r o o f i s e s s e n t i a l l y t h e same ( i n t h i s case, t h e Hi
# G do
n o t c o n t a i n L-generators o f G, i m p l y i n g t h a t K does n o t c o n t a i n L-generators o f G)
.
T h i s p r o p o s i t i o n g i v e s some hope f o r 3, and J 1 t o be M - r a d i c a l s f o r "good"
. However,
relations M
a t l e a s t f o r J1, t h i s hope i s destroyed by t h e f o l l o w i n g
P r o p o s i t i o n 3 : Let p be a Hoehnke-radical on a universal c l a s s U of near-
.
rings If U contains a t l e a s t one near-ring N s a t i s f y i n g pN # 0 and p(pN) = 0 then p i s not an M-radical on L1 f o r any r e l a t i o n ifl s a t i s f y i n g ( b l l ) t o (M4)
.
Proof : Suppose t h a t p i s an M - r a d i c a l ; t h e n by (p5M), pN # 0 i m p l i e s t h e e x i s t e n c e o f some A M N w i t h pA = A i n pN
.
Now, ( M 4 ) i m p l i e s A M pN
assumption
. Because o f
, yielding
p(pN)
(M3) and ( p l ) , A i s c o n t a i n e d
# 0 i n contradiction t o our
.
C o r o l l a r y 3 : For most of the universal claeses U df near-rings (including that of a l l zero-symmetric near-rings with conanutative addition and a l l larger universal c l a s s e s ) , J 1 i s not an M-radical f o r any r e l a t i o n M s a t i s f y i n g (MI) t o (M4)
.
,
R. MIitz
196
Proof : F o l l o w i n g K a a r l i ([41), i t s u f f i c e s t h a t , f o r some f i n i t e group G w i t h a proper nonzero subgroup H, t h e c l a s s U c o n t a i n s t h e n e a r - r i n g o f a l l mappings from G i n t o G which preserve 0 and H (which t h e n f u l f i l s t h e r e q u i r e -
.
ments o f P r o p o s i t i o n 3 above)
F o r Jo t h e s i t u a t i o n i s more complicated: u n t i l now t h e r e i s o n l y one example ( g i v e n by K a a r l i [41) showing t h a t J o i s n o t idempotent
. Unfortunately, t h i s . Nevertheless, i t
example does n o t f u l f i l t h e requirements o f P r o p o s i t i o n 3 m i g h t be used t o o b t a i n a p a r t i a l answer f o r Jo
. To
c o n s t r u c t h i s example,
K a a r l i s t a r t s w i t h a c y c l i c group A o f o r d e r 4 ( g e n e r a t o r : a),
and t h e group
B o f a l l {O,lbsequences w i t h a f i n i t e number o f nonzero e n t r i e s . On t h e d i r e c t sum G o f these groups he considers t h e (zero-symmetric) n e a r - r i n g o f t r a n s f o r mations S generated by a t r a n s f o r m a t i o n so s a t i s f y i n g 2so(a+b)f! B f o r a = a. ( f o r more d e t a i l s see [41) . B i s t h e n an S-subgroup o f G, S/(B:G)S # 0 i s 2 1 - p r i m i t i v e , (B:G)S = 0 , i m p l y i n g Jo(S) = (B:G)S . F u r t h e r , K a a r l i considers t h e (zero-symmetric) n e a r - r i n g T o f a l l t r a n s f o r mations t on G which s a t i s f y :
[ o tg =
I
f o r g e 6
a(t)ao
1 B(t)ao
t
or
g = o
b ( t ) E 2AtB
E 2A
f o r a l l g E BLC
,
f o r a l l g E CLCOI
where C denotes t h e subgroup o f B c o n s i s t i n g o f a l l elements w i t h an even number o f nonzero e n t r i e s
.T
i s shown t o be 0 - p r i m i t i v e
.
The n e a r - r i n g N d e f i n e d on t h e d i r e c t sum o f t h e a d d i t i v e groups S+ and Tt by t h e f o l l o w i n g m u l t i p l i c a t i o n has S as i t s J o - r a d i c a l : (s+t)(s'tt') = with k(s',t) =
1
0
SS'
t
k(s',t)so
f o r s ' G f! B
B(t)so
+
or
f o r s'G E B
tt'
s' = 0
, s' #
0
.
Now, i f L i s any J o - r a d i c a l l e f t i d e a l i n N, by P r o p o s i t i o n 2 i t i s c o n t a i n e d i n Jo(N) = S and hence i n J o ( S ) = (B:G)s some t E T w i t h B ( t ) = 2 (stt)(mtt')
-
, we
( s + t ) t ' = sm
.
I f we t a k e some m # 0 i n (B:G)S and
obtain
+
2so ;
. Thus, - ( s + t ) t ' i s n o t i n (B:G)s and ( I ~ : G ) ~ c o n t a i n s no nonzero l e f t o f N . It follows t h a t N i s a near-ring without Jo-radical l e f t - i d e a l s i s n o t Jo-semisimple . Using t h i s example, we o b t a i n :
sm belongs t o t h e i d e a l (B:G)s o f S, b u t 2so(a+b) f B f o r a = a, (stt)(m+t') ideal which
P r o p o s i t i o n 4 : For most of the universal classes U of near-rings ( i n c h -
ding t h a t of a l l zero-symetric near-rings with c o m t a t i v e addition and a l l larger universal c l a s s e s ) , Jo is not an M-radical f o r any r e l a t i o n M s a t i s f y i n g
(MI) t o (M4) und contained i n the r e l a t i o n of being a nonzero l e f t ideal
.
197
Are the Jacobson-radicals of near-rings M-radicals?
The q u e s t i o n whether Jo i s an M - r a d i c a l f o r some n a t u r a l r e l a t i o n M remains thus open. I n connection w i t h P r o p o s i t i o n 3 i t would h e l p knowing t h e s o l u t i o n o f the following
Is t h e r e a (zero-symmetric) n e a r - r i n g N ( w i t h commutative a d d i -
Problem :
t i o n ) s a t i s f y i n g Jo(N) # 0 and J o ( J o ( N ) ) = 0 ?
REFERENCES 1
2
3 4 5
6
Anderson, T. , K a a r l i , K. and Wiegandt, R. , On l e f t s t r o n g r a d i c a l s o f near-rings, Manuscript . Betsch, G. and K a a r l i , K. , S u p e r n i l p o t e n t r a d i c a l s and h e r e d i t a r i n e s s o f semisimple classes o f n e a r - r i n g s , C o l l o q u i a Math.Soc.Janos B o l y a i 38 (Radical Theory, Eger 1982), 47-58, North-Holland Pub1 .Camp., Amsterdam/ Oxford/New York , 1985 Hoehnke, H.-J. , Radikale i n allgemeinen Algebren, Math.Nachr. 32 (1966), 347-383 K a a r l i , K. , On Jacobson t y p e r a d i c a l s o f n e a r - r i n g s , Acta Math. Hung., t o appear . MBrki, L., M l i t z , R. and Wiegandt, R. , A general Kurosh-Amitsur r a d i c a l t h e o r y , Manuscript . M l i t z , R. , R a d i c a l s and semisimple classes o f Q-groups, Proc. Edinburgh Math.Soc. 23 (1980), 37-41
.
.
.
SUMMARY
.
An example o f a non idempotent &radical o f near-rings i s given The behaviour of the Jacobson-radicals of near-rings with respect t o the concept of M-radical is studied I t turns out that J 2 and J are M-radicals f o r s e u e m l natural relations M, t h a t on most of the uniuersaj classes of near-rings J is not not an M-radical f o r relations M contained in that of being a nonzero ? e f t ideal and t h a t J1 i s in general not an M-radical f o r any r e l a t i o n M
.
.
This Page Intentionally Left Blank
Near-rings and Near-blds, C. Betsch (editor) 0 Elsevier Science Publishers B.V (North-Holland), 1987
199
ON MEDIAL NEAR-RINGS
S i l v i a P e l l e g r i n i Manara # ABSTRACT We study t h e n e a r - r i n g s N i n which x y z t = x z y t f o r a l l x , y , z , t ~ N and we c a l l them medial. We prove t h a t a l l such n e a r - r i n g s have a s u b d i r e c t s t r u c t u r e very s i m i l a r t o t h a t o f t h e s t r o n g l y I.F.P. nearr i n g s "61). I n t h e p a r t i c u l a r case o f medial r e g u l a r n e a r - r i n g s , we observe t h a t they g e n e r a l i z e t h e 8-near-rings o f L i g h [ 5 ] and t h e sem i - r i n g s o f Subrahamanyam [ 131 Moreover we o b t a i n a necessary and s u f f i c i e n t c o n d i t i o n f o r a nearr i n g t o be medial, r e g u l a r and s u b d i r e c t l y i r r e d u c i b l e . Afterwards we have s t u d i e d t h e s e t o f n i l p o t e n t s elements o f a medial n e a r - r i n a . We f i n d c o n d i t i o n s f o r t h e s e t Q(N) o f t h e n i l p o t e n t elements o f N t o be an i d e a l . I f N i s a medial r i n g , Q(N) i s an i d e a l and i t c o i n c i des w i t h t h e r a d i c a l s o f N. I n t h i s case N/Q(N) i s a s u b d i r e c t sum o f i n t e g r a l domains.
.
1. INTRODUCTION The i d e n t i t y " x y z t = x z y t " has been developed i n t h e t h e o r y of quasi-groups:
an
impulse t o t h e i r i n v e s t i g a t i o n goes back t o t h e ' 4 0 s w i t h Toyoda's theorem (see Toyoda, On axioms o f l i n e a r f u n c t i o n s , Proc. Imp. Acad. Tokyo, 1941, 17). We study t h e n e a r - r i n g s N i n which x y z t = x z y t f o r a l l x,y,z,t medial.
For n e a r - r i n g s , t h e
t y " and i t s study
f i t s into
E N and we c a l l them
m e d i a l i t y i s a c o n d i t i o n o f " p a r t i a l commutativithe
set
of
studies o f the near-rings w i t h
p r o p e r t i e s o f p a r t i a l commutativity. We see t h a t t h e medial n e a r - r i n g s general i z e t h e 8 - n e a r - r i n g s o f L i g h [ 5 l a n d t h e semi-rings o f Subrahamanyam [ 13
1
.
The necessary and s u f f i c i e n t c o n d i t i o n f o r a n e a r - r i n g t o be medial, r e g u l a r and s u b d i r e c t l y i r r e d u c i b l e p u t s i n evidence t h e connection between t h i s s t r u c t u r e and t h e weakly commutative n e a r - r i n g s . I n t h e study o f t h e s e t o f n i l p o t e n t elements, t h e m e d i a l i t y works i n a s i m i l a r way as t h e I.F.P.
We f i n d
c o n d i t i o n s f o r t h e s e t Q ( N ) o f n i l p o t e n t elements o f N t o be an i d e a l and show t h e connection t o t h e various prime i d e a l s o f N. I f N i s a medial r i n g , Q(N) i s an i d e a l , and i t c o i n c i d e s w i t h t h e r a d i c a l s o f N. I n t h i s case t h e med i a l i t y presents i t s e l f again as a c o n d i t i o n o f weak c o m m u t a t i v i t y : i n f a c t #Address:
S i l v i a P e l l e g r i n i Manara, F a c o l t B d i I n g e g n e r i a d e l l ' U n i v e r s i t a d i Brescia, V i a l e Europa 39. 25060 BRESCIA. Work c a r r i e d o u t on b e h a l f o f M.P.I.
S.Pellegrini Manara
200
N/Q(N) i s a s u b d i r e c t sum o f i n t e g r a l domains.
2. PRELIMINARIES N w i l l i n d i c a t e a l e f t n e a r - r i n g ; f o r t h e d e f i n i t i o n s and t h e fundamental nomention. An element e E N i s idempo-
t a t i o n s we r e f e r t o [ l o ] w i t h o u t e x p l i c i t e 2 t e n t i f e = e; i f x e N we i n d i c a t e A (N) = d r i n g N = N + N "mixed" i f No# { O ) and Nc# o c i m p l i e s xzy = 0 f o r a l l x , y , z ~ N , s t r o n g l y I.F.P.
I01
N / x =
and r e g u l a r i f W x a N 3
x ' x a r e i d e m p o t e n t elements f 4 ]
t y ~ /N x y
.A
=
0 1 and we c a l l a n e a r -
n e a r - r i n g N i f I.F.P.
if xy=O
I.F.P.
i f e v e r y homomorphic image i s
XX'X.
In a r e g u l a r n e a r - r i n g x x ' and
.
We call a near-ring N medial if xyzt = xzyt for all x,y,z,t in N. 3.
SUBDIRECLY IRREDUCIBLE MEDIAL NEAR-RINGS P r o p o s i t i o n 1. A weakly commutative $ near-ring is medial.
I n f a c t if N i s weakly commutative x y z t
=
zxyt
=
x z y t and N i s m e d i a l .
P r o p o s i t i o n 2. Let N be medial:
1. If e # 0 is an idempotent element of N, A (el is an ideal of N. d 2. I f N has a left identity, it is wakly commutative.
k
k k
3. For all x, Y E N and k integer, ixyl = x y
.
1. We know t h a t A ( e l i s a r i g h t i d e a l o f N. For ncN and y6A (N) we have: d d eny = eeny = eney = 0 and A ( e ) i s a l e f t i d e a l . d 2. I f u i s a l e f t i d e n t i t y o f N, xyz = uxyz = u y x z = y x z , f o r a l l x , y , z ~ N and N i s w e a k l y commutative. 3. T r i v i a l . P r o p o s i t i o n 3. I f N is a mixed medial near-ring, N
N N = {O 1 and NoNc = c o I f N i s m e d i a l , V XEN, WnotNo, Onn
=
OOnn
=
OnOn
0
0
o f N. I t f o l l o w s t h a t N N = I 0 1 and N N = N c o o c c
.
0
=
is an ideal of N,
On
= 0
0 and N
0
i s an i d e a l
P r o p o s i t i o n 4. The homomorphic images of a mediaZ near-ring are medial. The p r o o f i s t r i v i a l . The c o n s t a n t and t h e z e r o - n e a r - r i n g s
a r e m e d i a l . I n t h e f o l l o w i n g we w i l l
e x c l u d e t h e s e t r i v i a l cases. Theorem 1. If N is a non trivial medial subdirectly irreducible near-ring,N
satisfies one of the following properties: §
a. n e a r - r i n g N i s w e a k l y commutative i f x y z = y x z
f o r a l l x,y,z E N , [ 1 0 1
.
20 1
On medial near-rings
1. N is zero-symmetric, simple and each non-zero
idempotent is a Left identity;
2. N is zero-symmetric and the intersection of the non-zero ideals of N is non-
zero and without non-zero idempotents; 3. N is mixed and when the intersection I of the non-zero ideals of N contains an
idempotent, then I = N
.
L e t N be medial, s u b d i r e c t l y i r r e d u c i b l e and non t r i v i a l . I f N i s simple and zero-symmetric, N s a t i s f i e s 1.:
i n f a c t , i f i t has an idempotent e, Ad(e) = I 0 1
because A ( e ) i s an i d e a l . Therefore e ( e n - n) = 0, W n e N , i m p l i e s e l e f t i d e n d t i t y . L e t N be zero-symmetric b u t non simple and l e t P be t h e i n t e r s e c t i o n of t h e non zero i d e a l s o f N. Now P # t O 1 and we suppose t h a t
P c o n t a i n s an idempotent
e. I f t h e r e i s an element n E N such t h a t en # n, we have A d ( e ) # 1 0 1
,
hence
n - e n E A ( e l . Therefore P i s c o n t a i n e d i n Ad(e) and t h u s e E A ( e ) , t h a t i s d d 2 = e, a c o n t r a d i c t i o n . Otherwise e i s l e f t i d e n t i t y c o n t a i n e d i n P and P = N
e
because N i s zero-symmetric,
b u t t h i s i s excluded. Therefore P d o e s n ' t c o n t a i n
idempotents and N s a t i s f i e s 2. Suppose N i s mixed
. We know t h a t
No i s an i d e a l
o f N and so N i s non simple. I f P i s t h e i n t e r s e c t i o n o f t h e non zero i d e a l s o f N, we have P c N
0
. If P has an idempotent e #
0, i t cannot be en # n f o r n E N,
because
otherwise e E A ( e ) and t h i s i s excluded. Therefore, i f P has an idemd p o t e n t e # 0, t h i s element i s l e f t i d e n t i t y , P = N and N s a t i s f i e s 3. 0
C o r o l l a r y 1. A non trivial, medial near-ring is isomorphic to a subdirect sum of near-rings that satisfy one of the properties of Th. 1.
I t f o l l o w s immediatly from Prop. 4 and Th. 1. I n f a c t each n e a r - r i n g i s i s o -
morphic t o t h e s u b d i r e c t sum o f s u b d i r e c t l y i r r e d u c i b l e n e a r - r i n g s ( s e e [ 2 1 ) .
4.
SUBOIRECTLY IRREDUCIBLE MEDIAL REGULAR NEAR-RINGS A z e r o - n e a r - r i n g i s medial b u t n o t r e g u l a r , a c o n s t a n t n e a r - r i n g i s medial
and r e g u l a r ;
i n t h e f o l l o w i n g , an MR-near-ring i s a medial r e g u l a r
n e a r - r i n g , and i t i s non t r i v i a l i f i t i s n o t constant. I n general a r e g u l a r n e a r - r i n g i s n o t s t r o n g l y I.F.P.,
t h i s happens i f N i s
medial: i n f a c t : P r o p o s i t i o n 5. An MR-near-ring is strongly I.F. P. L e t N be an MR-near-ring and I an i d e a l o f N: i f x y e I then f o r a l l n E N , xny = x x ' x n y = x x ' n x y c I . P r o p o s i t i o n 6. Let N be a non trivinl MR-near-ring; then
202
S. Pellegrini Manara
1. N has no nilpotent elements. 2. A (xi is an ideal of N, for alZ Z E N. d 3. I f x = x x ' x t N , then A (xx') = { O ) [ A I x ' x ) ={OI] i f f x x ' [ x ' x l is a LefL d d identity of N. k 1. If n i s n i l p o t e n t , n = 0 f o r some i n t e g e r k. Now n = n n ' n i m p l i e s n ' n = k k k ( n ' n ) = n ' n = 0 (Prop. 2.3) and hence n = 0.
A ( x ) , xny = x x ' x n y = x x ' n x y = 0. d 3. I n f a c t x x ' n = x x ' x x ' n [ x ' x n = x ' x x ' x n 1 and hence x x ' ( n 2. For n
t
N, and y
[x'x(n - x'xn) = 0
1
E
.
I f A (xx') = i 0 )[Ad(x'x)
d i d e n t i t y . The v i c e v e r s a i s t r i v i a l .
=
101
1,
XX'
-
xx'n)
=
0
[ x ' x ] i s a left
P r o p o s i t i o n 7. Let N be a r.on trivial subdirectly irreducible MR-near-ring: 1. I f z is a non zero idempotent of N, A ( z l # ( 0) i f f z is constant and in
d
.
this case A l z ) = Iv d 2. If x is not constant, and x = xx'x, then xx' and x'x are not constant and non-zero idempotents and A (x) =
d
3.
0 1=
A
d
(xx') = A fx'x). d
N has a left identity.
1. I f z i s constant, A ( z ) # I 0 1 . Viceversa,let Ad(z) # { O 1 and A = n Ad(h) d w i t h A ( h ) # I 0 1 . Since N i s s u b d i r e c t l y i r r e d u c i b l e , t h e r e i s some W E A , w # 0. d Then a l s o W ' W E A (ww'w = 01, hence A i s an i d e a l . I f t h e r e i s an element 0 # YEN, 0, we g e t A (w'w) # { 0 I and so w ' w w ' w = 0, b u t t h i s i s excluded d Therefore f o r a l l Y E N ( y # 0 ) w'wy # 0 and A d ( w ' w ) = { O 1 . From
such t h a t w'wy (Prop. 6.1). Prop. 6.3,
=
w ' w is a l e f t i d e n t i t y . So f o r a l l idempotent z o f N such t h a t
A ( z ) # { O 1 and f o r a l l Y E N we have zy = z ( w ' w ) y = ( z w ' w ) =~ Oy. I n p a r t i c u l a r d i f y = z, z z = z = Oz,and z i s c o n s t a n t . F i n a l l y , i f z i s constant,A ( z ) ? N . i f d 0' y t A ( z ) , zy = 0 = Oy and y t No. d 2. L e t x=x +x be a n o t c o n s t a n t element o f N. I f x x ' i s c o n s t a n t , we o c have x = x x ' ( x + x ) = x x ' x t x w i t h x x ' x = 0 because o f N N = ( 0 1 Thereo c o c 0 c o f o r e x must be constant, a c o n t r a d i c t i o n . I f x x ' = 0, we have Ox = x and x i s
.
again constant. I n a s i m i l a r way we prove t h a t x ' x i s not c o n s t a n t . F i n a l l y from (XI { O ) = A (xx') = A (x'x). d d d 3. This f o l l o w s from t h e Th. l.,keeping i n mind t h a t now N i s r e g u l a r and
t h e above p o i n t 1.,A
cannot be o f t y p e 2 (see above p o i n t 2. and Prop. 6.2) Theorem 2. A non-trivial subdirectZy irreducible MR-near-ring with a right distributive element h is a commutative field.
203
On medial near-rings Obviously h i s zero-symmetric and so Ad(h) Under our hypotheses, AS(h) i s an i d e a l o f N
= {
. We
0 1
. Let
A ( h ) = I y e N / yh=Ol. S
suppose A ( h ) # t 0 1 and we S
A = n A ( x ) ( A ( X I # I 0 1 ) . Since N i s s u b d i r e c t l y i r r e d u c i b l e , d d AS(h) n A = L # t 0 } I f w e L we have wh = 0 and t h e r e f o r e Ad(w) # t 0 1 and w
let
.
constant, b u t w e A ( x ) = N and t h i s i s excluded. Consequently A ( h ) = { 0 ) . Mo0 S d reover N i s zero-symmetric: i n f a c t f o r a l l a € N, ( 0 a ) h = 0 = OaOh = 0 and
A ( h ) = { O 1 i m p l i e s Oa
=
S
and x # 0,
we g e t
0. I f N i s zero-symmetric i t i s i n t e g r a l : i f xy = 0
x'xy
=
0 (where x = x x ' x ) and x ' x i s a zero-symmetric
idempotent element, t h e r e f o r e y = 0. Also
-
(n
n x ' x ) h = nh
n
=
=
xx'xyxx'
-
nx'xh
=
N has i d e n t i t y : i n f a c t
0 because x x ' i s l e f t i d e n t i t y , b u t A ( h ) = S
{
0 1 and so
n x ' x f o r a l l nE N. In our case N i s medial and so f o r a l l x, Y E N , xy = =
xx'yxxx'
=
y x . So N i s a commutative f i e l d .
Theorem 3. A non trivial MR-near-ring is isomorphic to a subdirect sum of subdirectzy irareducible near-rings that are fields or near-rings with left identity in which each non constant element x = x x l x is such that xx! and z'x are left identities. I t f o l l o w s f r o m Th. 2, Prop. 7, keeping i n mind t h a t each n e a r - r i n g i s i s o -
morphic t o t h e s u b d i r e c t sum o f s u b d i r e c t l y i r r e d u c i b l e n e a r - r i n g s ( [ 21 ) . I f a n e a r - r i n g i n which each elemnt i s power o f i t s e l f ( [ 1 1
,[ 7 1 )
i s called
MP-near-ring, we observe t h a t an MP-near-ring i s r e g u l a r and t h e r e f o r e Th. 3 app l i e s . The MP-near-rings g e n e r a l i z e t h e 8 - n e a r - r i n g s o f L i g h [ 5
1 because
a wea-
k l y commutative n e a r - r i n g i s medial and t h e 8-near-rings Consist o f idempotent elements. I n t h i s way t h e r e s u l t s o f [ 5 We g i v e now a (see [ 8
characterization
1
of
are c o r o l l a r i e s
to
Th. 3 ( s e e [ 7
I).
s u b d i r e c t l y i r r e d u c i b l e MR-near-rings
I).
Theorem 4. A uear-ring N is a non trivial subdirectly irreducible MR-nearring iff: 1. it is a zero-symmetric, weakly commutative, N-simple and integral near-ring; 2. it is a weakly commutative near-ring,
with N
0'
regular, minimal ideal,
N -simple, integral; with N unique N-subgroup of N. Moreover the right c' annihilator of each non-constant element is zero.
L e t N be a non t r i v i a l s u b d i r e c t l y i r r e d u c i b l e MR-near-ring.
It s a t i s f i e s t h e
hypotheses o f Th. 1, b u t n o t t h e p r o p e r t y 2 because N i s r e g u l a r . I n f a c t , if
I i s t h e i n t e r s e c t i o n of t h e non zero i d e a l s o f N, f o r i € 1 , t h e s i s some i ' c I
204
S. Pellegrini Manara
such t h a t ii' and i ' i a r e non zero idempotent i n I and t h i s i s excluded. Theref o r e , i f N i s zero-symmetric i t i s simple and a l l i t s non-zero idempotents a r e i s a weakly commutative n e a r - r i n g (Prop. 2.2) and f o r a l l
l e f t identities; N
X E N , A ( x ) = { O l (Prop. 6.2).Hence N i s i n t e g r a l . Moreover N i s N-simple: l e t d R be an N-subgroup o f N; we have R N c R . I f X E R, t h e r e i s some X ' E N such t h a t
xx'x
=
x with xx'
being a l e f t i d e n t i t y o f N, belonging t o R. Then N C R .
V i c e v e r s a , i f N i s a weakly commutative n e a r - r i n g i t i s medial (Prop. 1 . ) and obviously
i t i s s u b d i r e c t l y i r r e d u c i b l e because i t i s simple. Moreover i t i s
r e g u l a r : i n f a c t f o r a l l a e N aN i s some
YE
N because now N i s N-simple and hence t h e r e
=
N such t h a t ay = a and f o r a l l Z E N , ayz
i n t e g r a l . Now W aEN, 3 a ' e N / aa'
=
=
az hence y z
=
z,because N i s
y and a a ' a = y a = a i m p l i e s t h a t N i s r e g u l a r .
Otherwise, f r o m t h e Th. 1. N i s mixed and No i s t h e minimal i d e a l o f N. We N t N . L e t R be a proper non-zero N-subgroup o f N and R n N # R; o c C then t h e r e a r e a c R \ N and a ' such t h a t aa' i s a non zero and non constant i -
know t h a t N
=
C
dempotent o f N (Prop. 7 . 2 ) .
Therefore A d ( a a ' ) = t 0 1
Nc. L e t us prove t h a t N 0
RCNc and so R = N
C
i s regular. If
0
group o f N i s N
i s a l e f t i d e n t i t y and
I f now RNcR, a a ' c R and aa'N
N i s weakly commutative (Prop. 2.2). which i s excluded.Therefore
, aa'
and i t cannot be xN
= (
XE
. The
No, xN
=
NcR
unique p r o p e r N-subgroup i s N because t h e unique N-sub-
0 ) because x i s non c o n s t a n t . Moreover
h + h such t h a t x ( h + h ) = y. o c o c Such y e x i s t s because xN = N, b u t now x ( h t h = y i m p l i e s xh t h = y and so o c o c y = 0. Therefore N i s N -simple, i t i s an i n t e g r a l zero-symmetric weakly comxN = N 0
C
0'
i n fact: for a l l Y E N
0
0
there i s h
=
0
= 10) , d f o r a l l n o t constant elements x. Viceversa, i f N i s a weakly commutative medical
m u t a t i v e n e a r - r i n g . From t h e above p o i n t 1, N i s r e g u l a r .
Finally,
0
n e a r - r i n g w i t h a minimal i d e a l ,
A (x)
i t i s s u b d i r e c t l y i r r e d u c i b l e . We prove t h a t
N i s r e g u l a r . If n = n + n i s a t y p i c a l element o f N and 0 # n ' e N 0 0' o c nn' # 0 because Ad(x) # I0 ) i f f x i s a c o n s t a n t element. Therefore ( x t 0 o = N because t h e unique N-subgroup o f N i s N and so t h e r e i s Y E N such C'
then x )N = c that
ny = n + n and f o r a l l Z E N we have nyz = nz t h a t i s yz = z. Otherwise t h e r e o c i s Y ' E N such t h a t n y ' = y and t h e r e f o r e n y ' n = n. I f n i s a zero-symmetric e l e ment t h e r e i s some n ' such t h a t n t h e a s s e r t i o n is tri v i a1
=
nn'n, so N
.
0
i s r e g u l a r ; i f n i s constant
C o r o l l a r y 2. The ideals of a non trivial subdirectly irreducible MR-nearring N are of the form N + H, with H being an idea2 of N C
I f N has i d e a l s , i t i s o f t y p e 2 of Th. 4.
.
Now i f I i s one o f i t s i d e a l s ,
205
On medial near-rings
i t i s of t h e form I = N +I w i t h I an i d e a l o f N .If I = Io+ I I = InN I =InN o c C C c' 0 0' c c Now N i s minimal, s o I = N and I cannot be an i d e a l b u t It i s normal i n N 0
0
0
.
C
and t h e n It i s normal i n Nc. C
Viceversa if I
I = N t I i s an i d e a l o f N. I n f a c t f o r C' o c n t i t i - n - n = n t ( n t i - n ) t ((11 t i c 0 c - c - 0 0 c 0 c c c - n ) - n - ( n t i - n ) 1 t ( n t i - n 1 = n f i w i t h 7 E I because 1; i s c 0 c c c c c c o c c c normal i n Nt Therefore It i s a normal subgroup o f Nt. Moreover N I = N(N t I c ) = I and nEN, n
i s
C
i s an i d e a l o f N
-
n = n
t
i
=
I and f o r a l l n, n ' c N and
t 0
C'
= NN
t
= (n
t
o
I LN
0
I
t
c o c i ) ( n ' t n ' ) - n(n' o c o
n') c
t
=
(n
i t
i ) n ' - n' o c
t
I : (n
-
t
i ) n ' - nn' =
nn' E N c I . 0
0-
C o r o l l a r y 3. A non trivial subdirectly irreducible MR-near-ring is a right duo-near-ring. § I f N i s o f t y p e 1 o f Th. 4, t h e a s s e r t i o n i s t r i v i a l . L e t N be o f t y p e 2 of
I,
i c
=I01
n
t
i ) n ' - nn'
For n
=
0, i n '
0
=
=
I nN # 0
{
0)
C
N I L I and t h i s i s excluded ( N = N
0
0
5.
t
, J i s r i g h t i d e a l o f N and so J
a r e t h e sum o f N
N and f o r
i ) ( n ' t n ' - n(n; t n ' ) = ( n t i ) n ' - n n ' E N n I = C 0 0 0 o c 0 and (Prop. 7 11, i c N C , t h a t i s I L N ; b u t t h e n I i s
(n
=
a l e f t i d e a l ,too,because
J
nNo= { 0 1 we have f o r n, n ' E
4 and I a r i g h t i d e a l o f N. I f
t h e Th
0
0
i s m i n i m a l ) . Therefore
. The r i g h t
ideals of N
and o f a normal subgroup o f Nc,and these a r e a l s o l e f t i d e a l s .
NIL RADICAL OF MEDIAL NEAR-RINGS We w i l l denote by N' t h e m u l t i p l i c a t i v e semigroup of N; if K i s a subset of N
and n is an element o f N, we w i l l We w i l l
denote
write
t h e prime r a d i c a l s o f t y p e v (
P (N), P (N), see f o r i n s t a n c e [ 9
1
P(K)
2
1,
[ 12
=GKAd(k) v
=
o,
and L(K)=N\(Ku P ( K ) ) .
1, 2 ) w i t h Po(N),
1.
P r o p o s i t i o n 8. If K is a maximal subsemigmup of N . , N is the disjoint union of
K, P ( K I and L I K ) . F i r s t o f a l l , P ( K ) n K =$because K i s maximal: i n f a c t i f y c K w i t h xy =
XE
P ( K ) n K, t h e r e i s
O E K and t h i s i s excluded. By d e f i n i t i o n L ( K ) n ( K u P ( K ) )
=
@
and t h e r e f o r e p r o p o s i t i o n holds. Theorem 5. If N is medial and K is a maximal subsemigroup of N', then for each
y
E
L(KI there is some x E K such that xy is nilpotent iff N is zero-symmetric.
L e t ' s suppose t h a t f o r
Y E
L(K) there i s
§ A r i g h t duo-near-ring (see[ 5 are ideals.
1)
XE
K such t h a t xy i s n i l p o t e n t . From
i s a n e a r - r i n g i n which a l l r i g h t i d e a l s
S.Pellegrini Manara
206
Prop. 8 we w i l l prove t h a t L ( K ) , K and P ( K ) d o n ' t have constant elements. L f K I c No:
L e t y be a constant element o f L(K); from t h e hypotheses t h e r e i s some x
E
K with
xy n i l p o t e n t , b u t f o r t h e constant element y we g e t xy = y f o r a l l x t N, a c o n t r a t y and xy a n i l p o t e n t element, w i t h x t K . Then t h e r e i s an o c k k k i n t e g e r k , such t h a t ( x ( y o t y,)) = ( x y o t y c ) = 0. We o b t a i n ( x y t y ) = By +y = o c o c = 0 w i t h R element o f N. Now Byo€ No because N i s an i d e a l of N, y E Nc b u t
d i c t i o n . Let y = y
C
0
N n N = o c
K n N =$ c
(
0 1 and t h e r e f o r e yc= 0, so L ( K ) has o n l y zero-symmetric elements.
:
L e t k be a c o n s t a n t element o f K; from t h e Prop. 3, A d ( k ) l N
0
= @
and A d ( k ) n L ( K )
=
, b u t t h i s i s absurd because o f t h e c o n d i t i o n L ( K ) c N o above.
PIKIn N =
4:
I n f a c t , t h e elements o f P(K) are r i g h t a n n i h i l a t o r s o f some element o f K, and i f a constant element y belongs t o P(K), we have xy Viceversa, l e t N be zero-symmetric,
. Since y
nerated by K L J t y l i n K there are Put x1x2.
".
x .s t a i n x y ' l x yl* 1 2 k = il+ i2t
x =
~
...
x 2, , 1' x; o b v i o u s l y
... xsyi'
... t
=
y, f o r a l l x t N.
L(K) and M t h e subsemigroup o f N ' ge-
K, M c o n t a i n s K p r o p e r l y . Therefore O E M . Now i xs such t h a t xlyi1x2yi2 xsy = 0.
. ..
XE
x1x2
K. Applying t h e m e d i a l i t y k - 1 times we obilti2t...i 5 = 0, t h a t i s xyk = 0 where
...
SY
is. I f xyk = 0, a l s o x k - l x y k
Therefore xy i s n i l p o t e n t and
6.
Y E
=
=
xkyk = ( x y l k
and xy # 0 because of Y
=
o
(Prop. 2 . 3 ) .
4 P(K).
ZERO-SYMMETRIC MEDIAL NEAR-RINGS
I n t h e f o l l o w i n g our n e a r - r i n g s w i l l be zero-symmetric.
If N has subsemigroups in N', we define Q(NI = N\ mi&
U .€A
K a , where{ K' 1 is the f a =€A
of a l l maximal subsemigroups of IY-; if N has no multiplicative subsemi-
groups in N ' , we p u t Q1N) = N .
P r o p o s i t i o n 9. The set & ( N I is the set of a l l niZpotent eZements o f N. I f N has no subsemigroups i n N \ I O } Q(N) = N. I f N has subsemigroups i n N', subsemigroups i n N'.
, each element i s n i l p o t e n t . We have l e t { K a I a E h be t h e f a m i l y o f maximal
Since each Ka d o e s n ' t c o n t a i n n i l p o t e n t elements, Q ( N )
c o n t a i n s t h e n i l p o t e n t elements o f N.
Hence
i f x c Q ( N ) , x i s n i l p o t e n t be-
cause otherwise i t would generate a subsemigroup o f N ' contained i n a maximal subsemigroup K.
,
a contradiction.
