THE STRUCTURE OF BN by RODNEY SEUNARINE RAMBALLY B.A. (Hons.), University of Saskatchewan, 1969 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER- OF ARTS i n the Department of MATHEMATICS We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA August, 1970 I n p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f t h e r e q u i r e m e n t s f o r a n a d v a n c e d d e g r e e a t t h e U n i v e r s i t y o f B r i t i s h C o l u m b i a , I a g r e e t h a t t h e L i b r a r y s h a l l m a k e i t f r e e l y a v a i l a b l e f o r r e f e r e n c e a n d s t u d y . I f u r t h e r a g r e e t h a p e r m i s s i o n f o r e x t e n s i v e c o p y i n g o f t h i s t h e s i s f o r s c h o l a r l y p u r p o s e s may be g r a n t e d by t h e H e a d o f my D e p a r t m e n t o r by h i s r e p r e s e n t a t i v e s . I t i s u n d e r s t o o d t h a t c o p y i n g o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l n o t be a l l o w e d w i t h o u t my w r i t t e n p e r m i s s i o n . D e p a r t m e n t T h e U n i v e r s i t y o f B r i t i s h C o l u m b i a V a n c o u v e r 8, C a n a d a Supervisor: Professor T. Cramer i i ABSTRACT Our subject matter consists of a survey of the major results concerning the topological space gN-N where N represents the space of natural numbers with the discrete topology, and gN the Stone- • Cech compactification of N . We are mainly concerned with the results which were derived during the l a s t ten years. When there i s no advantage i n r e s t r i c t i n g our work to the space N we work with an arbitrary discrete space X and f i n a l l y formulate our results i n terms of gN-N . In some cases, pre-1960 results concerning . gN-N are obtained as special cases of the results we derive using an arbitrary discrete space X . The material presented i s divided into four chapters. In Chapter I, we discuss certain subsets of gN-N which can be C*-embedded i n other subsets of gN-N . This study leads to the conclusion that no proper dense subset of gN-N can be C*-embedded. In the second chapter we devise a general method of associating certain classes of points of gN-N with certain subalgebras of C(N) .. The P-points of gN-N form one of these classes. The answer to R. S. Pierce's question, "Does there exist a point of gN-N which l i e s simul-taneously i n the closures of three pairwise d i s j o i n t open sets" i s discussed i n Chapter I I I . F i n a l l y i n Chapter IV we present two proofs of the non-homogeneity of gN-N , without the use of the Continuum Hypothesis. I l l ACKNOWLEDGEMENTS I wish to express my appreciation and gratitude to my advisor, Dr. Timothy Cramer, for suggesting the topic and .offering invaluable assistance i n the preparation of this thesis. I also wish to thank Dr. J. V. Whittaker for reading the thesis and Mrs. M. Wang for s k i l l f u l l y typing i t . To National Research Council of Canada go my thanks for the i r f i n a n c i a l assistance. i v TABLE OF CONTENTS Page INTRODUCTION 1 CHAPTER I. ON THE EMBEDDING OF CERTAIN SETS IN BN . . . . . . 4 CHAPTER II. AN APPLICATION OF B-SUBALGEBRAS OF C(N) TO BN-N . 19 CHAPTER III. c-POINTS OF gN-N 32 CHAPTER IV. TYPES OF ULTRAFILTERS AND THE NON-HOMOGENEITY OF BN-N 41 BIBLIOGRAPHY 60 1 INTRODUCTION Our purpose i s to investigate the internal structure of gN-N . We do this with the help of the major results of the l a s t ten years concerning this topological space. Most of the pre-1960 results i n connection with gN can be found i n the Gillman and Jerison text, {4], and, when needed, we simply quote these results without proofs. Some of them w i l l be eas i l y obtained from certain theorems which we discuss here. In some cases we do not r e s t r i c t our considerations to the space N , simply because such a r e s t r i c t i o n does not decrease our work appreciably. In such cases we discuss an arbitrary discrete space and f i n a l l y apply our results to the space N . Much of the work i n the f r i s t three chapters depends heavily on the continuum'hypothesis. Thus i t i s expected that one of the major techniques used i n these chapters i s the induction process, as i s indeed the case. The terminology and notations we use are those of the Gillman and Jerison text, [4]. However we would l i k e to make special mention of the following notations: N* = gN-N (or, more generally, D* = gD-D where D i s a discrete space); c w i l l denote the power of the continuum; , for f a function on N , w i l l denote the continuous extension of f to gN . We sometimes denote this continuous extension by f* . Unless otherwise stated we denote "closure" by "c£" and " i n t e r i o r " by " i n t " . The set of a l l continuous functions from the topological space X into the topological space R (space of r e a l numbers) i s denoted by C(X) and the set of a l l bounded functions i n 2 C ( X ) i s d e n o t e d b y C * ( X ) . A s u b s p a c e S o f X i s s a i d t o b e C -e m b e d d e d i n X i f e v e r y f u n c t i o n i n C ( S ) c a n b e e x t e n d e d t o a f u n c t i o n i n C ( X ) . S i m i l a r l y S i s C * - e m b e d d e d i n X i f e v e r y f u n c t i o n i n C * ( S ) c a n b e e x t e n d e d t o a f u n c t i o n i n C * ( X ) . F o r f e C ( X ) , we l e t Z ( f ) = {x e X : f ( x ) = 0} , a n d we c a l l Z ( f ) a z e r o - s e t o f X . T h e s e t o f z e r o - s e t s i n X i s d e n o t e d b y Z ( X ) • T h e c o m p l e m e n t o f a z e r o - s e t i s c a l l e d a c o z e r o - s e t . C h a p t e r I . We i n v e s t i g a t e t h e n e c e s s a r y a n d s u f f i c i e n t conditions for subsets of 8N to bs C*-embadded. The results of t h e c h a p t e r a r e f o r m u l a t e d : i n t e r m s o f a c e r t a i n c o n d i t i o n d e f i n e d i n t h e c h a p t e r i t s e l f . I t i s f i n a l l y p r o v e d h e r e t h a t d e n s e s u b s e t s o f N * a r e n o t C * - e m b e d d e d . C h a p t e r I I . - H e r e we i n t r o d u c e a m e t h o d f o r a s s o c i a t i n g a s e t o f p o i n t s i n g X - X w i t h a t y p e o f s u b a l g e b r a A o f t h e a l g e b r a C ( X ) w h e r e X i s a c o m p l e t e l y r e g u l a r s p a c e . We c a l l t h e s e p o i n t s t h e A - p o i n t s o f g X - X . We t h e n p r o v e t h a t t h e s e t o f A - p o i n t s i s d e n s e i n g X - X a n d i s o f c a r d i n a l i t y 2° . We s h o w t h a t C * ( N ) i s o n e o f t h e s e t y p e s o f s u b a l g e b r a s o f C ( N ) a n d t h a t t h e P - p o i n t s o f g N - N a r e t h e C * ( N ) - p o i n t s , f r o m w h i c h we c o n c l u d e t h e w e l l k n o w n r e s u l t t h a t g N - N h a s a d e n s e s u b s e t o f 2 P - p o i n t s . C h a p t e r I I I . T h i s i s a s h o r t c h a p t e r i n w h i c h we p r o v e t h a t e v e r y p o i n t i n g N - N i s c o n t a i n e d s i m u l t a n e o u s l y i n t h e c l o s u r e s 3 o f c d i s j o i n t o p e n s e t s o f BN-N . C h a p t e r I V . We d i s c u s s t y p e s o f u l t r a f i l t e r s a n d i n t r o d u c e a s e t T , t h e e l e m e n t s o f w h i c h we c a l l t h e t y p e s o f t h e f r e e u l t r a f i l t e r s o n N . We t h e n d e f i n e a c e r t a i n r e l a t i o n (j) o n T c a l l e d t h e p r o d u c i n g r e l a t i o n . U s i n g t h i s c o n c e p t a n d w i t h o u t a s s u m i n g t h e c o n t i n u u m h y p o t h e s i s we p r o v e t h e n o n - h o m o g e n e i t y o f BN-N . \ / 4 C H A P T E R I ON T H E E M B E D D I N G O F C E R T A I N S E T S I N gN 1 . 1 . T h e p u r p o s e o f t h i s c h a p t e r i s t o c o n s t r u c t , c e r t a i n s e t s i n gN t h a t c a n b e C * - e m b e d d e d i n o t h e r s u b s e t s o f gN . We s h a l l a l s o • p r o v e t h a t i f p e 3N-N t h e n g N - N - { P } i s n o t C * - e m b e d d e d i n 3N-N f r o m w h i c h we s h a l l b e a b l e t o c o n c l u d e t h a t p r o p e r d e n s e s u b s e t s o f g N - N a r e n o t C * - e m b e d d e d . We u s e t h e c o n c e p t s o f F - s p a c e s a n d z e r o -d i m e n s i o n a l s p a c e s i n m o s t o f o u r p r o o f s . A l l s p a c e s c o n s i d e r e d a r e a s s u m e d t o b e c o m p l e t e l y r e g u l a r . T h e r e s u l t s h e r e a r e d u e m a i n l y t o F i n e a n d G i l l m a n . 1 . 2 . D e f i n i t i o n . X i s c a l l e d a n F - s p a c e i f e v e r y c o z e r o - s e t i n X i s C * - e m b e d d e d . I t i s p r o v e d i n [ 4 ] , 1 4 . 2 7 , t h a t i f . a s p a c e X i s -l o c a l l y c o m p a c t a n d a - c o m p a c t b u t n o t c o m p a c t t h e n g X - X i s a c o m p a c t F - s p a c e . N i s a n e x a m p l e o f s u c h a - ' s p a c e . We n o w p r o c e e d t o d e f i n e z e r o - d i m e n s i o n a l . L e t a n d S h e f a m i l i e s o f s e t s . We s a y t h a t ~V r e f i n e s H i f f o r e v e r y V e t h e r e e x i s t s U e "It s u c h t h a t V ^ U . L e t b e a c o v e r o f X . We d e f i n e t h e o r d e r o f *LL- t o b e t h e l a r g e s t i n t e g e r n f o r w h i c h t h e r e a r e n + 1 m e m b e r s o f "IL w i t h n o n - e m p t y i n t e r s e c t i o n . B y a n o p e n - c l o s e d p a r t i t i o n o f S C X w e m e a n a f i n i t e c o l l e c t i o n o f d i s j o i n t o p e n a n d 5 c l o s e d s u b s e t s o f S w h o s e u n i o n i s S . T h u s a p a r t i t i o n o f S i s a c o v e r o f S o f o r d e r z e r o . 1 . 3 . D e f i n i t i o n . A s p a c e X i s s a i d t o b e z e r o - d i m e n s i o n a l i f e v e r y b a s i c c o v e r o f X h a s a r e f i n e m e n t b y a n o p e n - c l o s e d p a r t i t i o n o f X . R e c a l l t h a t t w o s u b s e t s A a n d B o f a s p a c e X a r e c o m p l e t e l y s e p a r a t e d i n X i f t h e r e e x i s t s a f u n c t i o n f i n C 5 ' C ( X ) s u c h t h a t 0 _< f _< 1 , f ( x ) = 0 f o r a l l x e A , a n d f ( x ) = 1 f o r a l l x e B . I t i s p r o v e d i n [ 4 ] , 1 6 . 1 7 , t h a t t h e f o l l o w i n g a r e e q u i v a l e n t : a n y t w o c o m p l e t e l y s e p a r a t e d s e t s a r e c o n t a i n e d i n c o m p l e m e n t a r y o p e n a n d c l o s e d s e t s f o r a s p a c e X ; i s t o t a l l y d i s c o n n e c t e d ; - X i s z e r o - d i m e n s i o n a l . I t i s a l s o p r o v e d i n [ 4 ] , 6 . 1 0 , t h a t gN i s t o t a l l y d i s c o n n e c t e d . F o r a n y s p a c e Y , l e t r Y d e n o t e t h e s e t o f a l l p o i n t s p i n gY s u c h t h a t e v e r y z e r o - s e t i n gY t h a t c o n t a i n s p a l s o i n t e r s e c t s Y . 1 . 