207
On medial near-rings
We recall that: P2(N) 1 P ( N ) 1 1 (see [12 1 Def. 3.1, 3.2, 3 . 3 , 3.4) (a
1
PO(N)
In [ 1 1 lit is proven that P O ( N ) has only nilpotent elements. Therefore: Po
( 6 )
(N)
Q(N)
Theorem 6. If N i s a medial near-rings t h e n ( Y
1
P
-2
( N l 2 P (Nl 2 & I N ) 2 P o f N l 1
We have to prove that each prime ideal of type 1 elements o f N: from of
(a )
and ( 6
)
it will follow
( y
contains the nilpotent 1. Let P be a prime ideal
type 1 and .a a nilpotent element with order of nilpotency k . If aoP P , the-
re is t0E N such that a 1 = a0t0a0 f P; otherwise a0c P since P is prime of type 1. In this way we are able to construct two sequences of elements t a and It I n n such that a = a t a f o r n = 1, 2, ... and an f P. Moreover each a # 0, n n-1 n-1 n-1 n because a = a t a = 0 implies a E P and this is excluded. The element ak n n-1 n-1 n-1 n-1 k . “contains“ the element a 2 times; applying the mediality k-1 times, we 0 have a = 0, a contradiction. Therefore ac P. k Corollary 4. If N i s a medial near-ring i n which each prime i d e a l of t y p e 0 i s a prime i d e a l o f t y p e 1, t h e n &(hi) i s an i d e a l .
Corollary 5 . If N i s medial r i n g ,
Q ( N ) is an i d e a l .
7. MEDIAL R I N G S Theorem 7. If N i s a medial r i n g t h e n N\K subsemigroup K of N’
i s an i d e a l of N , for each maximal
.
I f y l , y are elements of N \ K , there are two elements m
m EK, such that 1’ 2 2 mlyl and in2y 2 are zero or nilpotent. Therefore there are two integer nl”l and n >1, such that (m y In’= (in y 0. If (m y )‘I= 0, applying n times the me2‘ 1 1 2 2 1 1 diality, we obtain ( m m y Inlt1 = 0. If (rn2y2ln2= 0, applying n times the me1 2 ln 2 diality we obtain ( m m y ) 2= 0. Let now n + n t n + 1. One of the terms in the 1 2 2 1 2 n . expansion of (mlm2y1 - m m y ) i s o f the form (m m y 1’1 (m m y )i2 (mlm2yl)i3 1 2 2 121 1 2 2 ... ( m 1m 2y 2 i where i 1+ i 2t ... t i k = n. Applying the mediality, we have: i ( m rn y 1’1 ( m m y 2 ... (in m y lik= ( m m y (m1m 2y 21’2. NOW i l s n l + 1 or 121 1 2 2 122 121 ‘i2bn and this term i s zero. Therefore the element m m (y - y ) is nilpotent 2 1 2 1 2 and y - y E N\K. Moreover, for all Y E N \ K and n c N, it is easy to verify that 1 2
-.-
ny and yn belong to N \ K .
208
S.Pellegrini Manara Theorem 8. I n a medial r i n g N , an i d e a l P i s minimal prime of t y p e 2 i f f
P = N
\
K f o r some m m i m a l subsemigroup K i n N'
.
x o r y belong t o N \ K because o t h e r w i s e x y t K and
If x y N~ \ K ,
L e t P = N'K.
t h i s i s excluded. Therefore N \ K i s an i d e a l o f t y p e 2. Moreover i t i s minimal of t y p e 2. We suppose t h a t H i s a prime i d e a l o f t y p e 2 contained i n N \ K . we prove t h a t
y t N\K:
YE
H. If
YE
from Th. 5 t h e r e i s some
N\K,
xy i s zero o r n i l p o t e n t . I f xy = 0, then x y H~ and t h e r e f o r e prime o f t y p e 2; b u t now x F H because x t K and then (xy)
S E
X E
XE
K such t h a t
H since H i s
H. I f xy i s n i l p o t e n t , s-1 H f o r some i n t e g e r s . Since H i s prime o f t y p e 2, xy t H o r ( x y ) E H.
A f t e r a f i n i t e number of steps
4H
y e H and s i n c e x
Y E
we o b t a i n i n any case x y H~ and t h e n
X ~ Y KE
Hence K
because otherwise x y e P and s i n c e P i s prime o f t y p e 2,
is
a
subsemigroup o f N',
subsemigroup K ' o f N'. f o r e N x K ' c P . So P
X E
H or
i t i s ~ E H Therefore . H = N\K.
Viceversa l e t P a minimal i d e a l prime o f t y p e 2. We p u t K = N \ P . also
Let
Now N \ K '
X E
YE
K,
P o r y e P.
t h e r e f o r e c o n t a i n e d i n a maximal
i s a minimal i d e a l prime o f t y p e 2 and t h e r e -
and P
N\K'
=
and
If x,
=
N\K'.
P r o p o s i t i o n 10. I f N i s a mediaZ r i n g , N\K
i s a minima2 prime idea2 of
type 1 , for a maximal subsemigroup K of N'
From Th. 7.,
N\K
of t y p e 2 and t h e r e f o r e prime o f t y p e 1.
i s prime
L e t us suppose t h a t a prime i d e a l H o f t y p e i s c o n t a i n e d i n N \ K . therefore there i s me i n t e g e r s , ( x y )
X E
S
=
K
LetyE N\K,
such t h a t xy i s zero o r n i l p o t e n t . I n any case, f o r so-
0. Applying t h e medial p r o p e r t y s times,we o b t a i n ( x n y )
= 0 f o r a l l n EN.
From Th. 6
, H
contains
the
s + l--
n i l p o t e n t elements
o f N, t h e r e f o r e x n y e H, f o r a l l nE N. As H i s prime o f t y p e 1, xNyc H i m p l i e s x c H o r y t H , s i n c e x # H, y e H and N \ K i s a minimal prime i d e a l o f t y p e 1. Now we are a b l e t o prove: Theorem 9. I n a medial r i n g ,
&fly)
From t h e d e f i n i t i o n , Q(N) = N\*:I\K,
= P in) = P f ~ ) . 2 1 =.:AN\
K a ) and t h e r e f o r e f r o m t h e Th. 8,
Q(N) i s t h e i n t e r s e c t i o n o f t h e prime i d e a l s o f t y p e 2. But P ( N ) c n ( N \ K , 2 =€A and t h e r e f o r e Q(N) = P 2 ( N ) = P 1 ( N ) . Theorem 10. A medial r i n g N
)
without n i l p o t e n t elements is a commutative r i n g .
We observe t h a t i n a medial r i n g N w i t h o u t non zero n i l p o t e n t elements, f o r 2 a l l x, Y E N , y x y = ( y x I 2 . I n f a c t f o r a l l x, Y E N xyxy = xxyy because N i s med i a l and t h e r e f o r e xyxy
-
xxyy = 0, t h a t i s x ( y x
-
x y ) y = 0. Since N has no non-
209
On medial near-rings z e r o n i l p o t e n t elements, x ( y x
-
xy)y = 0 i m p l i e s y x ( y x
I t ' s easy t o prove t h a t f o r a l l x, Y E N , n i l p o t e n t elements,xy
-
x y ) = 0.
L
( x y - y x ) = 0 and s i n c e N i s w i t h o u t
= y x and N i s a commutative r i n g .
C o r o l l a r y 6. If N is medial, N/&(NI is a subdirect sum of integral domains. Trivial.
REFERENCES H.E. B e l l , Near-rings i n which each element i s power o f i t s e l f , B u l l . Aus. Math. SOC. 2 (19701, 363-368. G. Gratzer, U n i v e r s a l Algebra, Van Nonstrand 1968. H.E. Heatherly, Near-rings w i t h o u t n i l p o t e n t elements, Publ. Math.Debrecen, 20 (19731, 201-205. H.E. Heatherly, Regular n e a r - r i n g s , J . I n d . Math. SOC. 38 (19741,345-354. S. Ligh, The s t r u c t u r e o f a s p e c i a l c l a s s o f n e a r - r i n g s , J . A u s t r . Math. SOC. 13 (19721, 141-146. S. Ligh, A s p e c i a l c l a s s of n e a r - r i n g s , J . A u s t r a l . Math. SOC. 18 (1974), 464-467. S. P e l l e g r i n i Manara, Sui q u a s i - a n e l l i m e d i a l i i n c u i ogni elemento e potenza d i se stesso, R i v . Mat. Univ. Parma, ( 4 ) ll (1985) S. P e l l e g r i n i Manara, Sui q u a s i - a n e l l i m e d i a l i r e g o l a r i , ( t o appear) S. P e l l e g r i n i Manara, Sul r a d i c a l e n i l d i q u a s i - a n e l l i m e d i a l i , ( t o appear) G. P i l z , Near-rings, North-Holland, Amsterdam, N.Y.(1977). D. Ramakotaiah, R a d i c a l s f o r n e a r - r i n g s , Math. Z. 97, 45-46. 0. Ramakotaiah, G.K. Rao, I.F.P. Near-rings, J . A u s t r a l . Math. SOC. ( 2 7 1 , (19691, 365-370. N.V. Subrahamanyam, Boolean semirings, Math. Ann. 148 (19621, 395-401.
SUMMARY
I quasi-gruppi mediali sono stati a lungo studiati; noi introduciamo nei quasi-anelli la 'xyzt = xzyt' per ogni x, y, z , t del quasi-anello e chiamiamo Mediali, in analogia a quunto visto nei quasi-gruppi, i quasi-anetli con questa propriet6. Nei quasi-anelli La Medialit6 si presenta come una condizione di "commutativitd parziale" e il suo studio si inserisce nella serie di studi di quasi-anelbi con propriet6 di commutativitci parziale. In generale dimostriamo che tali quasi-anelli hanno una struttura sottodiretta molto simile a quella dei quasi-anelli fortemente I.F.P. Se, in particolare sono anche regolari, sono isomorfi ad una somma sottodiretta di quasi-anelli sottodirettamente irriducibili che sono campi o quasi-anelli con identitci sinistra in cui ogni elemento non costante x = xx'x B tale che x'x ed xx' sono identitci sinistre. In questo mod0 vediamo che i quasi-anelli mediali generalizzano i 8-quasi-anelli di Ligh ed i semianelli di Subrahamanyam. Otteniamo anche una condizione necessaria e sufficiente affinchd un quasi-inello sia mediale regolare e sottodirettamente irriducibile. Questo risultato wette in evidenza le analogie fra questa struttura ed i quasi-anelli debolmente commutativi. Infine abbiamo studiato l'insieme degti elementi nilpotenti di un quasi-anello mediale. Anche qui la medialitci agisce in mod0 simile alla I.F.P. Inoltre troviamo legami tra 1 'insieme degZi elementi nilpotenti e i vari radicali primi e condizioni affinch2 tale insieme sia un ideate.
This Page Intentionally Left Blank
Near-ringsand Near-fElds, G.Betsch (editor) 0 Elsevier Science Publishers B.V.(North-Holland), 1987
21 1
NEAR-RINGS AND NON-LINEAR DYNAMICAL SYSTEMS Gunter F. PILZ I n s t i t u t f u r Mathematik, Johannes-Kepler-Universitat Linz A-4040 Linz, A u s t r i a ABSTRACT
In t h i s paper we study l a r g e classes o f n o n - l i n e a r systems a d m i t t i n g a t r a n s f e r f u n c t i o n which completely describes t h e i r input-output behaviour. Many concepts which are widely b e l i e v e d t o be p a r t i c u l a r t o l i n e a r systems t u r n over t o t h i s c l a s s o f "separable" systems. These are systems X = (Q,A,B,F,G) i n which t h e map F,G can be separat e d as F(q,a) = a(q) + a(a), G(q,a) = y ( q ) + 6 ( a ) , where Q,A,B a r e groups, a,y maps and a,6 homomorphisms. I f we study separable systems w i t h i d e n t i c a l i n p u t - and output groups then i t turns o u t t h a t t h e p a r a l l e l - ( t h e s e r i e s - ) connections o f systems correspond t o the a d d i t i o n ( t h e composition, r e s p e c t i v e l y ) o f t h e i r t r a n s f e r functions. Hence these systems form a near-ring ( t h e "non-linear analogue o f a t o p a r a l l e l - and series connections. Many systemr i n g " ) w.r.t. t h e o r e t i c questions can be studied i n t h e framework o f near-rings. I n p a r t i c u l a r , we study feedbacks, r e a c h a b i l i t y questions, i n v e r t i b i l i t y w i t h delay L and close w i t h a discussion o f t h e r e a l i z a t i o n problem o f separable systems. D e f i n i t i o n 1: A ( d i s c r e t e , dynamical , t i m e - i n v a r i a n t ) nyntem ( o r aLLtomaton) I i s a quintuple I = (Q,A,B,F,G), where Q i s a s e t ( o f n R a t e ~ ) ,A a s e t ( o f LnpU%), B a s e t ( o f o u * p u t ~ ) ,F a f u n c t i o n Q x A + Q ( t h e n t a t e .thanbLtLon d u n d o n ) and G a f u n c t i o n Q
x
A
-
B ( t h e ou*plLt d u n d o n ) .
The d e s c r i p t i o n o f I i n D e f i n i t i o n 1 i s u s u a l l y c a l l e d the " l o c a l descript i o n " o f X. I n order t o o b t a i n t h e " g l o b a l " d e s c r i p t i o n , we do n o t consider a s i n g l e i n p u t , b u t a s e r i e s o f i n p u t s i g n a l s aiy
which e n t e r the system " a t time i E Z " . Hence w e ' l l consider i n p u t sequences (ai)i z . It i s generally assumed t h a t the i n p u t s d o n ' t come i n "since e t e r n i t y " ; so we assume t h a t t h e r e e x i s t s an index k € Z such t h a t ai = O f o r a l l i < k . Sequences o f t h i s type are u s u a l l y c a l l e d "formal Laurent series": D e f i n i t i o n 2: For any set X containing 0, l e t L(X) be t h e s e t o f a l l sequences (xi)i
, for
which t h e r e i s some k E Z such t h a t xi = O f o r a l l i < k.
The elements o f L(X) a r e c a l l e d (60hmaL) Lament h e h i e 6 o f elements o f X. I f X i s a f i e l d then L(X) i s j u s t the q u o t i e n t f i e l d (see e.g. r i n g X [ [ z l l o f ,jotunaL powetr h e h i e s (xo,x,,
...)
[ I l l ) o f the
( w i t h xi E X ) . I f we d e f i n e
212
G. Pilz
z:= (0,l ,O,O,...)
then u s i n g t h e usual a d d i t i o n and m u l t i p l i c a t i o n o f power
t 1
series, we can w r i t e X [ [ z l l =
1
L(X) =
xiz1
I
and i n the same s p i r i t we w r i t e
xiziIxiEX>,
i20 k c Z , xi EX}.
i>k Hence z i s n o t a "symbol" o r "indeterminate", as f r e q u e n t l y believed. The r e p r e s e n t a t i o n o f s e r i e s (xi)
i s u s u a l l y c a l l e d t h e z-hnbdozrn
by sums B i z '
o r discht-te Laplace t ~ n ~ d o l u nIt . was n o t r e a l i z e d from t h e beginning t h a t t h i s transform i s ( a t l e a s t a l g e b r a i c a l l y ) merely a t r i v i a l i d e n t i t y . t h e r e i s a l s o no need whatsoever t o w r i t e xaiz-i
(Actually,
i n s t e a d o f raizi.)
I n t h i s context, t h e i n t e r p r e t a t i o n i s as f o l l o w s . A t a c e r t a i n "time" k E
Z
(hence k can be negative), t h e system i s i n s t a t e qk when t h e f i r s t i n p u t ak a r r i v e s . The system produces an output bk=G(qk,ak) and changes i t s s t a t e qk i n t o qk+l
a r r i v e s , and so on.
= F(qkyak). Then ak+l
L e t us remark t h a t f o r continuous systems, t h e sequences (ak,ak+l which can be considered as f u n c t i o n s from
Z into A
-
,...) -
have t o be replaced by a
c o n s i s t e n t and causal f a m i l y o f i n p u t f u n c t i o n s from some time s e t T i n t o A.
t, as i n D e f i n i t i o n 1 , i s c a l l e d f i n e m i f Q,A,B f i e l d K and F,G a r e l i n e a r maps on t h e product space G can be decomposed i n t o l i n e a r f u n c t i o n s a:Q
+
a r e v e c t o r spaces over some Q x
Q, B:A
A.
+
Q,
I n t h i s case, F and y:Q +
B and 6:A
-+
B
A. If the v e c t o r spaces i n question a r e f i n i t e dimensional, a,B,yy6 a r e u s u a l l y r e presented by matrices. such t h a t F(q,a) = a ( q ) + B(a), G(q,a) = y ( q ) + 6 ( a ) h o l d f o r a l l (q,a) E Q
x
I t i s n o t t r u e , however, t h a t these decompositions are o n l y p o s s i b l e f o r
l i n e a r systems. We a r e going t o
i n t r o d u c e "separable" systems now. They a r e
much more general than l i n e a r ones, a l l o w h i g h l y n o n - l i n e a r t r a n s i t i o n and output f u n c t i o n s , b u t we can do w i t h these systems most t h i n g s we can do w i t h l i n e a r ones. D e f i n i t i o n 3: I (as i n D e f i n i t i o n 1) i s c a l l e d oepmabbe i f Q,A,B
a r e groups
( w r i t t e n a d d i t i v e l y , b u t n o t n e c e s s a r i l y a b e l i a n ) and i f t h e r e a r e maps a:Q -. Q, y:Q +
B, and homomorphisms 8:A
+
Q, 6:A
+
B such t h a t F(q,a) = a ( q ) + B(a),
G(q,a) = y ( q ) + 6 ( a ) h o l d f o r a l l q E Q , aEA. We then denote Z by (Q,A,B,a,p,y,6) o r simply by (a,B,r,6). C l e a r l y , each l i n e a r system i s separable. I f Q = A = B = ( R , + ) , = q2 + s i n q t 3a, G(q,a) = eq
- 1 + na
determine
F(q,a)
=
a n o n - l i n e a r separable system.
213
Near-rings and non-linear dynamical systems
Separable systems f i t i n t o t h e classes o f non-linear systems described by D.L.
Casti (131, p. 19).
The system M(G) o f a l l maps from an ( a d d i t i v e ) group G i n t o i t s e l f i s a "near-ring" w . r . t . pointwise a d d i t i o n and composition. A neat-hing N i s a gener a l i z e d r i n g i n which we do n o t assume t h a t a d d i t i o n i s commutative and i n which j u s t one d i s t r i b u t i v e law ( f o r t h i s paper: (a+b)c = actbc) i s assumed t o hold. We say t h a t n E N i s zetto-nymm&,ic
i f nO = 0 holds;
N i s zeho-nymmethic i f every
n c N has t h i s property. Mo(G) : = { f € M ( G ) l f ( O ) = O ) i s an example of a zerosymmetric near-ring. See [14] f o r t h e theory o f near-rings. For d i s c r e t e systems ( o r automata) i n which the s t a t e s e t Q i s a group, t h e r e i s a n a t u r a l associated n e a r - r i n g
N(z);
i t i s t h e subnearring o f M(Q)
generated by a l l f u n c t i o n s fa ( a E A ) w i t h f a ( q ) =F(q,a) and by t h e i d e n t i t y f u n c t i o n i d (because we d o n ' t want the system t o change i f no inputs a r r i v e ) . For separable systems, N(Z),
i s generated by i d and a i f Q i s abelian. This
construction o f a " s y n t a c t i c near-ring" i s the obvious e x t e n t i o n o f the monoid o f an automaton (see e.g.
[lll)i n t h e presence o f a group s t r u c t u r e on Q. N(,)
i s t y p i c a l l y n o t a r i n g , n o t even i n t h e case o f l i n e a r systems o r automata (see [61). I n t h i s paper, however, we show t h a t separable systems i t s e l f form a nearr i n g by means o f s e r i e s / p a r a l l e l connections o f these systems. For l i n e a r systems we do get a r i n g , see 1161. I n f a c t , our treatment f o l l o w s p a r t l y t h e ideas i n Sain's book [16].
I n order t o perform t h i s c o n s t r u c t i o n , we have t o
represent a system 1 by i t s g l o b a l input-output map .,f
This i s always possible
f o r separable systems (see below). Even if A i s "only" a group, we may w r i t e (ak,a ktl,...)
( w i t h k € Z ) as
1 aizi. I t i s one o f t h e fundamental ideas o f a l g e b r a i c systems theory t h a t i zk t h e s h i f t - t o - t h e - l e f t operator ( " u n i t d e l a y " ) (ak,aktl ) (ak,akt, .,) ( w i t h ak now a t t h e ( k - 1 ) s t place, etc.) j u s t corresponds t o the m u l t i p l i c a t i o n o f zaizi by z-'. Hence we denote the s h i f t operator by 2. Note t h a t L(A) (see D e f i n i t i o n 2 ) i s an a d d i t i v e group, too, under component-wise a d d i t i o n ; i n B f a c t , L(A) i s a subgroup o f A
,...
-.
,.
.
The map a:Q
-+
Q can be extended t o a map ( a l s o denoted by a) from L ( Q ) i n t o Also, we can extend 0 t o a map from L(A)
L ( Q ) by a(.l qizi) := 1 a(qi)zi. ilk iLk i n t o L ( Q ) , and so on.
214
G. Pilz P r o p o s i t i o n 1: For a separable system I: = (Q,A,B,~,B,Y,~),
- a + f : L ( Q ) + L ( Q ) i s always b i j e c t i v e . Also, - a +
f
t h e map
i s zero-symmetric i f f a i s
zero-symmetric.
1
1
qizi) = 1 -a(qi)zi + qiz i-1 = q k z k - l + ilk irk i>k qizi) = (-a+ Z ) ( , q/zi) i m p l i e s t h a t + 1 (-a(qi)+qi+,)zi. Now ( - a + f)( it k ibk ibk q.z1 and q;zi b o t h have t h e f i r s t non-zero c o e f f i c i e n t a t t h e same index n
Phaob: We have (-a+;)(
1
1
(say) , and qn = q,; -a(qn) + qn+l = -a(qA) + qA+l , -a(qn+l 1 + qn+2 = a(qA+l) + t q;\+2y... i m p l i e s qi = q i f o r a l l i. Hence - a + : i s i n j e c t i v e . I f qiziEL(Q) itk then i t i s easy t o see t h a t ( - a + i ) ( 1 q;zi) = 1 qizi if t h e q; a r e given i n it k ibk d u c t i v e l y as q i : = 0 and q;+l :=a(q;) + qi f o r i > k. This shows t h a t -a+; i s b i jective.
I f a ( 0 ) = 0 , i t i s c l e a r from the formula f o r -a+? t h a t -at;
zero element
x ozi
o f L ( Q ) i n t o i t s e l f . Conversely, i f (-a+Z)(
then we g e t a ( 0 ) = 0.
x 0zi)
maps the =
I 0zi
0
Now we a r e i n a p o s i t i o n t o g i v e a formula which d i r e c t l y r e l a t e s t h e output sequence o f a system t o i t s i n p u t sequence. Observe t h a t a l l c a l c u l a t i o n s have nothing t o do w i t h l i n e a r i t y . A l l we need i s s e p a r a b i l i t y . Theorem 1: I n a separable system
x
= (Q,A,B,a,a,y,6),
hold, i f t h e f i r s t non-zero i n p u t a r r i v e s a t time k E
t h e f o l l o w i n g formulas
Z i n which t h e system i s
i n s t a t e 0: (i)
1
qizi
1
bizi
= (-a+i)-l
(ii)
1
aizi)
= [y(-a+f)-lp+6](
1
aili)
it k
ihk
mod:
B(
il k
ilk
( i ) Since qi+l
= a(qi) + B(ai)
we g e t
1
qi+lzi
i>k = z
+ B(.I aizi). But i 1 qi+lzi+l ilk ibk = z(a( 1 qizi) + whence -a( 1 qiz i>k i>k = B( 1 aizi). I f q k = O then we can w r i t e (-a+z)( i>k P r o p o s i t i o n 1 , -a+; i s b i j e c t i v e , hence t h e r e s u l t .
(ii) f o l l o w s from (i), s i n c e bi=y(qi)+6(ai), + 6(xaizi).
qi+,zi,
: = y ( - a + Z ) - l ~ + 6 from L(A) i n t o L(B) i s c a l l e d t h e Observe t h a t f,
1
q i z i1 +
i>k hence
b i z i = y(.x qizi)+
hence
D e f i n i t i o n 4: I n a separable system I: = (a,B,y,d),
f,
= a(
it k
i >k
the f u n c t i o n ~ J L ~ F I A ~ C&ztLon VL
o f X.
i s zero-symmetric i f t h i s a p p l i e s t o a and y (by Proposi-
215
Near-rings and non-linear dynamical systems
t i o n 1 and t h e f a c t t h a t each group homomorphism (hence B and 6) i s automatic a l l y zero-symmetric). fx completely characterizes t h e input-output behaviour o f 1 i f X s t a r t s i n
s t a t e 0. I f 1s t a r t s i n a s t a t e q # 0 which i s reachable from 0 by means o f an we simply s t a r t a t time k-r, and then i n s t a t e 0. Since
i n p u t sequence al...ar,
i t does n o t make much sense t o s t a r t from non-reachable s t a t e s , f x " c h a r a c t e r i -
zes" X i t s e l f , We c a l l
x
zeho-nqmrnc&tic
Now we connect two systems tion
x1 tt z2
xi
=
i f t h i s a p p l i e s t o f,.
(9, ,Ai ,Bi ,Fi ,Gi)
r e q u i r e s B1 = A2. Then X1
( i = 1,2). The nenied cannec-
X2 : = (Q1 x Q2,A1 ,B2,F,G)
with
F( (4, ,q2) ,a) = (F1(ql ,a) ,F2(q2,G1 (4, ,a) 1 and G( (ql ,q2) ,a) = G2(q2 ,GI (ql ,a) 1. The pm&d connection E l / t2 works w i t h A1 = A2 =: A, B1 = B2 =: B and gives Xl//x2:=(Q1
x
Q2,A.B,F',G')
G ' ( (ql ,q2) ,a) = GI (ql ,a)
w i t h F'((ql,q2),a) +
I f Xi = (Qi ,Ai ,Bi ,ai ,Bi ,yi ,ai) F((ql
,a) = (al(ql
= (Fl(ql,a),F2(q2,a))
and
G2(q2 ,a). a r e separable then i n El
1 + ol(a),a2(q2)
+
tt
E~ we g e t
b2(y1(q1 1 + 6 1 ( a ) ) ) =
is
= (al(ql),a2(q2)+
B2ul(q1))+ (Bl(a),B26,(a)). Note t h a t a + (Bl(a),B26,(a)) a homomorphism. S i m i l a r l y , G((ql,q2),a) =y2(q2) +a2(y1(q1) +61(a)) = = (y2(q2)+62yl(q1))+626,(a). Hence El ++I2is again separable. I n I l / / X 2
we
g e t F ' ( (ql .q2) ,a 1 = (al (4,) + ( a ) ,a2(q2) + b2(a) 1 = (al (4,) ,a2(q2) 1 + ( a ) .b2(a)), and G'((qlyq2),a)=y,(ql)+61(a)+y (q ) + 6 2 ( a ) . I f B i s abelian, t h e l a t t e r 2 2 can be w r i t t e n as ( y l ( q 1 ) + y 2 ( q 2 ) ) + ( 6 1 + 6 2 ) ( a ) , and 6 , + 6 2 i s again a homomorphism. I n summary, we have shown P r o p o s i t i o n 2 : I f El and X2 a r e separable systems, t h e same a p p l i e s t.o tl *t2 and ( i f t h e output groups a r e abelian) t o X 1 / / X 2 - I f X1 and x2 a r e zerosymmetric then t h i s a l s o a p p l i e s t o
x1 ttz2 and rx1//x2.
Theorem 2 : If1 1 + are separable then (i)
fq//x2 = ftl
( i i ) fXl*x2
=
+
fX2 ( i f t h e output groups a r e a b e l i a n )
fx2 a fxl
Phood: These formulas a r e already motivated by t h e d e f i n i t i o n s o f and
tl*x2.
xl//rx2
The equations bi+l = y ( q i ) + 6 ( a i ) etc. o n l y r e l a t e t h e i n p u t a t time
i t o t h e output a t t h e same time i. But i n p u t s a t time i a l s o have i n f l u e n c e on l a t e r outputs ( w e ' l l i n v e s t i g a t e t h i s i n t h i s paper as w e l l ) , and t h i s i n f l u e n c e
i s described by .,f
So we g i v e d i r e c t proofs. Note t h a t t h e p r o o f o f Proposi-
216
G. Pilz
217
Near-rings and non-linear dynamical systems
I o f separable systems a t
Now l e t us assume t h a t we have a f a m i l y
hand which a l l have t h e same a b e l i a n group A as i n p u t and o u t p u t group. Under t h e s e assumptions we g e t t h e f o l l o w i n g C o r o l l a r y t o Theorem 2. C o r o l l a r y : The s e t S o f t r a n s f e r f u n c t i o n s o f those systems which can be constructed from t h e
zi
( i E I ) by means o f s e r i e s and p a r a l l e l connections
c o i n c i d e s w i t h t h e seminear-ringgenerated by {f, I i E 11 i n i i M(L(A)). I f A i s o f f i n i t e exponent t h e n S i s t h e n e a r - r i n g generated by t h e 's.
f 'i
T h i s h o l d s because i f e i s t h e exponent o f A t h e n eA=O, whence ef,
i s the
zero map, and so ( e - l ) f t = -fz can be o b t a i n e d by (e-1) p a r a l l e l connections o f
Note t h a t t h i s a p p l i e s i n p a r t i c u l a r t o f i n i t e groups A. O f course, S = t h e n e a r - r i n g generated by I f ] iE I} i f I c o n t a i n s w i t h each ti = (a,B,y,d) 'i a l s o t h e system -ti = (ayBy-y,-6) which y i e l d s as can be e a s i l y seen the
-
-
t r a n s f e r f u n c t i o n f - z =-fE. Also, t h i s can b e achieved b y " e x t e r n a l " devices which t r a n s f e r elements o f A i n t o t h e i r negatives. Hence i t makes sense t o impose t h e f o l l o w i n g Conventicxn f o r t h e r e s t o f t h i s paper:
I i s a family o f separable
systems which have t h e same a b e l i a n i n p u t - and o u t p u t group A and which cont a i n s w i t h each
xi a l s o -xi.
l i E I> i s a s u b n e a r - r i n g o f M(L(A)). If each fI i s zero-symmetric, 'i i SO i s < f I i E I> (as a s u b n e a r - r i n g o f Mo(L(A))). I f each f i s an endomor'i 'i phism on L ( A ) then < f / i E I> i s even a r i n g . Hence < f
'i
I f we i d e n t i f y a separable system 1 with i t s t r a n s f e r f u n c t i o n fy ( a s desc r i b e d e a r l i e r i n t h i s paper), we can r e s t a t e o u r r e s u l t s as f o l l o w s .
218
G. Pilz
Theorem 3: L e t S ( A ) be t h e s e t o f a l l separable systems w i t h a b e l i a n i n p u t -
,* )
and output groups A. Then ( S ( A ) , J /
i s a near-ring.
By standard means o f u n i v e r s a l algebra (see e.g.
[41),
I i E I>
=
I
Mf,
I . . . ’fXi ) t E T n ) , where Tn i s t h e s e t o f a l l terms (polynomials) i n il n t h e v a r i e t y o f a l l near-rings.
=
I i E I > , we t r y t o f i n d a more e x p l i c i t e ‘i formula f o r t h e a c t i o n o f fx. This can be achieved by the f o l l o w i n g I n order t o l e a r n more about
D e f i n i t i o n 5: I f X = (a,B,y,6), t h e n - t h Mahhov nymboL ( w i t h n 2 - 1 ) [a,B;nl i s the map from L ( A ) i n t o Q defined i n d u c t i v e l y by
[~,B;-II [a,~;n+11
1
aiz’)
:= o,[~,B;oI(
1
aizi)
:= a ( [ a , ~ ; n ~ (
i>k i>k
aiz’)
1
aizl))+B(ak+n+l)
i>k
=: a [ a , ~ ; n l
1
Note t h a t [a,p;nl
1
i2k
aizi
.I
i>k
aiz
i
:= a ( 0 ) + B(ak)
+
B(ak+n+l)
o n l y depends on ak,
i>k w r i t e [a,Q;nl
(ak,
...,ak+ n ).
...,a k+n .’ h
Obviously we g e t
Proposition 3: I f a i s zero-symmetric then each [a,B;nl and [a,B;nl
ce we can a1 0
aiz i = a ( a ( . . . a ( O ( a k ) + B ( a k + l ) . . . ) ) +
i s zero-symmetric
B(ak+n-l))+B(ak+n).
ibk
Since [a,B;O]
(a,)
d e f i n i t i o n o f [a,b;n]
= a ( 0 ) + p(ak) i s the s t a t e qk+l
Proposition 4: I f a separable system
r =(a,p,y,6)
s t a t e q k = 0 and i f the i n p u t sequence (aklak+l,...) states qk+l ,qk+*,.
( i f qk = 0 ) , we g e t by the
the following
.. such t h a t ,
f o r n E IN
qk+n = [ a , ~ ; n + l l (ak,
...,a k+n-1 )
From t h i s , we g e t the desired formula f o r f,: Proposition 5 : I f f x (
1 ibk
aizi)
s t a r t s a t time k i n the a r r i v e s then
=
1 ibk
bizi
then
x
assumes new
219
Near-rings and noelinear dynamical systems
Note t h a t Theorem 2 could a l s o be proved by i n d u c t i o n using t h i s formula. Now we i n v e s t i g a t e , which f u n c t i o n s f E M(L(A)) can a r i s e as f = ft f o r some separable system
x = (a,p,y,6).
The determination o f a,p,y,6
from f i s then
known as " r e a l i z a t i o n problem". I t i s u s e f u l t o e s t a b l i s h t h e f o l l o w i n g Notation: Lk(A) : = {
1
aizi /ai E A).
i>k
u Lk(A). Furthermore, l e t , f o r B,CEX, (B:C) be t h e s e t k€Z o f a l l f u n c t i o n s f from X t o X such t h a t f ( C ) c B . This coincides w i t h t h e O f course, L ( A ) =
"noetherian q u o t i e n t s " as d e f i n e d i n 1.41 o f [141. Since f z ( L k ( A ) ) E L k ( A ) f o r each separable system t, we g e t Proposition 6: L e t ti ( i E I ) be separable systems.
(ii)N(L(A)) i s a near-ring. l i E I> i s a subnear-ring o f N(L(A)).
( i i i )
'i (iv)
I f each ti i s zero-symmetric theni s a subnear-ring o f t h e 'i zero-symmetric p a r t No(L(A)) o f N(L(A)).
I f t i s zero-symmetric then i t makes no d i f f e r e n c e i f we consider
1
1
aizi) o r f ( aizi) w i t h ai = O f o r i < k . Hence i n t h i s case ft can be izk iEZ regarded as a member of If E M(Az I f ( i ) = 0 f o r a l l ic k, where k i s some i n t e ger (depending on f ) ) .
f(
Notation: I f (A,+) t h e s e t of a l l f',
i s an a b e l i a n group and (Q,+) any group, l e t N(A,Q) be
where I = (a,8,yY6)
i s any separable system on (A,A,Q).
Let
n(A,Q) be t h e subnear-ring cN(A,Q)> generated by N(A,Q) i n t h e n e a r - r i n g N(A) o f a l l separable systems on (A,A,Q) a l s o by f
f o r some Q. I f
x
i s as above, we denote fx
a, 8 ,Y $6'
Recall t h a t we g e t N(A,Q)cfi(A,Q) S N ( A ) SN(L(A)) SM(L(A)). t h i s chain (except t h e f i r s t one) a r e always near-rings.
A l l members o f
220
G. Pilz 0 1 ,... ,...
Now i f f E M ( L ( A ) ) should c o i n c i d e w i t h t h i s ,f we can w r i t e f = ( f ,f ) such t h a t , f o r a l l 1 aiziEL(A), f(Taizi)=(fo(Eaizi),f 1( x a i z i ) ). T h i s imposes q u i t e a number o f r e s t r i c t i o n s on t h e f u n c t i o n s fn (n E
No )
and hence
on f i t s e l f . By comparison o f components, we see t h a t
( 0 ) f o ( . l aizi) = y ( 0 ) + 6 ( a k ) . Hence fo i s an a f f i n e f u n c t i o n depending o n l y on i>k ak, and we can s o l v e f o r 6, s i n c e 6 = f g = t h e zero-symmetric p a r t o f fo. Also, y(0) =f: i s a l r e a d y f i x e d a t t h i s stage. (1)
fl(.l aizi)=u(a(0)+
~ ~ a k ~ ) + 6 ( a k + l ) = u ( a ( O ~ + ~ ( a k ) ) ~ Hence f~~ak+l~.
iSk
1 t h e r e must be a f u n c t i o n g EM(A) such t h a t f ' (
1
aizi)
1 = g ( a k ) + f:(aktl).
i>k We can always s o l v e if a ( 0 ) B I m 8 , and a l s o i n case a ( O ) E I m p i f we have g 1 (a,) =f:(ak) =y(O) f o r each ak such t h a t p ( - a k ) = - p ( a k ) =a(O).
.. .. .. .. .. . . . . . ( n ) The f u r t h e r equations determine a&,y moreandmore, and show t h a t each fn i s a f u n c t i o n j u s t depending on t h e f i r s t n + l c o e f f i c i e n t s ak,...,ak+n
1
aizi.
I n f a c t , t h e r e must be f u n c t i o n s gn : A
-,A
in
such t h a t
irk
"In p r a c t i c e " i t t u r n s o u t t h a t a,p,y
(and 6, anyhow) a r e a l r e a d y determined
a f t e r c o m p a r a t i v e l y few steps. T h i s i s t h e r e a l i z a t i o n procedure o f c a l c u l a t i n g out o f a given f = f
We i l l u s t r a t e t h i s i n t h e case Q = Z , . Z2, p:A + ZZ,, y : Z + A, 6:A -,A, and t h a t y(0) and 6 a r e f i x e d i n s t e p ( 0 ) as 6 = f i , y(0) =f:. I n s t e p ( 1 ) we g e t 1 y c o m p l e t e l y u n l e s s (case ( i i ) ) y(a(O)+p(a,))=g ( a k ) . T h i s f i x e s (case (i)) a(0) = 0, p = 0 ( = zero map) ( i n which case we must have g 1 (a,) = f: f o r a l l a k ) . a,p,y,6
a,B,~,6'
Recall t h a t i n t h i s case a:Z2
4
Note t h a t i n case ( i f a and p may remain "hidden" i f e.g. y i s c o n s t a n t ( t h i s i s o n l y p o s s i b l e i f each gn i s c o n s t a n t ! ) . T h i s case w i l l appear i n t h e Example below. I n case ( i i ) , a ( 0 ) . p and 6 a r e f i x e d , and we g e t t h a t y(O)= - 0 - f c = g n , so gn must be c o n s t a n t ; t h e n y ( 1 ) can (and need) n o t be determined. We see: Remark: The e q u a t i o n f = f i s s o l v a b l e i f f f E N(A). I n t h i s case, y(0) a d3 ,Y ,6 and 6 a r e always u n i q u e l y determined, w h i l e t h i s m i g h t n o t be t h e case f o r a,B, and t h e o t h e r values o f y ( q ) ( Q i s thought t o be known and f i x e d ) . l i n e s a f t e r Theorem 7.
C f . also the
221
Nearrings and non-linear dynamical systems
How b i g i s N(A,Q ? Although L ( A ) , hence M(L(A)), and a l s o N ( L ( A ) ) i n Propos i t i o n 6 a r e always i n f i n i t e ( u n l e s s A = EO}), we g e t P r o p o s i t i o n 7: N A,Q) i s f i n i t e i f A and Q a r e f i n i t e . T h i s f o l l o w s from t h e f a c t t h a t t h e r e a r e o n l y f i n i t e l y many f u n c t i o n s a,B, y,6 which determine separable systems on (A,A,Q).
Z,)
We now determine N(Z,,
i n d e t a i l ; t h i s w i l l a l s o show how one can
proceed i n more general cases. Example 1: I n o r d e r t o d e s c r i b e N : = N(Z2,ZZ2 ) , we need t h e f o l l o w i n g f o u r
).
f u n c t i o n s o f M(L(Z,)
- -1
z := z
j :
( " u n i t s h i f t t o the r i g h t " )
1
aizi
1 1
aizi
+
s
:
+
lzi
1 1
(ak+aktl
i21
i2k
i :
1
i>k
ilk
aizi+
t
... +ak+i-l)zk+i
("summation o p e r a t o r " )
i z k+ i
i20
i>k
Now B and 6 can each be t h e zero map 0 o r t h e i d e n t i t y map i d , w h i l e a and y can be 0,l
( = c o n s t a n t map w i t h values l ) , i d , o r 1 + i d . I t t u r n s o u t t h a t y
plays a c r u c i a l role. I f y = O t h e n each f
Ify
=l
a ,B ,Y .6
= 6
then f a , B , 1 , 6 = j + 6
( i n particular, fo,O,O,o=O, ( i n particular,
1 = f0 $0,I
fo,O,O,id=idEN) € N)
9 0
I f y = i d , t h e n i t i s easy t o see t h a t f O ,B , i d ,6 = z o p + a
fl,B,id,6=Zo f i d ,B , i d ,6
( i n p a r t i c u l a r , z =fo,id,id,OEN)
(ItBIt6
= S o
fit6
( i n p a r t i c u l a r , S = fid,id,id,O€N)
fl+id,B,id,6 = i + S 0 B + 6
IfY
= l t
fa,~,l+id,6'1t
( i n p a r t i c u l a r , i = fl+id,O,id,O€ N ) fa,~,id,6'
This, and t h e f a c t t h a t B,6€ I O , i d l shows t h a t N i s generated by O,id,l,z,S,
i. Since i d t i d = O , S O l = i , we can even say t h a t
222
G. Pilz
Note t h a t N(Z2, Z 2 ) i s n e i t h e r c l o s e d w . r . t .
a d d i t i o n ( s i n c e e.g.
composition ( z o z $ N ( Z 2 , Z 2 ) ).
z + S ( N ( Z 2 , Z 2 ) ) , nor w . r . t .
We now g i v e a l i s t o f t h e f u n c t i o n s i n N(Z2$Z 2 ) , along w i t h t h e i r values i n t h e f i r s t t h r e e components (which shows t h a t a l l o f them a r e d i f f e r e n t ) .
ak t 1
ak
t
1
0
i d z o l t i d ak * 'k+l*
ak ak+l
akt2-
aktl
ak akt 1 ak+l
1
i t i d ak aktl+ akt2
1
z+id
S t id
ak a k + ak t l
ak aktaktl
a k + l t ak+2 a k t a k t l
zo (jtid)tid
'k a k t aktl +
akt2
'kt2'
i+ S + i d ak
+ 1 ak + ak+l
aktl
'
+
1
aktak+ltak+2
ak * aktl* aktE'
Hence we g e t t h e f o l l o w i n g C o r o l l a r y : N(Z2,
Z,) has 28 elements and m(Z2, Z2 ) can be generated by
f o u r elements. Hence every separable system c o n s t r u c t i b l e by s e r i e s / p a r a l l e l connections o f separable systems on (Z2, Z2, Z2 ) can be o b t a i n e d by means o f s e r i e s / p a r a l l e l connections from
xz = (O,id,id,D), zs = (id,id,id,D),
xid
= (O,O,O,id)
, Z,
= (0,0,1
,O),
which a r e systems on (Z2, B 2 , Z,).
In o r d e r t o d e s c r i b e a more general s i t u a t i o n , we need t h e f o l l o w i n g
Nearrings and non-linear dynamical systems
223
D e f i n i t i o n 6: For n E N o and A = B = Q = K = a f i e l d , l e t
...
1
aizi -,~ o , ~ ~ , B ; o I ~ ~ s ~ , ~ ~ , ~ ; I) , I ~ ~ ~ ~ , ~ i>k (where t h e exponent n denotes t h e one w . r . t . t h e m u l t i p l i c a t i o n i n F) be t h e
M ~ s: =
~ , ~ ; ~ I ~ ~
n - t h h4mk.o~map on K. Theorem 4: L e t K be a f i n i t e (Galois) f i e l d and A = B = N(K,K) always contains id,z,l,S,i
(i)
Q = K. Then
o f the preceding Example.
(ii) IfN(K,K) i s contained i n t h e n e a r - r i n g generated by f u n c t i o n s fi (iE.some index set I)then i ( K , K ) = < f i I i ~ I > . ( i i i ) N(K,K)
i s always f i n i t e , a(K,K)
always i n f i n i t e .
where p = c h a r K and hk:x
f o r a l l x E K),
(iv)
N(K,K) =,
(v)
k being a p r i m i t i v e element (i.e. a generator o f t h e c y c l i c group (K*, The n e a r - r i n g h(K,K) f u l f i l l s t h e ascending chain c o n d i t i o n (ACC) on
+
kx
.)I.
subnear-rings.
Phood: ( i ) i s c l e a r since id=fo,O,O,idy z = fO,id,id,O' 1= f O , O , l ,id ,id,O' i = S o l . ( i i ) I f N ( K , K ) c < f i l i c I}S[(K,K) then=R(K,K) = < f i l i c I > .
,O'
S=fid
(iii) f o l l o w s from Proposition 3. ( i v ) a(K,K) i s generated by N(K,K), w i t h a,yEM(K),
B,dEHom(K,K)
i.e.
the s e t o f a l l functions y(-a+z)-lp+d,
(which i s K-isomorphic t o K). I f we take y = O , we
g e t a l l f u n c t i o n s i n Hom(K,K), and hence a l l y(-a+,?)-lO a r e a l s o i n R(K,K).
Now
a f i n i t e f i e l d i s known t o be "polynomially complete" (see [ t o ] o r [ I l ] ) , h e n c e
,.,.ykp-l
t h e r e a r e ko,kl
E K such t h a t y ( x ) = ko + klx + k2x2+
a l l xEK. Hence we can generate a l l y(-a+;)-lO
...
+ k xP-' f o r P-1 1 a d d i t i v e l y from kxn(-a+z)- 8 ,
which i n t u r n can be obtained as t h e composition o f x n ( - a + i ) - ' a w i t h t h e homomorphism hk:x [a,B;lI'(s)
1 aizi i n t o (O,[a,~;Oln((s), izk .), which means t h a t xn(-a+Z)- 8 =Mn. F i n a l l y , a l l elements i n
+
,..
kx. Now xn(-a+?)-'g sends s =
Hom(K,K) can be obtained from a s i n g l e map hk:x
-,kx,
i n which k i s a p r i m i t i v e
element i n K. ( v ) f o l l o w s from t h e f a c t t h a t a(K,K) i s f i n i t e l y generated. Remarks: ( i ) A s i m i l a r , b u t more complicated r e s u l t as f o r f i n i t e f i e l d s (Theorem 3 ) holds f o r f i n i t e vector spaces. I t does n o t seem t o be too w e l l known t h a t each map (we need i t f o r a,y)
-.
f r o m a f i n i t e vector space Q (over a
f i e l d F ) i n t o i t s e l f i s o f t h e form
,...
,...
,. ..
,...
q = (4, ,qn) (p, (ql ,qn) ,pn(ql ,s,f). where t h e pi a r e polynomial functions over K and n=dimFQ. This i s b a s i c a l l y due t o t h e polynomial completeness o f f i n i t e f i e l d s (see [ l o ] ) .
224
G. Pilz
( i i ) I f A i s a f i e l d , N(A) can be considered as n e a r - r i n g o f mappings on a f i e l d , namely on t h e q u o t i e n t f i e l d o f t h e i n t e g r a l domain A [ [ x ] ] o f f o r m a l power s e r i e s . Theorem 3 ( i v ) i n d i c a t e s t h e appearence o f a d d i t i o n , m u t t i p l i c a t i o n , as w e l l as composition. The s u i t a b l e frame f o r t h e s e c o n s i d e r a t i o n s is t h e one o f "composition r i n g s " , whose s t u d y was i n i t i a t e d i n [12] and [ l ] . ( i i i ) I f N(A) i s a r i n g and generated by fx ( i E I ) then every element on i N(A) ( = e v e r y system on (A,A,Q) f o r some Q ) can be o b t a i n e d by a p a r a l l e l connection o f (one o r s e v e r a l ) s e r i e s connections o f t h e xi.
This i s not the
case any more i f N(A) i s a " p r o p e r " n e a r - r i n g ( t h e n one has t o t a k e p a r a l l e l connections of s e r i e s o f p a r a l l e l o f
... c o n n e c t i o n s ) .
( i v ) I f N = N(A), o r N = B(A,Q), can be generated by f u n c t i o n s which a r e e i t h e r c o n s t a n t o r homomorphisms ( t h i s was t h e case f o r
[(Z,,72,) i n o u r
Example!) t h e n N i s an a b s t r a c t a f f i n e n e a r - r i n g (see [141, ch. 9 c ) ) . I n t h i s case, No i s a r i n g and Nc i s a module o v e r No. A f f i n e n e a r - r i n g s a r e t h e "most simple" ones among t h e p r o p e r n e a r - r i n g s . ( v ) Several o t h e r n e a r - r i n g t h e o r e t i c concepts have meaningful i n t e r p r e t a t i o n s i n t h e c o n t e x t o f systems t h e o r y . L(A) i s , i n a n a t u r a l way, an N = N ( A ) o r N = N(A,Q) - group ( = ( n e a r - ) module). A subgroup L of L(A) i s an N-subgroup
i f fz( xaizl)
E
L whenever f x E N and xaizl
monogenic N(A)- o r N(A,Q)-group i s o f t h e form xbizl
=fx(zaizl)
E L . Note t h a t L(A) i s never a
( t h e r e i s no xaizi f o r some f,cN(A)
a r e b a s i c a l l y subsets of N which a r e closed w . r . t .
w.r.t.
such t h a t each
x bizi
cL(A)
o r EN(A,Q)). I d e a l s I on N p a r a l l e l connections, and
p a r a l l e l connections w i t h a r b i t r a r y systems i n N. Elements i n N(A,Q) can
be n i l p o t e n t : take'e.9.
A = R = Q = Z 4 , 6(a) =2a, f = f 0 ,o ,o ,6' Then f which means t h a t t h e s e r i e s composition o f t h e corresponding system i t s e l f sends e v e r y i n p u t sequence i n t o t h e zero sequence (O,O,..
o
x
f
=0, with
.). The
appearance o f n i l p o t e n t ( o r n i l ) N-subgroups i n d i c a t e s t h a t t h e r a d i c a l o f N(A) o f N(A,Q) i s non-zero "141, e.g. 5.45). We now t u r n t o separable systems w i t h a ( s e p a r a b l e ) feedback. L e t T E S ( A ) have a hepahabee 6eedback 6unotion
@:QxA-,
A, ( q , a ) + $ ( q ) + @ ( a ) ,
q a homomorphism. Ifwe encode t h e s t a t e q E Q and t h e o u t p u t a E A t o @(q,a) and add i t t o t h e i n p u t , we g e t a new system . ' 1 see.
The f o l l o w i n g r e s u l t i s easy t o
225
Near-rings and non-linear dynamical systems
P r o p o s i t i o n 8: I f EES(A) has maps a.B,y,d
and i f 0 i s a separable feedback,
(a(q,a) = + ( q ) + q ( a ) , then t0 i s again i n S(A) and has maps a+!3Q,, p$, y + d $ , €I$.
I f E and 0 a r e zero-symmetric ( i . e .
+(O) = 0 ) then t h i s a l s o a p p l i e s t o . ' 1
Hence t h e c l a s s o f separable systems i s closed w . r . t
parallel/series
connections and separable feedbacks. We now t u r n t o questions o f r e a c h a b i l i t y . L e t A* be t h e f r e e monoid over A;
so A* c o n s i s t s o f a l l f i n i t e s t r i n g s o f elements o f A ( i n c l u d i n g t h e empty sequence A).
I f a = a l a 2...anEA*,
we w r i t e n =
151. I f z i s any system (separable
o r n o t ) such t h a t t h e s t a t e s e t i s a group ( Q , + ) , we l e t Ni(z) of
M(0)
N(E ,)
:=
generated by {f;laEA*,
.u
~ i } Then . N0(x)
be t h e subgroup
... and
are groups, too. I f q E Q then Ni(Z)q =: Reachi(q)
Ni(E)
consists
'wl
of a l l s t a t e s which can be reached from q by " ( m u l t i p l e ) i n p u t s t r i n g s o f
length Si".
C l e a r l y , each Reachi(q)
Reacho(q)cReachI(q)c
... holds.
t h e bubgtloup tleachable &om
i s a subgroup o f
u
Q and
i s called iEJN q. If ( Q , + ) f u l f i l l s t%e ascending chain c o n d i t i o n Reach,(q)
:=
Reachi(q) = N,(E)q
(ACC) f o r subgroups (which i s e q u i v a l e n t t o say t h a t (Q,+) i s f i n i t e l y generat e d ) then t h e r e i s some n E INo
such t h a t Reachn(q) =Reach,(q).
One might be i n t e r e s t e d how f a r one can g e t i n p r e c i s e l y i steps. L e t := CfaqI
Rchi(q)
(a1 = i } .
Then i t i s n o t t r u e f o r separable systems t h a t
(as i t i s f o r l i n e a r ones):
Rchi(q)
Example 2: L e t Q = A = ( Z , 2 y + ) , F ( q , a ) = ( q + l ) 2 + 4 a . Then Rchl(0) = tF(O,a) l a € Z12 1 = C1 + 4 a ( a Z,2 ~ 1 = 11,5,91 and Rchl(l) = Rchl(5) =Rch,(9) = C0,4,81. Rchl(0)nRch2(0)
Hence Rchl(0)$Rch2(0),
and we have even
=0 !
The t r o u b l e came from t h e f a c t t h a t t h e system i n Example 2 was n o t zerosymmetric: Proposition 9: L e t E be separable and zero-symmetric, and l e t Qi =Rchi(0). Then, f o r each i E INo
,
( i1
Qi+l
(ii)
Q o ~ Q...E , z Rch,(O)
= a(Qi)
+
I m B ?a($) :=
U
QisReachm(0).
iEINo ( i i i ) Qi+l (iv)
=a(Qi)+Q1
I f B i s an epimorphism then Q, = Rchoo(0) = Reachoo(0) =Q.
226
G. Pilz
Roolj:
Qi+,
(i)
= {F(O,a)l
la1 = i + 1)
=
= {F(F(O,Z),a)la"EA*,
= a(Qi
(ii)
1+ I m B
= i , aEA} =
la"l
= a({F(O,Z)lZEA*,
2a(Qi
= i l l + {B(a)laEAl =
1
We show Qi5Qi+,by i n d u c t i o n on i. Q o = {O), Q, = {a(O) + B(a) l a E A} = I m b 5 (Q,+) , s i n c e B i s a homomorphism and E i s zero-symmetric.
So
QosQ1. Now
l e t Qi-,5Qi. From ( i ) we g e t
Q i + l = a ( Q i ) + I m p ~ a ( Q i - , ) + 1 m B =Qi. ( i i i ) f o l l o w s f r o m ( i ) and ( i i ) . ( i v ) i s a consequence o f ( i i i ) and o f Q, = I m 0. Separable feedbacks do n o t change t h e r e a c h a b i l i t y behaviour: Theorem 4: L e t
x
be a system, and @ a feedback on E. Suppose t h a t E and @
a r e b o t h separable and zero-symmetric w i t h $J b e i n g an epimorphism. Then Rchi(0), and hence Rch,(O)
and Reach-(O)
coincide i n
x
and i n I@,
.
@ @ Ptoolj: L e t Rchi(0) =: Q.1 i n X and =: Q. in X We proceed a g a i n by i n d u c t i o n . 1 For i = O we g e t Q o = I01 =a!. L e t Q. =a:. From P r o p o s i t i o n 8 and P r o p o s i t i o n 9 1 @ ( i ) we g e t Qi+, = ( a + B@)Qf+ I m BJ,= ( a + B@)Qi + B W I ) = a(Qi) + B@(Qi) + B(A) =
= a(Qi)+ I m b=Qi+,,
the r e s u l t .
s i n c e B @ ( Q i ) c B ( A ) , which i s a subgroup o f (Q,+). Hence
0
The l a s t c o n s i d e r a t i o n s had t o do w i t h t h e s y n t a c t i c n e a r - r i n g s N(I)
(lines
a f t e r D e f i n i t i o n 3 ) ; c f . a l s o [151. Now we t u r n t o t o p i c s which i n v o l v e t h e n e a r - r i n g s N(A) = I f , l x E S ( A ) l I f a f u n c t i o n fI
i n "A)
o f t h e "second type". i s ( l e f t - , r i g h t - ) i n v e r t i b l e ( i n N(A)) t h e n t h e r e
i s a separable system X' such t h a t
fI+xl
= f Pf = i d ,,(f 1 ,
=f;fZI
=id,
r e s p e c t i v e l y ) , where i d i s t h e i d e n t i t y f u n c t i o n on t h e group L(A) o f a l l formal L a u r e n t s e r i e s on A. F o r many a p p l i c a t i o n s , however, t h i s i s t o o r e s t r i c t i v e and one r e q u i r e s a weaker k i n d o f " r e v e r s i b i l i t y " : D e f i n i t i o n 8: f EN(A) i s &wem56le w i X h delay L i f t h e r e i s a system 1' such t h a t f,f,= z , where z L ( r a i z i ) = zaizi+L.
f
Hence Z* tl i s a separable system whose t r a n s f e r f u n c t i o n maps (ak,ak+, ,...)
..
i n t o (bk = 0, ,,b, = 0,. ,b k + L - a = O ~ b k + L = a k s bk+L-l = a k + l s - . . ) . i n v e r t i b l e w i t h delay L = 0 i f f i t i s l e f t i n v e r t i b l e .
Also.
fx i s
227
Near-rings and non-linear dynamical systems
I n v e r t i b i l i t y , as defined so f a r , o n l y says something i f X s t a r t s i n s t a t e 0. I n order t o o b t a i n a more general s e t t i n g we f o l l o w 1161 by d e f i n i n g a s l i g h t l y more general form o f i n v e r t i b i l i t y : D e f i n i t i o n 9: 1 i s i n u e h t i b t e luith d&g
L dhom the d r n e q € Q i f t h e r e i s
a f u n c t i o n f which associates t o each sequence bkybk+, sequence sky...
. . ,b k+n+L a
such t h a t E, s t a r t e d i n s t a t e q, responds bk,...ybk+n+L
t o t h e i n p u t sequence ak 'k+n+l
,...,bk+,,.
,...yak+n,xk+n+, ,...,x k+n+L ( f o r some a p p r o p r i a t e
"k+ntL)'
y"'
L i k e i n a good f a i r y - t a l e , t h r e e wishes w i l l be f u l f i l l e d : we can r e s t r i c t
us t o k = 0, q = 0, and n = 0. The t h i n g w i t h k = 0 i s simply a consequence o f timeinvariance: i t does n o t matter a t which time k we s t a r t our engine. We s h a l l assume k = O from now on. The remaining two wishes are l e s s t r i v i a l and seemed a t f i r s t t o be p a r t i c u l a r t o l i n e a r systems. Theorem 5: L e t 1 be separable w i t h maps a,B,y,6
( i n s h o r t : t =(aY!3,y,6)),
and q E Q . ( i ) t i s i n v e r t i b l e w i t h delay L from q
...,bL) -,a. w i t h fE(a,,xl p r i a t e x,,x2 ,... . (bo ,
,x2,...)
( i i ) t i s i n v e r t i b l e w i t h delay L from q
there i s a f u n c t i o n f,: = ( bo y . . .
99
,bLy...)
BL+'
-D
A,
f o r some appro-
E i s i n v e r t i b l e w i t h delay L from 0.
Pmod: (i) e:
4
i s c l e a r ( t a k e k = q = O i n D e f i n i t i o n 9).
L e t bo,...,bn,...,bn+L
be given. We have t o show t h a t we can determine
...,an such t h a t we get G(q,aoa ,...anxn+,...xn+L) xn+, ,. .. ,xntL. We proceed by i n d u c t i o n on n . a,,
I f n = O we choose G(O,ao) =b;,
b i :=y(O)- y ( q ) + bo.
=b b
We get a. = f o ( b i , b ,
,..
.bn+L f o r some
,...,bL)
such t h a t
which means y(0) + 6(ao) =y(O) - y ( q ) + bo, whence 6(ao) = - y ( q ) + bo.
Hence G(q,ao) = y ( q ) + 6(ao) = b o y as desired. For 1 2 i 5 n, l e t b i :=y(O)-y(F(q,a,a aiel)) + bi. With ai =fo(b;,bi+, , b i + L ) we g e t ai-,))+biy from which we get G(O,ai) = b ; , hence y ( 0 ) + 6 ( a i ) = y ( O ) - y ( F ( q , a = G(q,a o...ai)y the l a s t output i n G(q,a ,...ai), as C(q,a ,...ai) = y(F(q,a ai-,))+6(ai) =bi. For i > n we can do t h e same, f o r an a r b i t r a r y ybn+Ly i n order t o g e t some appropriate x,,+~ ,x,,+~. choice o f ,,b, ( i i ) f o l l o w s from (i). n
,,..
,...
,...
h
,...
,...
,...
Now we g i v e some necessary and s u f f i c i e n t conditions f o r the actual c o n s t r u c t i o n o f a separable system E ' such t h a t fx++ = frl
o
fz transforms
228
G. Pilz
(ao,O,O
,...) =aoz0 .i n t o
...,O,a 0 ,o,...)
(0,
=aoz
L
.
I n o r d e r t o s i m p l i f y t h e f o l l o w i n g expressions, we r e t u r n t o t h e Markov f o r a separable system w i t h maps a,p,y,6 ( D e f i n i t i o n 5). The should be meaning of [a,p;nl ($1 ,$,,-,) f o r appropriate maps @1 Remember t h a t we can take k = O i n D e f i n i t i o n 5, and we s h a l l do so. c l e a r , too. symbols [a,p;n]
)...
,...
P r o p o s i t i o n 10: L e t E = (Q,A,B,F,G) symnet ric
= (Q,A,B,a,p,y,6)
.
( i ) I f t h e r e e x i s t zero-symmetric maps a ' : Q
d:B
-, Q,
6':B -, A such t h a t , f o r some L E
+
be separable and zero-
Q, y ' : Q + A, and homomorphisms
No,
( 0 ) 6 ' 6 = 0 ( = zero map)
( 1 ) y[a',p';Ol6+
6'yB = 0
-
(*
( i y 1 [a I ,D I ;i 1 I ( 6,yp ,yap
.,yai -2
,..
( L ) y'[a',p';L-ll (6,yp,yap,...,ycrL-2p) then 1' : = (Q,B,A,a',p',y',6')
( ** )
fx,+x,(ao,O,O
+ 6 lyai
Ptroob: L e t
x
+
(1 5 i5 L 1
f u l f i l l s f o r each a o E A :
,...) = ( 0 ,... ,0,xL,xL+, ,... ) w i t h Q, 6':B
-
p=o
+ 6'yaL-lb= i d
( i i ) Conversely, i f zero-symmetric maps a ' : Q homomorphisms d:B
-
+
+
xL =ao.
Q, y ' : Q + A, and
A f u l f i l l (**) then they f u l f i l l (*).
.
s t a r t a t time 0 i n s t a t e qo = 0. The i n p u t sequence aoOO..
gives a sequence o f s t a t e s qo = 0, q1 = a ( O )+ p(ao) = p(ao), q2 = a p ( a o ) + p ( 0 ) =
,...,qi
( i 2 1 ) and produces t h e output sequence i-1 b o = v ( 0 ) + 6 ( a 0 ) = 6 ( a o ) , bl = v ( a ( a o ) ) + 6 ( 0 ) =uD(ao) bi = y a e(a,) (1 l i ) . Suppose t h e r e a r e maps a',p',y1,6' (as s p e c i f i e d i n t h e statement) then we g e t = ap(ao)
(ao,O,O,
fl,+l
= a i-lp(ao)
,...,
...I
=(xoyxl,
...) w i t h x o = y ' ( 0 ) + 6 ' ( b o ) = 6 ' 6 ( a 0 ) ,
xi = u ' [ a ' , B ' . i - l l (bo,b l , . . . ,bi-l) + 6 ' ( b i ) = ( u ' [ a ' , p ' , i - l I i-1 B)(ao). Hence (**I i s f u l f i l l e d i f f (*) holds.
+ 6'ya
and, f o r 1
(6,y0,yaB,
...,ya
0
Remark: I f L = O then (*) reduces t o 6 ' 6 = i d . This i s s o l v a b l e i f f 6 i s a monomorphism such t h a t I m 6 i s a semidirect summand ( r e t r a c t ) o f B y i.e.
if
N o f B such t h a t I m 6 + N = B y I m 6 n N {O). I n t h i s case, every b E B can uniquely be w r i t t e n as b = i + n w i t h i E I m 6, n E IN.
t h e r e e x i s t s a normal subgroup
This holds, because i f 6 has t h i s p r o p e r t y then 6 i s an isomorphism from A t o
I m 6 , having an inverse 6':Im 6 + A .
If B = I m 6+N,
I m 6 n N = { O ) , we can
extend 6 ' t o a homomorphism from 6 t o A by d e f i n i n g 6 ' = 6 ' ( i + n ) : = 6 ' ( i ) . Conversely, i f 6 ' 6 = i d holds f o r some homomorphism 6 ' : B
+
A then
p)+
229
New-rings and non-linear dynamical systems B = I m 6 t K e r 66' v i a b = 6 a 1 ( b ) t ( - 6 6 ' ( b ) + b ) , and I m 6 n K e r 6 6 ' = { 0 1 . L e t us now f i x t h e f o l l o w i n g
Notation: L e t z , ~ 'be as i n P r o p o s i t i o n 10, w i t h ( * ) holding. I f L 2 1 , l e t L lpl1 1 x1 : = tQ .B,A,a , ,y ,d ) w i t h 1 a (bl ,...,bL) : = (b2, bL,O) 1 13 (b) : = (0, 0,b) 1 Y (b, 9 . SbL) 1 6 (b) : = 6 ' ( b 1 Low l e t , f o r L 2 1, I f L = 0, take x : = ({O},B,A,O,O,O,6'). L- 1 L + 6'ya Then x2 : = (Q,A,A,a-B?&,-?,id : = y'[a',p',L3 (y,ya ya
...,
...,
--
.
,...,
1 1 : L e t x be as above. I f t s t a r t s i n s t a t e q, = ( 0 Proposition 1 i t assumes s t a t e s q1 = (0,. ,O,bo) .qL = (bo,.. qLtl = (bl,..
..
,...
.
under t h e i n p u t sequence (bo,bl,b2,...), c1 = y ' [ a ' , p ' , L - l ] t
6'(bLtl),
cLtl
(0
,...,O,bo) 1
PUJO~:We g e t q1 = a ( 0 1
t p (bl) = ( 0
t
=y'[a',B',L-lI
,... ,bo,bl),
c 0 = y ' [ a ' ,p' ,L-1 ] ( 0
(0
c, =y'[a',B',L-l]
,... ,O)
. , b L b L t l ,-.
and produces outputs co = 6 ' ( b o ) ,
6'(b,),
..., cL = y ' [ a ' , P ' , L - l l
(b
,bLtl)
+
,... ,0) then
1
13 (b,)
=
(0
+
(bo
,... ,bL)
+
6'(bL+2),...
,...,O,bO),
1
q2 = a ( 0
,...,O,bo)
+
and so on. The outputs a r e
,...,0) 6 ' ( bo) = 6 ' (b,) , ,...,O,bo) + 6 ' ( b l ) , etc. t
0
Before t h e n e x t p r o o f i t i s good t o have t h e f o l l o w i n g Lemma: I f 1 and 1' a r e as above, we g e t f o r a l l i €IN and bo, [a',~',L-ll
(0
,...O,bo, ...,b.1-1 )
= [a',p',i-I]
(bo
,...,b i - 1 )
... ,bi-l:
Pmob: This can e i t h e r be shown by i n d u c t i o n o r by r e c a l l i n g t h a t [a',p',L-l]
,...,O,bo,..,
(0
ybi-l)
i s t h e s t a t e o f 1' a f t e r s t a r t i n g a t "time"
...,bi-,.
L - i t 1 and i n p u t sequence b o y
Due t o time-invariance,
t h i s i s t h e same
as t h e s t a t e o f Z ' a f t e r s t a r t i n g a t time 0 and i n p u t sequence bo,...,bi-l. Theorem 6: L e t 1, 1' , x2 be as above. Then we have f o r a l l a o E A and L E IN L + ,f E l i t 12 (ao,O,Oy. ..I = (0 ,O,ao,O ) = aoz
,...
P4.006: We i n s e r t b o = 6 ( a o ) , bi = y a
,...
i-I
.
p(ao) (see P r o p o s i t i o n 10) i n t o t h e
formulas f o r ci i n P r o p o s i t i o n 11 and g e t
0
230
C.Pilz
c0 = 6 ' ( b0 ) = 6 ' 6 ( a o ) = 0 , and, f o r l < i < L ,
,... ,O,bo,...
ci =y'[a1,!3';L-11
(0
,b i - 1 ) t 6'(bi)
i-1 )+6'(bi)
( u s i n g t h e Lemma) i-2 i-1 !3(ao))+ 6 ' y a e(ao)
=y'[a',B',i-11
( bo,...,b
=v'[a',!3',i-11 o(ao)=O
(6(ao),v8(ao),ua~(ao,... ,ya
.....
i
I f , however, i > L, we g e t
,...,
ci = y ' [ a ' , B ' , L - l ] (biTL bi - 1 ) + 6 ' ( b . ) = y' [a' ,B' ,L-1 I (ya1 -L- 1e(ao) , ,yai'2fi(ao)) = (Y'[~',P',L-II
(ya
or, with i= L + j ( j > O )
i-L-1
,. .
+ 6'yai-lg(ao)
i-1 ~ , . . . , y a ~ - ~ ~ ) + b ' y ae)(ao),
,...,yaL+J-')+6'ya L+j-1 ) ( D ( a o ) ) = ,.y aj - 1 p(ao). 2 The s t a t e s o f X a r e given by qo = ... =qL = 3, qL+j = aj-'o(a0) f o r j 2 1. We c ~ = +( u ' ~[ a ' , B ' , L - l l
j-1 j (ra ,ya
show t h i s l a s t statement by i n d u c t i o n on j. For j = 1 we g e t qL+l = (a-R?)(qL)
+
!3(cL) = (a-@?)(O) + B(ao) = B(ao) =aog(ao). The i n d u c t i o n s t e p from
+ o ( c ~ + =~ (a-b<)aj-'P(a0) ) + which was t o be shown.
j t o j+li s g i v e n by qL+j+, +
B$-l 8(ao) = aJe(a,),
= (a-&)(qLcj)
,. .. o f E x1* x2 . We
F i n a l l y , we can now compute t h e outputs do,dl course, do = dl =
... = dL-l
j 2 1, dL+j = -;(qL+j)
+t
= 0, and then dL = -(;q).,
+ id(cL+j)
= -
get, o f
+ i d ( c L ) = 0 + cL = a,
and f o r
+ $aJ-'@(a0) = 0. Hence t h e r e s u l t .
Remark: ( i ) I f L = O then we g e t already fz,+E1
(ao,O,O
,... )
,...I= (ao,O,O
( i i ) I t i s easy t o show t h a t , i f L = 0,
fX*Xi,+Ez
(ao,al ,a2,... 1 = (ao,al ,a2,.. . I holds. Hence fEl+tx2 i s a l e f t i n v e r s e o f fE.