4 . D e f i n i t i o n . I f r Y = Y , t h e n Y i s s a i d t o b e r e a l c o m p a c t . I n C h a p t e r I I , w e s h a l l f i n d i t m o r e c o n v e n i e n t t o d e f i n e r e a l c o m p a c t n e s s a s f o l l o w s : Y i s r e a l c o m p a c t i f f o r e v e r y p e Y * t h e r e e x i s t s Z e Z ( g Y ) s u c h t h a t p e Z <Z Y* . We s h a l l n o w s h o w t h a t t h e s e t w o d e f i n i t i o n s a r e e q u i v a l e n t . L e t t h e f i r s t d e f i n i t i o n b e I a n d t h e 6 / second be I I . I implies II : I f p e Y* , then p £ Y . Thus p £ rY . so for some zero-set Z containing p , Z 0 Y = <j> . Hence Z c Y* and so p e Z c Y* . II implies I : Let y e rY . Then y £ BY - Y for i f i t were then there would e x i s t a zero-set Z such that y e Z c Y* . Hence Z n Y = <j> , a cont r a d i c t i o n . Now i f y e Y then obviously every zero-set containing y i n t e r s e c t s Y . Hence Y C rY . Thus Y = rY . Hence the two d e f i n i t i o n s are equivalent. The following lemma i s needed both i n t h i s chapter and i n the next. 1.5. Lemma. I f Y i s l o c a l l y compact and realcompact, then, each zero-set i n BY-Y i s the closure of i t s i n t e r i o r . Remark. I t s u f f i c e s to prove that a non-empty zero-set Z i n BY-Y has non-empty i n t e r i o r ; f o r then i f there exists p e Z - cl i n t Z, then from l o c a l compactness there e x i s t s a zero-set Z^ , d i s j o i n t from i n t Z , such that p e Z^ and i n t ( Z ft Z^) = <j> . Since zero sets are closed under countable i n t e r s e c t i o n we then have a non-empty zero-set, Z^ f) Z , with empty i n t e r i o r , contradicting the f a c t that a non-empty zero set has non-empty i n t e r i o r . Proof: Since Y i s l o c a l l y compact then BY-Y i s compact (this follows from the'fact that Y i s open i n BY i f and only i f Y i s l o c a l l y compact. Thus Y i s open i n BY thereby making BY-Y a closed subset of a compact space, hence compact). Thus BY-Y i s C*-embedded i n 7 3Y by [4], 6.9(b). Thus for Z an arbitrary zero-set i n 3Y-Y , Z = Z(f) - Y for some f e C * ( 3 Y ) . Let p e Z . By realcompactness p £ rY . Thus there exists a function g e C*(3Y) such that g(p) = 0 but g(y) ^ 0 for any y e Y . Let h = |f| + |g| . Then h(p) = 0 so p e Z(h) . Also Z(h) C Z . Thus p e Z(h)C Z . We s h a l l show that this non-empty set Z(h) contains a non-empty open set. Let be a sequence of d i s t i n c t points i n Y on which h converges to zero. By l o c a l compactness we can choose d i s j o i n t compact neighborhoods of y such that |h(y) - h(y ) l < — for y e V . Let u e C * ( 3 Y ) n n ' n n b e s u c h t h a t u(y^) = 1 for every n u(y) = 0 , y e Y - V n (such a function u exists because for every n there exists a function f^ which i s equal to 1 at y^ and equal to 0 everywhere on Y-V . Let u = \ f ). Suppose q £ 3Y-Y i s such that u(q) 4 0 . n , This, means that every neighborhood of q meets i n f i n i t e l y many of the compact sets . Hence h(q) = 0 ; one may j u s t i f y this as follows: since every neighborhood of q meets i n f i n i t e l y many sets , then we can find a net {S } such that each S i s i n V and {S } converges n n n n ° to q . Since h i s continuous we have lim h(S ) = h(q) . Now n-*° |h(S ) - h(y )| < — . So lim h(S ) = 0 . Thus h(q) = 0 . Hence for ' n w n 1 n n n N any point q e 3 Y-Y such that u(q) $ 0 , we have h(q) = 0 . Thus (BY-Y) - <£ Z(h) ,. Also ( 3 Y-Y) - Z(u) i s a non-empty open subset 8 of BY-Y . Since Z(h)<r Z we conclude that Z has non-empty i n t e r i o r . 1.6. Remark. Local compactness i s a c r i t i c a l property of the space Y i n the above lemma. We can prove this by i l l u s t r a t i n g a non-empty zero-set i n 3 Q - Q with empty i n t e r i o r . Define a real-valued function h on R as follows: h(x) = , X _< TT , X _> TT . Let h have an extension g to BQ • Then (1) Z(g)/J Q = (j> (2) Z(g) =f ij) (there exists a sequence { a n} converging to IT such that a e Q . Then Cl„Au }^Z(g)) . n 3Q n We would l i k e to show that Int ~Z(g) = <j> . Assume x e Int ~Z(g) . B Q - Q 3 Q - Q Then there exists some basic open set (3Q - cl Z) i n BQ where Z e Z (Q) such that x e (3Q - c i l Z) - Q C Z(g) . (See remark at end for the existence of BQ - ci Z) . Since (3Q - cSL Z) i s open and since Q i s dense i n BQ then (BQ - cl Z) /? Q 4 <j> . Hence Q - Z f cj> so there exists an element y such that y \ Z but y £ Q . Thus we can find some s > 0 such that y e (y-e, y+e) and (y-e,y+e) f) Z = (j> Note that this implies that cAQ.(y-e,y+e) P c% Z = cfi pQ From above g(z) = 0 for every z E BQ - cA Z - Q . Hence g{ [cH (y-e,y+e) ] - Q} = 0 since we'have shown that 9 c£ _(y-e,y+£) f) c£ Z = <j> . But by d e f i n i t i o n of h this i s a contradiction. (For y ^ ir since y e Q . Hence we can fi n d a 6 > 0 such that ir sj: (y-S,y+6) and (y-5,y+S) C (y-£,y+e) . Now min {h(z) : z £ (y-<5,y+S)} > 0 by d e f i n i t i o n of h so 0 \ g{ [c£gQ(y-<5,y+S) ] - Q} . Hence our contradiction since we said that g{[c£3Q(y-£,y+£)] - Q} = 0). Remark. We would l i k e to show that a basic open set i n gQ i s the complement of a basic closed set. From [4], 6.E(1), we know that basic closed sets i n gQ are of the form c^gg z where Z e Z(Q) . Now l e t O be an open set and p £ O . Let Q1 denote the comple-ment of O . Then OX i s closed i n gQ and p £ Q' SO there exists, a closed set F such that F = c&gQ z f° r some Z e Z(Q) and p £ F _ t and O' c F (this i s by d e f i n i t i o n of basic closed sets). Thus O D / and so we have, since p ^ F , p £ gQ - c^^^ Z c 0 . Hence the complements of basic closed sets are basic open sets. 1.7. Remark. I f Y i s l o c a l l y compact and realcompact but not compact, then gY-Y i s not b a s i c a l l y disconnected, i . e . not every cozero-set has open closure. For i f every cozero set S has open closure then the complement of S , Z say, has open closure. Thus i n t Z i s closed and so Z i s open since by Lemma 1.5 Z = c£ i n t Z = i n t Z . Since every zero-set i s open, then gY-Y i s a P-space by [4], 4J. We already know that gN-N'' i s not b a s i c a l l y disconnected (see [4], 6W) and gN-N i i s not a P-space. 10 Given a space X , l e t S C X and p e X - S . Our other results' w i l l be formulated i n terms of the following condition which we denote by (p, S) . (p, S) : There exist a neighborhood V of p and a cozero-set HC S such that (S D V) - H has empty i n t e r i o r . Remark. This condition holds i f p ^ c& S , and i f S i s a cozero-set, then the condition holds for a l l p \ S . 1.8. Lemma. Let F be a compact set i n X such that (p, X - F) holds for a l l p e F . Then i n t Z C F c Z for some zero-set Z . Proof: We s h a l l show that there exist a f i n i t e open cover {V.,, V } of F and zero-sets Zn , Z„, .... Z containing F i n 1 2 ' n such that i n t Zfc - F C X - Vfc for k = 1, ..., n . Since (p, X - F) holds then for every p e F there exists a neighborhood of p , V p , and a cozero-set H P C x - F such that i n t [ (X - F) n V P - H P] = <}> . Consider a l l such sets V p for p e F . Without loss of generality we can l e t {V P : p e F} be an open cover of F so there exists a f i n i t e P l P n subcover, V , V . We s h a l l now c a l l this f i n i t e subcover P k V^, V n . S i m i l a r l y l e t H = , k = 1, n . Let the complement of H^ be , k = 1, .. ., n . Then Z^ .-5 F since H k C X - F . Also i n t [ ( X - F) 0 Vfc O Zfc] = <J> . Hence no element of V, i s an element of i n t Z, - F , i . e . X - (int Z, - F) 3 V, and so k k k k i n t Z. - F C (X - V, ) . Let Z 0 = D Z, . Since Z, 3 F for a l l k k v k , k k k then Z 0 i s a zero-set containing F and i n t Z Q - F c X - (j V . k k 11 But X - i / V . c X - F since F C U v . Thus k k i n t Z 0 - F c r X - l ' V X - F . Hence c£(int Z Q - F) >') F = $ . Since k k F i s compact, then from [4], 3.11, we conclude that F and c&(int Z 0 - F) are completely separated. This 'means that F and i n t Z 0 - F are contained i n d i s j o i n t zero-sets, i . e . there exists a zero-set •Z^ containing F and d i s j o i n t from i n t Z Q - F . Hence i n t ( Z 0 fi Z f) C F C (Z 0 P) Z f) . So the required zero-set i s Z 0H Z . • Theorem. Let X be an F-space, S c X and p e cA S - S . I f S fi V i s open for some neighborhood V of p , and i f (p, S) holds then S i s C*-embedded i n S O {p} . Proof: We s h a l l show that there exists a neighborhood of p and a cozero-set H C S such that S fl i s open and d i s j o i n t from int(X - H) . We know that there exists a neighborhood V of p such that S fi V i s open. Also there exists a neighborhood V 0 of p and a cozero-set H c S such that int[S O V 0 - H] = <j since (p, S) holds. V 0 contains an open set U which contains p and O /Q S ^ <j> since p e cJl S . Clearly Q /T V ^ <j> and O n V i s a neighborhood of p . Also (S Pi V) Pi O i s open, being the intersection of two open sets, and non-empty since If) O Is a neighborhood of p . The required neighborhood i s V / ? Q Now l e t f e C*(S) and g e C*(H U {p}) be an extension 1 2 of f|H where f|H i s the r e s t r i c t i o n of f to H (note that X i s an F-space so every cozero-set i s C*-embedded). Define a function h on S 0 {p} such that • r f(x) , x e S h(x) = \ ^ g(p) , x = p . Since H /? V i s dense i n (S U {p})/0 V± = (S /"> £/ {p} (follows from the f i r s t part of the proof), then h i s a continuous extension of f . Thus S i s C*-embedded i n S O' {p} . 