( i i i ) The same holds i f L = 1 and 6 = i d , f o r instance, b u t n o t i n general. ( i v ) Also, i t can be seen t h a t Theorem 6 i s n o t t r u e i n general i f zero-symmetri c. The c o n s t r u c t i o n i n Theorem 6 o f an i n v e r s e X1*X2
x
w i t h delay L o f
i s not
x
can be
used f o r decoupling procedures (see [161) i f A i s t h e d i r e c t product o f some o t h e r groups. I f o n l y one component i s e x c i t e d , u s u a l l y a l l components respond. Decoupling i s t h e procedure t o p l u g another system (z 1*x2, f o r instance) a f t e r X so t h a t o n l y one output channel responds t o t h e a c t i v a t i o n o f j u s t one i n p u t channel. For more on t h i s , see e.g. [161. I n v e r t i b i l i t y has t o do w i t h r e g u l a r i t y i n t h e n e a r - r i n g N(A). We c a l l a system z c S ( A ) keg&,
i f t h e r e i s some o t h e r system Y € S ( A ) such t h a t Z * Y * x
has t h e same i n p u t - o u t p u t behaviour ( = t r a n s f e r f u n c t i o n ) as 1 i t s e l f . S(A) i s
23 1
Nearrings and non-linear dynamical systems
r e g u l a r i f every E E S ( A ) i s regular. A system t E S ( A ) i s LdempoZent, i f t*t has t h e same behaviour as
z. We say t h a t N(A) i s A h n g &
hqgCLeah
i f f o r each
S(A) t h e r e i s a zero-symmetric YES(A) such t h a t
zero-symmetric system
t i t Z i + Y behaves l i k e 1. I n t h i s case, S(A) i s r e g u l a r and
\y
i s invertible with
delay L = O . Now we g e t from 9.155,
9.156,
9.158,
9.159 ( c ) , ( f ) and 9.62 o f [141:
Theorem 7: I f S(A), A f i n i t e , i s r e g u l a r then )
and W E (as defined above) a r e idempotent systems.
t*Y
The s e t o f a l l systems X*X
i)
( X E S ( A ) ) coincides w i t h t h e s e t X + E f o r
an idempotent system E. i i ) I f t h e zero system R ( w i t h f R = O ) and t h e i d e n t i t y system I ( w i t h fI = i d ) a r e t h e o n l y idempotent systems i n S(A) then every system i n S(A) i s i n v e r t i b l e with d e l a y
L = O (and conversely).
The s i t u a t i o n i n ( i i i ) happens i f f f o r each E E S ( A ) t h e r e i s no TES(A)
(iv)
such t h a t ttt T = 62.
I f S(A) i s even s t r o n g l y r e g u l a r then every t c S ( A ) i s t h e p a r a l l e l
(v)
connection o f a constant and f i n i t e l y many systems which are i n v e r t i b l e w i t h zero delay. The same statements hold, o f course, f o r subnear-rings o f S(A) i f they happen t o be r e g u l a r ( w i t h t h e obvious changes). F i n a l l y , we b r i e f l y mention t h e r e a l i z a t i o n problem. For a given f u n c t i o n f:L(A)
+
L(B) (A,B groups) we want t o f i n d a group Q and "separable" f u n c t i o n s
F:Qx A
+
Q,
G:Qx
A
-,8
such t h a t f o r 1 : = (Q,A,B,F,G),
f =f,.
This i s c e r t a i n l y
a n o n - t r i v i a l problem, and i t i s n o t always solvable, s i n c e we know from Example 1 t h a t n o t every f u n c t i o n f : L ( A ) separable systems
+
L(B) i s o f t h e form f = f xf o r some
x.
I f f i s l i n e a r and A,B a r e a b e l i a n then t h e usual methods and r e s u l t s o f
l i n e a r r e a l i z a t i o n theory, as developed by Kalman i n [81, apply. I n t h e o t h e r case, one might t h i n k about t h e Arbib-Manes c a t e g o r i c a l approach (see [21 and [13]). The obvious candidate i s t h e category C c o n s i s t i n g o f groups and maps between these groups. Since every s e t can be made i n t o a group (see e.g. [91), C has products ( = Cartesian products), coproducts ( = d i s j o i n t unions w i t h some group s t r u c t u r e s ) , and. f r e e o b j e c t s , hence a l s o f r e e dynamics i n t h e sense o f [2].
Hence we g e t a f r e e r e a l i z a t i o n ( l e f t a d j o i n t ) w i t h s t a t e group A* (any
group s t r u c t u r e on A*) i n the sense o f [21 (p. 680). b u t t h e maps
F,G a r e n o t
separable i n general. Also, s i n c e f ( A ) i s i n general n o t a group i f f i s n o t a
232
G. Pilz
homomorphism, we g e t d i f f i c u l t i e s w i t h c o e q u a l i z e r s and minimal r e a l i z a t i o n s ( c f . [131, p. 726). Also, t h e r e s t r i c t i o n s t o a f f i n e maps F,G ( i . e .
sums o f a
homomorphism and a c o n s t a n t map) and a b e l i a n A does n o t work: t h e c a t e g o r y C ' o f a b e l i a n groups and a f f i n e maps between these does n o t c o n t a i n (co)products.
So t h i s remains an open problem.
REFERENCES
[l] A d l e r , I.,Composition Rings, Duke Math. J. 29 (1962), 607-625. [21 Bobrow, L.S. and A r b i b , M.A. , D i s c r e t e Mathematics, Saunders, P h i l a d e l p h i a , 1974. [3] C a s t i , J.L., N o n l i n e a r System Theory, Academic Press, New York, 1985. [41 Gratzer, G., U n i v e r s a l Algebra, 2nd e d i t i o n , Springer-Verlag, New YorkH e i d e l b e r g - B e r l i n , 1979. [5] Hofer, G. and P i l z , G., Near-rings and automata, Proc. Conf. Univ. Algebra, K l a g e n f u r t , A u s t r i a , 1982, 153-162. The s y n t a c t i c n e a r - r i n g o f a l i n e a r s e q u e n t i a l machine, 161 Holcombe, W.M.L., Proc. Edinbg. Math. SOC. 26 (1983), 15-24. [7] Holcombe, W.M.L., A r a d i c a l f o r l i n e a r s e q u e n t i a l machines, Proc. Royal I r i s h Acad. 84A (1984), 27-35. [8] Kalman, R.E., Falb, P.L. and A r b i b , M.A. , Topics i n Mathematical System Theory, McGraw-Hill , New York, 1969. [91 Kertesz, A., E i n f u h r u n g i n d i e t r a n s f i n i t e Algebra, B i r k h a u s e r , Basel, 1975. [I01 Lausch, H. and Nobauer, W . , Algebra o f Polynomials, N o r t h H o l l a n d / American E l s e v i e r , Amsterdam, 1973. [Ill L i d l , R. and P i l z , G., A p p l i e d A b s t r a c t Algebra, S p r i n g e r - V e r l a g (Undergraduate Texts i n Mathematics), New Y o r k - H e i d e l b e r g - B e r l i n , 1984. [12] Menger, K., Algebra o f A n a l y s i s , N o t r e Dame Math. L e c t u r e s , No. 3, Notre Dame, I n d i a n a , 1944. [I31 Padulo, L. and A r b i b , F1.A. , System Theory, Saunders, P h i l a d e l p h i a , 1974. [14] P i l z , G., Near-Rings, 2nd e d i t i o n , N o r t h Holland/American E l s e v i e r , Amsterdam, 1983. [151 P i l z , G., S t r i c t l y connected group automata, submitted. 1161 Sain, M.K., I n t r o d u c t i o n t o A l g e b r a i c System Theory, Academic Press, New York, 1981.
Near-rings and Near-frlds, C. Betsch (editor) 0 Elsevier Science Publishen B.V.(North-Holland), 1987
233
REDUCED NEAR-RINGS
D. RAMAKOTAIAH and V. SAMBASIVARAO Department o f Mathematics, Nagarjuna U n i v e r s i t y , Nagarjunanagar-522 510, A.P. I n d i a ABSTRACT I n t h i s paper we introduce a p a r t i a l o r d e r r e l a t i o n i n a reduced nearr i n g and show t h a t t h e s e t o f a l l idempotents o f a reduced n e a r - r i n g w i t h i d e n t i t y forms a Boolean algebra under t h i s p a r t i a l ordering. F u r t h e r we introduce t h e n o t i o n s hyper atom and orthogonal subsets i n a reduced n e a r - r i n g w i t h i d e n t i t y and show t h a t a reduced n e a r - r i n g with i d e n t i t y i s isomorphic t o a d i r e c t product o f n e a r - f i e l d s i f and o n l y i f i t i s hyper atomic and o r t h o g o n a l l y complete. I n t r o d u c t i o n : Abian [ l ] introduced a r e l a t i o n 2 i n a reduced r i n g R by d e f i n i n g a 2 b y a,b E R i f and o n l y i f a2 = ab and showed t h a t t h i s r e l a t i o n i s indeed a p a r t i a l order r e l a t i o n i n R. I n t h i s paper, we examine t h e p o s s i b i l i t y o f extending t h e above mentioned p a r t i a l order i n a reduced r i n g t o a reduced near-ring and o b t a i n some o f t h e consequences which o f course a r e generalizat i o n s o f t h e r e s u l t s which were already obtained f o r reduced r i n g s , This paper i s d i v i d e d i n t o t h r e e sections. I n s e c t i o n
5
1, we show t h a t t h e
s e t o f a l l idempotents o f a reduced near-rings w i t h i d e n t i t y forms a Boolean algebra and o b t a i n some i n t e r e s t i n g consequences o f t h i s r e s u l t . I n s e c t i o n
5
2 , we introduce t h e n o t i o n o f orthogonal subsets i n a reduced n e a r - r i n g and
show t h a t t h e f o l l o w i n g c o n d i t i o n s f o r a reduced n e a r - r i n g w i t h i d e n t i t y w i t h p a r t i a l order 5 a r e equivalent: (i)
(N,2) s a t i s f i e s t h e ascending chain c o n d i t i o n
(ii)
( N & ) s a t i s f i e s t h e descending chain c o n d i t i o n and N i s o r t h o g o n a l l y complete
( i i i ) Any orthogonal subset o f N i s f i n i t e (iv)
N s a t i s f i e s t h e ascending (descending) chain c o n d i t i o n on a n n i h i l a t o r s .
I n section
5
3, we d e f i n e hyper atoms i n a reduced near-ring w i t h i d e n t i t y
and show t h a t a reduced near-ring w i t h i d e n t i t y i s hyper atomic and o r t h o g o n a l l y complete i f and o n l y i f i t i s isomorphic t o a d i r e c t product o f n e a r - f i e l d s .
D. Ramakotaiah and V. Sambasivarao
234
P r e l i m i n a r i e s : We r e c a l l t h a t an algebraic system N = (N,+,.,o) n m - h i n g i f and only i f ( i ) (N,+,o)
i s a ILegt)
i s a ( n o t necessarily a b e l i a n ) group, ( i i )
(N,.) i s a semi-group, ( i i i ) a(b+c) = a b + a c f o r a l l a,b,c E N and ( i v ) oa = o f o r a l l a E N. A near-ring w i t h o u t Ron-zero n i l p o t e n t elements i s c a l l e d a d u c e d near-ring. For any x E N, we denote t h e i d e a l generated by x by <x>.
For o t h e r terminology t h e reader i s r e f e r r e d t o [21. We note the f o l l o w i n g w e l l known r e s u l t s i n the theory o f near-rings. L e m 0.1:
Let
N be a reduced near-ring. Then ab
= 0 , a,b C:
N i m p l i e s ba =
0.
Further f o r any a E N, A(a) = {x E N I ax = 01 i s an i d e a l o f N. L e m 0.2:
L e t N be a reduced near-ring. Then
(i) N has IFP, t h a t i s , a,b E N, ab = o implies anb = o f o r a l l n E N. ( i i ) ne = ene f o r every idempotent e i n N and n ’ i n N. ( i i i ) IfN contains an i d e n t i t y , a l l idempotents i n N are c e n t r a l .
A c a r e f u l observation o f ( i i ) i n lemma 0.2 says t h a t a reduced n e a r - r i n g has a r i g h t i d e n t i t y i f and o n l y i f i t has an i d e n t i t y .
Lemma 1.1: L e t N be a reduced near-ring. I f a,b E N, ab = 0, then a+b = b+a. Further f o r a l l idempotents e,f E N, e f = o i m p l i e s ( e + f ) e = e and ( e + f ) f = f. Proof: Suppose a,b E N such t h a t ab = 0. Then by lemma 0.1, ba =
0.
This
f a c t and our supposition imply t h a t a(a+b-a-b) = o and b(a+b-a-b) = 0. Again by lemma 0.1 (a+b-a-b?
, we
have (a+b-a-b)a = o and (a+b-a-b)b = 0. Consequently
= o and hence a+b-a-b = 0. This shows t h a t a+b = b+a. The r e s t o f
the proof f o l l o w s from lemma 0.2 ( i i ) . Lemma 1.2: L e t N be a reduced near-ring w i t h
d e n t i t y . Then f o r a l l idem-
potents e,f E N, e f and e + f - f e a r e idempotents. Proof: By lemma 0.2 ( i i i ) , a l l idempotents i n N a r e c e n t r a l and the r e s t o f the p r o o f f o l l o w s as i n the case o f r i n g theory.
Reduced near-rings
235
We now introduce an order r e l a t i o n i n an a r b i t r a r y n e a r - r i n g N by s e t t i n g a 2 b f o r a,b E N i f and o n l y i f a2 = ab. I f a n e a r - r i n g i s a Boolean r i n g , than t h e r e l a t i o n 6 i s a p a r t i a l o r d e r i n N. What i s t h e c l a s s o f near-rings f o r which t h e above i n t r o d u c t e d r e l a t i o n i s a p a r t i a l o r d e r ? We answer t h i s question i n t h e f o l l o w i n g . Theorem 1.3:
L e t N be n e a r - r i n g and 5 a r e l a t i o n on N as defined above. Then
N i s a reduced n e a r - r i n g i f and o n l y if5 i s a p a r t i a l o r d e r r e l a t i o n on This i s t h e case, ( N , . , < )
N.
i s a p a r t i a l l y ordered semigroup.
Proof: Suppose N i s a reduced near-ring. C l e a r l y 5 i s r e f l e x i v e . Suppose E N such t h a t a 5 b and b S a. Then a2 = ab and b2 = ba, t h a t i s , a(a-b) = o
a,b
and b(b-a) = (a-b)'
0.
By lemma 0.1,
we have (a-b)a = o and (b-a)b = 0. Consequently
= 0 . Since N i s reduced, a-b = o and hence a = b. So i i s antisymmetric.
Suppose a 5 b and b 5 c. Then a2 = ab and b2 = bc, t h a t i s , a(a-b) = o and 3 3 2 0 . This i m p l i e s aca = acb, a c = a b and a -ac = a(b-c). From these 2 f a c t s , we have a2(a2-ac) = o and ac(a -ac) = 0 . By lemma 0.1, we have b(b-c) =
(a 2 -ac)a2 = o and (a 2 -ac)ac = 0 . Consequently (a2 -ac)' = 0. Since N i s reduced, 2 we have a -ac = o and hence a2 = ac. So i i s t r a n s i t i v e . Suppose x 5 y and a 5 b. Then x(x-y) = o and a(a-b) =
0.
This i m p l i e s xa(x-y) = o and ax(a-b) = o ,
t h a t i s , xax = xay and axa = axb. From these f a c t s , we have xa 5 yb. Therefore
(N,.,i)
i s a p a r t i a l l y ordered semigroup. Conversely,suppose Ii s a p a r t i a l
o r d e r r e l a t i o n on N. I t i s enough i f we show t h a t a2 = o i m p l i e s a = 0. It i s c l e a r t h a t o 2 a. Since a2 = o = oa, we have a 5
0.
Hence a = 0.
We introduce suprema and infima o f any subset o f a reduced n e a r - r i n g under t h e p a r t i a l o r d e r r e l a t i o n 2 i n t h e usual way. For any subset S o f N, sup S ( i n f S ) denotes t h e supremum (infimum) o f S. From now on,
N stands f o r a
reduced n e a r - r i n g with i d e n t i t y , unless s t a t e d otherwise, and 2 always stands f o r t h e p a r t i a l o r d e r defined above. Proposition 1.4: For any subset S of idempotents o f N, i f sup S ( i n f S ) e x i s t s , then i t must be an idempotent. Proof: Suppose sup S = x e x i s t s . Then s = sx f o r a l l s E S. This i m p l i e s s x 2 y n d s =< x2. Since sup S = x, we have x s x2 and hence x2 = x3. Now x(x-x2) = o and x 2 (x-x 2 = 0. Consequently (x-x 2 ) 2 = o and hence x = x 2 Suppose i n f S = x. Then x2 = xs f o r a l l s E S. Now x2 = xs = xs2 = x 2s = x 3
.
.
By an argument s i m i l a r t o the above, i t can be shown t h a t x i s an idempotent, too.
D. Ramakotaiah and V. Sambasivarao
236
I n p r o p o s i t i o n 1.4, i f x i s a l o w e r bound o f S, t h e n x i s a l s o
Remark 1.5: an idempotent.
F o r a l l idempotents e,f E N, e + f - f e = sup { e , f )
Theorem 1.6:
and
Thus any two idempotents commute w i t h each o t h e r a d d i t i v e l y .
e f = i n f [e,fl.
T h i s r e s u l t can be v e r i f i e d i n t h e usual way. The p a r t i a l o r d e r r e l a t i o n S i n N induces a p a r t i a l o r d e r r e l a t i o n which we a l s o denote by 6, i n B, t h e s e t o f a l l idempotents o f N. I n view o f theorem 1.6, we have
e
A
Theorem 1.7:
(B,<)
Theorem 1.8:
Define
i s a l a t t i c e w i t h o and 1. A,
v and
'
on B as f o l l o w s . L e t e,f
E B. D e f i n e
f = e f , e v f = e + f - f e and e ' = 1-e. Then (B,A,v,',o,~)
i s a Boolean
a1 gebra. L e t a E N. Then aN = A ( l - a ) i f and o n l y i f a i s an idempotent.
L e m 1.9:
T h i s i s t h e case, eN i s an i d e a l f o r each e E B and <e> = eN. Proof: By lemma 0.2 ( i i i ) , a l l idempotents i n N a r e c e n t r a l and i t i s a l s o c l e a r t h a t f o r each idempotent e, 1-e i s a l s o an idempotent i n N. Suppose a i s an idempotent i n N. L e t x E A(1-a).
Then ( 1 - a ) x = 0. By lemma 0.1,
we have
x = xa = ax E aN. So A(1-a) 5 aN. L e t x E aN. Then x = an f o r some n E N. T h i s i m p l i e s x ( 1 - a ) = 0. By lemma 0.1,
(1-a)x = o and hence x E A(1-a).
So
A(1-a) = aN. Conversely, suppose a E N such t h a t A(1-a) = aN. Since a E aN, we have (1-a)a = 0. By lemma 0.1,
a ( 1 - a ) = o and hence a i s an idempotent. By
lemma 0.1 and by t h e above argument, f o r each e E B y eN i s an i d e a l and <e>=eN. Lemma 1.10: Proof:
F o r any e,f E B, eN n f N = efN.
By lemma 0.2 ( i i ) , efN _c eN r l fN. L e t x E eN
n f N . Then x
= en and
x = f n ' f o r some n,nl E N. Now x = en = een = e f n ' E efN. So en n f N = efN. Lemma 1.11:
F o r any e,f
Proof: By lemma 1.2,
E B y e N + f N = (e+f-fe)N.
e + f - f e i s an idempotent and by lemma 1.9,
eN, f N and
( e + f - f e ) N a r e i d e a l s o f N. Since e,f E B y we have e E eN and f - f e E f N . So e + f - f e E eN+ f N and hence ( e t f - f e ) N _c eN+ f N . On t h e o t h e r hand, e = e + f e - f e = = e ( e + f - f e ) = ( e + f - f e ) e E (e+f-fe)N
show t h a t f N
and hence eN _c (e+f-fe)N.
5 (e+f-fe)N. So eN t f N _c (e+f-fe)N. Hence eN + f N
S i m i l a r l y we can = (e+f-fe)N.
231
Reduced near-rings Lemma 1.12: The s e t S = {eN
Ie
Proof f o l l o w s from lemmas 1.9, Theorem 1.13:
L e t S = {eN
Ie
E B1 i s a l a t t i c e under s e t i n c l u s i o n . 1.10 and 1.11.
E 6). Then the mapping a:B
a(e) = eN i s a l a t t i c e isomorphism and hence Proof: By lemmas 1.10 and 1.11,
(S,s)
+
S defined by
i s a Boolean algebra.
a i s a ,+and v-homomorphism. I t i s easy t o
v e r i f y t h a t a i s onto and f o r any e,f
E B, e 2 f i n B i f and o n l y i f a ( e )
5 a(f).
Suppose e,f E B such t h a t a(e) = a ( f ) . Then e = f n and f = e n ' f o r some n,n'
E N. Now f = e n ' = een' = e f = f n f
= e. So a i s one-one.
Therefore
a:B + S i s a l a t t i c e isomorphism and hence S i s a Boolean algebra. We r e c a l l t h a t a near-ring N i s b L t e g d a h i f and o n l y i f f o r each a E N, t h e r e e x i s t s a c e n t r a l idempotent e E N such t h a t = <e>. Theorem 1.14:
I f N i s a b i r e g u l a r reduced n e a r - r i n g w i t h i d e n t i t y and i f
S = CeN I e E B I and S ' i s the s e t o f a l l p r i n c i p a l i d e a l s o f N, then S = S '
and hence S ' i s a Boolean algebra. Further, the c a r d i n a l i t y o f t h e s e t o f a l l idempotents i n N i s the same as the c a r d i n a l i t y o f the s e t o f a l l p r i n c i p a l i d e a l s o f N. Proof: By lemma 1.9,
S 5 S ' . Since N i s b i r e g u l a r , f o r each a E N, t h e r e
e x i s t s an idempotent e E N such t h a t = <e> and hence by lemma 1.9,
=eN.
So S ' 5 S. Hence S = S ' and S ' i s a Boolean algebra by theorem 1.13. Since t h e c a r d i n a l i t y of B i s t h e same as t h e c a r d i n a l i t y o f S, by the above argument, t h e c a r d i n a l i t y o f B i s the same as the c a r d i n a l i t y o f 5 ' .
9 2 Throughout t h i s section,
N stands f o r a reduced n e a r - r i n g w i t h i d e n t i t y .
For any subset X o f N, A(X) denotes t h e s e t {n E N Proposition 2.1: c =
SUP
= o for all x
E XI.
L e t X be any subset o f N. Then t h e f o l l o w i n g are e q u i v a l e n t
f o r any c E N. (i)
I xn
X
( i i ) c i s an upper bound o f X and A(X) 5 A(c). I n f a c t , A(X) = A(c).
D. Ramakoraiah and V. Sambasivarao
238
Proof: ( i ) i m p l i e s ( i i ) : Suppose sup X = c. Then x S c f o r a l l x E X. L e t 2 f o r a l l x E X. This i m p l i e s c+d i s an
d E A(X). Then xd = o and x(c+d) = x
upper bound o f X. Since sup X = c, we have c 5 c+d. Consequently cd = o and 2 hence d E A(c). So A(X) c - A(c). L e t d E A(c). Then cd = o and x d = xcd = o f o r a l l x E X. Consequently xd =
So d E A(X) and hence A(X) = A(c).
0.
( i i ) i m p l i e s ( i ) : Suppose c i s an upper bound o f X and A(X)
5 A(c). L e t d be
any upper bound o f X. Then f o r a l l x E X, x2 = xc and x2 = xd. This i m p l i e s
5 A(c). Consequently c(c-d)
c-d E A(X)
C o r o l l a r y 2.2:
=
o and hence c 2 d. So sup X = c.
I f X i s a subset o f N such t h a t sup X = b e x i s t s , then f o r
a l l a E N, sup aX e x i s t s and equals ab. Proof: L e t a E N. Since x 5 b f o r a l l x E X, we have x ( x - b ) = 0. Then axa(x-b) = o and hence ax 2 ab f o r a l l x E X. So ab i s an upper bound o f ax. L e t d E A(aX). Then axd = o and hence xda = o f o r a l l x E X. This i m p l i e s da E A(X). Since sup X = b y by p r o p o s i t i o n 2.1, we have da E A(b). Consequently d E A(ab). So A(aX) _t A(ab). Hence by p r o p o s i t i o n 2 . 1 , sup aX = ab. D e f i n i t i o n 2.3:
A subset X o f non-zero elements o f N i s s a i d t o be o h t h o g o -
m d i f xy = o f o r a l l x,y E X w i t h x
$
y.
I
i E I}and Y = Cy. I j E J) be two orthogonal J subsets o f N such t h a t sup X = x and sup Y = y. L e t Z = { x . Y . I i E I, j E J I .
P r o p o s i t i o n 2.4:
L e t X = {xi
Then Z i s an orthogonal subset of N and sup Z = xy. Proof:
1 J
It i s c l e a r t h a t Z i s an orthogonal subset o f N. Since xi
5 x, we E I. This i m p l i e s x.y.(xi-x) = o and hence 1 J X . Y . X . = x.y.x f o r each i E I and f o r each j E J. Consequently x . y . 5 xyj f o r 1 3 1 1 J 1 J each i E I and f o r each j E J. Since y < y, we have y j ( y j - y ) = o f o r each j = j E J. Then xy.x(y.-y) = o and hence xy < xy f o r a l l j E J . Hence by t h e
have xi(xi-x)
= o f o r each i
J J j = above two arguments, x . y . 5 xy f o r each i E I and f o r each j E J . So xy i s an 1 J upper bound o f Z. L e t d E A(Z). Then x . y . = o f o r each i E I and j E J . This J J J J J f o r each j E J and dx E A(Y). Hence dx E A(y) and d E A(xy). So A(Z) 5 A(xy).
i m p l i e s y . d E A(X) and hence y . d E A(x) f o r each j E J. Consequently y.dx = o
Hence by p r o p o s i t i o n 2.1,
sup z = xy.
We note the f o l l o w i n g lemma Lemma 2.5: d E N.
I f f o r any x,y E N with xy = 0 , then (x+y)d = x d + y d f o r a l l
Reduced near-rings
239
Proof: L e t x,y E N be such t h a t xy = 0. Then by lemma 0.1, these f a c t s , we have x((x+y)d-yd-xd) lemma 0.1,
yx = 0. From
= 0 and y((x+y)d-yd-xd)
= 0. Again by
we have ((x+y)d-yd-xd)x = o and ((x+y)d-yd-xd)y = o and hence = 0. Consequently ((xty)d-yd-xd)'
((x+y)d-yd-xd)(x+y)
= 0. Since
N i s reduced,
(x+y)d-yd-xd = o and hence ( x t y ) d = xd+yd. As a consequence o f lemma 2.5,
we can deduce the f o l l o w i n g .
I f X i s an orthogonal subset o f N, then
P r o p o s i t i o n 2.6:
1 xN i s a d i r e c t XEX
sum o f N-subgroups o f N.
The proof o f t h i s p r o p o s i t i o n i s easy and w i l l be omitted.
I n order t o o b t a i n the main r e s u l t o f t h i s s e c t i o n , we need t h e f o l l o w i n g . Lemma 2.7:
I f {ail
i s a s t r i c t l y ascending chain o r a descending chain o f
elements i n N , then {ai-ai+ll Proof: L e t {ai) ai(aj-ajtl) = a.a 1
= aitlaj
i s an orthogonal subset o f N.
be a s t r i c t l y ascending chain. Suppose i j and i < j. Now = ai 2 - ai2 = o and ai+l( a . aj+l) = 2 2 = ai+l - aitl = 0. Consequently ( a . - aj+l ) a . = o and J 1
- aiajtl j
- ai+lajtl
-
= 0. So (a - a j + l ) ( a i -ai+l) = o and hence {ai -aitl} i s an ( a j ajtl )aitl j orthogonal subset o f N. By a s i m i l a r argument, t h e o t h e r case can be disposed
-
of. Lemma 2.8: sup {al,a2,
...,an]
I f {al,a2,
...,an}
i s an orthogonal subset o f N, then
e x i s t s and equals a +a +...+ a., 1 2
Proof: W r i t e a = a +a +...+an. Then aia = ai2 1 2 f o r a l l 1 5 i 5 n. Suppose ai 2 d f o r a l l 1 5 i 2 2.5, ad = a l d t a d t + and = al2 + a22 t +an. 2 orthogonal subset o f N, by lemma 2.5, we have a'
...
...
f o r a l l 1 5 i 5 n. So a . 5 a 1
5 n. Then a:
= aid.
,. ..,a ...+a$
Since {al ,a2 = a,2 t a;+
By lemma
1
i s an
So a'
= ad
and hence a = sup {al,a2,...,anl. Lemma 2.9: I f { a i l i s an orthogonal subset o f N, then {sup { a . J a s t r i c t l y descending chain (provided t h a t the suprema e x i s t ) and {SUP { a i 1 1 5 i 5
jll
i s a s t r i c t l y ascending chain.
1
j 2 ill i s
D. Ramakotaiah and V. Sambasivarao
240
Ij
Proof: W r i t e ci = sup {a.
J
2 i } . Then ci 2 a j f o r a l l j 2 i and ci+l
2 aj
f o r a l l j 5 i + l . This i m p l i e s ci 2 a . f o r a l l j 2 i + l . Therefore ci i s an
Ij
upper bound o f {a.
J
J
2 i + l I . So ci+l
descending chain. I f ci = ci+l aici+l
= ai
sup {a.
J
1j
5 ci f o r each i and hence {cil
f o r some i, then aici
:s
2 i+l) = sup {a.a.
orthogonal subset o f N. This shows tha:
Ij =
2
= ai.
is a
By c o r o l l a r y 2.2,
2 i + l ) = o since { a i l i s an
o and hence ai = 0 , which i s a
c o n t r a d i c t i o n . So {ci)
i s a s t r i c t l y descending chain. Write It i s clear t h a t d . = sup {a. I 1 5 i 5 j } . By lemma 2.8, d = a +a +...+a J 1 j 1 2 j' = 0 , which I d j } i s an ascending chain. Suppose d j = dj+l f o r some j . Then a j+
i s a c o n t r a d i c t i o n . So I d . ) i s a s t r i c t l y ascending chain.
J
The proofs o f t h e f o l l o w i n g lemmas 2.10,
2.11 and 2.12 a r e easy and w i l l be
omitted. Lemma 2.10:
The descending c h a i n and t h e ascending chain c o n d i t i o n s on
a n n i h i l a t o r s i n N are equivalent. Lemma 2.11:
I f {ail
i s an orthogonal subset o f N , then {A(al,a2,..
,ai
1I
i s a s t r i c t l y descending c h a i n o f a n n i h i l a t o r s . Lemma 2.12:
If {Ii) i s a s t r i c t l y ascending chain o f a n n i h i l a t o r s
n A(Ii) then Ii+,
*
( 0 ) and
i f o 9 ai E Ii+l n A(Ii),
then {ai}
n N, i s an orthogo-
nal subset o f N. D e f i n i t i o n 2.13:
A n e a r - r i n g N i s said t o be o h t h o g o m U y cornpLeAe i f every
orthogonal subset o f N has supremum. We now s t a t e t h e main r e s u l t o f t h i s section. Theorem 2.14: The f o l l o w i n g c o n d i t i o n s on N a r e e q u i v a l e n t : s a t i s f i e s the ascending chain c o n d i t i o n
(i)
(N,S)
(ii)
(N,5) s a t i s f i e s t h e descending chain c o n d i t i o n and complete
N i s orthogonally
( i i i ) Any orthogonal subset o f N i s f i n i t e (iv)
N s a t i s f i e s the ascending (descending) chain c o n d i t i o n on a n n i h i l a t o r s
The proof o f t h i s theorem i s a consequence o f the lemmas 2.7 t o 2.12.
24 1
Reduced near-rings 9 3
I n t h i s s e c t i o n we i n t r o d u c e t h e n o t i o n o f hyper-atoms and study some o f t h e consequences. Throughout t h i s s e c t i o n , N stands f o r a reduced n e a r - r i n g w i t h identity
.
D e f i n i t i o n 3.1: only i f x
j
A non-zero element a E N i s s a i d t o b e a hypm-atam i f and
a i m p l i e s e i t h e r x = o o r x = a and i f an
+
o f o r any n E N, t h e r e
e x i s t s s E N such t h a t ans = a. Example: L e t N ' be a n e a r - f i e l d . Then N x N ' i s a reduced n e a r - r i n g w i t h i d e n t i t y and S i s a p a r t i a l o r d e r i n g on N x
N'. Now f o r each non-zero element
a E N ' , (o,a) i s a hyper-atom i n N x N ' . Lemma 3.2:
L e t a be a hyper-atom i n
N. For any n E N,
i f an
+
0,
then an i s
a hyper-atom i n N. Proof: Since an
* o and a
i s a hyper-atom i n N, t h e r e e x i s t s s E N such t h a t
ans = a. Suppose x S an. Then xs 5 ans = a. Since a i s a hyper-atom, e i t h e r xs = o o r xs = a. Case ( i ) : I f xs = 0 , then xans = o and hence xa = 0. Since x 6 an, we have xL = xan = o and hence x = 0. Case ( i i ) : Suppose xs = a. x 5 an i m p l i e s x ( x - a n ) = 0. Then xs(x-an) = o and hence a ( x - a n ) = o and an(x-an) = 0. By lemma 0.1,
( x - a n ) x = o and (x-an)an = 0. Consequently
(x-an)2 = 0. Since N i s reduced, x-an = o and hence x = an. Suppose a n t t E N. Then a n t s = a f o r some s E
+
0,
N s i n c e a i s a hyper-atom. Now
( a n t ) s n = ( a n t s ) n = an. Hence an i s a hyper-atom. Lemma 3.3:
L e t a be a hyper-atom i n N. Then t h e r e e x i s t s an element s E N
such t h a t a = asa, t h a t i s , a i s r e g u l a r i n t h e sense o f Von Neumann.
*
o and t h e r e e x i s t s an element Proof: L e t a E N be a hyper-atom. Then a' s E N such t h a t a 2 s = a. T h i s i m p l i e s a(a-asa) = o and asa(a-asa) = 0. By lemma 0 . 1 , (a-asa)a = o and (a-asa)asa = o and hence (a-asa)2 = 0. Since N i s reduced, we have a = asa. Lemma 3.4:
L e t a be a hyper-atom i n N. Then t h e r e e x i s t s an element s E N
such t h a t as i s an idempotent hyper-atom i n N. Proof: Follows f r o m lemma 3.2 and lemma 3.3.
D. Ramakotaiah and V. Sambasivarao
242
The proofs o f the f o l l o w i n g lemmas 3.5 and 3.6 a r e easy and w i l l be omitted. Lemma 3.5:
The set o f a l l idempotent hyper-atoms i n N i s an orthogonal sub-
N.
set o f
Lemma 3.6:
I be the s e t o f a l l idempotent hyper-atoms i n N. Then n E N) i s a subnear-field o f N and
L e t IeiIi
f o r every i E I, Ni = b e i
Ni n N. = (0) i f i
J
D e f i n i t i o n 3.7:
9
I
j.
N i s s a i d t o be hgpe,t-uLotnic provided f o r each non-zero
element n E N, t h e r e e x i s t s a hyper-atom 3. E N such t h a t a 5 n. Lemma 3.8:
L e t N be a hyper-atomic near-ring and l e t IeiIiEI
be the set o f
a l l idempotent hyper-atoms i n N. Then f o r each non-zero element q E N, there e x i s t s an idempotent hyper-atom ek such t h a t qek the sup {nei
I i E I1 exists
* 0. Moreover,
f o r each n E
N,
and equals n.
Proof: Since N i s hyper-atomic, i f q a E N such t h a t a 5 q. Then a2 = aq
* 0.
* 0 , then
t h e r e e x i s t s a hyper-atom
By lemma 3.4,
there e x i s t s s E N such
*
t h a t as i s an idempotent hyper-atom. Now we w i l l show t h a t qas 0. Suppose qas = 0. Then aqas = 0. This i m p l i e s a2 as = 0 and hence as = 0, a contradiction. So qas
qek
* 0.
* 0.
Thus there e x i s t s an idempotent hyper-atom as = ek i n
Since (nei)*
= nein
N such t h a t
f o r a l l i E I,n i s an upper bound o f h e i
I
i E I}.
Then nnei = nuei f o r each i E I . Suppose u i s an upper bound o f h e i I i E I}. 2 I f n $ u, then n -nu 0. By the above argument, t h e r e e x i s t s an idempotent hyper-atom ei i n N such t h a t (n2-nulei 0 , which i s a c o n t r a d i c t i o n . Hence
*
sup b e i
I i E I1
Lemma 3.9:
= n.