1.10. Theorem. Let X be an F-space and l e t S C X be a union of cozero-sets S i n X . Then a (a) i f G C S and i f G fl S^ i s a cozero-set i n S for a l l a , then G i s C*-embedded i n S . (b) S i s an F-space. Proof: We s h a l l use a well-known technique of induction i n this proof. We can assume that S = S and that S 0 C Sn c . . . • a 1 Consider any S . Since every cozero-set i n S i s C*-embedded then a J a S i s an F-space for a l l a . Let g e C*(G) . Let g = g| (G /') S ) . For given a < assume that g^ has been extended to a function t e C"(S^) for every £ < a , and that f 0 f^ <£_ ... (note that we can make such an extension because (G H s ) i s a cozero-set i n S and so (G H S ) i s a cozero-set i n S because there exists some function h which takes S-(G f~i S ) to zero and (S - G) C" s - (G /) S ) . Thus ^ S ^ 13 h takes (S _ - G) to zero also. Nov? h (S^ - G> makes (G O S _) a cozero-set i n . Now use the fact that i s an F-space to extend the bounded function g on the cozero-set G /] S to f on S Q> s s s> Also note that we are treating each function f_^ as a set of ordered p a i r s ) . The function U f [i e i s well defined and continuous on £ a £<a the cozero-set L; S_ U (G /) S ) . To show that i t i s continuous we £<a K consider any open set 0 i n the range of O f (J g . Then £<a s ( L-1 £ ;j g ) "*"(o>) = (V f iJ g "*"(£>) which i s an open set since the £ a £ a £<a £<a functions f and g are continuous. Also L/' S l) (G S ) i s a £ a £ a £<a cozero-set because i t i s the countable union of cozero-sets. In fact i t i s a cozero-set i n the F-space S ( r e c a l l that S 0 c Sn c ... ). a 1 Hence ( U f n g ) has a continuous extension to a function >- £ a £<a f e C"(S ) . Now U f i s a continuous extension of g to a l l of a a a S . Thus G i s C*-embedded i n S . \ (b). That S i s an F-space follows quite easily from (a) since i f G i s any cozero-set i n S , then G f) S i s a cozero-set a i n S for a l l a and so G i s C*-embedded which implies that S i s an F-space. We now apply this theorem to gN-N to conclude that a l l open subsets of".gN-N are F-spaces. The following i s a proof of this statement. We can show that gN-N has 2 zero-sets (see remark at the end) and so by the continuum hypothesis i t has cozero-sets. 14 Every closed set i n gN-N can be written as an intersection of a set of closures of zero-sets i n N . Thus every open set can be written as a union of cozero-sets. As already mentioned, gN-N i s an F-space. Hence our theorem applies with X = gN-N and S an open set i n gN-N . Remark. Let Z be a zero-set i n gN-N . Since gN-N i s C*-embedded i n gN there exists a bounded continuous function f* on gN such that f*(Z) = 0 and so Z i s a zero-set i n gN also. By [4], 6E(3), Z = f) ci Z± , where Z ± e Z(N) . Hence there are Ar zero-lew sets i n gN-N since there are jV^ sets Z^ i n Z(N) . 1.11. Theorem. Let X be an F-space having just 2 zero-sets. Let S be open and l e t p e c£ S - S , and suppose that (p, S) f a i l s . Then S i s not C*-embedded i n S U {p} . Proof: Let {S_}_ * be a family of cozero-sets i n X whose £ £<w1 union i s S (such a family exists because for every p e S there exists a function f such that f (p) = 1 and f (x) = 0 for a l l x e X - S. P P P Consider the set of Z(f ) ) . Let {V>.K be a base of zero-set-p 5 5<a) 1 neighborhoods of p . We proceed inductively as follows. For each a < a)^ , assume that cozero-sets A and B , contained i n S , have been defined for every £ < a . Since (p, S) f a i l s then int[S f) V - U. (A B r U S J ] + a .?«* 5 5 K 15 (note that S i s also a cozero-set i n S) . Thus we can choose d i s j o i n t . non-empty cozero-sets A and B contained i n S n V - c7 (A,. U (/ S J . Define A =.W A , B = l> B , a >- ? 5 5 a c G = A L> B . Hence, by construction, for each £ < a , G 1 S = (A L' B) /) S„ = <-• (A B ) /) which i s a cozero-set. So by I • K a ± K a a K Theorem 1.10, G i s C"-embedded i n S . Note that A and B are d i s j o i n t since the sets A and the sets B are. Hence A and B a a are complementary open sets i n G and by construction of the sets A A and B we conclude that both A and B meet every neighborhood of p. a Thus A and B are completely separated i n the subspace G but not i n G U {p} since p i s i n the closures of both A and B (they are completely separated i n G because A i s contained i n the zero-set B' and B i s contained i n the zero-set A' and A' /? B' = <j> i n the subspace G ; hence A and B are contained i n d i s j o i n t zero-sets). Thus from Urysohn's Extension Theorem, [4], 1.17, G i s not C*-embedded in G V {p} . Hence S i s not C5'c-embedded i n S 0 {p} ( r e c a l l that G i s C"-embedded i n S). / We now apply Theorem 1.9 and Theorem 1.11 to the space 3N-N. to conclude the following: 1.12. Let S C (BN-N be open and p e ciL S - S . Then S i s C"-embedded i s S V {p} i f and only i f (p, S) holds. 16 1.13. Theorem. Let K be a compact F-space such that K = 3Y-Y where . . . , o Y i s l o c a l l y compact and realcompact and such that K has just 2 zero-sets. (Such a space K i s 3N-N). Then (a) The following are equivalent for an open set, S : (i) (p, S) holds for a l l p e K - S . ( i i ) S i s a cozero-set i n K . ( i i i ) S i s C"-embedded i n K . (b) No proper dense subset i s C*-embedded. Proof: ( i ) implies ( i i ) : By Lemma 1.5 we see that for any zero-set Z i n K , ck i n t Z = Z . Since (p, S) holds for a l l p e K - S , then (p, K - (K - S)) holds for a l l p e K - S , i . e . (p, K - F) holds for a l l p z F where F = K - S (note that F i s compact). We now apply Lemma 1.8 to conclude that for some zero-set Z i n K , i n t Z c F C Z . Hence ci i n t Z c. cZ F c cJl Z and so Z C F c Z . Thus F i s a .zero-set, i . e . K - S i s a zero set, from which we conclude that S i s cozero-set i n K . That ( i i ) implies ( i i i ) follows from the d e f i n i t i o n of F-space. That ( i i i ) implies ( i ) follows from Theorem 1.11. (b) Since no point of K i s isol a t e d , then no point i s a zero-set; for i f a singleton were a zero-set then, as already proved, this 17 singleton would have non-empty i n t e r i o r contradicting the fact that K has no isolated points. Thus the complement of any point i s not a cozero-set which implies by part (a) that the complement of any point i s not C*-embedded. Now l e t D be a proper dense subset of K . We must show that D i s not C*-embedded. Assume that D i s C*-embedded. Since D i s a proper subset there exists an element p e K - D . We have proved above that K - {p} i s not C*-embedded so there exists f , which i s a bounded continuous function on (K - {p}),i.e. f e C*(K - {p}). such that f has no continuous extension to K . Consider f|D . This i s a continuous function on D so i t has a continuous extension f to K . K But since D i s dense then f |(K - {p}) = f . This i s a contradiction since we are now saying that f has an extension to a l l of K . Thus no proper dense subset i s C*-embedded. Applying this theorem to gN-N we conclude that for p e gN-N, gN-N-{p} i s not C*-embedded i n gN-N and that proper dense subsets of gN-N are not C*-embedded. Before ending this chapter we make the following remarks. 1.14. Remark. Let X be a compact F-space and l e t S C X be such that 2' 0 card(X - S) < 2 . Then S i s pseudocompact. 18 Proof: Assume S i s not pseudocompact. Then S admits an unbounded real-valued function and so N i s a closed subspace of S (see [4], 1.20). By [4], 14 N(5), every countable set i n an F-space i s C"-embedded so N i s C*-embedded i n X . Hence c£ N = gN . Now A'" 2 ° X - S D gN-N and since card(gN - N) = 2 we have a contradiction. Hence S i s pseudocompact. Again this remark can be applied to gN-N . 1.15. Each pseudocompact subset P of gN which contains N i s of c a r d i n a l i t y j> 2 ° . Proof: Let <j> be a mapping from N onto . W , where W i s the set of rationals i n £0, 1] . By the compactification theorem cj> can be extended continuously to a mapping (j>* of gN into [0, 1] . Since N C P then W C <}>*(?) • Since the continuous image of a pseudocompact space i s pseudocompact - see [4], 1G - then <J>*0?) i s pseudocompact. But <j>*(P) i s closed i n [0, 1] (since pseudocompact i s equivalent to countably compact i n a normal, T^ space. But since [0, 1] i s f i r s t countable then countably compact i s equivalent to compact and since [0, 1] i s Hausdorff, compact implies closed). Now since W c <}>-c(P) then c£ W C c£ <}>*(P) = <S>*(P) , i. e . 10, 1] c «|>*(P) . Hence [0, 1] = , and so card P > 2 0 19 CHAPTER I I AN APPLICATION OF g-SUBALGEBRAS OF C(N) TO gN-N 2.1. In [10], W. Rudin has shown, using the continuum hypothesis, that gN-N has a dense subset of 2 P-points. Our purpose now i s to prove a more general result than t h i s . For a completely regular space X we s h a l l define a class of subalgebras of C(X) , called g-subalgebras, and with each g-subalgebra A , we s h a l l associate a set of points i n gX-X cal l e d A-points. Then with some re s t r i c t i o n s on A and X we s h a l l show that gX-X has a dense subset of 2 A-points. Rudin's theorem will then be ea s i l y obtained by proving that C*(N) i s a g-subalgebra and that the P-points of gN-N are the C*(N)-points. I t i s not necessary to r e s t r i c t our work to the space N since this r e s t r i c t i o n does not simplify our work much. Hence we work with an arbitrary completely regular space X . Most of the results i n this chapter are due to D. Plank (see [8]). Note that under pointwise operations C(X) and C*(X) are algebras over TR (real numbers). We note also that i f f e C(X) and a l denotes the one-point compactification of IR , then by the compactification theorem there i s a unique continuous function f* : gX -y a l such .that f * | x = f . By a subalgebra of C(X) we s h a l l mean a subalgebra i n the usual sense which contains the constant functions. 20 We s h a l l denote by. f' and Til respectively the set of x\ xi prime ideals and maximal ideals i n A , where A i s a subalgebra of C(X) . We now define the hull-kernel topology on )° and M . In xi xi general the hull-kernel topology on a c o l l e c t i o n o7 of prime ideals i n a commutative ring A with i d e n t i t y i s defined by taking a = {P e J : P 3 f) ft- } to be the closure of the subset 61 of <J i 2.2. Proposition. The sets E.7(a) = {P ec7 : a e P} where a e A , are o / closed and constitute a base for the closed sets i n vJ (A i s any subalgebra of C(X)). Proof: ( i ) To show that the sets of the form E.(a) are C/ closed i t suffices to show that E^(a) C. E 7(a) . E^Ca) = {P z'3 : Flfl E ( a ) J . Let P e E. (a) . We must show that a £ P . Since PO/TE. (a) =/HP £ iJ : a £ P} and since a E f] {P zrj : a £ P} then a e P and so P £ E. (a) . Hence the sets E^(a) are closed. ( i i ) To show that these sets form a base for the closed sets consider P 0 £ J and S closed such that P Q £ S . Since S i s closed then S = ~S and so S =' {P e J : PO/0 S} ;. I f fl S = {a} , a £ A., then we are finished since there would then exist an element C which i s closed and of the form E (a) (take C = S) such 2 1 that SC C and P D fj: C (note that P 0 | C because {a} =/? S c|- p°) . If fi S i s not a singleton then consider one element a' £/?S such that a' £ P 0 . Note that there must exist one such element; otherwise P 0O .OS and so P 0 would be an element of S , a contradiction. Now consider the family of sets P i n ,// such that a' e P . C a l l this set D . Then S C D since for any Q £ S , Q D f] S and so a' e Q thereby giving Q e D . We have already shown that D i s closed. Clearly P 0 £ D because a' £ P 0 . The proof i s now complete. We s h a l l put the hull-kernel topology on , and we s h a l l introduce a family .ofy of prime ideals i n A , which w i l l reduce to '))'") when A = C(X) and A = C*(X) (A i s again a subalgebra of C(X)). F i r s t we state the following characterization (which was f i r s t introduced by Gelfand and Kolmogoroff, [5]; i t i s also discussed i n [4], 7.3 and 7 D(l)) of the maximal ideal i n C(X) associated with the point P e gX ( c a l l this maximal i d e a l ). Then =• {f e C(X) : ( f g ) * ( p ) = 0 for a l l g e C(X)} . Gillman and Jerison have also proved i n [4], 7.2, that M L = {f e C*(X) : (fg)*(p) = 0 for all g e C*(X)} and that the L i * mappings p and p M^. are one-to-one. I t i s proved i n [5] that the mappings p -> and p -> are homeomorphisms of gX onto the maximal-ideal spaces r and >7'1 ,. . We s h a l l prove l a t e r that the mapping p -> M^ 1 i s continuous for any subalgebra A <^ C(X) . The 22 above characterizations of MP and MP5,. suggest the following d e f i n i t i o n . 2.3. D e f i n i t i o n . For any subalgebra A of C(X) and for p e gX l e t MP = {f E A : (fg)*(p) = 0 for a l l g E A} . 2.4. D e f i n i t i o n . Let .*#A =' {MP : p e gX} . We now prove the following theorem i n order to show that .o^ may be given the hull-kernel topology. By the results already mentioned we see that = M 5 &i.'rJ. =))','.,. • 2.5." Theorem. For p £ gX , MP i s a prime ide a l i n A Proof: For p £ gX , <j> f MP $ A because OE MP and I i MP . A A A I t i s clear that MP i s an i d e a l i n A . To show that i t i s a prime idea l consider f, g e A such that f \ MP and g \ MP . Then by d e f i n i t i o n of. MP there exist h, k s A such that (fh)*(p) ^ 0 and (gk)*(p) i 0 . Thus (fghk)*(p) i 0 . Hence fg \ MP Thus i f fg £ MP then either f e MP or g e MP . Hence MP i s a prime i d e a l . Define T a : gX -> '-^ by T A ( p ) = MP 2.6. D e f i n i t i o n . A subalgebra A of C(X) i s said to be a g-subalgebra of C(X) i f T a i s a homeomorphism of gX onto JTI . 23 Recall that we said that and t J( are homeomorphisms of gX onto r,ir and ))1 , thereby making C(X) and C*(X) g-sub-Ci 0*^ algebras of C(X) . Before we introduce the concept of A-points we would l i k e to define a certain class of sets and show that they are a base for the closed sets i n gX . Let T.^ denote the inverse map. For f e A l e t A A = {p e gX : f e M^ } . I t makes sense to speak of E^p (f) because we have shown that <z$'^ i s a c o l l e c t i o n of prime ideals. " A Now by d e f i n i t i o n , {p £ gX : f £ M^ } = Z((fg)*) where, as usual, g e A Z denotes a zero-set. But zero sets being closed i n gX , implies that S A ( f ) i s closed i n gX . 2.7. Theorem. Let A be a subalgebra of C(X) . Then (a) T : gX i s continuous A A / (b) T. i s a closed mapping i f and only i f i s a Hausdorff space. A A Proof: (a) Recall that we have shown that the sets of the form - 1 E.t,-. (f) are basic closed sets for Q5\ . Also T . , A A A closed i n gX . v Hence T i s continuous. A (f) = S A ( f ) i s 24 (b) Let be a closed mapping. Suppose f . Then F,= {x £ gX : T A(X) = T A(p)} and K = {x e gX : T A(X) = T A(q)} are d i s j o i n t compact subsets of gX (compact because F and K are closed sets being zero sets i n gX ; M A = {f e A : (fg)*(p) = C> g e A} so F = {x : (fg)*(:x) =0} , and therefore F and K are closed subsets of a compact space gX ). Now l e t U and V be d i s j o i n t open neighborhoods of F and K respectively and l e t U 0 = gX - ? A V A ( g X - U)] and V 0 = gX - ^ [ ^ ( g X - V) ] . Since T A i s closed, then U 0 and V 0 are open i n gX and p e U 0 U , q e V 0 c V , U 0 = T^t^CUo)] , V c = ^ [ ^ ( V o ) ] . Then c l e a r l y T.(U 0) and T (V 0) are d i s j o i n t open neighborhoods of and A A A A respectively. Hence ^ i s Hausdorff. Conversely, i f i s Hausdorff then we use the fact that T i s continuous to prove that t i s closed. Let B be a A A closed subset of gX . Since gX i s compact then B i s compact. Hence T^^) - ^ s compact (T a i s continuous). Being a compact subset of a Hausdorff space T.(B) i s closed. ( I t i s easy to show that a A compact subset of a Hausdorff space i s closed). Thus for a g-sub-algebra A , i s a Hausdorff space. Remark. I f T : gX A' i s a homeomorphism then {S (f) : f. e A} A A forms a base for the closed sets i n gX . This i s so because -1 V f> - ^A (f) and the sets of the form E (f) , f £ A , form a 25 base for the closed sets i n A * / . > a s already shown. A Now l e t A be a 8-subalgebra of C(X) . This means that x i s a homeomorphism and so {S.(f)' : f e A} i s a base for the A A closed sets i n 8X . For f e A , define S*(f) = S (f) /) X* , where A A X* = 8X - X . The c o l l e c t i o n # = {S*(f) : f e A} i s c l e a r l y a base for the closed sets i n X* . Let fiS*(f) denote the boundary of S*(f) i n X* . 2.8. D e f i n i t i o n . Let A be a 8-subalgebra of C(X) . A point p e X* i s called an A-point of X* i f , for a l l f e A , p { SS*(f) . Thus A the set of A-points i s the set P (X* - 6S*(f)) , x^hich i s an i n t e r -f e A A section of a family of card A , dense, open subsets of X* . We s h a l l now prove an existence theorem for A-points. F i r s t we need the following d e f i n i t i o n and propositions. 2.9. A space X i s said to have the G„-property i f every non-void G r-subset of X has a non-void i n t e r i o r , o 2.10. Proposition. Let Y be a non-void, l o c a l l y compact, T 0 space with the Gg-property. I f 73 i s a family of at most co^ dense, open subsets of Y , then f ) D i s dense i n Y . Also i f Y has no isolated u l points, then card 26 Proof: Let AJ = {U : a < co.. } . I t suffices to show that for a 1 an arbitrary non-void open set G i n Y , 0 Z > '1 Q, f cj) . We proceed inductively as follows. Let a < a>^ , and suppose that there i s defined a c o l l e c t i o n {V^ : B < a} of non-empty open sets i n G with compact closures such that for any B < ct c£ y V G c U G fl V t for a l l x < ( i ) I f a = y + 1 then we s h a l l show ^.V„ = V (and of course B<a Y V i if). We have V. C cJL, V 0 c U 0 O V c V , for a l l x < 8 . Let Y B Y g B x x ' x e C] V„ . Then x e CJL. V 0 for a l l B < a . Thus x e V for a l l 8<a 6 Y • 3 x < B. Hence x s , x < a which implies that x £ since y < a • Hence H V„ c V B<a 3 Y Now l e t B < ct . This implies 8 _< y • I R B < Y then V.D V .. I f B = Y then V„ = V . Hence we have V n 3 V for a l l B Y B Y B Y B < a which implies that 0 V: D V . Thus ^ V„ = V . 8<a Y 8<a 3 Y ( i i ) We s h a l l now show that i f a i s a l i m i t ordinal then N V = ^.c£ V . Since V C cj> V , then fl V c O cJL. V R B<a 3 B<a Y 3 3 Y 3 B<a 3 B<a Y 3 Now i f x £ cJL, V„ for a l l 3 < a , then x £ V for Y 3 x a l l x < 3 • But x < a implies that x<x+1 < a and so x £ V for x a l l T < a , i . e . X E V for a l l 3 < a . Hence ^ cl V C f) V . 3 3<a 3 8<ct 3 27 H e n c e H V . = 0 C J L _ V . . B u t . ' { c J L , V „ : 3 < a } , 3 < a 3 3 < a • Y 3 Y 3 b e i n g a . d e c r e a s i n g f a m i l y o f c o m p a c t s e t s , h a s n o n - v o i d i n t e r s e c t i o n ( t h a t {eft V 0 : 3 < o t } i s a d e c r e a s i n g f a m i l y i s e a s i l y s e e n a s f o l l o w s : Y P l e t 3 n < 3 0 • T h e n c £ v V 0 C V 0 . B u t V 0 <r c L V 0 . H e n c e L ' 2 1 1 1 * H e n c e , i n e i t h e r c a s e , Si V 0 i s a n o n - v o i d G . - s u b s e t 3 < a 3 6 o f Y a n d b y h y p o t h e s i s h a s a n o n - v o i d i n t e r i o r w h i c h i n t e r s e c t s t h e d e n s e o p e n s e t . A n d s i n c e Y i s l o c a l l y c o m p a c t t h e n t h e r e e x i s t s a n o n - e m p t y o p e n s e t i n Y s u c h t h a t cl^. i s c o m p a c t a n d c L V < c U D i n V J C U n G . T h e r e f o r e ' {V : a < co, } i s d e f i n e d Y a a n 3 a a 1 3 < a i n d u c t i v e l y i n s u c h a w a y t h a t ( c i , V : a < w, } i s a c o l l e c t i o n o f x a 1 c o m p a c t s e t s w i t h t h e f i n i t e i n t e r s e c t i o n p r o p e r t y ( f o r c o n s i d e r cl^V^ a n d c J L r V w i t h a , < a „ . T h e n 2 c L V C U 0 ( Si V c ) C V C cJL. V ) , a n d s u c h t h a t a2 a 2 3 < a 2 3 \ a l a l c J L , V C U H G f o r a l l a < to, . H e n c e Y a a 1 (OH)n G ^ o c £ v v + <j, . Y a a < w 1 T h u s /? ^ i s d e n s e i n Y . N ow i f Y h a s n o i s o l a t e d p o i n t s t h e n a t e a c h s t a g e o f t h e c o n s t r u c t i o n o f t h e s e t s V w e c a n f o r m t w o V ' s w i t h d i s j o i n t a . a c l o s u r e s . T o s e e t h i s c o n s i d e r t h e V t h a t we h a v e a l r e a d y c o n s t r u c t e d a 2 8 a n d a n e l e m e n t x e . B y l o c a l c o m p a c t n e s s t h e r e e x i s t s a s e t V , , s u c h t h a t x e V ' .. a n d V ' 5" v • C o n s i d e r V - V , . j ct+1 crl-1 cri-1 fo. a a+1 p i c k y e V - V ' a n d a n o p e n s e t V s u c h t h a t y e V ' a n d a crt-1 cd-2 J a+2 ci V , „ C V . T h e n ' V , „ s e r v e s a s t h e o t h e r V t h a t we m e n t i o n e d , cn-z a arZ • a ^ 1 H e n c e i f Y h a s n o i s o l a t e d p o i n t s t h e n c a r d f)Z> _> 2 . T h e p r o o f i s n o w c o m p l e t e . R e c a l l t h a t a s p a c e X i s r e a l c o m p a c t i f f o r a l l p e X * t h e r e e x i s t s Z e Z ( B X ) s u c h t h a t p e Z C X * . T h i s d e f i n i t i o n w a s a l r e a d y d i s c u s s e d i n C h a p t e r I . 2 . 1 1 . P r o p o s i t i o n . I f X i s l o c a l l y c o m p a c t a n d r e a l c o m p a c t t h e n X * h a s t h e G g - p r o p e r t y . P r o o f ; We h a v e , i n f a c t , p r o v e d t h i s i n L e m m a 1 . 5 w h e r e we p r o v e d t h a t a n o n - e m p t y z e r o - s e t Z h a s n o n - e m p t y i n t e r i o r . N o t e t h a t e v e r y n o n - e m p t y G ^ i n X * m u s t c o n t a i n a c o m p a c t s e t , C , ( e . g . e v e r y p o i n t i n t h i s n o n - e m p t y G ^ i s a c o m p a c t s e t ) a n d b y [ 4 ] , 3 . 1 1 ( b ) , t h e r e e x i s t s a z e r o - s e t w h i c h i s a s u b s e t o f t h e G . a n d w h i c h c o n t a i n s 6 C . Now t h i s z e r o - s e t h a s n o n - e m p t y i n t e r i o r , s o e v e r y G ^ - s u b s e t o f X * h a s n o n - e m p t y i n t e r i o r . 2 . 1 2 . P r o p o s i t i o n . I f X i s r e a l c o m p a c t t h e n X* h a s n o i s o l a t e d p o i n t s . P r o o f : S u p p o s e p i s a n i s o l a t e d p o i n t i n X * . L e t b e a 29 zero-set neighborhood of p i n gX such that z ^ / 0 X* = {p} (such a set Z^ , exists because every neighborhood of a point i n a completely regular space contains a zero-set-neighborhood of the point - [ 4 ] , p. 3 8 ) . Since X i s realcompact there exists a zero-set Z^ £ Z(gX) such that p E Z 2 C X* . Thus ' {p} = Z fi Z 2 (Z fi Z 2 C Z' 0 X* = {p}) and Z^ 0 Z 2 e Z(gX) , which i s impossible by [ 4 ] , 9 . 6 . This completes the proof. We can now prove very e a s i l y the existence theorem we are seeking. 2 . 1 3 . Theorem. Let X be l o c a l l y compact and realcompact but not compact. I f A i s a g-subalgebra of C(X) with card A = c , then X* has a dense set of 2 A-points. Proof: Let Z> = {X* - 6 S * ( f ) : f e A} . From the l a s t two propositions we conclude that X* has the G^-property and X* has no isolated points. Now ^ i s a family of c dense open subsets of X* . Hence by Proposition 2 . 