L e t N be hyper-atomic and l e t Ieili
I be the s e t o f a l l idemN onto a Ni o f n e a r - f i e l d s Ni.
potent hyper-atoms i n N. Then a : n + (neiIi subnear-ring o f the d i r e c t product
n
€ I i s an isomorphism o f
iEI Proof: C l e a r l y a i s a homomorphism. Suppose a ( n ) = 0. Then (nei)i This implies sup h e i [ i E I ) = 0. By lemma 3.8,
I =
n = 0. Hence a i s a mono-
morph ism. We now s t a t e and prove the main r e s u l t o f t h i s section.
0.
243
Reduced near-rings Theorem 3.10: N i s isomorphic t o a d i r e c t p r o d u c t o f n e a r - f i e l d s i f and o n l y i f N i s hyper-atomic and o r t h o g o n a l l y complete. Fi, where each Fi i s a n e a r - f i e l d . L e t o Proof: Suppose N = T'I iE1 I and t h e r e e x i s t s j E I such t h a t n . 0. Now Then n = (ni)i x = (o,o
J
,...,o,n.,o,....o) J
+nE
N.
*
i s a hyper-atom i n N and x 5 n. L e t X by any
orthogonal subset o f N. L e t ki:N
+
Fi be t h e c a n o n i c a l p r o j e c t i o n . W r i t e
*
o f o r some x E X and ai = o o t h e r a = {aili wise. Since X i s orthogonal, i t f o l l o w s t h a t a = sup X. Conversely, suppose
I , where ai
= ki(x)
i f ki(x)
N i s hyper-atomic and o r t h o g o n a l l y complete. By lemma 3.9, i t i s enough i f we Ni o f lemma 3.9 i s onto. L e t (niei) E n Ni. + n i€1 i€1 By lemma 3.5, t h e s e t hiei I i E I1 i s a n orthogonal subset o f N. Since N i s o r t h o g o n a l l y complete, sup h i e i I i E I 1 = h e x i s t s . Now by c o r o l l a r y 2.2 and show t h a t t h e mapping a:N
lemma 3.5, f o r each j E I , we have he = e.h = e . sup {n.e. J J J 1 1 Now a ( h ) = (heiIi So a i s onto. I = (nieili
I
i E I1 = n.e J j*
REFERENCES
[I] Abian, A .
, Direct
product decomposition o f commutative semisimple r i n g s ,
Proc. Amer. Math. SOC. 24 (1970).
€21 P i l z , G., Near-rings, North-Holland, Amsterdam 1977. [31 Raphael, R. and Stephenson, W . , Orthogonally complete r i n g s , Can. Math. B u l l , Vol. 20, No. 3 (1977).
This Page Intentionally Left Blank
Near-ringsand Near-fields, C. Betsch (editor) 0Elsevier Science PubMers B.V. (North-HOW), 1987
245
NON-COMMUTATIVE SPACES AND NEAR-RINGS INCLUDING PBIRD's PLANAR NEAR-RINGS AND NON-COMMUTATIVE GEOMETRY Maic SASSO-SANT
INTRODUCTIOY T h i s work shows some r e l a t i o n s between n e a r - r i n g s , non-commutet i v e g e o n e t r x and c o m b i n a t o r i c s (PBIBD's and RIBD's). Many i n t e r e s t i n g problems f o r f r u i t f u l f u t u r e r e s e a r c h i n t h i s a r e s may be found i n [ 2 ] o r [4] We s t a r t w i t h some p r e r e q u i s i t e s i n Eon-commutative geometry.
.
I NOX-COMMUTATIVE GEOMETRY 1.1.DEFINITION: A space i s a s t r u c t u r e ( X , 0 ) where X i s a ncnempty s e t ( e l e m e r t s a r e c a l l e d p o i n t s ) and b i s a s i l t s e t o f t h e power s e t X. 0 7 :
1.2.DEFINITION: An L-space i s a s t r u c t u r e (X,) non-empty s e t X and a nnspping:
c o c s i s t i n g of
3
possessing t h e following p r o p e r t i e s : (LO)
x u x
=
I4
t X U Y (12) z ~(x-y)\(x)
(L1)
X,Y
II)
f o r all x g X f o r a l l x,y 4 X x u y = x-z
(incidence condition) for a l l x , y , z t X , x 9 y (exchang;e s o n d i t i c n )
The ? o i n t x i s c a l l e d b a s e p o i n t (Aufpunkt) of xuy d e n o t e d by x f x u y . A p o i n t s e t o f t;he f o m xu;.' with x y is called l i n e (Linie).
+
-
I.~.DEFINITION: An LG-space i s a n L-space w i t h (L3)
xuy
= y u x =
xuc 4
x u 2
=
zux
(X,Y,Z €XI
M. Sasso-Sant
246
L i n e s w i t h (L3) a r e c a l l e d e t r a i g h t l i n e e (Geraden). 1.4.DEPINITIOX: L e t (X,,) An e q u i v a l e n c e r e l a t i o n
11
s
be an L-space and := { x U y / x , y c X ) on is called parallelism i f
3
1
xux YUY f o r a l l x,yd X I f L e 9 and x S X t h e n t h e r e e x i s t s e x a c t l y one L ' such
(PO)
(PI)
t h a t x 6 L ' and 1,' 11 L ( E u c l i d e a n p a r e l l e l i s m c o n d i t i o n ) (P2) I f xuy xIuy' then y u x uy",~' (x,y,x',y'e X) (symmetry c o n d i t i o n )
I
1.5.DEPINITIOIT:
i s a t r i p l e (X,,,
An LF--pace
1)
with t h e
p1'3 p *?rt i a s
(X,-)
(a)
1
(b1
i s an b s p a c e is 3 paralle~sm
1.G.DEFZNITION: An LP-space i s c a l l e d s e l f a d j o i n t ( s e l b s t a d j u n gie7.t) i f (P2) x u y yux f o r a l l x,y E X ( s t r o n g synrnetry c o n d i t i o n ) A 3 p q u i - a l s n c . ' c l a s s xu? is c a l l s d direction.
u
1
I.?.DEFIiITTION: subspace, i f :
(U1) (u2)
X,
r,y
eu 3 X U Y SU ~ 1 ,3L o u
XLU,
{XI
A s u b s e t Us X o f a n LP-space
(X
(X,,,
1)
is c:alied
UL)SU
( x a x ) , 0 3ce c a l l e d t r i v i a l subspRce3 o f ( X , U ,
1).
An LP-spsce i s c a l l e d p-irnitive, i f c n l y t h e t r i v i a l subspaces e x i s t ; o t h e r v i s e i t i s c a l l e d imprirnitive.
I.E.DEFIRITION: A s k e w a f f i n e space i s a n LP-space where t h e l'amaschke-condition h c l c l s : ?am: 1: (x,y,z} +, x u y I I x ' u y ' ( x ' , y ' € X) t h e n t h e r e s x i s t z ' C X such t h a t xuz U x l u z ' an11 yue y1-z'
(x,,l
)
The f o l l o w i n g - l o s u r e c o n d i t i o n i s c a l l e d p a r a l l e l o F r a m n c l s s u r e condition :
-
PGM: If x , y , z a r e p a i r w i s e d i f f e r e n t p o i n t s o f a n LP-spacs and (X,U, ther. t h e r e ? x i s t s w r X such t h a t x ~ y zl u w
xL1z
I
a)
jruw
.
247
Non-commutative spaces and near-rings
An LP-space i s c e l l e d Desarguesien i f t h e f o l l o w i n g D e s a r g u e s i e n c l o s u r e cor,dition holds:
&
a r e p a i r w i s e d i f f e r e n t p o i n t s o f t h e s e t X end
I f u,x,y,z
i f u-x = u-x' (XI( X) then t h e r e e x i s t d i f f e r e n t p o i n t s y ' , z ' ( X such t h a t u 4 y = u-y' , u u z = I i u z ' , xUy)x'-y'
xuz
1 x ~ u z ' , yuz[
y ~ u z l .
A s u b s e t FCX o f an L-space
I.9.DEFINITION: if
(X,)
is called
flat,
(FI) (F2)
x , y ( L e t , x,g6F, x x,yCF, x
+
jr;
+
y II) L G F x C L € g , y 4 M r 9 ; L,M$F
+ lLnM[
6 I
An i m p r i m i t i v e FkewJffine space i s c e l l e d s e m i e f f i n e , i f e v p r y subspace i s f l a t .
I1 GEOMETRIC AND PLANAXR NEAR-RINGS 11.1 .DEFINITION: A n e a r - r i n g ( N , + , - ) w i t h t h e d i s t r i b u t i v e l a w (s+b)c = a c + bc i s c a l l e d g e o m e t r i c 5 6 i f : (")N i s zerosym-
L HI
m e t r i c , i . e . O R =ao = o f o r 311 EI E N. ( 2 ) F o r all a € N t h e r e e x i s t s 1 E N such t h a t a'! = a . ( j ) I f o = a = bc t h e n aN = bN ( a , b , ( . € N ) .
.
F o r t h e d e f i c i t i o r , o f 3 p l a n a r n e s r - r i n g , s ? e [7] The f o l l o w i c g theorem:; 3 r e t r u e . Some o f them were proved independectly by R.Scapellato. THEOREM 1 : I F N i s t h e n Na THEOREM 2 : ~f m i s Na i s a THEOREK 3 : I f N i s then N l I THEOREM 4 : (And&):
3 geometric n e a r - r i n g 2 n d Pa + ( 0 ) ( a E N ) , i s a l s o a geometric near-ring. a plans? n e o r - r i n g and 2 3 (s E N ) , t h e f i P planar near-rlng, too. 6 g e o m e t r i c neRr-ring, I 8 p r o p e r i d e a l of N , i s a geometric near-ring. Every f i n i t e p l a n a r n e a r - r i n g i s g e o n e t r i c .
IN./-[
THEOREM 5: Every i n t e g r a l p l a n a r n e a r - r i n g
i s geoaetric:.
The l a s t theorem g l v e s 3 g c s n e t r i c i n t e r p r e t a t i o n of thi. B a t s i n t h e s t r u c t u r e t h e o r e n of 3 p l a n a r n e a r - r i n g :
-
x u y = By-x
-
+ x , where By-x
= By-x~[o)
M. Sasso-Sant
248
I11 PBIBD’s, PLANAR NEAR-RINGS AND NON-COMMUTATIVE GEOMEZRY
.
F o r t h e d e f i n i t i o n o f a t a c t i c a l c o n f i g u r a t i o n s e e [3’J III.1.DEFINITION: S t a r t with 3 t a c t i c s 1 c o n f i g u r a t i o n ( v , 8 , E ). Let 7 = { A c V / I A l = 2 ) and l e t eC = {A,,, A2, Am} bs a p a r t i t i o n o f T. Then & i s a n s s s o c i a t i o n schame on V if, g i v e n ( x , y j € A h , the number z C V such t h a t { x , z ) € A i and { ; J , z ) ~ A depends ocly 31: h , i 3r,d j and not on x and/or y. That i s , t h e r e h h i s a number p i j such t h a t f o r f x , g ) € A h t h e r e a r e e x a c t l y p L.l d i s t i n c t e l e m e n t s z € V , such t h a t { x , z J c A i and { y , x ) eAj. A s s o c i g t i o n schemes w i t h m = 1 o r m = v(v-1)/2, v = I V l 3re declared ”uninteresting”.
......,
III.2.DEFINITION: Suppose ( V , @ , E , K ) i s a t a c t i c a l c o n f i g u r a t i o n w i t h a s s o c i a t i o n schem?. T h i s s t r u c t u r e i s a PBIBD ( p a r t i n l l y b a l a n c e d i n c o m p l e t e block d e s i g n ) , i f : a . ) f a r eech A i € & , t h e r e e x i s t s a number Ai such t h a t { x , y } € Ai i m p l i e s x , y belong Yo e x a c t l y blocks of t h e r e e x i s t s a numbe? ni, s u c h t k z f f o r b.) for each Aie e s c h x E V , t h e r e a r e e x s - t l y ni d i s t i n c t e l e m e n t s y E V , such t h e t { x , y ) 6 Ai. SO a PSIBD h3.5 p a r a m e t e r s p hi j , l j ,ni i n a d d i t i o n t o th o s e 3f a ts c t i c 3 1 con fip r a t i on
a ,
L
x.
.
Now t h e q u e s t i o n i s : “Which p l a n a r n e a r - r i n g s y i z l C PBIED’s ? ” *
a
III.3.LEN;uA 1 : Let N be a f i n i t e p l s n a r n e a r - r i n g , a € N ( N = N+u f i x e d , b Q N. Then [art + S b a N ) g i v e s a t a c t i c a l configuration
1
ITI.4.LEMMA 2 : Let N be 3 f i n i . t e p l a n a r n e a r - r i n g . Then { (&aN) + b b 6 N } f o r A C N: g i v e s a t a c t i c a l c o n f i g u r a t i o n , if ~t l e a s t one a # C , ( N ) , where C1(N):= ( x 6 N / xN = -xN + x i
1
III.5.DEFINITION:
.
A n e a r - r i n g N i s c a l l e d symmetric, i f f XN = -xN for a l l x E N
1II.G.TEEOREM: Suppose (N, +, . ) i s a f i n i t e , symmetric, p l a n a r n e a r - r i n g and @ i s t h e s e t o f a l l b l o c k s [ aN + b a 6 N: b 6 N ) Then t h e f o l l o w i n g i s t r u e :
I
I.>( N , m )
i s a PBIBD
+
b ) i s a PBIBD
2.)
(N, aN
3.)
(N, ( g A a N )
a
C1(N))
+
(a0U* fixed)
b ) f o r an A c N ” i s a PBIBD ( a t l e a s t one
.
249
Non-commutative spaces and neawings
REMARK: If C,(P!)
i n 6.1.),
= to)
t h e n t h e PBTBD is a BTBD and
( N , (8) i s a Desarguesian, s k c w a f f i n e space w i t h p a r a l l e l c l o s u r e c o n d i t i o n (PGM). If C,,(N) f { o ) t h e n ( N , @ ) i s a Desarguesian, s e r n i a f f i n e s p a c e w i t h PGM. I f cne c o r l s i d e r s 6.2.) t h e n t h e b l o c k s R N + b a r e l i n e s o f one d i r e c t i 3 n o f t h i s space an2 i n 6 . 3 . ) t h e b l o c k s a r e t h e unior? o f l i l i e s wi%h d i f f e r e n t d i r e c t i o n s , b u t f i x e d b a s e p o i n t . F o r p a r a m e t e r s ni, we have :
Ai,
a.)
ni = l x y y l
b.)
1.)
2.)
p 9 j o i a PBIBD o v e r a p l a n a r n e a r - r i n g
-
-
1 = k
ph 3 1 + p i i ;j pii = 0 , pg
63"
=
1
,......,m
7 and p1 z I hj = 'I, where k = 3
4.)
pg = k - 2, i f k ag pig = 0, i g
5.)
p6. =
6.:
p Jh h = k - 3 ,
5.)
for a11 i
1
3
+
JJ
0,
j
+
g,
pEj
I
j + g + h ,
(semiaffine space) j f h ,
k a 3
The i n d e x g d e n o t e s 3 s t r a i g h l ; l i n e d i r e c t i o n , t h a t incans i f 5 = x-y , then x u y = yux.
EXAMPLES: 1 . ) Consider
(z,-, +)
with t h e u s u a l modulo a d d i t i o n a n 3 t h e f i x e d p o i n t f r e e automorphism group = {id,'f) with ' f ( x ) = -x. T h i s i s a v e r y i n t e r e s t i n g example, because t h e b l o c k s a r e c i r c l e s o f a MBbius p l a n e , t h e y form a B I B D , t h e y a r e Lines o f a Desarguesian s k e w a f f i n e space w i t h PGM, and a l l t h e l i n e s form a PEIBD. 2.) We c c n s i d e r +> a g a i n with t h e u s u a l a c d u l c 3 d d i t i 3 r snd t h e f i x e d p o i n t f r e e autonorphlsm group = [ i d , ' f ) , with y(x)= - x . The b l o c k s form 6 PBIED snd t h e y a r e l i n e s o f a Desargues'Lan s a i a f f i n e space w i t h PGM. See a l s o [2]
(z9,
4
9
.
S t a r t i n q with Leinma 1 one may a s k t h e q u e s t i o n , which g l a n e r n e a r - r i n g s l-ead t o symmstric b l o c k d e s i g n s ? The snswer i s s t i l l unknown. The r e s d e r may f i n d examples of s p e c i s l symmetric b l o c k d e s i g n s , Hsdamard d e s i g n s and p r o j e c t i v e p l a n e s i n L4J
.
M. Sosso-Sant
250
IV NON-COMMUTATIVE GM)METRY OVER
(r,M(r))
We b e g i n w i t h a n s d d i t i v e , n o t n e c e s s a r i l y a b e l i a n group ( L e t M( ) d e n o t e a s u b s e t o f t h e s e t o f a l l f u n c t i o n s from
r to r.
--
r , +).
We make t h e f o l l o w i n g d e f i n i t i o n : u :
rxr (x,Y)
yr) xuY
:=
~ ( r ) ( ~ - x+ )x,
and we a s k t h e q u e s t i o n , under which’ c o n d i t i o n s t h e s p a c e ( , M( )), , becomes a non-commutativa space? I t ’ s a g e n e r a l i z a t i o n of t h e c o n c e p t of Theobald r53 [6] I f one c o n s i d e z s t h e s p e c i a l f u n c t i o n s M( ) : = {yseW( ) / y a ( n ) = m } where ( , +, i s a n e a r - r i n g and n , a € , t h e n we g e t t h e d e f i n i t i o n : xuy = ( y - x ) r t x. Among m s n y o t h e r r e s u l t s t h e f o l l o w i n g i s i m p o r t a n t :
r
r
r
r
a >
r
r
.
IV.’!.fHEOREM: I f (N, i s 3 f i n i t e , zero-symmptric, quasii n t e g r a l n e a r - r i n g , t h e n (E, + , * ) i s a g e o m e t r i c n e a r - r i n g , where q u a s i - i n t e g r a l m e a n s a - b = o =? a = o v b = o . + , a >
T h i s s e c t i o n m o t i v a t e s t h e f o l l o w i n g s e c t i o r , V.
V STElONGLY GEOMETRIC PJEAR-RINGS FROM AUTOMORPHISM GROIJPS V.1 . D E F I N I , T I O N : A g e o m e t r i c n e a r - r i n g N i s callecl. s t r o n p l g geom e t r i c , if N c o n t a i n s a r i g h t u n i t 1 ( i . s . n . 1 = n f o r a l l n C N ) .
9
V.2.DEFINITION: C o n s i d e r a n automorphism group on a group , +). A n o r b i t ( y > , Y E % i s c a l l e d p r i o c i p a l , i f l + ( y o t h e r w i s e i t i s c o l l e 3 secondar:r o r b i t . We denote wLth H t h e f a m i l y o f a l l p r i n c i p a l o r b i t s and with N t h e f a a i l y o f a l l sec o n d a q o k b t t s . From e v e r y p r i c c i p a l o r b i t we choose r e p r e s e n t a t i v e s . N o w we d e f j n e a p r o d u c t on +) i n t h e f o l l o w i n g way:
(r
4
r
(r,
]=l+l,
25 1
No n-commu ta live spaces and near-rings
(r ,
With t h e l e s t d e f i n i t i o n +, ) becomes 8 s t r o n g l y g e o m e t r i c n e a r - r i n g , with d i s t r i b u t i v e l a w ( a t b ) * c = a'c + b-c.
V.3.THECREM:
V.4.THEOREM:
?he n e a r - r i n g
(r , +,
0 )
i s quasi-integml.
The p r o o f s o f a l l theorems i n t h i s s e c t i o n may be found i n (4)
.
REFERENCES
"73 [2]
And& J.
B e t s c h G.-Clay
Z.ur Geometrie d e r Frobeniusgruppen. Mathem. Z e i t s c h r i f t 154 (1977) S.159
-
15R
Block d e s i g z s from Frobefiius g r o u 2 s J.R. and planar iear-rings.
Proc. Conf. f i n i t e g r o u p s (Psr.1: C i t y , Utzh) Acad. Press (1976), pp. 475 - 502
[3]
P i l z G.
Near-Rings The Theory and i t s A p p l i c a t i o n North Holland (1983)
[4]
Sasso-Sant M.
Nichtkoqmutstive R&me und Fastringi:. Diplomarbeit U n i v r r s i t a t Saarbriickpn (1986)
[5] Theobsld E.
Nich$kommutstive Geometrien iiber F a s t r i n g e n . D i p l o a a r b e i t U n i v e r s i t a t S a a r b r u c k e n (198'1 )
[6] Theobald E.
IJear-.-ings and non-commutative geometry. I n : Proceed. Conf. ' o n N a a r - r i n g s and N e a r - f i e l d ? , ed. by G. F e r r e r o and C. F e r r r r o C o t t i , S a n Senedetko d e l Trorito, pp. ?',,I - 218 (1981).
Author's address Maic Sasso-Sant I V . Gartenreihe I 8 663 S a a r l o u i s West
-
Germeny
This Page Intentionally Left Blank
Near-rings and Near-fields, G.Betsch (editor) 0 Elsevier Science Publishers B.V. (North-Holland),1987
253
ON GEOMETRIC NEAR-RINGS
Raffaele SCAPELLATO Dipartimento d i Matematica, Universith d i P a r m a , Parma, I t a l y ”
In t h i s p a p e r we g i v e a short account of o u r r e s u l t s on geometric near-rings.
A zero-symmetric
near-ring
is s a i d geometric if i t s a t i s f i e s the fol-
lowing conditions:
(G1) for e v e r y a € N t h e r e i s a n e E N such t h a t e a = a ; ( G 2 ) for every a , b e N, if ab#O then Nab=Nb.
An L-space
( s e e ( 1 1 ) is a p a i r ( S , U ) , where S is a set a n d U is a
map from S x S to P ( S ) , such t h a t , i f U ( x , y ) i s denoted by xUy, the following conditions hold for every x , y , z
t
S:
(LO) x U x = ( x ) ;
(L1) x , y
(L2) i f z
E
t
xuy; xUy \ i x ) t h e n xUz=xUy.
Theobald [ 6 ] proved t h a t , i f N is a near-ring
a n d U i s defined b y
xUy=N(y-x)+x, then the p a i r (N,U) is a L-space iff N is geometric. This is a link between near-rings
a n d non-commutative geometry i n the sense
of Andrk ( s e e [ 11 ) . For geometric a p p l i c a t i o n s of n e a r - r i n g s see a l s o ,
e. g.,
[21, l31,[41.
The following r e s u l t s (except Th.3) have been proved i n [ 51.
L E M M A 1 . I f I i s a proper i d e a l of a geometric near-ring
N , then
1N=O and N/I is geometric. THEOREM 2. A f i n i t e non-constant near-ring
is geometric i f f i t is
strongly monogenic.
I n the next statement we w i l l denote by F the set of a l l non-zero maps x * a x .
THEOREM 3. A near-ring
N is both geometric a n d s t r o n g l y monogenic
iff F i s a g r o u p . *Work supported by t h e i t a l i a n M P I .
R. Scapellato
254
PROOF. Suppose t h a t N is both geometric a n d s t r o n g l y rnonogenic. Let a
E
N with aN=N, let e
E
N with e a = a . If b
E
N we get b=ac for some c , so
eb=eac=ac=b, a n d e i s a left i d e n t i t y . By (G2) we h a v e e t Ne=Nae, so e=dae for some d . Since d a i s c l e a r l y a left i d e n t i t y , i t i s now e a s y to conclude t h a t F i s a group. Conversely, i f F is a g r o u p , let a
6
N such t h a t a N # O .
There is b
t
N
such t h a t a b is a left i d e n t i t y : now N=abN c aN a n d N is s t r o n g l y rnonogenic. Finally, let a , b
t
N a n d ab#O. For c
E
N , i f cb=O it is c l e a r t h a t
c b E Nab; otherwise the map x c c x belongs to F a n d i t i s e q u a l to the map x - d a x
for a s u i t a b l e d . We get c b = d a b , t h u s c b t Nab, a n d ( G 2 )
follows. This completes t h e proof.
REFERENCES
[ 11 Andre, J . , Nichtkomrnutative Geornetrie und verallgerneinerte HughesEbenen. Math. Z . 177 (1981), 449-462. 121 Ferrero, G . , Sterns p l a n a r i e BIB-disegni, Riv. Mat. Univ. Parrna, ( 2 ) 11 (1970), 79-96. 131 Ney, H . , P l a n a r n e a r - r i n g s a n d t h e i r r e l a t i o n s t o some non-commutative s p a c e s , i n Proc. Conf. on Near-rings a n d Near-fields, Harris o n s b u r g , 1983. [ 41 Sasso-Sant, M., Nichtkommutative Raume und F a s t r i n g e , to a p p e a r . [ 51 Scapellato, R . , On geometric n e a r - r i n g s , Boll. Un. Mat. I t . , ( 6 ) 2-A ( 1983) , 389-393. [ 61 Theobald, E., Near-rings a n d non-commutative geometry, i n Proc. Conf. on Near-rings a n d Near-fields, 5. Benedetto del Tronto, 1981.
Near-rings and Near-fslds, G.Betsch (editor) 0 Elsevier Science PublishersB.V. (North-Holland), 1987
A
TERNARY
255
INTERPRETATION @F THE
TNFRA-NEAR
RINGS
Mlrela Stefhescu 1.
INTRODUCTION
In studying some g e n e r a l i z a t i o n s of r i n g s a s infra-near rings (see S t e f h e s c u
[lll ),weak
r i n g s (see Climescu [3,41
,
Cupona [5J ) , p r e r i n g s (see Janln [6,7] ),we have noted t h a t cert a i n facts can be b e t t e r explained by means of a ternary i n t e r p r e t a t i o n o f t h e a d d i t i v e composition law. T h i s i n t e r p r e t a t i o n might be of considerable use i n studying a f f i n e infra-near r i n g s ( Stefgnescu LlO3 ) a s well a s i n studying i d e a l s o f those alge-
b r a i c systems. Given t h e development of t h e theory of t e r n a r y groups and t e r n a r y rings,such an i n t e r p r e t a t i o n using a t e r n a r y operation Instead o f t h e binary a d d i t i o n may be i n t e r e s t i n g I n itself. F i r s t we r e c a l l some d e f i n i t i o n s and p r o p e r t i e s o f generali-
z a t i o n s of r i n g s involved I n our considerations. (1.1) Definition. A l e f t h f r a - n e a r r i n g is a triple (I,+,.),
where I i s a nonempty s e t
, + and
(addition and m u l t l p l i c a t i o n )
.
a r e binary operations on I
, such t h a t
t h e following condi-
tions are fulfilled: (1) (I,+) ie a group (generally,noncommutative) i
(il) (I,.)
I s a semigroup;
( i 1 i ) t h e m u l t i p l i c a t i o n is l e f t i n f r a - d i s t r i b u t i v e w i t h r e s p e c t t o t h e a d d i t i o n , 1.e.
x
. (y + z ) = x.y
- x.0
+
X.Z
, for
a l l x,y,z t I .
The concent o f r i g h t infra-near ring is analogous. If x.0 = 0 f o r a l l x ~ 1 , t h e nwe o b t a i n a l e f t near-ring.
If
(I,+,.) i s a l e f t and a r i g h t infra-near ring w i t h a commutative
M. Stefanescu
256
addltion,we obtain an equivalent d e f i n i t i o n o f a weak rhg, t h e general case. We mention t h a t t h e weak rings were first Introduced by A1.Cllmescu 131 in 1961 i n a p a r t i c u l a r case,when x.0
= 0.x = x f o r a l l
X E
I
In 1964,Climescu [41 has noted a ge-
n e r a l possibil.%ltyQf o b t a i n i n g weak r i n g s s t a r t i n g from a g i v e n ring. The p a r t i c u l a r case was rediscovered by G.Cupona
(53 I n
1971, who c a l l e d it "quasi-ring".
Let ( I , + , . ) be a l e f t infra-near ring. One can e a s i l y check that f o r each
- x.0
, is
X G I t h e mapping f x r
I--+I,given by fx(y)
an endomorphism of the a d d i t i v e group (I,+)
I=
.
-
x.y
This a l -
most obvious remark may help us t o study t h e l e f t infra-distribut i v e a s s o c i a t i v e multiplications over an a d d i t i v e group (see Stefgnescu (123 ) , t h e discussion being unexpectedly i n t e r e s t i n g .
One of t h e first examples o f r i g h t near-rings was the s e t of a l l a f f i n e transformations over a vector space or over an Abelian a d d i t i v e group endowed w i t h the pointwise addition and the mapping composition. This is a l s o a l e f t infra-near ring and a
few p r o p e r t i e s of t h i s s e t come from i t s infra-near r i n g struc-
ture
. It is s u r p r i s i n g how much algebra can be obtained only by
considering t h i s weaker s t r u c t u r e (see ClO'J
,a
?aper i n which we
have studied and generalized the infra-near ring of a f f l n e trans-
formations from t h i s point o f view ). If a l e f t infra-near ring s a t i s f i e s the condition I 0.x = 0
, for
all xcI,
then we c a l l it a l e f t C-infra-near 0.x = x
, for
all x
rinp;
, while
in the case
I
61,
we c a l l it a l e f t Z-infra-near
ring.
we give two examples o f f i n i t e l e f t C-infra-near rings: .'1
The c y c l i c group (&,,+),with
1.x = 2.x = x , f o r a l l x e Z 3
.
t h e multiplication
I 0.x
= 0,
A ternary interpretation of the infra-near rings
.'2
The symmetric group S3
3:
where 3a = 2x = 0 and x+a = b
257
{O,a,b = 2a,x,y = x+a,z
+
, with
x
a
x+b],
t h e multiplicatlon given
by one of the following t h re e p o e s l b i l l t i e s r
(I) 0.8 = 0,a.s =
x.8
= y.s =
z.8
= a
, b.8
=
(11) 0.8 = 0,a.s = x . 6 = y.6 = a, b.s = z . s =
(111) 0.8 = 0,a.s = x . 8 = a , b.8
P
y.6
P
z.8 =
B
, for a l l
BC
%'
8
,for a l l s e S 3 1
8
,for a l l B E + .
Many o t h e r eaamples of l e f t infra-near rings and a general s t u d y of such algebraic systems can be found in [llJ
mark that
, If
. Let u s re-
(I,+) is a group,then (I,+,+)I s a l e f t and a
right
lnfra-near ring. I n a l e f t lnfra-near r i n g we do not obtain 0.0 = 0. For examp l e , i f we consider I = 25 and the binary operations I x
+ y = (x1+91,72+92,X?)+Y3
,XI++Yl+,x5+x2 .Y3+Xl+.Yl+Y51#
x.y = ( X I .Y1,'2*92,a~ ,a&, w i t h a3,a4,a5 fi x e d elements in Z,then we obtain a l e f t infra-near ring for which
0.0 = (0,0,a3,a4,a5),whlle 0 = (0,0,0,0,0).
The r e a l difference between t h e general case (see L83
in
which th e multiplication is an "independenttt operation on an ad-
d i t i v e group,
- that means the multlplication
cond!tions w i t h respect t o t h e addition,
does not f u l f i l any
- and our case
( I n which
t h e multiplication I s l e f t l n fra -d i st r i b u t i v e over t h e addition) is perceptibly I n the theory of congruences,hence i n defining homomorphisms and Ideals of such st ru ct u res . (1.2) Definition. A fno-Slded Ideal of a l e f t infra-near ring (I,+,.)I s a nonempty subset J o f I s a t l s f y l n g the conditions: (I(J,+) ) is a normal subgroup of t h e group
- x.0 e J
(ii) x . j (ill ) ( j
+
x).y
, for
a l l x E I and j E J
- x.yeJ , f o r
(I,+)i
i
a l l x , y E I and
j C J
.
If J o n l y satisfies t h e conditions (Iand ) (ii) (reSpeCtiVdg, (Iand ) (ill )) , J I s c a l l e d a l e f t Ideal (respectlvely,a right
I d e al) of
I.
M. StefGnescu
258
If (I,+,.)is a l e f t near-ring,hence x.0 = 0 for a l l x e 1 , t h e n
we obtain t h e d e f i n i t i o n s o f the two-sided,left and r i g h t idealB of a l e f t near-ring.
The similitude is a m a s i n g , B u t i n f a c t the
s i t u a t i o n 1s more complicated
, since
even t h e two sided i d e a l s
of a l e f t infra-near ring a r e not closed under the multiplloa-
tion,hence generally they a r e not subinfra-near r i n g s o f I. For l e f t near-rings a s i m i l a r assertion is t r u e only f o r the r i g h t ideal8 (see P i l z L9l
, 1.28
and 1.33).
O f c o u r s e , i f J is a two-sided i d e a l of the l e f t Fnfra-near
ring I
, then
I/J = f x + J / x
E
I) i s a l e f t infra-near r i n g with
respect t o t h e usilal operations beween cosets.
(1.3) Definition. Let (I,+,.)and (It,+,.) be t w o l e f t infra-near r i n g s . A mapping y t 141’ i s c a l l e d a homomorphism of l e f t infra-near rings i f t h e following two conditions a r e f u l filled : (1)
‘P(x + Y) = \ p ( x ) + \ P ( Y )
,
.
fix). y(y) , f o r a l l x , y c I , k e r ‘4 = jx E I / yJ (x) = 0’5 , i s a two-sided
(ii) .p(x.y) =
The kernel of ‘f
i d e a l of 1,while t h e image o f p
, Im y
=
(p(x> / x E I
a subinfra-near ring of 1’. Moreover,the %-sided
f , is
i d e a l s of I
a r e exactly t h e kernels of the homomorphisms from I t o another l e f t infra-near ring
.
(1.4) Definition. Let I be a l e f t C-bfra-near be a group. If t h e r e e x i s t s a mapping
= m.x
, such
(I) m.(x
MxI4M
that t h e following axioms a r e f u l f i l l e d
+ y) = m.x
(ii) (m.x).y
= m.(xy)
- m.0
+ m.y
,p ( m , x )
(M,+,p) a right
, f o r a l l m h M and x,y 0 I 4
, I-group.
Obviously I is i t s e l f a r i g h t I-group and each right i d e a l of I i s a r i g h t I-subgroup
, while
=
I
, f o r a l l m cM and x,ye I ;
(111) 0.x = C , f o r a l l x C I
then we c a l l
,441
r i n g and (M,+)
a r i g h t I-subgroup of I i s
259
A ternary interpretation of the infra-near rings
n o t a r i g h t ideal of 1,even in th6 cam when it i e a normal oubgroup of (Il+).
Denote by D = i d G I / d.0 = 0 5 and W
wrI / w.0 =
o
w),
t h e s u b s e t of a l l l e f t d i s t r i b u t i v e elements of I and t h e s u b s e t of
a l l weakly l e f t d i s t r i b u t i v e elements of I , r e s p e c t i v e l y . I f I
i s a l e f t C-infra-near
x,yr:I
, hence
+
ring and (x
y).O
e
x.0
+
y.0 f o r
0 i s a r i g h t d i s t r i b u t i v e element of I
all
, then
D is
a l e f t subinfra-near r i n g of I which i s a normal eubgroup of (Il+)
.
and W i s a r i g h t I-subgroup of I
r e c t decomposition I = D + W x = (x
- x.0)
2. TERNARY
+
, where
x.0
x
;
Moreover w e have t h e semidi-
indeed,for any x E I ,we have
- x.OED
.
and x . O E W
GROWS
Let u s note t h a t t h e l e f t i n f r a - d i s t r i b u t i v i t y of t h e m u l t i p l i c a t i o n w i t h r e s p e c t t o t h e a d d i t i o n reminds one of a l e f t distri-
b u t i v i t y of the binary multiplication over a ternary composition defined by
8
( y , t , z ) := y
-t+z
(y,@,z) = y
Indeed,we have then for a l l x l y l z 6 1
.
for all x , y , t , z 6 1
for a l l y , t l z GI.
+
Moreover we have x.(y
.
= (x.y1x.Olx.~)
z and x.(y,O,z)
-t
+ z)
= xdy-x.t+x.z,
B u t such t r i p l e s were obtained by Dijrnte
and by Baer i n t h e t w e n t i e s i n connection w i t h some g e o m e t r i c a l f a c t s
( t h e Erlangen Program on groups) ; In
[11
,R.Baer had n i c e l y ex-
plained the geometrical and algebraic reasons to take a s Fnva-
riants f o r a group of transformations t h e expressions of t h e form x
.
-y+z
S t a r t i n g f r o m t h i s point
Certaine
was
guided t o consider a ternary operation on a group (G,+) defined by
(x,y,z)
I=
x
-y+z
f o r a l l x,y,z E G (we t r a n s l a t e i n t o an
a d d i t i v e notation t h e paper written by Certaine)
.
By considering
t h e a b s t r a c t properties of such an o p e r a t i o n , C e r t a h e has defined ternary groups.
M.Stefdnescu
260
(2.1) Definition. Let I! be a nonemptg eet endowed with a tar-
nary composition
,.) : T
(.?.
%
T x T 4 T and With a f i x e d element
O d T such that t h e following conditions a r e f u l f i l l e d : (1) (x,O,(y,Z,u)).= ((x,O,y),z,u)
, for a l l x t T , for a l l x e T
(2) (x,O,O) = x
(3)
(x,x,O) = 0
, for
a l l x,y,z,u e T ;
;
.
Such an a lg ebra i c system i s called a ternary group of first type
(Certaine 121
, Definition
3
By defining a binary operation on T
(4) x
+ y r = (x,O,y)
, for a l l
x,yeT
,
we obtain a group (T,+) w i t h 0 a s i t s n eu t ral element ( 123
Theorem 4 )
.
,
?%en t h e binary and t h e ternary operations a r e r e l a t e d by t h e
eq u ali ty :
(5)
(x,y,z) = x
- y + z , f o r a l l x,y,z
E T
,
then we say t h a t t h e ternary group i s reg u l ar
, the
In a binary group
n e u t ra l element is unique
statement does n o t hold f o r a ternary group satisfy (11,
( 2 ) and ( 3 ) i n D e f i n i t i o n ( 2 . 1 1 ,
, but
.
A s i mi l ar
if 0 and O1 both
then t h e cor-
responding binary groups by (4) a r e isomorphic a s it i s proved in
(21
( t h e isomorphism i s given by x ~ ( x , O , O 1 )
, for
xe!&
Moreover it i s possible t h a t d i s t i n c t ternary groups correspond t o the same binary group,aa it is shown
in c2’J f o r t h e
c y c li c group w i t h two elements. To i l l u s t r a t e this,we consider th e cy c li c group o f order 3 3a = 0
,
(T = {O,a,b) ,+) w i t h b = 2a and
. We f i n d three p o s s i b i l i t i e s t o define t ern ary group8
o f f i r s t type,namely
t
(I) the r egu l a r ternary gmup (x,y,z) = x X,Y,Z
-y+
, f o r my
6T 3
(11) the s p e c i a l products a r e (a , a , a ) = (O,b,b) = b
, (b,b,b)
=I
26 1
A ternary interpretation of the infra-near rings
P
(O,a,a)
a
L
, (b,a,a)
(a,b,b)
o
, the
0
other ternary pro-
ducts being r e g u l a r ;
(a,a,b) = (b,b,a) = 0
(111) the s p e c i a l products a r e
, (O,a,b)
(O,b,a) = (b,a,b) = a
= (a,b,a) = b
, the
,
other pro-
ducts being regular, ~e obtain anotkz concept of ternary group, which we call
of
seccpld
type,
type, while a ternary group of Certain's type is called of f & s t type.
(2.2) Definition. A couple ( T , ( . , * , . ) )
.,
set and (., .)
I
Tx Tx T
,where T i s a nonempty
, if
called a t e r n a r y group of second type
it s a t i s f i e s the
following axioms f o r a l l x , y , z E T and a fixed element O € f
=
( 6 ) ((X,Y,O),Z,O)
(7)
(O,X,O)
(8)
((O,O,x),x,o>
= x
(X,(YtZ,O>,O~
, then
I
,
, = 0
(2.3) Proposition. If ( T , ( . , . , . ) ) cond type
, is
i s a t e r n a r y operation on T
3T
(T,+)
, where
i s a t e r n a r y group o f set
i s t h e binary composition
defined by (9)
+
x
y
I=
(x,y,O)
, for
a l l x,yC:T
,
is a group.
Proof. The a s s o c i a t i v i t y
comes from (6)
, by
using the defi-
n i t i o n o f t h e binary operation. The other two axioms t r a n s l a t e d into the addition complete the definition of a group. We a l s o have the equalities:
,for
(7')
(x,O,O)
(8')
(x,(O,C?,x),C) = 0
= x
all x%T
, for
,
all xbT ,
(2.4) Definition. A t e r n a r y group of second type i s c a l l e d
regular
,
i f the ternary operation is connected w i t h t h e binary
composition (9) by the following equality!
(lo)
(x,y,z) = x
+y
-z
, for
all x , y , z E T
.
As in the case of t e r n a r y groups of first type, t h e r e
are
M. Stefanescu
262
many t e r n a r y groups of second t y p e a s s o c i a t e d t o t h e same group by t h e e q u a l i t y ( 9 ) . For example, for t h e c y c l i c group of o r d e r 2 , (T = { O , a ) , + ) , we may d e f i n e a t e r n a r y composition: ( O r O r O ) = = (a,O,a) = (a,a,O) = ( a , a , a ) = 0, (O,a,O) = (O,O,a) = ( a t O , O )
=
= ( O , a , a ) = a , which p r o v i d e s T w i t h t h e s t r u c t u r e of a t e r n a r y g r o u p of second t y p e w i t h (T,+) a s t h e a s s o c i a t e d group. T h i s t e r n a r y group is n o t r e g u l a r , s i n c e ( a , a , a ) = 0 # a = a + a a.
-
The following ternary compositions on T = \O,a)
,namely (7), (8)
only two of the three axioms I n Definition (2.2) for (tl),
(61,031 for (t,) and (6),(7) for (t,)
satisfg
I
o
{
(O,O,O) = (a,O,O> = (O,a,a) = (a,a,o) =
,
(O,o,a) = (a,O,a) = (O,a,O) = ( a , a , a ) = a ; (0,0,0)= ( O , O , a )
= (o,a,o) = (O,a,a) = 0
,
= (a,O,a> = (a,a,O) = ( a , a , a ) = a ;
lo ,o,o)
= (a,O,a) = (O,a,a) = 0
(t3) (O,O,a) = (a,O,O) = (o,a,o)
b
= (a,a,O) = ( a , a , a ) = a
.
These uodels prove t h e following
(2.5) Proposition. The system of axioms in Definition (2.2) is independent. (2.6) Definiti0n.A ternary group
(of first or of
(I,(.,.,.))
second type) together w i t h an a s s o c i a t i v e binary m u l t i p l i c a t i o n which i s l e f t d i s t r i b u t i v e w i t h respect t o the t e r n a r y composi-
t i o n , i . e . which s a t i s f i e s the e q u a l i t y (11) x.(y,z,w)
= (x.~,x.z,x.w)
, for
a l l x , y , z , w ~ I,
i s c a l l e d a l e f t t e r n a r y near-ring (of first o r of second type).
(2.7) Fropositlon.(l) If ( I , ( . , . , . ) , . ) is a l e f t t e r n a r y near-ring of first type
, then
(I,+,*), where + is defined by (4),
.
., ,
is a l e f t infra-near r i n g , (ii)If (I,(., .)
i s a repular
l e f t ternary near-ring of second type, t h e n (Il+, defined by (9)
, is
a l e f t infra-near ring.
,where + is
263
A ternary interpretation o f the infra-near rings
Proof. (i) Since (I,+) I s a group (Certalne 127, Theorem 4) , and (I,.) is a semlgroup by assumption, we have to check the left
i n f r a - d i s t r l b u t i v i t y of t h e m u l t i p l i c a t l m w i t h r e s p e c t t o t h e addition. We have
I
= ((x.y,0,0>,x.0,x.z) But x.0
+
(O,X.O,x.Z)
= (x.O,x.O,x.z) - - x.0 + X . Z y,zGI
,
+ z ) = x.(y,O,z) = (x.y,x.O,x.z)
x.(y
= (x.y,0,(0,x.0,x.z~)
= x.y + (O,X.O,X.Z).
= (x.O,O, (O,X.O,X.Z))
= x.(O,O,z) hence x.(y
=
+
=
= ((X.O,O,O) ,x.O,X.Z)
, therefore
X.Z
z ) = x.y
-
(O,x.O,x.z) f x.0 + xlz , f o r all x,
.
(ii)Here we obtain immediately
= (x.y,x.z,x.O)
= x.y
+
- x.0
X.Z
x.(y
t z)
, from
= x.(y,z,O)
=
t h e r e g u l a r i t y of t h e
t e r n a r y group. B u t taking y = 0 we obtain (12)
therefore
x.z
+
x.(y
x.C = x.0
+ z) = x.y
+
X.Z
- x.0
,for a l l x , z e I
+
X.Z
The o t h e r exioms i n the d e f i n i t i o n o f
,
.
, for
a l l x,y,z € 1
left
infra-near r i n g s
come from t h e hypotheses and t h e Proposition (2.3)
.
Ye note t h a t i n the g e n e r a l case t h i s c o m u t a t i v i t y g h e n
, as
by (12) does n o t h o l d f o r l e f t infra-near rings
one can
see from the exemple I = E5 i n Section 1.
T h i s way we obtain a s a p a r t i c u l a r case of t h e t e r n a r y near-
-rings the ttpreringsttdefined by Janin [6,73
.
Indeed, a p r e r h g
is a ternary near-ring of first type or of second type which satisfies for all x,y,zqI a commutativity condition of the ternary com-
position
, namely
(x,y,z) =(y,x,z)
(x,y,z) = (z,y,x)
- for
- f o r the
first type and
t h e sec-nd e T e ,which has a n idempotent
m u l t i p l i c a t i o n being l e f t and r i g h t d i s t r i b u t i v e w i t h r e s p e c t t o t h e ternary composition.
We t r a n s l a t e the d e f i n i t i o n of a two-sided i d e a l of a l e f t infra-near ring into ternary language. In a left ternary near-ring,
we say t h a t a nonempty subset J i s a n idea1,lf it t h e following conditionsr
satisfies
M.Stefdnescu
264
, ((x,O,a),x,O)e
J
(i) (a,b,O)(
J
for a l l a , b € J and ~ € 1 1
, for a l l a d J and x G I ; (ill)((x,O,a).y,x.y,O) € J , f o r all a E J and x , y e I (ii) (x.a,x.O,O)
E J
.
Let I be a l e f t ternary near-ring of first type i n which x.0 f 0 b u t 0.x = 0 f o r a l l x G 1 . In addition, we assume t h a t f o r any x,y( I
, (x,O,y).O
, hence
= (x.O,O,y.O)
i s r i g h t d i s t r i b u t i v e with respect t o the
t h e element 0
t r i p l e s o f t h e given
, a s defined above, which a r e semigroups under multiplj.catlon , a r e a l s o closed under t h e t e r n a r y operation. We note t h a t x.wEW , for any X E Iand w e W
form. Then t h e subsets D and W
In f a c t , t h e following proposition holdsi
Proposition. Let I be a l e f t ternary near-ring of f i r s t
(2.8)
type which s a t i s f i e s t h e conditions
= x
(I) 0.x
, for
(ii)(x,y,z).O
I
a l l XO ; I
, for
= (x.O,y.O,z.O)
a l l x,y,zeI ;
, for a l l X Q ; I (w.x,O,d.x) , f o r a l l d c D , w e W and
(iii) (O,x.O,x) = (x,x.O,O)
( i v ) (w,O,d).x =
/ deD,weW
Then : (i) I = (D,G,W) = i ( d , O , w ) x
f
xaI.
and foa? each
e I , t h e elements d t D and w c g W such that x = (d,O,w)
a r e uni-
quely determined; (ii)Using ternary expressions f o r two a r b i t r a r y elements
of I
,
the multiplication is given by t h e formula8
= (d.d',O,(w,O,d.w'))
(13) ( d , O , w ) . ( d ' , O , w ' > Proof. (i) For each -
x C I
belongs t o D and w = x.0 EW
, we
.
determine d = (x,x.O,O) which w i t h d and w j u s t
B u t x=(d,O,w),
determined above (one can verify this relation by straightforward calculations). w,wl t w
, we
have
If (d,O,w)
= (d',O,w')
, with
d , d l E D and
=
(((O,d',O),O,(d,O,wR)),O,(O,w,O))
= (((0,d',0~,0,(d',0,w~),0,~0,w,0~~ and after some skillful calcu-
l a t i o n s we obtain = (w',w,O)
(0,d:d) = ( w ' , w , O ) C W n D
= 0 , B u t then d1 = ( d ' , O , O )
, hence
(O,d',d)=
= (dl,O,(O,dl,d)) =
A ternary interpretation of the infra-near rings
= ((d',OiO),d',d)
(d',d',d)
p
(d',d',(O,O,d))
p
= (O,O,d) = d
t((d',d',O),O,d)
265 0
I n t h e same manner, we prove
that w' = w
(it) The formula (13) is obtained by using t h e condition (iv) the e q u a l i t y w.x = w f o r a l l x 6 1 and w t w .
t a k i n g i n t o account
It might be Fnteresting t o study ternary near-rings In an Independent context
. Here we
have used t h e t e r n a r y i n t e r p r e t a -
t i o n f o r showing t h e r e l a t i o n between infra-near rings and pre-
rings and f o r explaining t h e l e f t l n f r a - d l s t r i b u t i v i t y a s a ref l e c t i o n o f a l e f t d i s t r i b u t i v i t y of t h e m u l t i p l i c a t i o n with r e s p e c t t o a t e r n a r g composiflon. REFEFf ENCES
C11. Baer,R.
- Z u r EinfUhrung des Scharbegriffs, J.reine Math. ,160 (1929)
C23. Certaine,J.-
, 199-207.
The t e r n a r g operation (abc) = ab-lc of a
group, Bull.Amer.Math.Soc.,49
c3]. Climescu,Al.-
Anneaux f a i b l e s
(11) (1961)
ClFmescu,Al.
[4],
angew.
9
(1943)
, 869-877.
, Bul.Inst.Polit.Ia$i
,7
1-6
- A new c l a s s of weak rings , (Romanian) ,
ibidem, 1 0 ( 1 4 ) ( 1 9 6 4 ) , 1 - 4 .
[5]. Cupona,G.
- On
quasirings
Phys,ldaC$doine
, (Macedonian
, 20
, Eull.Soc.lyfath.
)
(1969) ,19-22 (1971) 4 MR
44 #I703
c63. J a n h , P . - Une g h 6 r a l l s a t i o n de l a notion d'anneau.Pr6-
anneaux.C.R.Acad.Scl.Paris,Sec.A
, 269
(1969)
,
62-64
[7].
Janln,P.
- Une g h 6 r a l i s a t i o n de l a notion Prealghbres. C.R.Acad.Sci,mris
(81
.
(1969)
, 120-122.
Murdoch,D.C. ,Ore,O.
d'algbbre.
, Sec.A , 269
- On generalized rings, Amer. J.Math. ,
M.StejZnescu
266
63 (1941) 973-78
0
[91. P i l e ,G.- Near-rings .The theory and its applications
North-Holland Mathematics Studies 23
, North-Holland
Pub1 .Comp. ,Amsterdam, 1977.
[lo].
Stefhescu,MFrela
- Infra-near
An.St.Univ.Al.I.Cuza
1111.
Stefhescu,Mirela
-A
r i n g s o f a f f i n e type,
I a s i , 2 4 (1978) ,5-14
generalization o f the concept of
n e a r r i n g rInfra-near r i n g s
45 [12].
- 56
,
ibidem, 25 ( 1 9 7 9 ) ,
9
StefZinescu ,Mirela
- Multiplications i n f r a - d i s t r i b u t i v e 6
s u r un groupe,Publ .hth.Debrecen, 255
.
- 262.
27 (198C3,
Near-rings and Near-fiilds,C. Bersch (editor) 0 Elsevier Science Publishen B.V. (North-Holland), 1987
267
ON TWO-SIDED IDEALS IN MATRIX NEAR-RINGS
ANDRIES PJ VAN DER WALT Department of Mathematics, University of Stellenbosch, 7600 Stellenbosch, Republic of South Africa
Matrix near-rings over arbitrary near-rings were introduced in [ 2 ] ,
where
some results about the correspondence between the two-sided ideals in the base near-ring
R
and those in the matrix near-ring
Mn(R)
were given. The pur-
pose of this paper is to give a f u l l e r account of this correspondence which turns out to be quite a bit more complex than in the ring case, although the overall picture is pleasantly similar. (R,+,-) be a right near-ring with identity 1.
Let
direct sum of
n
copies of
The elements of Rn
Rn
will denote the
(R,+), and similarly for subgroups of
(R,+).
are thought of as column vectors and written in transposed
form with pointed brackets, eg
.
The symbols
1.
1
and n ,
will
1
denote the jth coordinate injection and projection functions respectively, The elementary
-
n
X
r E R
s
R and
rs, all
s E R.
For typographical reasons f r .
[r;i,j] . The subnear-ring of
written
are the functions fr.: Rn -P Rn, I' i(r): R -P R is the left multiplica-
1
11
tion
n matrices over
fr. = liX(r)n.. Here
where
is sometimes
11
M(Rn)
generated by the
fr , 1.1
is the
near-ring of n X n matrices over R, denoted Mn(R). Now l e t I be a (two-sided) ideal o f R. There are two obvious ways in which one can let an ideal in Mn(R) correspond to I . The first is to define
I*
:= {A
E Mn(R) IAp E I n , vp E Rn}
as was done in [ 2 ] , where it was shown that the mapping I I+ I* is an injection from the set of all ideals of R into that of Mn(R) (see 4 . 3 of [2])
.
I+ where
The second is to define :=
Id({fa. ( a E I , 1
i,j
1.1
< n)),
Id(T) denotes the ideal generated by the set T .
for all
a E I , 14 i,j
PROPOSITION 1
7' 5 I*
PROPOSITION 2
The mapping
PROOF Suppose
1,
and
Since
fyj E I*
n, we immediately have the following for any ideal
7,
7
1 of R.
0
I+ is an injection. are ideals of
R
such that
1, # 1 2 ;
say
A.P.J. van der Walt
268
Il \ I,.
a E
E 1;.
Then l : f
Next, if J
J,
:=
{x E
E I;,
If l : f
is any ideai of
R-I
3
by Proposition 1.
< a , ~ ,..., O> f
=
r;.
(R) we define
for some A E
R ~ X = 71.~0
It was proved in [2] that J,
fI1 E 1;
then
..., O>
fIl <1,0,
But this is not the case because
J,
p
E R ~ 1, G j
is an ideal of
R.
nl.
The following is an important
result in this connection:
PROOF J iff
c_ (J,)* was shown in 4.7 of [2]. By 4.4 of the same paper fyl E J
a E J,.
This shows that
5 J.
(J,)'
0
In view of this result it is of interest to know whether the inclusion
I+ 5 I* of Proposition 1 can be proper. That this is indeed the case is shown in the next
EXAMPLE 4
Let
,
R = Z,[x]
the zerosymmetric polynomial near-ring in one
indeterminate over the integers (see Pilz [3] p 220 onwards). Consider the
7
ideal
:=
(2)[XI.
We proceed to show that in
2 end we prove that A = [x ;l,l] (f;l + f;2)
l2 for
Clearly, A= <2pq,0> E
A
$ I+
all
-
I+
M2(R)
2 [x ;l,l]
E R
2
I*,
and to this
2
-
,
f
[x ;1,2] E I* so A E I*.
\
I+.
TO see
that
it suffices to carry out the following steps:
(In the following discussion we let X,Y denote arbitrary matrices in MZ(R); B, B1
and
B2 will be elements of
arbitrary elements of (a)
R
2
;
I+;
, ,
are
a.,b.,c, are appropriate integers; and 1
1
1
Show that niX= a.p + b.q
+
2c.pq +
i E {1,21.)
. ..
(b) Show that n.(X( +) - X ) = (a. + 2 c . s )+~ (b. + 2 c . r ) ~ 2c.uv
1
1
+ ...
(c) Note that T.B= 2f.(p,q) for some polynomial f. in 1
(This is because (d)
I+
c -
2 variables.
I:.)
Prove that Tr,B= 2a.p + sb.q + 4cipq
+
.. .
for any
B E I+.
This is done recursively by showing the assertion (i) is true for
B = [2axt;i,jl;
(iii) if true for B,
is also true for
B1 + B2' BX (by (a) and (c));
(iv) if true for B,
is also true for
X(B
(ii) if true for B1,BZ, is also true for
For any ideal J
in
Mn(R)
+
Y)
-
XY
(by (b) and (c)1 .O
we therefore have the following diagram:
On two-sided ideals in matrix nearrings
Here
269
is the inclusion injection and, in general, not the identity function.
I
This relationship reminds one of the phenomenon of enclosing i d e a l s that one has in polynomial near-rings over commutative rings (see [ l ] or 1 3 1 ) .
However,
in that case one deals only with so-called full i d e a l s of the polynomial nearring, i.e. those ideals that are also ideals of the polynomial ring, whereas the situation we are concerned with here applies to any ideal
J of Mn(R).
The following proposition provides more information about the mappings and
(
PROPOSITION 5 (i) A
5B
(ii) If A
For ideals
+8
then
A+
(vi) PROOF
(i) This is clear.
(iv) (v)
A
and
implies A+ c - B+
(A n B)* = A* n B*. (A n B)+ c - A+ n B+ (A + 8 ) * 3 A* + 8* ( A + 8 ) + = A+ + 8+
(iii)
( )*
1'. 8
of
and A*
R we have:
5 B*.
8*. (In fact, this holds for arbitrary intersections.)
(ii) This was actually shown in the proof of Proposition 2. fiii) This is a standard result about noetherian quotients.
See [31, 1-44.
(iv) Obvious fran (I).
+
E An + 8" = ( A + 8)" for all p E Rn. (vi) It is clear that A+ + B+ 5 (A + 8)+. TO prove the reverse inclusion, a+b note that (A + 8)' is generated by the set if.. la E A, b E 8, 11 a+b - a 1 < i , j < n). Since f. , - fij + fyj E A+ + '8 we have our result. o (v) If A E A*
8*, then
Ap
11
Since O* = 0 we conclude from (iii) above and Propositim 3 .
COROLLARY 6 R
is subdirectly irreducible iff
Wn(R) is.
The next result gives some information on the behaviour of products: PROPOSITION 7
If BC
5A
for ideals A ,
8 and c of R, then 8+C* c_ A*.
270
A.PJ. van der Walt
c*,
PROOF We have to show that if B E 8+, C E p E Rn.
Since Cp E
all y E
en.
cn
An
then BCp E
for all
we need only show that By E b B is of the form f ,, , b E B ,
p E Rn
for all
This is easy enough when
17
An
for
and this
observation provides the starting point for a recursive proof based on the way b is generated by the set if. ,Ib E 8 , 1 G i,j G nl. 0 in which 8' 11
The ideal
COROLLARY 8
A
of
is nilpotent iff A+
R
The following result takes 4.11 of [ Z ]
P
PROPOSITION 9
P*.
So suppose
A+B*
$ P*.
P*
is prime, and let A,B
A8 $ P,
BY 7 this implies
treated similarly.
0
R
iff
P*
is a prime
M,(R).
P is prime (semiprime), then so is
In 4.11 of [ 2 ] it was proved that if
PROOF
Mn(R).
further:
is a prime (semiprime) ideal of
(semiprime) ideal of
is nilpotent in
so
$ P. P
Then
A+$*
is prime.
$ P*
by 5(ii), so
The semiprime case is
0
If we denote by rad S the prime radical of the near-ring
then we have
S,
the
5
COROLLARY 10 rad M n( R )
(rad R) * .
PROOF By S(iii) (rad R ) * =
n{P*[P
is prime in
R}
2
rad Mn(R).
0
It would be interesting to know whether the inclusion in this result can be proper, especially since it was proved in [41 that the J -radical of 2 is of the form (J2(R))*. We close with a result about
R.
generated near-ring
where
A*,
Mn(R)
is an ideal in a distributively
A
First we establish a purely group theoretic result
which is probably well known, and so we provide only a sketch of the proof: Let
LEMMA 1 1
that a = al
... +
a
m
+
+ bm
=
PROOF We have (a2 + a
m
-
... +
bm-l
be a group with
(G,+)
... + am, b a
+
b
+
b
1
+
c = a - a + c - b
... + am )
a
- (a3 +
- ...
- b + b
m
=
d, where
1
PROPOSITION 12 Let R
A 2 J1(R), where A
a,al,
... +
b = a
+ b2 +
-
(a
... +
2
+
... + am) + bl +
(amdl + am) - a
m
from which the result follows.
be a
dg
f G such
Then c := al + bl + a2 + b2 + b , . E J1(G), the commutator subgroup of G.
d
+
..., am, b,bl, ..., bm
+
bm-l
+
0
near-ring with identity, and suppose
is an ideal of
R
and
61(R) is the commutator sugbroup
27 1
On two-sided ideals in matrix near-rings of
(R,t). Then
t
... +
PROOF
A*
in
M2(R)
[al;l,l] + [b1;1,21+
the form
[c ;2,1] + [d ; 2 , 2 ] , 4
consists of all and only those matrices of
... +
where
q
In [2] it was noted that if
[am;l,l]
+
[bm;1,2] + [c1;2,1] t [d1 ;2,2]
Cai, Chi, Zci, Idi E A .
R
is dg, then so is
Mn(R). By applying
matrices in A*
to the vectors
matrices in A*
are of the given form. On the other hand, it follows from
<1,0> and
it is easy to see that all
Lemma 1 1 that every matrix of the given form is in
A*.
0
It will be obvious that an extension of this result to possible.
M,(R),
n > 2,
is
However, the formulation of the extension does not provide additional
insight, so we shall forego stating it.
ACKNOWLEDGEMENT Work on this paper was financially supported by the Council for Scientific and Industrial Research.
REFERENCES [l] [2] I31 [4]
Lausch, H and W Nobauer, Algebra of polynomials, North-Holland, 1973. Meldrum, JDP and APJ van der Walt, Matrix near-rings, Archiv der Math (Hasel). To appear. Pilz, G, Near-rings, Revised Edition, North-Holland, 1983. Van der Walt, APJ, Primitivity in matrix near-rings, Quaest. Math. To appear.
This Page Intentionally Left Blank
Near-rings and Near-fields, G. Betsch (editor) 0 Elsevier Science Publishers B.V. (North-Holland), 1987
273
SOME PATHOLOGY FOR RADICALS I N NON-ASSOCIATIVE NEAR-RINGS Stefan Veldsman Department o f Mathematics, U n i v e r s i t y o f P o r t E l i z a b e t h , P o r t E l i z a b e t h Republic o f South A f r i c a ABSTRACT. I t i s well-known t h a t every semisimple c l a s s i n t h e c l a s s m s s o c i a t i v e r i n g s o r a l t e r n a t i v e r i n g s i s h e r e d i t a r y , t h i s bei n g an easy consequence o f t h e Anderson-Divinsky-Sulifiski p r o p e r t y ( c f . [ 11). Dropping t h e a s s o c i a t i v i t y , one g e t degenerate r a d i c a l s i n t h e sense t h a t a r a d i c a l c l a s s has a h e r e d i t a r y semisimple c l a s s i f and o n l y i f i t i s an A - r a d i c a l , i . e . t h e r a d i c a l o n l y depends on t h e I n the s t r u c t u r e o f t h e u n d e r l y i n g a b e l i a n groups ( c f . [ 4 ] and [ 5 ] ) . class o f a l l near-rings ( o r abelian near-rings) t h e s i t u a t i o n i s not as bad, here we have b o t h h e r e d i t a r y and n o n - h e r e d i t a r y semisimple classes, see, f o r example [ 3 ] . Dropping t h e a s s o c i a t i v i t y i n t h i s case, t h e r e s u l t i s even worse than i n t h a t o f n o n - a s s o c i a t i v e r i n g s . We show t h a t i n t h e c l a s s o f a l l a b e l i a n , n o t - n e c e s s a t i l y a s s o c i a t i v e zero-symmetric n e a r - r i n g s t h e o n l y r a d i c a l s w i t h h e r e d i t a r y semisimple classes a r e t h e two t r i v i a l r a d i c a l c l a s s e s .
1.
PRELIMINARIES L e t W be t h e c l a s s o f a l l a b e l i a n n o t - n e c e s s a r i l y a s s o c i a t i v e zero-symmetric
near-rings.
Near-rings w i l l be r i g h t n e a r - r i n g s and we use t h e n o t a t i o n and
n o t i o n s o f P i l z [6].
W , being a u n i v e r s a l c l a s s o f a-groups i s hence s u i t a b l e
f o r t h e development o f a Kurosh-Amitsur r a d i c a l t h e o r y , basics o f which we r e f e r to
171. A c l a s s R 5 W i s a radical class (in t h e sense o f Kurosh-Amitsur) i f ii)
R i s homomorphically closed For every N E W , R(N) := c ( I I I an i d e a l i n N w i t h I E R) E R
ii)
R(N/R(N)) = 0 .
(i)
I R(N)
w i l l denote t h e semisimple cZuss o f R. SR Using standa r d techniques, i t can be proved, f o r any r a d i c a l c l a s s R : AS u s u a l , SR = {N E W
is
= 0)
h e r e d i t a r y i f from N E SR and I an i d e a l i n N, I E SR f o l l o w s .
(i) (ii)
I f K i s an i d e a l o f N and N / K E SR, then N’N)’I c _ R ( K). 5 R(N)’ f o r every i d e a l I o f N.
SR i s h e r e d i t a r y i f and o n l y i f R ( I )
( i i i ) R(N1 8N1) = R(NI) BR(N2). For N E W , N+ w i l l denote t h e u n d e r l y i n g a d d i t i v e gymup o f N and No w i l l be t h e zero n e a r - r i n g b u i l t on N+ by t h e m u l t i p l i c a t i o n ab = 0 f o r a l l a,b E N. The c l a s s o f a l l zero n e a r - r i n g s i n W w i l l be denoted by 2 .
N E W,
No E W holds.
I f I i s an i d e a l o f N, i t w i l l be denoted by I A N.
Obviously, f o r an)
214
S. Veldman
I n t h e sequel, a l l n e a r - r i n g s considered, w i l l be from W. 2.
RESULTS Betsch and K a a r l i [ Z ] have proved t h e n e x t theorem i n t h e c l a s s o f a l l near-
As i t makes no use o f t h e a s s o c i a t i v i t y , i t remains v a l i d i n W , and we
rings. have
5.1 Theorem L e t R be a r a d i c a l c l a s s w i t h h e r e d i t a r y semisimple c l a s s SR and assume SR c o n t a i n s a zero n e a r - r i n g N ( = No) w i t h N # 0.
Then 2
c_ SR.
Our main r e s u l t depends on t h e n e x t two c o n s t r u c t i o n s : For N
E W , we d e f i n i e two n e a r - r i n g s r(N) and h(N) by:
r(N)'
= NtON+4Nt
= A(N)+
w i t h m u l t i p l i c a t i o n d e f i n e d by (a,b,c)(x,y,z)
=
(ax
{
,az , 0 ) , az , c z )
(axtcy
in
r(A)
i n A(A).
Both T(N) and A(N) a r e a b e l i a n , r i g h t d i s t r i b u t i v e , zero-symmetric, not-necess a r i l y a s s o c i a t i v e (even i f N i s a s s o c i a t i v e ) and hence i n W. L e t B = {(a,b,O) B
N 4 No.
B
I a,b E
Then B i s an i d e a l i n b o t h T(N) and A(N) and
N}.
Furthermore,
r (B N ) "= No
and A(N) B
N.
2.2 L e t R be a r a d i c a l c l a s s w i t h h e r e d i t a r y semisimple c l a s s SR. No
E SR, then N
If
N E R and
= 0.
Proof Because SR i s h e r e d i t a r y , we have N = R(NBNo) = R(B)
zR( ( N ) )
A r(N).
On t h e o t h e r hand, R ( m ) = R(No) = 0 , N@N0
hence
R ( T ( N ) ) C_ R(NBNO) = N. By t h e above, we have N = R ( r ( N ) ) A
L e t a,b E N. (O,ab,O)
Because (a,O,O) E (N,O,O)
= (a,O,O)(O,O,b)
E N
Hence ab = 0 f o r a l l a,b E N. 2.3
r(N).
E
(N,O,O)
=
N A T(N), follows.
Thus N = No E R 11SR = 0, i . e . N = 0.
Lemma
L e t R be a r a d i c a l c l a s s w i t h h e r e d i t a r y semisimple c l a s s SR. No i
R, t h e n N
= 0.
If N E SR and
275
Some pathology for radicals in non-associative nearrings
Proof As i n t h e previous lemna, i t can be shown t h a t NO
= R ( A ( N ) ) A A").
For a,b E N, (O,b,O) (ab,O,O)
E (O,N,O)
= (O,U,a)[(a,O,O)
No A A(N).
t
(O,b,0)1
Hence
- (O,O,a)(a,O,O)
E
No
z (O,N,O).
Thus ab = 0 f o r a l l a,b E N.
E R I I SR
Hence N = No
= I ) , i.e.
N = 0.
We can now prove o u r main r e s u l t : 2.4
Theorem
L e t R be a r a d i c a l c l a s s i n W.
Then SR i s h e r e d i t a r y i f and o n l y i f R = 0 or
R = W. Proof Obviously o n l y t h e s u f f i c i e n c y needs v e r i f i c a t i o n . Indeed, i f Z
R, l e t N = No E Z w i t h N f R.
N/R(N) i s a zero n e a r - r i n g , Z 0 # B E R. Z c - R.
Then Bo t Z
For any N
5 SR
5 SR
I f R # I ) , then Z
5 R.
Hence 0 # N/R(N) E SR and because
f o l l o w s from 2.1.
Because R #
I),
let
and by 2.2 the c o n t r a d i c t i o n B = 0 f o l l o w s .
E W , we have N/R(N) E SR and (N/R(N)I0 E R.
€ Z
c_ R.
Hence
By 2.3,
N/R(N) = 0, thus N = R(N) REFERENCES
[I] Anderson, T., D i v i n s k y , A. and S u l i r l s k i , A., H e r e d i t a r y r a d i c a l i n a s s o c i a t i v e and a l t e r n a t i v e r i n g s , Canad. J . Math., 17, 1965, 594-603. [ 2 ] Betsch, G. and K a a r l i , K., S u p e r n i l p o t e n t r a d i c a l s and h e r e d i t a r i n e s s o f semisimple classes o f n e a r - r i n g s , C o l l . Math. SOC. J . B o l y a i , 38, Radical Theory, North H o l l and, 1.985, 47-58. [ 3 ] Setsch, G. and Wiegandt, R., Non-hereditary semisimple classes o f nearr i n g s , S t u d i a Sci. Math. Hungar., 17, 1982, 69-75. [ 4 ] Gardner, B.J., Some degeneracy and pathology i n n o n - a s s o c i a t i v e r a d i c a l theory, Annales Univ. Sci. Budapest, 22-23, 1979/80, 65-74. [ 5 ] Gardner, B.J., Some degeneracy and pathology i n non-associative r a d i c a l t h e o r y 11, B u l l . A u s t r a l . Math. SOC., 23, 1981. 423-428. [ 6 ] P i l z , G., Near-rings, (North-Holland, 1977). [ 7 ] Wiegandt, R., Radical and semisimple classes o f r i n g s , Queen's papers i n pure and a p p l i e d mathematics, No.37, Kingston, O n t a r i o , 1974.