1 0 , fi ^ i s dense i n X* and has c a r d i n a l i t y at least 2 . But since A i s a B-subalgebra of C(X) then > A cs.x"d A. c T : 3 X • * / v ' A i s a homeomorphism and so card X* <_ 2 = 2 Hence card fi& = 2 ° . The proof i s now complete since, as we have already seen, /?& i s the set of A-points of X* . F i n a l l y we formulate our results i n terms of 3N . 30 2.14. Theorem. A point i n X* i s a C*(X)-point i f and only i f i t i s a P-point .of X* . Proof: (We f i r s t r e c a l l that a point p of X i s a P-point of X i f any G^-subset of X containing p i s a neighborhood of p . Also, from the fact that every zero-set i s a G^ and from [4], 3.11(b), we conclude that p i s a P-point of X i f any zero-set of X containing p i s a neighborhood of p .) Thus a point i n X* i s a P-point of X* i f and only i f i t i s not an element of the X*-boundary of any Z where Z e Z(X*) . Recall that for a subalgebra A , S (f) = ,{p e gX : (fg)*(p) = 0 for a l l g e A} = (1 Z((fg)*) . In the g£A case of C*(X) we get S A ( f ) = Z ( f 3 ) by [4], 7.2 ( f 3 i s the continuous extension of f to gX) . So by d e f i n i t i o n S*.,;(f) = Z ( f 3 ) /I X* . Thus a point p of X* i s a C*(X)-point i f and only i f i t i s not on the X*-boundary of (Z fi X A) for any Z e Z(gX) . Hence every P-point of X* i s a C*(X)-point. v Conversely l e t p e where Z^ e Z(X*) . We s h a l l show that p i s then an element of the X*-boundary of a zero-set of gX (so p w i l l not be a C*(X)-point). Since Z^ e Z(X*) then there exists a G.-subset S of gX such that S f) X* - Z^ . We j u s t i f y o 1 CO this as follows. From [4J, 1.10,.we can conclude that Z = fl Om where 1 1 0 i s an open set i n X* for every i . Since O. i s open i n X* then 0. = 0 1 0 X* where £!X i s open i n gX . Hence Z = f) (o f> X*) , 1 1 31 CO CO oo and i t i s c l e a r that /I O x*) = <f)0 X)f) X* . 1 1 Now l e t /? Q X = S . 1 Then S i s the required G„-subset of gX . Now by complete regularity there exists an element Z 2 e Z(gX) such that p e Z- c: S (obtained from [4], p. 38 - every neighborhood of a point i n a completely regular space contains a zero-set neighborhood of that point). Hence p £ S(Z^H X*) . Hence a C*(X)-point of X* i s a P-point of X* . We can now f i n a l l y conclude that i f X i s l o c a l l y compact and realcompact but not compact and i f card(X) = c , then X* has a dense set of 2 P-points. Hence gN-N has a dense set of c 2 P-points. Remark. Local compactness i s an essential property. For the space Q (set of rationals)' s a t i s f i e s the property card C(Q) = c but Q* has no P-points (see [4], 60.5). 32 CHAPTER I I I c-POINTS OF BN-N 3.1. We s h a l l c a l l a point i n BN-N a c-point i f i t l i e s ' simultaneously i n the closures of c pairwise d i s j o i n t open sets of BN-N . We remark at once that i f a i s any cardinal greater than c then there does not exist any a-point i n BN-N , because there are at most c d i s j o i n t open sets i n BN-N . In the f i r s t part of the chapter we s h a l l prove that there exists a c-point i n BN-N . Then l a t e r on we prove that every point of BN-N i s a c-point, the proof of which i s independent of that of the f i r s t part. However we include the f i r s t part because the same proofs can be applied to draw the same conclusion for any discrete space of c a r d i n a l i t y m where 2 n _< m for a l l n < m . Following the notation of [ 4 ] we l e t , for any p e BN-N , A P be the free z - u l t r a f i l t e r on N with l i m i t p . 3.2. D e f i n i t i o n . Ap i s called uniform i f each element of AF has ca r d i n a l i t y 0 0 . Let uN = {p e BN : A P i s uniform} 3.3. Proposition. "* uN = BN-N 3 3 Proof: , Let p e gN-N and consider A P . In order to show that p e uN' . we must show that A P has no element consisting of a f i n i t e subset of N . Suppose there exists P e A*3 such that P = {n^, . .. , n } where n.^ £ N , 1 .1 i .1 j • Then {n^} ^ A P for otherwise A P would not be a free u l t r a f i l t e r . Hence (N - {n^}) e A P. Thus P f) (N - {n 1» e A P which means that {n 2, . .., n } . e A P . We proceed i n this way u n t i l we can f i n a l l y conclude that {n_. } e A P , a contradiction. Hence no such element, P , exists and gN-N C uN . Now p i t i s easy to see that uN C gN-N . Let p e uN , i . e . A i s uniform. I f p s N , then A P = {Z : Z i s a zero set containing p} . But this would mean {p} £ A P since any element i n N can be made into a zero set. The fact that {p} e A P i s a contradiction since A P i s uniform. In the sequel we s h a l l use the following facts. I f Z N , then c£.„ Z = {p £ gN : Z £ A P} . Also basic neighborhoods of gN p £ gN-N are of the form* (c^gN Z) - N where Z e A P . 3 . 4 . Lemma. Let Ci- be an i n f i n i t e c o l l e c t i o n of subsets of N such that card A = for every A £ £L , and cardCA^/O A^ .) < ° 0 for P such that card {A zC-L : card(Z D A) = J\} = card & for every Z £ A^ A., A. £ , i ^ j . Then there i s a uniform u l t r a f i l t e r A on N Proof: , For A £ CL pick x^ £ c ^ g N A D uN ; l e t B = {x A : A £ CL }. I f P, Q e CL such that P <f Q , then f x„ 34 because i f = x then A = A H and since P £ A ( r e c a l l that x for A c N , cJl A =' {P £ gN : A £ A P}) then P e A and so X0 >r P fl Q s A v . This means that card(P fl Q) = 0 , a contradiction. Thus card B = card . Since gN-N i s compact, there exists a point p e uN such that every neighborhood of p contains card Ct elements of B . For i f not then pick a neighborhood of every p £ gN-N , N^ , such that N contains less than card fl elements of B . Consider P {Np : p £ uN} so formed. Without loss of generality assume that this set i s an open cover for uN . Since uN i s compact, there exists a f i n i t e number of elements from this set, N , N , N , such that P l P2 P n n uN C V N . Note that B <^ uN . Since each N contains less than 1 p. p. card CL elements of B then card B < card CL , a contradiction. We claim that A P i s the required u l t r a f i l t e r . I t i s uniform since p e uN . Also i f card(Z /) A) < for Z £ A P then x A fj: Z /) yN (for XA XA X x. £ c£ A fj yN means A, £ A so Z % A ). Hence card(Z f) A) = J 0 for every Z £ A P and for every k e . f l . We digress a l i t t l e to prove the following theorem which w i l l be used i n the remainder of the chapter. 3.5. Theorem. Let m be a cardinal number such that 2 n _< m for every n < m and l e t .X be a set of "cardinality m . Then there exists a family Ct of subsets of X such that card Ci = 2 m , card A = m for every A £ CL and card (A f) B) < m for A, B £ Ct. and A r B . 3 5 Proof: We can well order X thereby l e t t i n g X = {x^ : a < a Q} X where a 0 i s the i n i t i a l ordinal such that card a0 = ra . Let Y = 2 and order Y lexicographically. Let Z = {y E Y : there exists a < a0 such that y(x )"= 0 • . v for a < 0 < a Q} We would l i k e to show that card Z = m . Indeed Z = {Z : a < a 0} a where Z i s the subset of Y whose functions take on the value zero a at every x for y > a ( i . e . Z = {y £ Y : y(x ) = 0 for y > a} . y a y card ct For every- a there are 2 such functions and so card Z = 2° a r^ < m by hypothesis. Thus card Z = Y. card Z a — J J L A a<a0 < m-m = m . Thus card Z = m . For each y E Y l e t be a sequence {z^ : 0 < a 0} of m d i s t i n c t elements which converge to y and where z^ e Z ^ Y such that r y(x ) , a < 0 z R ( x ) = { a U • • a > 0 . Now l e t f from Z onto , X be a one-to-one map. Then {f[Sy] : y E Y} i s the required family theorem. Note that N has the properties of X i n the above 3 6 3 . 6 . Lemma. Let Ct be a c o l l e c t i o n of subsets of N as given above and l e t A P be as obtained i n Lemma 3 . 4 , the elements Z of A? a being indexed by the ordinals less than to . Then we can choose f or each a < to, a subset X of Z so that card X = o and so that 1 a a • a card (X /) X ) < J \ i f a < Y < a, . a a ' • 1 Proof: Pick an element e /Z such that card Z^ fi A^ = J 0 ( r e c a l l that we have card (J such elements) . Let a < to^ . Assume that for a < a we have chosen A s such that card(Z /Ik) = J'0 CT CT a and that A 4- A for every v < cr . Since CT y •card{A z CI : card A fl Z = J Q} = to., (follows from lemma 3 . 4 ) then t h i s a 1 \r means that we have an element A e tl such that card (A /) Z ) = J Q a a a and A j- A for CT < ct ( r e c a l l a < ox.) . Let X = A ^ Z . Then cr a 1 a a a \r the sets X s a t i s f y the required properties since card X = <, and ot a for a < y < we have A f A so card (A ^ A ) < J\> % hence' the 1 a y a y second property i s s a t i s f i e d since X' -'1 X = A ^ Z r) A '"^ Z C A A . a Y a a Y Y a Y 3 . 7 . Theorem. There ex i s t s a c-point i n gN-N Proof: . Let (I , A P = {Z}' and X C z be as already a a < to., a a described. Since card X = 0 then again by Theorem 3 . 5 we have a ct v c o l l e c t i o n {X } of subsets of X such that card X = J 0 f o r xa x<to^ a Ta every T < to, and card ( X f~) X ) < J\ for T f CT . Now for every 1 TCI . CTOt T < to, l e t U = ( c Z n „ X ) ^ uN . Since ci „ X i s open i n 1 T gN ta gN Ta a<to^ gN then U i s open i n gN-N . We claim that {U } i s a family T T T<(0^ 37 of pairwise d i s j o i n t , open sets such that p e c£ for every T , i . e . p i s our c-point. That they are pairwise d i s j o i n t i s shown thus: l e t q e U fl U where a ^ x . Then for some a, , a~ we have a x 1 2 q e c£ D H X • fl uN , q e c£„_ X f] yN . Hence 3N aa^ 0N xc<2 q e (c£^7 X fl c£„,T X ) f) yN . In the case a, 4 a„ then 3N aa^ 3N xa 0 1 2 (X O X ) C (X /7 X ) and since card (X fl X ) < ''I then oa^ xa 2 a 2 1 a2 card (X X ) < J\ . I f a, = ou then we already know from ca^ xa 2 1 2 ^ properties of the sets {X } that card (X ff X ) < *'t . So i n TO^ aa 1 xcu/ either case card (X fl X ) < J\, . But q e c£„„ X f] c£^T X aa 1 r a 2 ^ 3N ac^ 3N xcu, means that X £ A q , X e A q and so (X ^ X ) e A q . Also cra-^ xa 2 aa^ xa 2 q E yN so A^ i s uniform. Hence card (X fl X ) = "l0 , a contra-aa^ xa 2 d i c t i o n . Hence our sets are d i s j o i n t . We now proceed to show that p i s i n the closure of U for every x . Let Z E A P ( i . e . T a p E c£ Z ). We know that ( c^ f l N 2 '^ yN) i s a basic neighborhood of p so we are finished i f we can show that for every x , C^3N ^ ViN Z"' ^ <|> since every neighborhood of p would then V' intersect U . Since card (X f> Z ) = ^ (X C Z and x xa a xa a card X = %) then there exists an element q i n •xa • n c£ X f) c£ 0 H Z /) yN<r c£ o W Z' U ^ yN . Hence 3N xa 3« a 3N .a x c£ O M Z /I U /I H • 3N a x y 38 Thus p i s a c-point i n BN-N . The purpose of demonstrating such an u l t r a f i l t e r A P i n Lemma 3.4 was to show i n Lemma 3.6 that we can choose, for each Z e A P a a subset X with those properties. Using this c o l l e c t i o n {X} a ^ ^ 6 a a<tu1 we proved that p was a c-point of BN-N . Hence a reasonable assertion i s that every point of BN-N i s a c-point i f we can show that for every free u l t r a f i l t e r A P there exists X C Z for every Z e A P such a a a Y" ' 'T that card X = „ and card (X O X j < J\ i f a T B • a a B 3.8. Lemma. Let A P be a free u l t r a f i l t e r on N and order i t s elements by the ordinals less than UJ_, . Then for each Z E A P there 1 a exists X C. Z such that card X = J \ and card (X /? X j < J*a i f a a a a 8 a ^ B • P A'" Proof: Recall that every element i n A has c a r d i n a l i t y „ since A P i s uniform. Let X^ be an i n f i n i t e subset of Z^ such that X1 £ A P . Fix a < ujn and assume that for each a < a we have chosen X C Z such that card X = ^ 0 , X i A P and card (X 0 X ) < K * r for a l l y < a . Let B = {CT < a : card (X fi Z ) = J\} . We now show CT a that, i f B i s f i n i t e then (Z - U{X : cr e B }) e A P . Consider a set a CT X' i A P . Then N - X e A P . Since Z e A P then (N - X ) fl Z £ A P a a a a a i.e. Z - X e A P . Thus i t suffices to show that O { X : CT e B} i A p . a a . CT ' Since X i A P for every cr e B then N - X e A p and so CT J CT 0 {N - X^ : CT £ B} s A P , i. e . N - ^{x^ : CT e B} e A P . Thus 39 V {X : a e B} i A P . Now l e t X be an i n f i n i t e subset of cr a (Z - L!