This Page Intentionally Left Blank
Newrings and Near-felds, G.Betsch(editor) 0 Elsevier Science Publishers B.V.(North-Holland), 1987
217
PARTIALLY AND FULLY ORDERED SEMINEARRINGS AND NEARRINGS Hanns Joachim WEINERT 1. INTRODUCTION
According to [ 1 1 I and [121, a (in this paper always right distributive) seminearring, is defined to be an algebra S = (S,+,*) such that (S,+) and (S,-) are semigroups and (a+b)c=ac+bc holds for all a,b,c E S. Seminearrings occur very naturally as transformations of semigroups ( r , + ) (cf. Thm. 3.1). They are also the common generalization of nearrings and semirings, and we refer to [I31 and [ I 4 1 for corresponding investigations of statements shared by both, nearrings and semirings. In a similar way we deal in this paper with seminearrings established with a partial order on S. These considerations will also provide new results on partially ordered (p. 0 . ) nearrings (cf. 151, [81 - I 1 1 1 1 and p. 0. semirings (cf. [21, [151), in particular in the fully ordered ( f . 0.) case. which is likewise In the papers just cited, a nearring ( S , + , * ) a p. 0. group (S,+,<) is called a p. 0. nearring iff the set P(S) of all positive elements of ( S , + , S ) is multiplicatively closed (hence a subseminearring). The latter turns out to be equivalent with P (S)S M r (S) (called IIIr in the f o l l o w i n g ) , where Mr ( S ) [Mi(S)1 denotes the set of all right [left] monotone elements of ( S , - , < ) (cf. (2.1)). By results of [51 , there are (proper) f. 0. nearfields S satisfying also P(S) CMk(S) (called 111%), as well as those for which IIIQ fails to be true. Taking also the sets N ( S ) for all negative elements of (S,+,<) and Wr(S) [W,(S)l of all right [left] antitone elements of (S,-,S) into consideration (cf. (2.2)), we introduce IVr by N ( S ) C Wr(S) and IVL by N(S) S Wi(S). For each ring (S,+,') which is also a p. 0. group ( S , + , S ) , we shall see that any of these four properties IIIr - IVi implies the other three (Remark 2.5). In the case of nearrings, merely IIIi c) IVa holds in general, IIIQ*IIIr and IIIQ 9 IVr for zero-symmetric ones, and 111, o IVr if S satisfies at-b) =-(ab) (cf. Thm. 2.4). More general, let (S,+,-) be a seminearring which is also a p. 0. semigroup ( S , + , S ) . We call such a structure ( S , + , - , S ) a w e a k l y partia l l y o r d e r e d I w . p. OJ s e m i n e a r r i n g and introduce the sets P(S), and the properties 111,' already used above for w.p.0. seminear-
...
...
218
H.J. Weinert
rings in 5 2.
The c o r r e l a t i o n s between t h e s e f o u r p r o p e r t i e s , t h e
main s u b j e c t o f t h i s s e c t i o n , t u r n o u t t o s t a g g e r from " t h r e e do n o t imply the f o r t h " t o " p a i r w i s e e q u i v a l e n t " , depending on f u r t h e r a s s u m p t i o n s on S. Some o f t h e s e r e s u l t s are mentioned above. They prove it meaningful t o d e f i n e nearring
-
a p.
0.
-
g e n e r a l i z i n g t h e c o n c e p t of a p.
s e m i n e a r r i n g a s a w. p .
0.
seminearring s a t i s f y -
0.
i n g I I I r y which i s now i n d e p e n d e n t o f and m o s t l y s t r o n g e r t h a n Those p .
P ( S ) ' P ( S ) C P(S).
t r a n s f o r m a t i o n s of p. Thm. 3 . 2 ) .
0.
0.
seminearrings occur n a t u r a l l y as
semigroups (r,+,S),
H e r e 111, i s i m p l i e d by
say
Mr(S) = S ,
I I I Q and IVR i s o n l y t r u e f o r a s p e c i a l c h o i c e of
C o n v e r s e l y , e a c h p.
0.
S =
(rry+,-,s) ( c f .
whereas e a c h of IV,,
seminearring s a t i s f y i n g
(r,+,S)
(Cor. 3 . 3 ) .
M (S) = S r
c a n be re-
p r e s e n t e d i n t h i s way (Thm. 3 . 4 ) . For s t a t e m e n t s on f .
n e a r r i n g s S, t h e c o n s t a n t s u b n e a r r i n g of
0.
S and t h e zero-symmetric
o n e , s a y S1 and S 2 , are of some i m p o r t a n c e
( c f . [ill). W e i n t r o d u c e i n 5 4 c o r r e s p o n d i n g s u b s e m i n e a r r i n g s S 1 and S 2 f o r s e m i n e a r r i n g s S w i t h a z e r o . If ( S , + ; , 6 ) i s a w. p . 0 . seminearring,
t h e same h o l d s f o r ( S 1 , + , ~ , S 1 and ) ( S 2 y + y - , < 2 )w i t h
t h e induced p a r t i a l o r d e r s S ( c f . Thm. 2 . 1 0 ) . For a f . 0. s e m i n e a r i r i n g S, t h e o r d e r o f S 1 + S 2 i s u n i q u e l y d e t e r m i n e d l e x i c o g r a p h i c a l l y by ( S 1 , S 1 ) and ( S 2 , S 2 ) , and, a p a r t from some e x t r e m e cases, a f . s e m i n e a r r i n g S such t h a t
Sl${o}+S2
0.
h o l d s i s non-archimedean
o r d e r e d (Thm. 4 . 2 ) . The a p p l i c a t i o n s t o f .
0.
n e a r r i n g s i n Thm. 4 . 3
p r o v i d e w e l l known a s w e l l a s new r e s z t s , i n p a r t i c u l a r t h a t Thm. 9 . 1 4 6 of
[ l l ] i s o n l y t r u e f o r zero-symmetric
nearrings. Since we
need v a r i o u s examples t o prove r e s u l t s i n 5 2 and § 4 and to show t h a t t h e y a r e f a i r l y s t r o n g , w e have c o l l e c t e d them i n
5.
There i s one more p o i n t c e n t r a l t o t h e d i s p o s i t i o n of t h i s p a p e r . To o u r own s u r p r i s e , a l l r e s u l t s on w. p . 0 . do n o t depend on t h e a s s o c i a t i v i t y o f
s e m i n e a r r i n g s (S,+,',S)
( S , + ) , which a l s o d o e s n o t
s i m p l i f y t h e i r p r o o f s . S i n c e one-sided d i s t r i b u t i v e a l g e b r a s ( S , + , * ) w i t h n o t n e c c e s s a r i l y a s s o c i a t i v e a d d i t i o n a r e of i n t e r e s t i n o t h e r contexts (cf. [4]
,
[6])
, we
f e e l t h a t w e should e x p r e s s t h i s gener-
a l i t y i n o u r p r e s e n t a t i o n . So, s i m i l a r l y a s i n [ 1 3 1 , w e g e n e r a l i z e t h e c o n c e p t of a s e m i n e a r r i n g a s f o l l o w s : DEFINITION 1 . 1 . An a l g e b r a ( S , + , ' )
i s c a l l e d a seminearring i f f ( S , + ) i s any g r o u p o i d , ( S ; ) a semigroup and ( a + b ) c= a c + b c h o l d s f o r a l l a,b,cES. T o a v o i d c o n f u s i o n , w e u s e t h e term n e a r r i n g a s u s u a l , a l t h o u g h a s e m i n e a r r i n g f o r which ( S , + ) i s a q u a s i g r o u p would now be t h e adequate concept.
279
Partially and fully ordered seminearrings and near-rings
Various statements on those seminearrings have been given in [I31 and [141. However, anyone not interested in this generalization, can read this paper without taking notice of Def. 1.1. In particular, all examples of seminearrings used to prove statements are of course additively associative. Finally, we need a few conventions. For subsets A,B of a seminearring (S,+,') we introduce A+B={a+b I a E A , ~ E B ]and AB={ab I a E A , b ~ B 1 , and use e. g. aB for {ajB. If a groupoid or a seminearring S has an additive neutral, this unique element is called the z e r o of S and denoted by 0. In the latter case, o is said to be l e f t [ r i g h t ] a b s o r b i n g iff o S = { o ) [ S o = { o } l holds. A seminearring with a (two-sided) absorbing zero is also called zero-symmetric.
2. BASIC STATEMENTS ON (WEAKLY) PARTIALLY ORDERED SEMINEARRINGS
According to 5 1 we do not assume additions to be associative (if not explicitely stated) and define the more general concept of a weakZy p a r t i a l l y o r d e r e d I w . p . 0 . ) s e m i n e a r r i n g S = ( S , + , * , = < ) by
I
(S,+,')
I1
a < b implies a + c < b + c and
is a seminearring, (S,<) a p.
0.
set, and
c+a6c+b for all
a,b,cES.
In particular, S is called a w e a k l y f u l l y o r d e r e d Iw. f . 0 . 1 s e m i is a f. 0. set, and trivially p. 0. iff the partial order equals equality. Note that a w. p. 0. seminearring S is just a seminearring (S,+,.) and a p. 0. groupoid (S,+,<) (cf. [I], Chap. X). By 11, a < b implies a+c c b+c if c is right cancellable in IS,+). So if (S,+) is cancellative, the monotony law I1 yields its strict version n e a r r i n g iff (S,<)
IISt a < b
implies a + c
I
a < a+p for all a E
we introduce the set seminearring S = (S,+,.,<), which are in fact the right positive elements of the p. 0. groupoid (S,+,S). Due to 11, p < s for P E Pry s E S implies s E Pry hence Pr(S) is an upper set of ( S , S ) and thus a subgroupoid of (S,+) if Pr$@. The same holds for the sets PQ(S) and P(S) =Pr(S) rl PR(S) of all l e f t p o s i t i v e or p o s i t i v e elements of S, respectively. If S has a zero 0, one has P =Pa = P = { p S~ 1 0 s p}. Dually with respect r to the relation <, we have the corresponding definitions and stateBy
Pr = Pr(S) = {p E
S
of all r i g h t p o s i t i v e elements of a w. p.
S]
0.
280
H.J. Weinert
ments f o r t h e sets N r ( S ) , NII(S) and N(S) of a l l [ r i g k t , l e f t ] n e g a t i v e e l e m e n t s of S.
briefly
Note t h a t
r* An element m of a w . p . 0 .
(2.1)
n < p holds for a l l
ncN
II
and
pepr,
NI16P
a cb
implies
seminearring
amsbm
for a l l
S = (S,+,',<) such t h a t
a,bES
i s c a l l e d r i g h t m o n o t o n e , and s t r i c t l y r i g h t m o n o t o n e i f f ( 2 . 1 ) h o l d s w i t h am
t h e s e sets, i f n o t empty, are subsemigroups of (S,.), and ML(S) i s a l s o a d d i t i v e l y c l o s e d . An element W E S such t h a t (2.2)
a
implies
awtbw
for a l l
a,bE S
i s c a l l e d r i g h t a n t i t o n e . W e d e n o t e t h e set of a l l t h e s e e l e m e n t s
by Wr(S). Likewise w e i n t r o d u c e t h e s e t W ( S ) of a l l l e f t a n t i t o n e 11 elements o f S and W ( S ) = W I I ( S ) n Wr(S). R u l e s l i k e W r W r E M r or W M CWr
are obvious. Note i n t h i s c o n t e x t t h a t (S,+,-,6) remains r r a w . p . 0 . s e m i n e a r r i n g r e p l a c i n g 5 by i t s c o n v e r s e r e l a t i o n S d . But whereas P ( S ) and N ( S ) change t h e i r r o l e s under t h i s p r o c e d u r e , Mr(S), M Q ( S ) , W r (S),and W II ( S ) remain i n v a r i a n t .
DEFINITION 2 . 1 .
L e t S be a w . p . 0. s e m i n e a r r i n g . W e d e n o t e , i n t h e
f o l l o w i n g o r d e r of s u c c e s s i o n , t h e p r o p e r t i e s (2.3)
P(S) C Mr(S),
N(S) C Wr(S),
P(S) C MII(S), N ( S ) C WR(S)
by 111,' I V r , I I I Q and I V R , and use a sequence a s e . g. ( l , l , O , l ) t o i n d i c a t e by "1" which of t h e s e p r o p e r t i e s h o l d s f o r S and by " 0 " which n o t . The c o r r e s p o n d i n g sequence i s c a l l e d t h e t y p e of S, and 1 S a p a r t i a l l y o r d e r e d ( p . 0 . ) s e m i n e a r r i n g i f f 111, h o l d s . Hencef o r t h w e w r i t e e . g . ( 1 1 0 1 ) = ( 1 3 ) f o r t y p e s i n an obvious and conv e n i e n t i n t e r p r e t a t i o n . As shown by examples c o l l e c t e d i n § 5 ( a l l with a s s o c i a t i v e a d d i t i o n ) w e s t a t e a t f i r s t : REMARK 2 . 2 .
The p r o p e r t i e s ( 2 . 3 ) a r e i n d e p e n d e n t , and t h e r e a r e
even f . 0. n e a r r i n g s and f . 0. s e m i r i n g s of t y p e ( l o l l ) = ( I l ) , f . 0. s e m i r i n g s o f t y p e (1101) = ( 1 3 ) and o f t y p e ( 1 1 1 0 ) = (14), and w. f . 0 . s e m i r i n g s o f t y p e (0117) = ( 7 ) (Expls. 5.3 and 5.91. Moreover, t h e r e a r e w . p . 0 . s e m i n e a r r i n g s S = ( S , + , * , < ) such t h a t a l l p o s s i b l e t y p e s from ( 0 ) t o ( 1 5 ) r e a l l y o c c u r .
Partially and fully ordered seminearrings and nearrings
28 1
On the other hand, additional assumptions on S cause implications between the properties ( 2 . 3 ) . For instance, there are no f.0. nearrings of type ( 1 3 ) or ( 1 4 ) , in fact not even p. 0. ones, and also no w. f.0. nearrings of type (7) (cf. Thm. 2 . 4 ) . For the corresponding investigations we need one more notion and call a w.p. 0. seminearring (S,+,-,6)with a zero p a r t i a Z l y o r d e r e d b y X=P(S) f r o m t h e l e f t iff the following holds for all a,bE S : (2.4)
a Sb
there is some X E X satisfying x+a = b.
Q
Likewise one defines S to be p . o . by X=P(S) f r o m t h e r i g h t with respect to a + x = b , and both concepts apply in fact to ( S , + , < ) . 1 ) Turning now to implications between the properties ( 2 . 3 ) for particular w. p. 0. seminearrings S = ( S , + , - , < ) , of course in the meaning of Def. 1 . 1 , we use the following abbreviations for additional assumptions on S: S has a left absorbing zero 0. (r) S has a right absorbing zero 0. (p) S is p.0. by its set P(S) of positive elements from one side. (n) S is a nearring (i. e.(S,+) a group), which yields ( 8 ) and (P). S is a nearring satisfying also the sign-rule a(-b) = -(ab). (s) (9.)
L W A 2 . 3 . Let
S = ( S , + , * , < ) be a w.p.0. seminearring and write P(S) = P etc. Then we have the following implications depending on the assumption on S fixed at each arrow:
-__ -__ P
111,
PP C P
NP C N
111,.
R
P
IVr
PNEN
IIIQ
NNEP
PP G P r
r
PN E N
S
-
IVr
IIIR
r
S
NP C N
IVQ
R
n
R
NN E P
IVg r
Proof. From n < o S p for each n E N , P E P and ( a ) we obtain NMrEN and PMrSP, hence IIIr implies N P E N and PPEP. All other implications indicated by ( R ) or (r) follow similarly. Now we assume P P E P [PNEN] and (p). Then ( 2 . 4 ) yields 111, [IV,] by right distributivity and the definition of P [N]. The implications for which
282
H.J. Weinert
( n ) o r ( s ) i s assumed a r e clear. The n e x t theorem y i e l d s i n p a r t i c u l a r s t a t e m e n t s on w . p .
0.
near-
o n e s ( ( n ) and ( r ) ) , and s u b n e a r r i n g s
r i n g s S ( ( n ) ) , zero-symmetric
of a n e a r f i e l d ( ( s ) and ( r ) ) ,which w e do n o t f o r m u l a t e e x p l i c i t e l y . I n t h i s c o n t e x t one a l s o c h e c k s t h e e q u i v a l e n c e s
7p E P 7n c N
(2.7)
IIIR
and
b-a€ P
implies
pb-pa
E
P
(2.8)
IVL
and
b-a
implies
nb-na
E
N,
clearly for a l l
a , b , n , p E S.
E
P
The r i g h t s i d e of
t o t h e supplementary axiom A 4 i n [ 5 ] f o r f . THEOREM 2 . 4 .
Let
S =
0.
(S,+,-,<) be a w . p . 0 .
and
( 2 . 7 ) corresponds
nearfields.
s e m i n e a r r i n g . Then,
u s i n g t h e above c o n v e n t i o n s , we have t h e f o l l o w i n g i m p l i c a t i o n s :
(2.9)
IVR
7I11
(2.10)
IVr
‘sc
== > I I I r
and
rYP
111,.
Moreover, a w. f . 0 .
n e a r r i n g s a t i s f y i n g I V r i s zero-symmetric. ( 2 . 9 ) and t h e e q u i v a l e n c e ( 2 . 1 0 )
Proof. The second i m p l i c a t i o n of
f o l l o w from ( 2 . 5 ) and ( 2 . 6 ) . To show r i n g S, a t f i r s t w e assume (-p) E N
IVQ
I V Q y P E P and
e,
I I I Qf o r a w . p . o . n e a r -
a < b , i . e . b-aE P. Then
and ( 2 . 8 ) imply ( - p ) b - ( - p ) a = - ( p b ) + p a E N , hence p a = p b + n
for some shows
IVr
nEN,
IVR
*
which y i e l d s
pa
by t h e d e f i n i t i o n of N . T h i s
I I I R , and s i m i l a r l y w e g e t
I I I Q=) I V R
using
(2.7).
For t h e l a s t s t a t e m e n t w e r e f e r t o Lemma 4 . 1 . From t h e above w e o b t a i n w i t h o b v i o u s p r o o f s : REMARK 2.5.
For a w . p . 0 .
r i n g S, from ( 2 . 9 ) and i t s l e f t - r i g h t
d u a l v e r s i o n it f o l l o w s t h a t each of t h e p r o p e r t i e s ( 2 . 3 ) i m p l i e s t h e o t h e r t h r e e and c h a r a c t e r i z e s S a s a p a r t i a l l y o r d e r e d r i n g . REMARK 2 . 6 . Def.
2.1
The c o n c e p t of a p . 0 . n e a r r i n g S i n t r o d u c e d by I I I r i n ( 2 . 5 1 , e q u i v a l e n t w i t h t h e u s u a l one which d e f i n e s
i s , by
a p a r t i a l o r d e r on ( S , + , - )a c c o r d i n g t o ( 2 . 4 ) by a s u b s e m i n e a r r i n g X=P,
assumed t o s a t i s f y t h e p r o p e r t i e s g i v e n i n f o o t n o t e 1 ) .
REMARK 2 . 7 .
Each p . 0 .
(1100)= (12),
n e a r r i n g S h a s one of t h e t y p e s
( l o l l ) =( 1 1 )
p o s s i b l e f o r p.
0.
and
(1111) = (15),
(1000) = (8). Only ( 1 5 ) and ( 1 2 ) are
n e a r f i e l d s , and t h e r e a r e even f .
0.
n e a r f i e l d s of
t h i s k i n d a s shown i n [ 5 ] . Moreover, f . 0 . n e a r r i n g s o f t h e t y p e s ( 1 5 ) and ( 1 2 ) are zero-symmetric, t h o s e of t y p e ( 1 1 ) n o t ( b y ( 2 . 9 ) ) , whereas b o t h c a s e s o c c u r f o r 5.1 and 5 . 5 ) .
f.0.
n e a r r i n g s of t y p e ( 8 )
( c f . Expls.
Partially and fully ordered seminearrings and near-rings REMARK 2.8.
The n e x t c l a s s a r e w . p . 0 . (0111) = ( 7 ) ,
s a t i s f y I I I r y hence w i t h and ( 0 0 0 0 ) = ( 0 )
283
n e a r r i n g s S which d o n o t (0011) = (3)
(0100) = ( 4 ) ,
a s p o s s i b l e t y p e s . The l a t t e r t h r e e o c c u r even a s
w. f . 0 . n e a r r i n g s ( E x p l s . 5.2 and 5 . 7 ) . By t h e l a s t s t a t e m e n t of Thm. 2 . 4 and ( 2 . 9 ) , t h e r e are no w. f . 0 . n e a r r i n g s of t y p e ( 7 ) . W e d o n o t know w h e t h e r t h e r e are w . p . 0 . There a r e w . f . 0 .
REMARK 2 . 9 .
(0111) = ( 7 ) ,
(1110) = ( 1 4 )
o n e s ( b u t c f . Expl. 5 . 3 ) .
s e m i r i n g s of t h e t y p e s
and
(1101) = ( 1 3 ) ,
( 1 0 1 1 ) = ( 1 1 ) ( c f . E x p l s . 5.3 and 5.9).
T h i s shows t h a t IIIr and I I I Q a s w e l l a s I V r and I V L a r e indepens e m i r i n g s . However, s i n c e s t a t e m e n t s on s e m i -
dent a l s o for w.p.0.
r i n g s are m u l t i p l i c a t i v e l y l e f t - r i g h t d u a l , one s h o u l d speak a b o u t
a p-
0.
semiring
S
o n l y i f I I I r and IIIR h o l d . T h i s s t r o n g e r con-
c e p t w a s i n t r o d u c e d i n [151 by
P ( S ) C M(S) = M r ( S ) fl M Q ( S ) .
W e close t h i s s e c t i o n d e a l i n g w i t h q u e s t i o n s on s u b s t r u c t u r e s :
U = (U,+,.) be a s u b s e m i n e a r r i n g of a w . p . 0 . (S,t,-,4). Then (U,+,.,<) i s a w . p . 0 . seminear-
THEOREM 2 . 1 0 .
a) Let
seminearring
S =
r i n g w i t h r e s p e c t t o t h e i n d u c e d p a r t i a l o r d e r , and one h a s (2.11)
C ( S ) n U E C(U)
for
C = P, N, Mry
Wry
M R and W Q .
b ) However, even i f S i s f u l l y o r d e r e d , a s s o c i a t i v e and commutative w i t h r e s p e c t t o b o t h o p e r a t i o n s and s a t i s f i e s I I I r , IV,,
I I I R and
I V k y none of t h e s e p r o p e r t i e s need t o be t r u e f o r U.
c ) S u f f i c i e n t c o n d i t i o n s such t h a t 111,. t r a n s f e r s from S t o U a r e Mr(S) = S ( c f . Thm. 3 . 2 ) o r P ( U ) C P ( S ) , which i s P(U) = P ( S ) t l U by ( 2 . 1 1 ) . The l a t t e r i s a l s o s u f f i c i e n t t o t r a n s f e r I I I Q . I n t h e same way work
Wr ( S ) = S
f o r IVr
and N ( U ) E N ( S ) f o r I V r and I V Q .
d ) A s s u m e t h a t U and S have a common zero, which i s a l w a y s t h e c a s e i f b o t h a r e n e a r r i n g s . Then
P(U) = P(S) fl U and N ( U ) = N ( S ) I l U hold, IIIQor IVQy the
and i f S s a t i s f i e s one of t h e p r o p e r t i e s I I I r , IV,,
same p r o p e r t y h o l d s f o r
U.
P r o o f . A l l s t a t e m e n t s of a ) are checked s t r a i g h t f o r w a r d , and c ) follows from ( 2 . 1 1 ) Example 2.11.
and i m p l i e s d ) . For b ) w e g i v e t h e f o l l o w i n g
a ) A semiring ( S , + ; ,
l i s t e d i n Thm. 2.10.
S = {o,a,b}
o < a < b
b ) i s g i v e n by
2)
s a t i s f y i n g a l l assumptions
+e b
b
.
H.J. Weinert
284
One also checks P(S) =M(S) = S and N(S) = W(S) = {o}. For the subsemiring U = {a,b} we have P(U) =M(U) = U , however, "U) = {a} and W(U) = {b}. So U satisfies 111,. and 111I?' but not IVr and IVI?' Another semiring ( S , + , - , 6 ) of this kind is given by
S = {a,b,c,d) a
b
b
b
b
b
b
b
b
d l b
b
c
c
d l b
b
a
a
.
Here one obtains P(S) = @ , M(S) = {a,b}, N ( S ) = {a} and W(S) = S . The subsemiring U = {a,b,c} satisfies P ( U ) = {c}, M(U) = {a,b}, but not 111, and I11 N(U) = {a) and w ( U ) = U, hence IVr and IV 2' R' 3 . TRANSFORMATIONS ON PARTIALLY ORDERED GROUPOIDS
The following characterization may be considered as known by its specializations bo additively associative seminearrings (cf. [33 , [ 1 2 ]) and nearrings. THEOREM 3.1. Let ( r , + ) be a groupoid and S = rr the set of all transformations f: r + r. For f,gES, define (3.1)
(f+g)( 5 ) =f(5)+9(5) and (f-g)( 5 ) = f(g(S)) for all
S E r .
Then S = (S,+,') is a seminearring such that (S,+) is associative, commutative, left [right] cancellative or a quasi-group, respectively, iff ( r , + ) has this property. Iff ( r , + ) has a zero w, S has a zero o given by o ( 5 ) = w, and of = o holds for all f E S. In this case, the sets S, of all constant transformations of r and S 2 = {f E S I f ( w ) = w) are subseminearrings of S (cf. 5 4 ) . can be represented in this Conversely, each seminearring (S,+,.) way. Take any groupoid ( I ' , + ) containing (S,+) properly as a subgroupoid, and define for all a € S
Then
a + fa
provides a monomorphism of (S,+,.)
into
(rr,+;).
Note in this context that there always exists such a groupoid (I',+)which has the same properties as (S,+) concerning associati-
vity, commutativity, left [right] cancellativity, or being a quasigroup. Apart from the fact that a quasi-group need not be embeddable into a loop, ( r , + ) may be assumed to have a zero in all cases under consideration. On the other hand, (I',+) = (S,+) will do the j o b
285
Partially and fully ordered seminear-rings and neur-rings ax = bx f o r a l l is r i g h t reductive. iff
x
a = b y i. e. i f f (S,.)
E S always i m p l i e s
t o a p . 0 . groupoid ( r , + , < )a s sumed t o be n o n - t r i v i a l l y p. o., and d e f i n e f o r f , g E S = r r
THEOREM 3.2. Now we a p p l y Thm. 3.1
(3.3)
f
s
g
f (5) 6 g(5)
w
Then we o b t a i n a p . 0 .
for all
seminearring
r
0. by
But (S,6)
which a l w a y s st I1 or is
S = (S,+,-,<)
s a t i s f i e s M ( S ) = S and Wr(S) = @ . I f p.
S E T .
(r,+,<) satisfies
P ( r ) from one s i d e , t h e same h o l d s f o r (S,+,<)
is never f . o . ,
sI
even i f
-
(I',S)
p ( r ) E p ( r )1,
and P ( S ) .
is. W e f u r t h e r s t a t e = { n E s,
I
n(r) 5 N(r)),
(3.4)
P(S) = tpE
(3.5)
MQ(S)= { h E S
h(a) 6 h ( B )
for a l l
a , @ r~} ,
(3.6)
WR(S)= { h ~ S l a < =,@h ( a ) 5 h ( P )
for all
a,PET}.
I
a
N(S)
and
r
h a s a z e r o and hence a l s o S, t h e zero-symmetric subseminearr i n g S 2 = (S2,+,*,S) i s a p . 0. one which s a t i s f i e s Mr (S2)= S 2 and If
Wr(S2) = (01,
and (3.4)
-
(3.6) remain v a l i d f o r S
2'
P r o o f . The g e n e r a l s t a t e m e n t s on S, a l s o M (S) = S (which y i e l d s IIIr) r and (3.4) are checked s t r a i g h t f o r w a r d . They imply t h e c o r r e s p o n d i n g ones on S2 by Thm.2.10. c o n t r a d i c t i o n , and
For Wr(S) = @ , assume
w(q) = 5
f o r some f i x e d
wEWr(S)
n
E
I-.
by way of
Define
f,gc S
by f ( 5 ) = a < p = g ( c ) f o r any p a i r a < B i n r , and f ( c ) = c , g ( c ) = c f o r a l l 5 $; 5 Then f < g and (fw) ( q ) < (gw) ( r l ) c o n t r a d i c t f w 5 gw
.
W,(S). I n t h e same way one o b t a i n s Wr(S2) = {o), s i n c e f o r 09w E S 2 t h e r e i s some rl E i- such t h a t w ( n ) = 5 w holds. T o show ( 3 . 5 ) f o r S, assume a t f i r s t h E M L ( S ) . For any p a i r a < @ of r , d e f i n e f , g e S by f ( 5 ) = a and g ( 5 ) = I3 f o r a l l 6 E r Then f < g i m p l i e s hf S h g , hence h ( a ) 6 h ( B ) by (3.3) and ( 3 . 1 ) . So M ( S ) i s c o n t a i n e d i n t h e s e t of a 1 1 i s o t o n e t r a n s f o r m a t i o n s of L! ( r , + , 5 ) d e f i n e d on t h e r i g h t s i d e of (3.5), and t h e o t h e r i n c l u s i o n i s c l e a r . The proof of (3.5) f o r S 2 i s s i m i l a r , b u t it n e e d s some c a r e f o r t h e p a i r s w < @ and a < w i n r , i f t h e r e a r e t h o s e . Likew i s e one shows t h a t W Q ( S ) [W, ( S 2 ) ] c o i n c i d e s w i t h t h e s e t of a l l a n t i t o n e t r a n s f o r m a t i o n s c o n t a i n e d i n S [ i n S21y i. e. (3.6). a s c l a i m e d by
W E
.
COROLLARY 3 . 3 .
Concerning I V r y I I I Q and IV,
f o r t h e p . 0. seminear-
r i n g s S and S2 c o n s i d e r e d i n Thm. 3.2 w e s t a t e :
NU) = @
(3.7)
N ( S ) C Wr(S)
o
(3.8)
P(S) E M ~ ( s )
C)
Ip(r)I
s
(3.9)
N(S) E WQ(S)
*
1N(r)l
5 1
I
H.J. Weinert
286 (3.10)
N ( S 2 ) C- Wr(S2)
w
N ( T ) ={&I}.
(3.11)
P ( S 2 ) E MQ(S2)
w
P ( r ) = {wf
or
P ( r ) = {w,n},
the l a t t e r
o n l y i n t h e c a s e t h a t t h e r e a r e no d i f f e r e n t
( r ,+ ,5 )
e l e m e n t s comparable i n (3.12) Proof.
N ( S 2 ) E WQ(S2) w
w < TZ.
except
N(r) ={w}.
( 3 . 7 ) and ( 3 . 1 0 ) a r e o b v i o u s . For ( 3 . 8 ) w e s t a t e t h a t t h e r e
i s some h E P ( S ) \ M Q ( S ) i f f a p a i r
i n P ( r ) e x i s t s . The l a t t e r n + n ' i n P ( r ) i m p l i e s nSn+n' and n ' Sn+n', hence n < n + n ' o r n ' < n + n ' f o r n + n ' E P ( r ) . I n t h e same way one o b t a i n s ( 3 . 9 ) . For ( 3 . 1 1 ) , P ( S z ) E M R ( S 2 ) f a i l s t o be t r u e i f f t h e r e i s some h E S2 s a t i s f y i n g h ( r ) C P ( r ) and h ( a ) >h(@) f o r some a < @ i n r . The l a t t e r i s i m p o s s i b l e f o r P ( r ) = {u}, and c l e a r l y t h e case i f I P ( r ) I 2 3. For P ( r ) = {w,n} such a mapping h h a s t o s a t i s f y h ( a ) = n > w = h ( B ) f o r some (y. < 8 , and h e S 2 f o r c e s a S w . C o n v e r s e l y , t h e r e i s such an h E S2 i f w + a < B e x i s t s i n r . E x a c t l y t h i s i s e x c l u d e d i f f w < n i s t h e o n l y n o n - t r i v i a l rel a t i o n . Following t h e same p a t t e r n , one shows ( 3 . 1 2 ) where t h e case
i s e q u i v a l e n t with IP(S) I L 2 ,
IN(r)
-
I
=2
IVQ
-
makes no d i f f i c u l t i e s .
A s a consequence of Cor. 3 . 3 ,
IVr
S < q
since
IVr
h o l d s f o r S and even
IVQ
f o r S 2 . So t h e r e remain t h e f o l l o w i n g p o s s i b l e t y p e s (1101) = (13),
( 1 1 1 1 ) = (15),
and t h e same l i s t i n c l u d i n g
( 1 0 1 0 ) = ( 1 0 ) and (1011) = ( 1 1 ) and
( 1 0 0 0 ) = ( 8 ) f o r S2,
(1001) = ( 9 )
f o r S.
A l l t e n c a s e s r e a l l y occur f o r a d d i t i v e l y a s s o c i a t i v e s e m i n e a r r i n g s ,
s i n c e t h e r e are p .
semigroups
0.
( r y + , 6 ) s a t i s f y i n g t h e correspon-
d i n g p r o p e r t i e s on t h e r i g h t s i d e s of
(3.10)
t h o s e w i t h a z e r o w s e l e c t e d by ( 3 . 7 )
-
r
semigroup
= (IN , + , S 0
-
(3.121, a s w e l l as
( 3 . 9 ) . E. g . ,
for the f .
0.
of " p o s i t i v e " i n t e g e r s i n c l u d i n g 0 one ob-
)
t a i n s a n S of t y p e ( 9 ) , b u t an S 2 of t y p e ( 1 3 ) whereas f o r e a c h p . 0 . group
( r , + ,2 )
t h e p.
0.
n e a r r i n g s S and S2 have t y p e ( 8 ) . W e c l o s e
w i t h t h e f o l l o w i n g r e p r e s e n t a t i o n theorem: THEOREM 3.4.
For each p.
( S , + , - , S ) which s a t i s f i e s M ( S ) = S t h e r e i s an order-isomorphism i n t o a p . 0 . s e m i r n e a r r i n g [ n e a r r i n g ] (rr,+,-,s) of a l l t r a n s f o r m a t i o n s of a p. 0. g r o u p o i d [ g r o u p ] ( r ,+ ,I) 0.
seminearring [nearring]
.
P r o o f . C l e a r l y , each p .
0.
groupoid
subgroupoid of a p . 0 .
proper p . 0 .
( r , + , < ) . So
a S b
c)
fa 5 f b
a +fa
for a l l a,bES.
(S,+,I) i s a
[semigroup, group]
w e can a p p l y t h e c o n v e r s e p a r t of Thm. 3.1
mains t o show t h a t t h e monomorphism satisfies
[semigroup, groupl groupoid
of If
and it re-
(S,+,* ) i n t o aSb
( r r ,+,
holds, we
.)
287
Partially and fully ordered seminear-rings and near-rings
o b t a i n by ( 3 . 2 ) f f
thus
a a
b5 = f b ( 5 )
(5)
= a[
S
(5)
= a
I b
= fb(S)
for a l l
[ E S = M ~ ( S )and
for a l l
5
by ( 3 . 3 ) . The c o n v e r s e f o l l o w s from t h e l a s t f o r m u l a .
f a s fb
4 . CONSTANT AND ZERO-SYMMETRIC
SUBSEMINEARRINGS with a zero o we d e f i n e
For any s e m i n e a r r i n g S = ( S , + ,
(4.1)
E h S ,
S1 = { s l e S I
s 1o = s l }
S2 = { s ~ E S s I 20=o}.
and
I t i s e a s i l y checked t h a t e a c h of t h e s e s u b s e t s i s e i t h e r empty o r
-
-
a s u b s e m i n e a r r i n g o f S. One a l w a y s h a s
s2+g
(4.2)
02=o
s1 n s 2
-
SSICS1
+0
and
sln s2
=
to},
s o = o * o = ( s +o)o= s o+02 = o2 shows t h e o n l y n o n - t r i v i a l 2 2 2 i m p l i c a t i o n . R e c a l l t h a t each n e a r r i n g S i s a d d i t i v e l y a s e m i d i r e c t
since
sum of i t s s u b n e a r r i n g s S1 and S 2 , e q u i v a l e n t l y , e a c h S E S h a s a s = s + s E S +S2. S e m i n e a r r i n g s , however, a r e
unique decomposition
1 2 1 f a r away from t h i s n i c e s i t u a t i o n . N e v e r t h e l e s s , e a c h s e m i n e a r r i n g S with a zero s a t i s f i e s
(4.3)
S1 =
(4.4)
o2 = o
(03
*
S
2
= S
0
s 1 s = s1
and
S o = to} for a l l
and s1 E S1,
S E S
0
o S = {o}.