{X : a e B}) such that X £ A P . Then by d e f i n i t i o n of B we have card (X 0 X ) < J \ for every a < a a a I f on the other hand B i s i n f i n i t e , then i t i s i n f i n i t e -CO l y countable by the continuum hypothesis so we write B = {<?n}^ • every k e N we can find an element x i n \ (X f) Z ) - 6 ' { X : j < k} because card (X fl Z ) = ~K while for o\ a a. a, a k j k j < k card (X n X ) < -J\ . Thus (X f) Z - 6/{X : j < k}) ^ <b . cr, o. cr, ot a. k 3 k 2 CO Consider the set {x }, . . Let X be an i n f i n i t e subset of ct^ k=l a oo "n {x }, . such that X" i A^ (note that X c Z ) . Now for any a e B, ot^ k=l a T a a A<" a = a, for some k e N so card (X' rt X ) = card (X rt X ) < k < " 0 k a cr a a, — k CO I from the way we have chosen {x }, , . I f however a I B then a k k=l * card (X rt X ) < card (X H Z ) < J\ by d e f i n i t i o n of B . Thus we a a — a a have shown that for every a there exists X c Z E A P such that a a card X = J I and card (X ^ x ) < J ) 0 i f a / g . ot a g 3.9. Theorem. For every p e gN-N , p i s a c-point of gN-N . Proof: Let A P = {Z } , X <Z Z as i n Lemma 3.8. Since C4 a«D^ a a card X^ = '\ for every a then we can use our Remark 3.5 to show that there exists a c o l l e c t i o n {X' \ } of subsets of X such that 40 card X = J'Q and such that card (X /> X ) < J , 0 i f x ^ a . We xa xa aa now form sets precisely as we did i n Theorem 3.7 and we proceed exactly as we did there to show'that p i s a c-point of 3N-N . 41 CHAPTER IV TYPES OF ULTRAFILTERS AND THE NON-HOMOGENEITY OF 6N-N 4.1. The major results of this chapter are due to Z. Froli'k. Here we discuss the concept of types of u l t r a f i l t e r s and using this concept we s h a l l prove the non-homogeneity of gN-N (this w i l l be proved x^ithout using the Continuum Hypothesis) . 4.2. D e f i n i t i o n . We say that two u l t r a f i l t e r s are of the same type i f they are isomrophic as p a r t i a l l y ordered sets. Theorem 4.3 w i l l show that not every two u l t r a f i l t e r s on N are of the same type. Remark. I f -rr i s a permutation on N and Q i s an u l t r a f i l t e r on N then TT(^) = {TT(E) : E e fi} Is an u l t r a f i l t e r on N . We show this as follows. (a) Let TT(E^) , TTCE^) e ir(^) . We would l i k e to show that TT(E 1 ) fl TT(E2) e Tr(n) . We claim that TT(E ) fl ir(E 2) = TT(E E ) . For l e t y £TT(E^) d T\(JLA . Then for some e^ e , Tr(e^) = y and for some £ E^ , T r( e2^ = v • ^ u t e ] _ = e2 s i n c e fr i s one-to-one. So y £ TT(E ;L f> EA . Thus TT(E ) fl ir(E 2) C TT(E ^ ' E ) . Obviously ir(E ^ E £) C TT(E ) ^ TT(E 2) . Hence 42 T T ( E 1 ) {) T T ( E 2 ) = T : ( E ; L / ' I E 2 ) . Since E ^ E 2 e Q then T T ( E ] L ) -0 T T ( E 2 ) E TT(O) . (b) Let TT(E) C E 0 . We must show that E Q e Tr(fi) . Since IT is a permutation on N then for some A <£* N TT(A) = E 0 (we are assuming that E 0 c N) . Then E c A since TT(E) <r i r ( A ) . Thus A E ft . Hence TT(A) e Tr(fi) , i.e. E Q E ir(fi) . Hence -rr(fi) is an u l t r a f i l t e r on N . 4.3. Theorem. If two free u l t r a f i l t e r s ft^ and Q. are of the same type then there is a permutation TT of N such that ' ^ 2 = ir(ft^) . c There are 2 types of free u l t r a f i l t e r s on N ; each type contains c u l t r a f i l t e r s . Proof: Since ft^ and Q,^ axe of the same type then there exists an isomorphism f taking ft^ to ^ 2 (note that i f A C B then f(A) C. f(B) and that f(N) = N) . Let H =' {q : q E N , q ^ n } . The sets H^ are maximal proper subsets of N which are permuted by f . That they are maximal proper subsets' is obvious. Now for any H^ , n i H . There must exist E E £L such that n i E since otherwise T n 1 T fl {E : E E ft^} f <f> , a contradiction since is free. Since E c- H n then' H n £ & and so f operates on H^ to give a set of the form similar to H , H say. Define ir such that -rr(n) = m i f f(H ) = H . n m n m Now let E E and let F = N - E . Then E =/7{H : n £ F} because 1 n the elements of E are the only elements in H^ for every n £ F . Let E 0 =/7{f(H ) : n E F} . Since E c H for every n then n n 43 f(E) c f (H ) f o r every n and so f (E) c: E Q . By symmetry we have f _ 1 ( E 0 ) C E . Thus f ( f _ 1 ( E 0 ) ) C. f(E) , i . e . E 0 <= f(E) . Hence f (E) = E 0 . Now N - E 0 = N -/?{f (H ) : n s F} =' {Tf(n) : n e F} because -rr(n) = m i f f(H ) = H and r e c a l l that m i s the only element of N n m that i s not i n H . Thus N - f(E) = TT(N - E) '. m For any A c N we have T T ( A ' ) = [ T T(A)]' (where primes denote, complements) . Hence TT(N - E ) = N - TT(E) . Thus, using the f a c t that N - f ( E ) = TT(N - E ) , we get f ( E ) = T T ( E) . Hence IT i s the required permutation. We have now established that a necessary condition f or two u l t r a f i l t e r s to be of the same type i s that there e x i s t a permutation ir on N taking the one u l t r a f i l t e r to the other. But there are only A1 2 = c permutations on N . Hence no type can contain more than c free u l t r a f i l t e r s . We can,now conclude that there are 2 types because the set of u l t r a f i l t e r s of the d i f f e r e n t types forms a p a r t i t i o n of gN-N which has 2 elements and each type has only c u l t r a f i l t e r s . F i n a l l y we must show that every type contains p r e c i s e l y c u l t r a f i l t e r s . From Theorem 3 . 5 we can f i n d a family c o n s i s t i n g of c i n f i n i t e subsets of N such that any two members of 3 have f i n i t e i n t e r s e c t i o n . Let • 0, be any free u l t r a f i l t e r and A e 0, be such that the complement of A i s i n f i n i t e . For each E e 3 l e t be an u l t r a f i l t e r containing E . Then there e x i s t s a permutation on N 44 carrying Q. to (take the' permutation which c a r r i e s A to E) . Since we have c such sets E then we can conclude that fi, i s of the same type as c other u l t r a f i l t e r s . (Note that we have already shown that no type can contain more than c free u l t r a f i l t e r s ) . 4.4. We now take a closer look at the set of types of elements of gN-N '. We l e t P denote the set of a l l permutations of N and p* with p e P the continuous extension of p to a homeomorphism of N* onto i t s e l f . Suppose _ f : N -> N i s a set isomorphism ( i . e . a permuta-t i o n on N). I f we consider f as operating on subsets of N then we can say that f induces a set isomorphism g : iP(N) ->£\N) (where P (N) denotes the power set of N). Now since an u l t r a f i l t e r on N can be considered as a subset of ^(N) then we can say that f induces a homeomorphism h from gN onto gN i n such a way that i f P l P2 fi^(= A , p^ e gN) and ft (= A , p^ e gN) are u l t r a f i l t e r s on N , then g(^ 1) = ft2 i f and only i f h(p 1) = p 2 . We have already proved that i f two u l t r a f i l t e r s and &2 a ^ e of the same type then there exists an element p e P such that a l a2 p(fi^) = ^2 (hence p*(a^) = ^ ^ ^± = A > ^2 = A ^" T ^ u s f r o m Theorem 4.3 and e a r l i e r remarks- i n t h i s s e c t i o n we conclude that a l ~' a2 £2^(= A , a^ E gN) and fi^(= A , a2 e gN) are of the same type i f and only i f p * ^ ) = a.^ f o r some p e P . 45 Let T be a set and T a mapping of N* onto T such that TO = T6 i f and only i f p*(a) = 8 for some p e P . Vie s h a l l c a l l the elements of T the types of u l t r a f i l t e r s on N , i . e . i f t = T0 then t i s the type of 0 . Note that we have defined our mapping T SO as to be consistent x^ith our d e f i n i t i o n of two u l t r a f i l t e r being of the same type. Note also that since we have proved that each type contains c u l t r a f i l t e r s and that there are 2 types of free -1 c u l t r a f i l t e r s we now have card T (t) = c and card 1=2 We now define a c o l l e c t i o n X of u l t r a f i l t e r s to be di s c r e t e i f there e x i s t s a d i s j o i n t family {M : x e X} with M e x . We know that c£ X i s homeomorphic to gN i f X i s i n f i n i t e . Thus for X a countable d i s c r e t e subset of gN-N and z e cl X - X the i n t e r s e c t i o n of the neighborhoods of z with X form an u l t r a f i l t e r on X . C a l l t h i s u l t r a f i l t e r z . We know that there e x i s t s a set X isomorphism k : ^ (X) ->?9(N) . Thus z i s taken by k to an u l t r a -f i l t e r C ^(N) . We now define the type of z r e l a t i v e to X , written T,TZ , to be T X y In the case when X = N we get and j^z . where z^ i s an u l t r a f i l t e r on N of type ' T^Z . We s h a l l now show that z^ = A , i . e . {N f] N : N i s a neighborhood of z} = A Z , thereby'giving T„TZ = T . Consider any N H N , where N i s a neighborhood of z . N z J z z Let N ^ N = B's-. Then z e B f o r otherwise there e x i s t s some neighbor-hood N* of z such that N* ft B = cb , i.e.' N* f) (N T) N) = A , i . e . z z r z z r 46 • (N A f) N ) fl N = d) . But this i s a contradiction since N* /I N i s a z z z z neighborhood of z E CX, N-N = 0N-N and so (N* /) N ) /) N ^ (J> . Thus — z z z e B and so B e A . Hence z„ C A N — Now l e t B e A Z . We would l i k e ' t o show that B = N O N z where N i s a neighborhood of z . We know that B = {gN - c£(N-B)}rt N. z '. But gN - cft(N - B) i s a neighborhood of z so we l e t N - gN - c£(N - B) . Thus B = N /? N . z z 2 Hence when X = N and Z E C £ X - X , Z N = A , and x , , z = T N z 4.5. D e f i n i t i o n . The producing r e l a t i o n <J> on T i s defined to be the set of a l l ( t , t') E T x T such that T Z = t' , T z = t for some z £ gN - N and X cr gN - N where X i s discrete and z E ci X - X . Thus we have <j)(t) = { T Z : z £ cZ X - X , T z = t , X C N*' and X discrete} Hence the domain of <f> i s a l l of T , i t s range i s a subset of T and when ( t , t') E <j> we say that t produces t ' or t' i s produced by t . 4.6-. Remark. Let { M ^ : n E u} be a countable decomposition of N and pick, for every n , x , y E C £ M - M . Let - n Jn n n 4 7 X = {x : n E to} and Y = {y : n e to} . We s h a l l show that i f , for n • • Jn every n , x f Y , then cZ X f) cZ Y = <p . Recall that a base for n n the closed sets i n gN i s {cZ A : AC N} . Now assume that there exists x e cZ X /7 cZ Y . Consider a base neighborhood A^ containing x_^ ( i e to ) and a base neighborhood A 1 containing y_^ such that A. fl A 1 = d> and M. = (M. /) A.) U (M. f) A 1) . Let M. f) A. = M. and i 1 1 1 1 ' i i i ^ M. /) A 1 = M. . Then M. = M. U M. . Let H, = c£t/{M. : i e to} and i ±2 • 1 1 1 12 1 1 H 2 = cJli^ML : i e to} . Then E± 0 H 2 = <j> . But contains the closure of M. for a l l i and so x e H . Sim i l a r l y x e EL . Thus i 1 we have a contradiction. Hence c% X cZ Y = A> Note that the converse holds as w e l l , i . e . i f ci X fl cl Y = 6 then for each n , x 4 y n Jn 4 . 7 . Lemma. Let y E gN - N . There exists a set x such that card x <_ 2 and whose elements are discrete countable subsets of gN-N and such that i f Y i s any discrete countable subset of gN-N , and i f y e cZ Y .- Y then Y 3 X' for some X' e x where y e cZ X' - X" . Proof: For each countable decomposition {^n} of N choose an x E cZ M - M (and l e t X = {x } ) such that y e cZ X - X , i f n n n n J ' possible, and consider a l l possible X'c X such that y e cZ X' - X . Let x = {X'} so formed as we vary our decompositions of N . Now x cer t a i n l y consists of discrete countable subsets of gN - N . We s h a l l 48 use Remark 4.6 to show that x has the asserted property. Let Y = {yn'} (where Y i s as in\the statement of the lemma). By the construction of x we have y e cH X - X for some countable discrete X . Let X = {x } . Since y i s i n the closures of {x } • and {y } n ,n n then by Remark 4.6 {x } and {y } have to agree for an i n f i n i t e J . n Jn & number of n's . So we l e t the required set X' be Y (1 X . Now y i s i n the closure of X' because i t i s i n the closures of Y and X (see [9], Lemma 2). There are at most 2 decompositions of N and K with each decomposition we get at most 2 sets X . Hence A' }< card x _< 2 > 2 =2 . The proof i s now complete. From the following theorem we s h a l l obtain a very simple proof of the non-homogeneity of gN-N . 4.8. Theorem. Any type i s produced by at most 2 types and any type produces 2^ types. Proof: We prove the second part f i r s t . Let t be any type. Recall that <j)(t) = {TZ : z e cH X - X , T z = t for some discrete A. X c N*} . Now l e t {M } be a decomposition of N where each M i s n n i n f i n i t e . We know that since M i s i n f i n i t e then 2 card (eft M - M ) = 2 for each n . Again using Remark 4.6 we n n y- & ° J o 2 conclude that there are 2 d i s j o i n t sets cJl X = c£ {x : x e M } n n n where each such c£ X contains a point z such that r z = t and X 4 9 o TZ e cj>(t) . Since at most 2 of these elements z could be of any o 2 one type (mentioned already) and since we have 2 such elements z then we can conclude that card <j)(t) = 2 . Before proving the f i r s t part we must prove the following. I f X C Y , Y discrete, and z e c£ X - Y , then T VZ = T z . We assume that both X and Y are i n f i n i t e . Let 6 be a one-to-one map from N onto X and x a one-to-one map from Y onto N . Let i be the inclusion map from X to Y . Define the permutation p on N as indicated: p : N y X c: Y x — s - N i. e . p = x ° i ° <5 • We define z and z as usual. We know X i Ci Y that z and z are set isomorphic to u l t r a f i l t e r s A and A on X Y N i n such a way that T a = T z and x y = T z . In order to show that A X T VZ = T z i t suffices to show that for some permutation q on X , A Y q*(a) = y .We claim that p , as defined above, i s the required permutation, i . e . p*(a) = y . Assume that p5'c(a) ^ y . Then we can f i n d d i s j o i n t neighborhoods A and B such that y e A , p*(a) e B , A f) N e A Y , B 0 N (j: A Y and p A~ 1(B) f) N e A K . Now [N n p''c-1(A)]/T [N/? p * " 1 ^ ) ] = $ . Also 6(p* _ 1(B)/? N) = (N n X) e where N^ i s some neighborhood of z . This i s so because 8 induces a set isomorphism between P (N) and <P (X) which takes Aa to z ; X 50 " p* (B) f) N e A and so i s taken by this isomrophism to N^/l X some neighborhood N^ of z . But considered as a subset of N 6 takes (p*~1(B)/'> N) to (N /? X) . Since [N /? p* _ 1(A)]/7 [N /? p*" L(B)] = <f> then S[E fl p* - 1(A)]/? <5[N f) p* _ 1(B)] = (J) . We s h a l l now show that [N /? Y ] /0 6[p" _ 1(A)/? N] = <)) . Assume this i s fa l s e . Then X[(N z/1 Y)/7 6(p* _ 1(A) N) ] = x&zr\ Y) f) X ( 6 ( P " _ 1 ( A ) fl N)) . Now x ( N /0 Y) £ A Y ( r e c a l l that x induces a set isomorphism between :f (Y) and 'f (N) which takes z y to A Y) . Also X(6(P" _ 1(A) n N)) = A/7 N (see remark later) and A P N e A Y . Thus xt(N ^ Y) /7 6(p* - 1(A) /I N) ] E A Y . Hence (N /) Y) /I S(p* - 1(A) /) N) e z y and so (N z Y) 0 6(p* """(A) N) = N^D where N' i s some neighborhood of z . Now z [(N ^ Y) n <5(p*_1(A) D N)] /IX = (N /) X) fl 6(p* - 1(A) fl N) z = cKp* - 1(B)/) N) /I o(p* - 1(A)/? N) ( r e c a l l that 5 (p* _ 1(B) f) N) = N /7 X). But we have already shown that <5(p* _ 1(B)rt N) /I 6(p* - 1(A) /) N) = <j) , i . e . [(N 0 Y) P 6(p* _ 1(A) /) N) ] /) X = <j> . Hence N' /) Y /) X = <J> which 2 2 implies that N' /? X = <j> . But this i s a contradiction since N' i s 5 1 a neighborhood of z . and z e c£, X . Thus ( N z rt Y) rt 6(p- - 1(A) n N) = 4 Hence • x(N rt Y)/? x(<5 (p* - 1(A) /> N)) = ^ . But x(N z/) Y) e A Y and x(6 ( p A _ 1 (A)/} N)) = A/} N e A Y (see remark l a t e r ) . Thus we have X ( N Z ^ Y ) x ( S ( p * - 1 ( A ) /? N)) = <f> and x ( N z n Y) f) x(6(P' c _ 1(A) ft N)) s A Y , a contradiction. Thus p*(a) = y _1 Remark. We must now show that X(<S(p* (A) rt N)) = A/? N . From our d e f i n i t i o n of p we get X(<5(p*-1(A) 0 N ) ) = p(p* _ 1(A) rt N) • = pt { y e BN : p * ( y ) e A} fl N ] = p{{y e N : p ( y ) e A}] = A f] N . We now continue with a proof of the f i r s t part of the theorem. Consider any type t' . Then (^ ""'"(t') = {t : <j>(t) = t'} . For any t e <}> "'"(t1) and for every X appearing i n the d e f i n i t i o n of <j)(t) there exists a Y e x such that X ~2 Y and x vz = T VZ = t (follows from Lemma 4 . 7 and the above observation). Hence with every 52 element t i n $ (t') we can associate an element of x . But there o o are at most 2 elements of x ; so there are at most 2 such elements t . This completes the proof. 4.9. We now prove the non-homogeneity of N* . Let us assume that N* i s homogeneous, i . e . for every pair x, y e N* , there exists a homeomorphism of N* onto N* taking x to y . Let, for any x e N* , T = <f> "''(TX) . I f f(x) = y for some homeomorphism f of N* onto N* then T X = xy by e a r l i e r remarks (see section 4.4). Hence Tx = Ty i n this case. Thus i f N* i s homogeneous then a l l sets Tx "' are equal for every x e N* . But card Tx _< 2 by our l a s t theorem and card T = 2 (shown already). Also the sets Tx for x e N* cover T . This i s a contradiction i f a l l sets Tx are equal. Hence N* i s non-homogeneous. The remainder of the chapter i s used to show that no homeomorphism of BN into N* has a fixed point.- a fact that w i l l enable us to give another proof of the non-homogeneity of BN-N without using the continuum hypothesis. F i r s t l y we need two lemmas. 4.10. Lemma. For a class X l e t f be a mapping of X into some other class such that fx T X for any x e X . Then there exist d i s j o i n t classes X Q, X±, X 2 such that • X = X Q U X;L 0 X 2 and f (X ) D X =. <j> for i = 0, 1, 2 . 53 Proof: We s h a l l associate with every x e X a class C(x) defined as follows: C(x) = {y e X : f m x = f n y for some m, n e N} We now show that any two classes C(x^) and C(x 2) either coincide or are d i s j o i n t . Assume that they are not d i s j o i n t . Then l e t y £ CCx^ fl C(x 2) . Hence (1) f x = f y ( 2 ) f x 2 = f y where m, m', n, n' £ N . We can assume that n' > n . Apply f to n n' f y u n t i l we achieve f y (while at the same time applying f to f m x ^ to maintain equality i n (1)). We then get ^irri-(n'-n) -^n' .m' f x^ = f y = f x 2 Hence m' m" (3) f x 2 = f x^ for some m" £ N . Now consider any element y 0 e C(x^) . We want to show y 0 £ C(x 2) . mo n° Then by symmetry we s h a l l be finished. We have f x^ = f y 0 where m0, n Q £ N . Assume mQ > m" . With (3) i n mind we apply f to m" f repeatedly u n t i l we get the following: _mT + (m0—m") _mQ _n 0 7T f x 2 = f x^ j^ = f y Q . Hence f m " x 2 = f n°y 0 where m" £ N . Hence y 0 £ C(x 2) . (Note that i f mQ < m" then we would apply f to fm°x^ u n t i l we get a si m i l a r equation). By symmetry we get C(x^) = C(J 54 We now show that f(C(x)) C C(x) for x s X . C(x) = {y : f m x = f n y , m,n e N} . Let fy e f[C(x)] . We must show that fy e C(x) . Since y C(x) then f m x = f n y and so f ( f m x ) = f ( f n y ) . Thus f m + 1 x = f n ( f y ) which means that fy E C(x) . Thus f(C(x)) C C(x) . Since the union of the sets C(x) , x E X , i s X then i t i s s u f f i c i e n t to prove the lemma for each class C(x) . So we l e t X i n the hypothesis of the lemma be of the form C(x) . Thus for any x and y i n X we may assume that there are m,n E N such that f m x = f n y . Now pick an element a e X . For x E X l e t m(x) be the smallest m E N such that f m a = f n x for some n E N . Then l e t n(x) be the least n £ N such that f n x = f m ^ x ^ a . We assert that for x E X with n(x) > 0 m(fx) = mx . For assume that m(x) = q . Then q i s the smallest integer such that f^a = f n ^ X ^ x which i s the same as saying that q i s the smallest integer such that f^a = f n ^ x ^ "*"(fx) (note that i t makes sense to speak of n ( x ) - l since n(x) > 0 ) . Thus m(fx) = mx and n(fx) = n(x) - 1 . We now show that there exists at most one b £ X such that n(b) = 0 and m(fb) T m(b) + 1 . Assume that we have d ^ J D 2 e ^ such that b 1 ^ b 2 , nO^) = 0 , n(b 2) = 0 and m(fb;L) r mOt^) + 1 , m(fb 2) r m(b2) + 1 . Now since n(b^) = 0 then m(b ) m(b )+l f. " a = b 1 = f°b1 and f a = f ft^) . Since m(fb;L) f m(b1) + 1 55 n(fb ) then we have a number p = m(fb ) such that f a = f (fb ) where m(b )+l 0 _< p < m(b^) + 1 (since we already have f = fb^ then m(b 1)+l p < m(b^) + 1). Hence f(b^) = f a = f a where m(b^) + 1 > p >_ 0. m(b 2)+l Similarly f (b^) = f a = f q a where ra(b^). + 1 > q _> 0 . Assume without loss of generality that m(b^) > m(b2) . Then m(b^) > m(b^) + 1 m(b ) m(b )+l because i f m(b1) = m(b 2)+l then b^ = f a = f a = f a . But m(b1) = m(b2) + 1 > q means that m(b^) > q . This contradicts the d e f i n i t i o n of m(b^) since i t must be the smallest number such that m(b1) b^ = f a . Thus m(b ) > m(b2) + 1 . Hence m(b-) m(b )-m(b„) m(b„) m(b )-m(b ) m(b, )-m(b,)-l \= f a = f X Z ( f a) = f 1 1 (b 2) = f 1 (fb m(b )-m(b ) - l m(b )-(mb +l-q) = f • (f q a ) = f (a) . But m(b ) > m(b )+l-q > 0. This contradicts the d e f i n i t i o n of m(b^) . Thus there exists at most one b e X such that n(b) = 0 and m(fb) 4 m(b) + 1 . I f such an element b exists then put X Q = {b} . I f not then l e t X 0 = <j> . For every x e X - X Q we have n(x) > 0 and, as we have already shown, m(fx) = m(x) , n(fx) = n(x) - 1 . Hence m(fx) + n(fx) = m(x) + n(x) - 1 . Let X^ = {x £ X - X 0 : m(x) + n(x) i s odd} , and X 2 = {x £ X - X Q : m(x) + n(x) i s even} . Obviously the sets X , i = 0, 1, 2 are d i s j o i n t and X = X^ U X^ U . To prove the second assertion of the lemma we f i r s t consider i = 1 . Let y £ f(X^) . Thus y = f x ^ for some x^ £ X^ . We must show y £ X^ . I f y £ X^ then m(y) + n(y) i s odd. But m(fx^) + n(fx^) = mx^ + nx^ - 1 , and 56 m(fx^) + n(fx^) i s odd because f x ^ e X^- . This i s a contradiction since m(x^) + n(x^) i s also odd. Thus y ^ X ^ . Sim i l a r l y f ( X ^ ) f) X ^ = <j) . The case i = 0 i s t r i v i a l because b r fb by hypothesis. The proof i s now complete. 