0 s = o * s s = ( s 0 ) s = s ( 0 s ) = s l y and ( 4 . 3 ) 1 1 1 For a s e m i n e a r r i n g S w i t h a l e f t a b s o r b i n g z e r o , w e
The l a t t e r f o l l o w s from from
SSIES1.
call S
1
t h e c o n s t a n t and S2 t h e z e r o - s y m m e t r i c s u b s e m i n e a r r i n g of S
( t h e i r o c c u r r e n c e i n Thm. 3.1 may j u s t i f y t h e f i r s t term). So w e
s t =s 1 1 1 groupoid (S,+)
c a l l a s e m i n e a r r i n g S, a c o n s t a n t one i f f i t h a s a z e r o and holds
s l S t lE
for a l l
S1.
Note t h a t e a c h ( p .
w i t h a z e r o p r o v i d e s such a ( w . P . Now l e t (S,+,*SS)be a w . p . 0 . b i n g z e r o . Then, by Thm.2.10,
w. p . that
0.
0.)
0.)
c o n s t a n t seminearring.
seminearring w i t h a l e f t absor-
and (S2,+,.,S2) a r e seminearrings w i t h r e s p e c t t o t h e induced r e l a t i o n s s i s u c h
P ( S i ) = P(S) n Si
p r o p e r t i e s 111,'
and
(S1,+,*,S1)
N(Si) = N(S) fl Si
h o l d and e a c h of t h e
I I I R and I V R t r a n s f e r s from S t o S1 and t o S2.
IVr,
I n t h i s c o n t e x t w e o b v i o u s l y have: LEMMA 4 . 1 .
Each w . p .
0.
c o n s t a n t seminearring ( S l , + , - , S l )
M r ( S 1 ) = M R ( S 1 )=W,(S1) = S I S whereas
is not t r i v i a l l y p. (1011) =(11)
0.
Wr(S1) = @
holds i f f
satisfies (S1,S1)
I n t h e l a t t e r c a s e , t h e t y p e of S1 i s always
s i n c e O E N ( S , ) . Hence, i f a w. p.
0.
(S,+S*,S)w i t h a l e f t a b s o r b i n g z e r o s a t i s f i e s I V r ,
seminearring
i t s subseminear-
H.J. Weinert
288
s1
ring
satisfies
s1 = t o }
( c f . (4.3)) or i s t r i v i a l l y p.
For t h e n e x t theorem r e c a l l t h a t a
f.0.
0.
semigroup (S,+,<)
z e r o o i s c a l l e d a r c h i m e d e a n i f f f o r each
a E
with
S t h e subsemigroup
of S g e n e r a t e d by a , s a y [a1 , i s n o t p r o p e r l y bounded from above i f o c a and n o t p r o p e r l y bounded from below i f a < o ( c f . [ l ] , Chap. X I ) . Obviously, t h i s c o n c e p t a p p l i e s a l s o t o a (S,+,<)
f.0.
w i t h a z e r o and i t s s u b g r o u p o i d s [ a ] g e n e r a t e d by
THEOREM 4.2.
L e t (S,+,*,<)
be a f . 0 .
groupoid a c S.
seminearring w i t h a l e f t a b -
S1=f { o l s S 2 h o l d s f o r t h e s u b s e m i n e a r r i n g s i n t r o d u c e d above. Then t h e induced o r d e r on t h e s u b s e t S1+S2 of S
s o r b i n g z e r o such t h a t
-
i s u n i q u e l y d e t e r m i n e d by (S, (4.5)
s1+ s 2 < t l + t 2
s 1 + s 2 $ t l + t 2of
for a l l
(4.6)
,S 1 )
according t o
and ( S 2 , S 2 )
s1 <1 t ,
or
s1 = t l ,
n1 < S 2 < p 1 f o r a l l n1 e N ( S 1 ) \ {o} and p1 E P ( S ~ ) \ { ~ } .
Moreover, a p a r t from t h e extreme c a s e s t h a t
S1 5
h o l d , t h e o r d e r of
(S,+,',2) i s non-archimedean.
Proof. M u l t i p l y i n g
s1+ s 2 < t l + t 2 by
yields
s 2 <2 t 2
S1+S2. T h i s y i e l d s i n p a r t i c u l a r
s1 S t l .
If
o
E
P ( S ) G Mr ( S )
s1 = t l h o l d s , s 2 2 t 2
s 1+ s 2 2 t l + t 2 by 11, hence w e o b t a i n
o2
S2
or
S2< 0 5 s
1
from t h e r i g h t
implies t h e contradiction For t h e c o n v e r s e i m -
s2 < t,.
(4.5) , assume a t f i r s t s1 < t l . Then w e have s1+s2< 1t +t2 f o r a l l s 2 , t 2E S 2 , s i n c e s1+s22 t l + t 2 y i e l d s a s above s1 2 t l . I f s1 = t l and s2 < t 2are assumed, a g a i n by I1 w e g e t s 1+ s 2 <= t +t2.
p l i c a t i o n of
of (4.5) which i s o n l y c l a i m e d f o r s 1 + s 2 ft l + t 2 .Now (4.6) f o l l o w s by n l + o < o+s2 < p 1 +o f o r a l l s 2 c S 2 .
This completes t h e procf
I f t h e extreme c a s e s a r e e x c l u d e d , t h e r e are
tl
E
S1 and a 2 E S 2
s a t i s f y i n g o < t , and o < a 2 , o r t l < o and a 2 < o . The f i r s t i m p l i e s S2 < t l by (4.6) and hence [a,]
tl
c
[a,].
SO
< t,
,
the l a t t e r likewise
(S,+,<) i s n o t archimedean o r d e r e d . I n t h e e x t r e m e t h e o r d e r of S may be
c a s e s , which r e a l l y o c c u r ( c f . E x p 1 . 5.61, archimedean or n o t .
Note t h a t (4.5) c o r r e s p o n d s t o t h e u s u a l l e x i c o g r a p h i c o r d e r of S1xs2 i f f
s 1 + s 2 = s l + t 2 always i m p l i e s
s2 = t 2 . Moreover,
t h e r e are
v a r i o u s seminearrings S s a t i s f y i n g S = S1+S2, i n p a r t i c u l a r t h o s e f o r which (S,+) i s a s s o c i a t i v e and c a n c e l l a t i v e . The l a t t e r are s e m i d i r e c t sums of
( S l y + ) and ( S 2 , + ) ,
and a c o n s t r u c t i o n method
s t a r t i n g from g i v e n n e a r r i n g s S1 and S 2 due t o [ l o 1 c a n b e g e n e r a l i z e d c o r r e s p o n d i n g l y , a l s o i n c l u d i n g p a r t i a l o r f u l l o r d e r relat i o n s . The most s i m p l e case ( c f . P r o p . S . 4 ) i s t o d e f i n e t h e opera-
289
Pariially and fully ordered seminear-rings and neaerings
t i o n s on
S1xS2
by
( a , . a 2 ) + ( b l , b 2 ) = ( a l + b l , a 2 + b 2 ), ( a l , a 2 ) ( b l y b 2 )= ( a l b l , a 2 b 2 ) .
(4.7)
A s u s u a l f o r r i n g s and n e a r r i n g s ( c f . [ill, 1 . 5 5 ) , we t h e n c a l l
t h e ( e x t e r n a l ) d i r e c t sum of t h e s e m i n e a r r i n g s S1
S = (S,xS2,+,*)
and S 2 . I f b o t h have a m u l t i p l i c a t i v e l y i d e m p o t e n t z e r o (as assumed i n our context) ,
al
(a, ,02)
-+
and a 2 -+ (ol , a 2 ) p r o v i d e monomorphisms
o f Si i n t o S. So one can change from ( a l , a 2 ) t o t h e n o t i o n a l + a 2 and t o t h e c o r r e s p o n d i n g i n t e r n a l d i r e c t sum. W e need t h i s f o r t h e n e x t theorem and d e a l w i t h r e f e r e n c e s i n i t s p r o o f . THEOREM 4 . 3 .
a)
(S,+,.,6) b e
Let
a f.
0.
nearrinq
and assume
S l = f { o l = / = S 2 f o r t h e c o n s t a n t and zero-symmetric s u b n e a r r i n g S1 and S of S. Then t h e o r d e r of S i s non-archimedean and u n i q u e l y d e t e r 2 mined by t h e i n d u c e d o r d e r s on Si a s t h e l e x i c o g r a p h i c o r d e r on
.
.
a c c o r d i n g t o ( 4 . 5 ) I n p a r t i c u l a r , w e have ( 4 . 6 ) 1 2 b ) There a r e f . 0 . n e a r r i n g s S which a r e t h e d i r e c t sum of t h e i r S = S +S
subnearrings
Sl$.{o}$S2
such t h a t b o t h have no z e r o d i v i s o r s .
c) L e t (S,+,*,S) be a f . 0 . d e t e r m i n e s t h e o r d e r of More p r e c i s e l y , a f .
0.
nearring.
(S,+,-),
Then t h e p o s i t i v e cone P = P ( S )
b u t by no means ( S , + , - , < )
seminearring
(P,+,',S)
cone P ( S ) = P ( T ) of non-isomorphic f . 0 . Proof. P a r t a ) i s a c o r o l l a r y are i m p o s s i b l e f o r a f . 0 .
itself.
may be t h e p o s i t i v e
n e a r r i n g s S and T .
of Thm. 4 . 2 ,
s i n c e t h e extreme c a s e s
n e a r r i n g . I t c o r r e s p o n d s t o [ill, 9.141
and g e n e r a l i z e s Lemma 1 of [91. P a r t b ) i s shown by Expl. 5 . 5 . c o n t r a d i c t s Thm. 1 of S S
1 2
[a]
( [ 11I
, 9.146) , where
It
it i s used t h a t
= S S = to) h o l d s f o r a d i r e c t sum of n e a r r i n g s Siy which i s 2 1
n o t i m p l i e d by t h e above d e f i n i t i o n . For c ) c f . Expl. 5.1 a ) . THEOREM 4 . 4 .
Let (S,+;,<)
s o r b i n g z e r o . Then, f o r a l l
be a f . 0 . s2 E S 2
s e m i n e a r r i n g w i t h a l e f t aband p l c P ( S 1 ) , t h e p r o d u c t
s2p1 i s a d d i t i v e l y i d e m p o t e n t . Hence s 2 p 1 = o h o l d s i f no i d e m p o t e n t s e x c e p t o., i n p a r t i c u l a r i f (S,,+)
( S l y + )h a s is l e f t [ o r r i g h t ]
cancellative. P r o o f . A t f i r s t w e assume
s2s1c s1
s2 = p2 f P ( S 2 ) .
Then I I I r , ( 2 . 5 ) and
imply p2p1 E P ( S ) n S1 = P ( S 1 ), hence
p2p1 = o o r o < p 2 p l .
I n t h e l a t t e r case w e can a p p l y ( 4 . 6 ) t o g e t p 2 + p2 < P2P1' BY 111, and p l p l = p l y t h i s y i e l d s p p + p 2 p 1
H. J. Weinert
290
t a i n e d i n S2, from p, = e R p l = o it f o l l o w s t h a t a w i t h a l e f t i d e n t i t y i s zero-symmetric
f.
0.
nearring
( c f . [ l l l , 9.136 and 9 . 1 3 7 ) .
5. EXAMPLES The p u r p o s e of t h i s s e c t i o n i s t o c o m p l e t e s t a t e m e n t s and p r o o f s in
2 and § 4 ,
n o t t o g i v e i m p r e s s i v e examples. I n p a r t i c u l a r , w e
show t h a t a l l t y p e s o f w . p . 0 .
s e m i n e a r r i n g s r e a l l y o c c u r and t h a t
o u r r e s u l t s c o n c e r n i n g t h e p r o p e r t i e s (2.3) a r e f a i r l y c o m p l e t e . So
we p r e s e n t e a c h t y p e by a w . p . 0 . s e m i n e a r r i n g S chosen a s s t r o n g a s w e c o u l d do, e. g , a s a n e a r r i n g , w i t h a f u l l o r d e r , o r w i t h a ( l e f t ) a b s o r b i n g z e r o . For t h e same r e a s o n , a l l s e m i n e a r r i n g s c o n sidered i n t h i s section are ~ l s a cdditively associative. W e c l e a r l y u s e t h e f . 0 . n e a r f i e l d s of t h e t y p e s ( 1 1 1 1 ) = ( 1 5 ) and t h a t e a c h c o n s t a n t w. f . 0 . s e m i n e a r r i n g h a s The f o l l o w i n 9 examples a r e , a s f a r a s p o s s i -
and (1100) = ( 1 2 ) , t y p e (1011) = ( 1 1 ) .
b l e , arranged according t o § 2. 5 . 5 , 5.6 and 5.7,
Expls.
The i n s e r t e d Prop. 5.4
prepares
where t h e f i r s t b o t h c o n c e r n
4 . A l l proofs
have been o m i t t e d . They a r e s t r a i g h t f o r w a r d even i f sometimes a
l i t t l e b i t tedious. Example 5.1.
a ) L e t (2,+,-,6) be t h e f . 0 . r i n g of i n t e g e r s i n t h e
u s u a l meaning. Define a new m u l t i p l i c a t i o n by S= (X,+,o,S)
is a f.
0.
zero-symmetric
and W ( S ) = M R ( S ) = Wk(S) = { O } ,
r
aob = a . I b l .
Then
n e a r r i n g s a t i s f y i n g Mr(S) = S
hence of t y p e
(1000) = ( 8 ) . N o t e
t h a t t h e r i n g Z a s w e l l as S have t h e same f . 0 .
seminearring
( N o , + , - , < ) a s p o s i t i v e cone ( c f . Thm. 4.3 c ) ) .
b ) The zero-symmetric satisfies of t y p e
f.
0.
Mr(U) = M k ( U ) = U
subseminearring U = (-No,+ and
Wr(U)
=W,(U)
={03.
,o
,I )
of S
Hence it i s
(1010) = ( 1 0 ) .
Example 5.2.
be t h e f . 0 . g r o u p o b t a i n e d as t h e d i r e c t (Z,+,<),o r d e r e d l e x i c o g r a p h i c a l l y : a l < b l o r a l = bl , a 2 5 b 2 . D e f i n e a p r o d u c t by
L e t (S,+,<)
sum of two c o p i e s of (al ,a2) S (bl,b2)
**
Then ( S , + , - , S ) i s a zero-symmetric w. f . b u t n o t 111,. Example 5.3.
This implies type a ) A zero-symmetric
0.
nearring satisfying I V
(0100) = ( 4 ) w.f.0.
by Thm. 2 . 4 .
s e m i r i n g (S,+,-,S) which
Mr(S) = W r ( S ) = N ( S ) = {o} and MR(S)= Wk(S) = P ( S ) = S g i v e n on S = { o , a , b , c , d } by satisfies
is
r’
29 1
Partially and fully ordered seminearrings and nearrings
o < a < b < c < d
Hence S i s of t y p e
b
b
b
d
d
o
b
b
b
d
d
d
d
d
o
a
a
a
(0111) = (7).
R e c a l l t h a t t h e r e a r e no
n e a r r i n g s of t h i s t y p e , and t h a t S c a n n o t b e p . 0 . by s i d e . By t h e way,
is a w.f.0.
U = (a,b,d)
b
a
.
w. f .
0.
P ( S ) from one
s u b s e m i r i n g of S w i t h
t h e element a a s l e f t absorbing z e r o , p . 0 .
by
P ( U ) and of t y p e
(0011 ) = ( 3 ) .
d i s t r i b u t i v e , w e can s w i t c h t o i t s m u l t i p l i -
b ) S i n c e S i s two-sided
c a t i v e l y l e f t - r i g h t d u a l v e r s i o n , s a y ( S ' , + , - , S ) , a zero-symmetric f.0.
s e m i r i n g of t y p e
a f.0.
(1101) = (13).
n e a r f i e l d of type
which i s a l s o of t y p e
(1100)
Note t h a t t h e p o s i t i v e c o n e of
is a f . 0 .
s e m i n e a r f i e l d ( c f . [131)
(1101) = ( 1 3 ) .
PROPOSITION 5.4. L e t ( S 1 , + , ' , < ) be a c o n s t a n t and (S2,+,',S2) a 1 zero-symmetric w . p . 0 . s e m i n e a r r i n g , b o t h n o n - t r i v i a l l y P . o . , and S = S 1 x S 2 t h e d i r e c t sum o f ( S l , + ) and ( S 2 , + ) .
W e assume b o t h t o be
c a n c e l l a t i v e semigroups. D e f i n i n g a l s o t h e m u l t i p l i c a t i o n a s i n ( 4 . 7 ) ,
i s t h e d i r e c t sum of
(S,+,')
a) (S,+,*,S)
is a w . p . 0 .
graphic order
( S l , + , - ) and ( S 2 , + , * ) .
seminearring with r e s p e c t t o t h e l e x i c o -
( a l , a 2 ) I (bl , b 2 )
iff
P ( S ) c o n s i s t s of a l l ( p , , s 2 ) s u c h t h a t that
a l <1 b l 0,
or
< p1
a l = b l , a 2 I2 b 2 ,
a n d a l l (ol , p 2 ) such
and N ( S ) i s d e s c r i b e d s i m i l a r l y . I f
02Sp2,
P ( S 1 ) = (o.,},
p(S1) 9 t o l l , IIIr f o r S f a i l s t o be t r u e . I I I Q f o r S
I I I r f o r S and I I I r f o r S2 a r e e q u i v a l e n t . I f
f o r S holds i f f holds i f f
M (S2) = S 2 .
r
v a l i d , and I V Q i f f b ) (S,+,',<)
r
P ( S 1 ) = { o , } and
S2S2 = {021 or
is also a w.p.0.
antilexicographic order alIlbl.
IV
S 2 S 2 = {02) o r b o t h ,
P ( S 2 ) S 2 ={ 0 2 ] , are
N ( S 1 ) = { o l } and
N ( S 2 ) S 2 = {02}.
seminearring with r e s p e c t t o t h e
( a , , a 2 ) 5 ( bl , b 2 )
iff
a2
c2
b2 o r
For e a c h (S,+,-,S) o b t a i n e d i n t h i s way, I11
r
a2=
and I V
b2' al-
ways f a i l t o be t r u e , whereas I I I R h o l d s f o r S i f f it h o l d s f o r S2, and l i k e w i s e w i t h I V Q . T o g e t h e r w i t h o u r examples f o r S1 and S2, v a r i o u s examples, i n p a r t i c u l a r of w. f . 0 .
Prop. 5 . 4 p r o v i d e s
s e m i n e a r r i n g s S of t h e
t y p e s ( 1 1 ) t o (8) and ( 3 ) t o ( 0 ) . Moreover, one c a n u s e o t h e r m u l t i plications for
S = SlxS2.
A r a t h e r s i m p l e one i s g i v e n by
H.J. Weinert
292
1 y a2b2) ( a l b l y a2b2) = (a ( a , , a 2 )( b l , b 2 ) = { ( a l b l . 02) = ( a l y02)
But w e u s e h e r e o n l y Prop. 5.4 Example 5.5. F o r any f .
for for
=ol 1 b l $;ol.
b
t o g i v e some needed examples:
c o n s t a n t n e a r r i n g S l and any f . 0 .
0.
zero-
symmetric n e a r r i n g S2 s a t i s f y i n g Mr(S2) = S2 and S2S2$. { 0 2 } , t h e d i r e c t sum S a c c o r d i n g t o Prop. 5.4 a ) i s a f . 0. n e a r r i n g of t y p e In p a r t i c u l a r , S
(1000) = ( 8 ) .
chosen so ( c f , Expl. 5 . 1 ) . Example 5.6. A
1
h a s no z e r o d i v i s o r s and S2 may be
T h i s p r o v e s Thm. 4.3 b )
seminearring
f.0.
S such t h a t
.
S 1 5 0 2 S2
hold ( c f .
Thm. 4 . 2 ) i s o b t a i n e d by Prop. 5.4 a ) e . g . for S1 = ( - R o y + ,w
From Prop. 5.4 b ) w e o b t a i n w. f . 0 .
(0011) = ( 3 )
types
r i n g S1 and a f . 0 .
and
zero-symmetric
(1100) = ( 1 2 ) .
(1111) = (15) o r
(0010) = (2)
of t h e t y p e s
and
n e a r r i n g s S of t h e
u s i n g any c o n s t a n t f . o . n e a r -
(OOOO)=(O)
n e a r r i n g S 2 e i t h e r of t y p e
L i k e w i s e , w. f . o , (0001) = ( 1 )
seminearrings S
( t h e r e a r e no n e a r r i n g s
of t h i s k i n d ) a r e o b t a i n e d u s i n g zero-symmetric s e m i n e a r r i n g s S 2 with types ( . . l o )
e . g . t h o s e of E x p l s . 5.1 b ) and 5.8.
o r (..Ol),
,F
Example 5.8. W e d e f i n e two s e m i n e a r r i n g s S and T on t h e same set {n,o,a,p} with the f u l l order
0
n
n
a
p
n
o
a
p
P
P
P
P
n < o < a < p by t h e t a b l e s o
a
p
n
n o
na
n p
0
a P
n
0
0
0
a
n
a
a
p
P
n
P
P
P
c l e a r l y b o t h zero-symmetric. The f i r s t one,
(S,+,*,<)
0
0
0
0
,
is a f . 0 .
s e m i n e a r r i n g such t h a t M ( S ) = WL(S) = S, b u t Wr(S) = S x { n } and r MQ(S) = Sx{p} h o l d , t h u s o f t y p e ( l o o t ) = ( 9 ) . Similarly, the other one i s a w. f . 0 . s e m i n e a r r i n g ( T , + , - , < ) of t y p e ( 0 1 1 0 ) = ( 6 ) . Example 5 . 9 .
Each p.
multiplication
0.
semigroup
s t = s a p.
0.
(S,+,<)
p r o v i d e s by t h e t r i v i a l
s e m i n e a r r i n g ( S , + , - , S ) such t h a t
M ( S ) = M (S) = W ( S ) = S and W r ( S ) = @ h o l d .
r
II ( l o l l ) = (11)
R
iff
N(S)$. @
( c f . Lemma 4 . 1 ) .
Thus S h a s t h e t y p e C l e a r l y , one o b t a i n s
(S,+,*,S) is a i s i d e m p o t e n t . S i n c e t h e r e a r e v a r i o u s even semigroups (S,+,I) o f t h i s k i n d s a t i s f y i n g N ( S ) +$, e . g - w i t h
f . 0 . n e a r r i n g s S i n t h i s way. On t h e o t h e r hand, p. o s e m i r i n g i f f.
0.
(S,+)
Partially and fully ordered seminearvings and nearrings
293
a zero, we get f.0. semirings S with a left absorbing zero of type (11). Now we can turn to the multiplicatively left-right dual f.0. semiring S ' which has the type (1110) = (14). Example 5.10. For the f. 0. group ( S , + , S ) of Expl. 5.2 define a product by (al,a2) (bl,b2) = (alb2+ a2bl,a2b2). Then (S,+, * , S ) is a commutative w. f.0. ring of type (0000), whereas its positive cone U=P(S) is a zero-symmetric w. f.0. semiring of type (0101) = (5). The latter was the last type needed to prove Remark 2.2. NOTES AND REFERENCES Footnote 1) The importance of these concepts is that one can introduce partial orders on groupoids (S,+) by (2.4) using suitable subsets X E S . For semigroups, this was extensively investigated in [151, and we note here only the following special result. Let (S,+) be a semigroup with a zero o and X E S . Then (2.4) defines a relation on S such that ( S , + , < ) is a p. 0. semigroup iff x+y = o for x,y E X implies x = y = o and c+X S X+c holds for all C E S . In this case, X coincides with the set of positive elements of ( S , + , < ) . Bibliography Fuchs, L., Partially ordered algebraic systems, Pergamon Press, Oxford 1963, Vandenhoeck & Ruprecht, Gottingen 1966. Hanumanthachari, J., K. Venu Raju and H. J. Weinert, Some results on partially ordered semirings and semigroups, Proc. First Intern. Symposium on Ordered Algebraic Structures, Marseilles, June 1984, Heldermann-Verlag, Berlin 1986. Hogewijs, H., Semi-Nearrings-Embedding, Med. Konink. Acad. Wetensch. Lett. Schone Kunst. Belgie K1. Wetensch. 32 (1970) 3 - 11. Karzel, H., Zusammenhange zwischen Fastbereichen, scharf zweifach transitiven Permutationsgruppen und 2-Strukturen mit Rechtecksaxiom, Abh. Math. Sem. Univ. Hamburg 32 (1968), 191 - 206. Kerby, W., Angeordnete Fastkorper, Abh. Math. Sem. Univ. Hamburg 32 (1968), 135 - 146. Kerby. W. and H. Wefelscheid, Bemerkungen iiber Fastbereiche und scharf zweifach transitive Gruppen, Abh. Math. Sem. Univ. Hamburg 37 (1972), 20 29.
-
Pilz, G., Geordnete Fastringe, Abh. Math. Sem. Univ. Hamburg 35 (19701, 83 - 88. Pilz, G., Direct Sums of nrdered Near-Rings, J. Algebra 10 (1971), 340 - 342.
H.J. Weinert
294
[9] Pilz, G. , Zur Charakterisierung der Ordnungen in Fastringen, 253. Monatsh. Math. 76 (1972), 250
-
[lo] Pilz, G., A construction method for near-rings, Acta Math. Acad. Sci. Hungar. 24 (1973), 97 - 105. [ll] Pilz, G., Near-Rings, North-Holland, Amsterdam 1977. [12] Weinert, H. J., Related representation theorems for rings, semirings, near-rings and semi-near-rings by partial transformations and partial endomorphisms, Proc. Edinburqh Math. SOC. 20 (1976 77), 307 315.
-
-
1131 Weinert, H. J., Seminearrings, seminearfields and their semigroup-theoretical background, Semigroup Form 24 (1982), 231 - 254. 1141 Weinert, H. J., Extensions of seminearrings by semigroups of right quotients, Lecture Notes in Mathematics 998 (19831, Proceedings Oberwolfach (1981), 412 - 486. 1151 Weinert, H. J., Partially ordered semirings and semigroups, Proc. First Intern. Symposium on Ordered Algebraic Structures, Marseilles, June 1984, Heldermann-Verlag, Berlin 1986.
Institut fur Mathematik Technische Universitat Clausthal ErzstraBe 1 D-3392 Clausthal-Zellerfeld
Near-ringsand Near-fzlds, G . Betsch (editor) 0 Elsevier Science Publishers B.V. (North-Holland), 1987
295
ON SUBDIRECTLY IRREDUCIBLE NEAR-RINGS WHICH ARE FIELDS Richard WIEGANDT Mathematical I n s t i t u t e o f t h e Hungarian Academy o f Sciences Budapest, P f . 127, H-1364, Hungary Herrn H. J. Weinert zum 60. Geburtstag am 26.01.1987
gewidmet
I n t h i s n o t e we s h a l l prove under c e r t a i n c o n d i t i o n s t h a t i f a l e f t i n v a r i a n t subset o f a s u b d i r e c t l y i r r e d u c i b l e n e a r - r i n g s a t i s f i e s a permutation i d e n t i t y , then t h e n e a r - r i n g i s a f i e l d . I n t h e sequel a near-ring w i l l always mean a r i g h t n e a r - r i n g , t h a t i s , i t s a t i s f i e s t h e r i g h t d i s t r i b u t i v e law ( x + y ) z = xz L e t us r e c a l l t h a t an i d e a l group o f
A
I
t
yz.
o f a near-ring
A
i s an a d d i t i v e normal sub-
such t h a t
I A r I and x(y holds f o r a l l
x, Y E A and
symmetric n e a r - r i n g
A
t
i)
-
. By P i l z
iE I A
[31 1.34 ( a ) any i d e a l
A IE I
satisfies also
subdirectly i r r e d u c i b l e , i f
xy E I
. A near-ring
c o n t a i n s a unique minimal i d e a l
i s r e f e r r e d t o t h e heart o f A . We s h a l l c a l l a subset
Zeft invariant subset, i f A L c L we s h a l l denote by Sk t h e s e t s k = Is = tl
holds. For a subset
... tk : tl
f o r any i n t e g e r
k 2 2
L
I . . . ,
tk E
A
I of a 0 i s s a i d t o be
H # 0
which
o f a near-ring
S
o f a near-ring
A
a A
SI
.
I n the f o l l o w i n g k w i l l always stand f o r a f i x e d i n t e g e r 2 2 , and (7 w i l l stand f o r a f i x e d permutation o f k elements such t h a t a ( 1 ) # 1 The n e x t Lemma i s p u r e l y semigroup t h e o r e t i c a l , and i t i s a m o d i f i e d s p e c i a l
.
case o f [ 2 1 Lemma 2. For the sake o f completeness we g i v e i t w i t h p r o o f . LEMMA. If a l e f t invariant subset mutation i d e n t i t y (PI
t,
L
o f a near-ring
... t k = t o ( , ) ... t o ( k l
tiEL,
A
s a t i s f i e s t h e per-
R. Wiegandt
296
then
... tk = y x t l ... tk tlY...'tkEL .
xytl
holds f o r a l l
x, Y E A and
roof. L e t r = yxtl
... tk
i d e n t i t y ( P ) t o the elements
and
xtl,
o(i) = 1 t2,
...) tk
u(1) = j
. Now we apply
and then m u l t i p l y by
y.
the Thus
we g e t
... to ( .-l)Xtu( i)to( i + l )
r = ytu(l)
i
* ''
to( k )
'
Now we use the i d e n t i t y ( P ) i n reverse on the elements Y t 0 ( 1) 'tU(2)
..ta(
Y..
1' - 1 1
SXtu(i
y t u ( i + l
' . * * I
tu(k)
t o obtain
... t k . ,...,tjm1,ytj,
... t J-1 .
y t. t. J J+1 A t t h i s p o i n t we use again ( P ) on the elements tl r=xtl
tj+l
,...,t k
t o obtain r = x y t F i n a l l y , we use ( P ) i n reverse on
o ( 1 ) %(2
t o obtain
tu(,],.,. t o ( k )
r = x y t
1 '
*
tk
Y
and the desired i d e n t i t y has been established. Another proof o f the Lemma has been given by H. J. Weinert ( p r i v a t e commun i c a t i o n ) . The s t r u c t u r e o f near-rings s a t i s f y i n g ( P ) has been i n v e s t i g a t e d by R. Scapellato [41.
be a subdirectly irreducible near-ring w i t h heart H , and A such t h a t H Lk # 0 , If L s a t i s f i e s t h e permutation i d e n t i t y (P), then A i s a commutative ring. If, i n addition, H2 0 , then A i s a f i e Z d . A 0 - s y m e t r i c , subdirectly i r r e d u c i b l e near-ring with commutative idempotent heart i s a f i e l d . THEOREM. Let
A
L be a l e f t invariant subset of
+
Proof. By the Lemma we have
... t k = y x t , ... t k ,
x y t , that is, (xy x, y E A
for a l l
and
t,,
-
y x ) tl
... tk = 0
...'tk E L . This means t h a t t h e element
i s contained i n t h e a n n i h i l a t o r k k ( 0 : L ) = {a E A : aL I = 0 k o f Lk One can e a s i l y check t h a t ( 0 : L ) i s an i d e a l i n A . k k subdirectly irreducible, either (0 : L ) = 0 or H G (0 : L ) k case we would get H L = 0 c o n t r a d i c t i n g t h e assumption. Thus
.
holds, implying
.
xy = y x
for a l l
x, y E A
. This means t h a t i n
xy
Since
- yx A
is
I n the l a t t e r k (0 : L ) = 0
A
the multi-
291
On subdirectly irreducible near-rings which are fields A
p l i c a t i o n i s commutative, and hence
i s a d i s t r i b u t i v e n e a r - r i n g which i s
a l s o c a l l e d a r i n g w i t h n o t n e c e s s a r i l y commutative a d d i t i o n . As one can prove (see f o r i n s t a n c e [l],151 o r [61), each a d d i t i v e commutator
(x, y E A)
,
i s contained i n the a n n i h i l a t o r
A
i s an i d e a l i n
and
A
( 0 : A)
of
-x
-
y
t
. Since
A
i s subdirectly irreducible, i t follows
x
t
y
,
( 0 : A)
,
( 0 : A) = 0
otherwise namely we would g e t k O # H L S ( O : A ) A = O . Thus we have x t y = y t x x, y E A
for all
A
and t h e r e f o r e
i s a commutative r i n g .
The h e a r t o f a s u b d i r e c t l y i r r e d u c i b l e r i n g i s e i t h e r simple o r a z e r o - r i n g . If
ideal A
, then
H2 # 0
H of
. Since
A
A
i s a s i m p l e commutative r i n g , t h a t i s , a f i e l d . Hence t h e
H
has a u n i t y element and t h e r e f o r e
H
i s a d i r e c t summand o f
i s subdirectly irreducible, i t follows
H = A
.
The l a s t a s s e r t i o n i s s t r a i g h t f o r w a r d . For r i n g s we can prove a s t r o n g e r v e r s i o n o f t h e Theorem. A be a s u b d i r e c t l y i r r e d u c i b l e ring w i t h idempotent heart i s a l e f t i n v a r i a n t subset of A , then H Ln # 0 holds f o r
PROPOSITION. Let H
. If
every
L # 0
.. .
n = 1,2,.
We prove t h e statement by i n d u c t i o n . F i r s t , assume t h a t ( 0 : H)r
right annihilator
( 0 : H)r
Further, yields
of
H in A
i s an i d e a l o f
2 0 # H = H E H(O : H)r = 0
.
n = 1 Next, assume t h a t
valid for
(H Ln-')L = H Ln = 0 (H L"')
. Hence
This y i e l d s f o r every
H L"'
n = 1,2,
,
, so
L
we conclude
, and
H L = 0 hence
. Then t h e (0 : H),#
H C ( 0 : H)r
0.
which
a c o n t r a d i c t i o n . Thus t h e a s s e r t i o n i s
for
n 2 2
the l e f t annihilator
(0 : L )
H L = 0
= 0
A
contains
.
Suppose t h a t
H L # 0
... .
. Then by
(0 : L)
i s a nonzero i d e a l o f
, contradicting
H Ln = 0
.
o f L i n A contains A and so we have H r ( 0 : L ) Thus H Ln # 0 has been proved
The Theorem and t h e P r o p o s i t i o n y i e l d immediately t h e f o l l o w i n g COROLLARY. Let
A
be a s u b d i r e c t l y i r r e d u c i b l e ring w i t h idempotent h e a r t .
I f a l e f t i n v a r i a n t subset (P), then
A
L # 0
of
A
s a t i s f i e s t h e permutation i d e n t i t y
is a f i e l d .
Rermrk. The P r o p o s i t i o n and hence t h e C o r o l l a r y a r e v a l i d a l s o f o r d i s t r i b u t i v e near-rings,
t h a t i s , f o r r i n g s w i t h n o t n e c e s s a r i l y commutative a d d i t i o n .
REFERENCES [ 11 Furtwangler, Ph. and Taussky, Olga, Ober S c h i e f r i n g e , Sitzungsber. Akad.
der Wiss. Wien, Mathem.-natum.
K l a s s e , 145 (1936), 525.
.
296
R. Wiegondt
[ 2 1 Parmenter, M. M., Stewart, P. N. and Wiegandt, R., On the Groenewald man s t r o n g l y prime r a d i c a l , Quuest. Math. 7 (1984), 225-240. [31 P i l z , G. , Near-rings, North-Holland, 1977.
-
Hey-
, Sui q u a s i - a n e l l i v e r i f i c a n t i i d e n t i t a semigruppale C-mobili, BoZZettino U.M.I. (6) 4-8 (1985), 789-799. [5] Taussky, Olga, Rings w i t h non-commutative a d d i t i o n , BUZZ. C Q Z C U t t Q MQth.
[4J Scapellato, R.
SOC., 28 (19361, 245-246. [6] Weinert, H. J., Ringe m i t nichtkomnutativer A d d i t i o n I , Jber. Deutsch. &th.-Verein. 77 H. 1 (1975). 10-27.