4 . 1 1 . Lemma. Let f : K -> N and f the continuous extension of f to BN . Then the set of fixed points of f i s the same as the closure of the set of fixed points of f . Proof: Let X = {x e N : fx = x} . Then by Lemma 4 . 1 0 N - X = X Q U X± U X 2 where f ( X ) fl X = <f> for i = 0 , 1 , 2 . Also c£ X U cl X 0 V cH X U ca X 2 = gN and obviously c£ X 0 c£ X = <j> for every i and c£ X_^ 0 c£ X = <j> for i r j • Now l e t A = {x s BN : fx = x} . Then A C cl X . Let x Q E c£ X . Consider a net } i n X converging to x D • Since f i s continuous then the sequence ( f C S ^ ) } converges to f x 0 . But f(S ) = S^ for every n and so {f(S )} = {S } which converges to x Q . Thus f x Q = x 0 and n n so x0. £ A . Thus A = cZ X and this ends the proof. The following theorem i s required for our second proof of the non-homogeneity of N* . 4 . 1 2 . Theorem. Let f be a homeomorphism of BN into N* . Then f has no fixed point. 57 Proof: Decompose N into sets G s a t i s f y i n g ( i ) "^ Gi c^ i s a f a m i l y °f d i s j o i n t sets \j {G^ .} = N ( i i ) k £ G^ for any k e N ( i i i ) {fk} U Gfc i s a neighborhood of fk. i n N L> f(N)' . (At the end of this proof we s h a l l construct such a decomposition.) For n e N l e t hn = fk i f n E G, . Thus h i s a k mapping from N onto f(N) . Let g = f ^ ° h . Hence g i s a mapping from N to N such that gn =f n for any n because gn = f ''"(hn) = f "'"(fk) for some k E N such that k £ G^ but n £ G^ (note that k E gn). Hence by Lemma 4.11 g 3 has no fixed point. From condition ( i i i ) we have fk £ G. i n N U f (N) and hn = fk for every k n £ G^ (hence one such element n i s f k ) . Thus h(fk) = fk and so h 3 ( f k ) = fk . We now have h 3x = x for x £ c£ f(N) . Suppose there exists x e gN such that fx = x . Since x £ 3N then x E ci N and i t i s very easy to j u s t i f y that f ( c i N) c ci f(N) . Thus fx £ f(ci N) implies that fx e ci f(N) . Hence from above h 3 ( f x ) = fx = x . Now g 3x = f ^"(h^x) = f ^(x)'= x , contradicting the fact that g 3 has no fixed point. Thus f has no fixed point. We now construct a decomposition {G, } 'of N s a t i s f y i n g ( i ) , ( i i ) and ( i i i ) . Consider any decomposition {Hn} of N . Then we 58 know that for j e N , f i s i n the closure of Hn for one and only-one n . Thus we l e t f. e cl H. . for some i . e N . I f 1 e H. , 1 i 1 1 x x then consider H. -{1} (place 1 i n some other component of the decom-1 1 position). Let G = H. -'{1} . I f 1 i H. then l e t G = H. . 1 \ \ • 1 \ For every j e N , repeat this process to get f_. e cl G. and j £ G. . Thus we have UG, = N , k \ G ' for any k e N , and fk e cl G . We want to show that fk U G^ i s a neighborhood of fk f k i n N o f(N) . Since fk e cl G^ then G^ e A which implies that i f k N - G k ij: A and so fk ij: c£(N - Gk> . Thus BN - cl (N - G ) i s an open neighborhood of fk i n BN . Now [BN - cS,(N - Gfc)] rt (f(N) O N) = cl G k C\ (f(N) U N) = [cl G k 0 N] L? [c£ G rt f (N)] = G, 0 fk . ' k (Note that cl G, f (N) = fk because, by construction, f. e c£ G. f k 3 3 every j and cl G^ rt c£ G^ = tj> for i ^ j ; so f £ cl G^ for j 5* k) . Thus fk U G k i s a neighborhood of fk i n N C f(N) . or 4.13. Remark. The above theorem implies that no type i s a r e l a t i v e type of i t s e l f , i . e . ( t , t) e <p for no type t . We prove this by contradiction. 59 Assume there exists some type t such that ( t , t) e ty Then t = xx = x^x and t = x xx for some x e gN - N and .X c: gN - N , where X i s discrete and x e c£ X - X (this i s a l l by d e f i n i t i o n of <J)(t) = t) . As before we l e t {N f) X : N a neighborhood of x} = x,r , x x & X ' an u l t r a f i l t e r on X which i s set isomorphic to A y , an u l t r a f i l t e r on N . Then x^ .x = Ty . But TX i s the type of A X . Hence p*x = y for some p £ P . Thus, whenever T x = Ty , we have an element p £ P A such that p*x = y . Since T VX = TX then for some p e P , p*x = x , A. a contradiction. 4 . 1 4 . We are now prepared to give another simple proof of the non-homogeneity of N* . As already mentioned hx = y implies that ty "''(xx) = ty "'"(xy) i f h i s a homeomorphism of N* onto N* . From —1 —1 Remark 4 . 1 3 , xx \ ty (xx) so xx £ tj> (xy) . Thus whenever hx = y —1 —1 then xx £ <$> (xy) . Hence i f xx e ty (xy) then hx £ y for any homeomorphism h of N* onto N* . In other words i f the type of x i s a r e l a t i v e type of y then hx = y for no homeomorphism h . But this concludes the proof since their always exist' some x,y e gN - N , x i= y , such that the type of x i s a r e l a t i v e type of y ( r e c a l l that we have proved i n Theorem 4 . 8 that for any x £ gN - N , xx produces jC 2 ° 2 types). 60 BIBLIOGRAPHY [1] N. J. Fine and L. Gillman, Extensions of continuous functions i n gN , B u l l . Amer. Math. Soc. 66 (1960), 376-381. [2] Z. F r o l i k , Fixed points of maps of gN , B u l l . Amer. Math. Soc. 74 (1968), 187-191. [3] , Sums of u l t r a f i l t e r s , B u l l . Amer. Math. Soc. 73 (1967), 87-91. [4] L. Gillman and M. Jerison, Rings of continuous functions, Van Nostrand, Princeton, 1960. [5] I. Gelfand and A. Kolmogoroff, On rings of continuous functions on topological spaces, Dokl. Akad. Nauk SSSR 22 (1939), 11-15. [6] N. Hindman, On the existence of c-points i n gN-N , Proc. Amer. Math. Soc. 21 (1969), 277-280. [7] M. Katetov, A theorem on mappings, Comment Math. Univ. Carolinae 8 (1967), 431-433. [8] D. Plank, On a class of subalgebras of C(X) with applications to gX-X , Fund. Math. LXIV (1969), 41-54. [9] M. E. Rudin, Types of u l t r a f i l t e r s , Topology Seminar Wisconsin, 1965, 147-151. [10] W. Rudin, Homogeneity problems i n the theory of Cech compactifica-tions, Duke Math. J. Vol. 23 (1955), 409-420. Ill] A. Tarski, Sur l a decomposition des ensembles en sous-ensembles presque d i s j o i n t s , Fund. Math. 12 (1928), 188-205.
- Library Home /
- Search Collections /
- Open Collections /
- Browse Collections /
- UBC Theses and Dissertations /
- The structure of βN
Open Collections
UBC Theses and Dissertations
Featured Collection
UBC Theses and Dissertations
The structure of βN Rambally, Rodney Seunarine 1970
pdf
Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.
Page Metadata
Item Metadata
Title | The structure of βN |
Creator |
Rambally, Rodney Seunarine |
Publisher | University of British Columbia |
Date Issued | 1970 |
Description | Our subject matter consists of a survey of the major results concerning the topological space βN-N where N represents the space of natural numbers with the discrete topology, and βN the Stone-Čech compactification of N . We are mainly concerned with the results which were derived during the last ten years. When there is no advantage in restricting our work to the space N we work with an arbitrary discrete space X and finally formulate our results in terms of βN-N . In some cases, pre-1960 results concerning βN-N are obtained as special cases of the results we derive using an arbitrary discrete space X . The material presented is divided into four chapters. In Chapter I, we discuss certain subsets of βN-N which can be C*-embedded in other subsets of βN-N . This study leads to the conclusion that no proper dense subset of βN-N can be C*-embedded. In the second chapter we devise a general method of associating certain classes of points of βN-N with certain subalgebras of C(N) . The P-points of βN-N form one of these classes. The answer to R. S. Pierce's question, "Does there exist a point of βN-N which lies simultaneously in the closures of three pairwise disjoint open sets" is discussed in Chapter III. Finally in Chapter IV we present two proofs of the non-homogeneity of βN-N , without the use of the Continuum Hypothesis. |
Subject |
Linear topological spaces. |
Genre |
Thesis/Dissertation |
Type |
Text |
Language | eng |
Date Available | 2011-06-14 |
Provider | Vancouver : University of British Columbia Library |
Rights | For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use. |
DOI | 10.14288/1.0080516 |
URI | http://hdl.handle.net/2429/35409 |
Degree |
Master of Arts - MA |
Program |
Mathematics |
Affiliation |
Science, Faculty of Mathematics, Department of |
Degree Grantor | University of British Columbia |
Campus |
UBCV |
Scholarly Level | Graduate |
AggregatedSourceRepository | DSpace |
Download
- Media
- 831-UBC_1970_A8 R34.pdf [ 2.98MB ]
- Metadata
- JSON: 831-1.0080516.json
- JSON-LD: 831-1.0080516-ld.json
- RDF/XML (Pretty): 831-1.0080516-rdf.xml
- RDF/JSON: 831-1.0080516-rdf.json
- Turtle: 831-1.0080516-turtle.txt
- N-Triples: 831-1.0080516-rdf-ntriples.txt
- Original Record: 831-1.0080516-source.json
- Full Text
- 831-1.0080516-fulltext.txt
- Citation
- 831-1.0080516.ris
Full Text
Cite
Citation Scheme:
Usage Statistics
Share
Embed
Customize your widget with the following options, then copy and paste the code below into the HTML
of your page to embed this item in your website.
<div id="ubcOpenCollectionsWidgetDisplay">
<script id="ubcOpenCollectionsWidget"
src="{[{embed.src}]}"
data-item="{[{embed.item}]}"
data-collection="{[{embed.collection}]}"
data-metadata="{[{embed.showMetadata}]}"
data-width="{[{embed.width}]}"
data-media="{[{embed.selectedMedia}]}"
async >
</script>
</div>
Our image viewer uses the IIIF 2.0 standard.
To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0080516/manifest