• Aucun résultat trouvé

by gm G. sup (cf. 1 ch. 5, w 1). A problem on the union of Helson sets

N/A
N/A
Protected

Academic year: 2022

Partager "by gm G. sup (cf. 1 ch. 5, w 1). A problem on the union of Helson sets"

Copied!
4
0
0

Texte intégral

(1)

A p r o b l e m o n t h e u n i o n o f H e l s o n sets

E. GALA~S

Institut Mi~tag-Leffler, Djttrsholm, Sweden

Let G be a locally compact abelian group and let G be its dual group.

Definition 1. A compact subset E c G is called Kroneeker if for every con- tinuous function f on E of modulus identically one (If(x)l =~ 1, V x E E ) and for every ~ > 0 there exists g E G such t h a t

sup

I f ( x ) - x(x)l ~_ ~

(cf. [1] ch. 5, w 1).

x E E

We shall denote b y M(G) the set of all bounded complex valued R a d o n measures on G and b y M(E) the elements of M(G) with support in a compact subset E of G.

W e shall denote b y C(E) the set of all continuous complex valued functions on E.

Definition 2. A compact subset E of G is called a Helson o~-set (Ha-set) ff there exists a constant ~ > 0 such t h a t

tl~lt~ = sup l,~(x)i ~ ~ tl~lt

for every ~u E M(E), (observe t h a t then 0 < ~ ~ 1).

I f K is a compact subset of G we shall write Gp (K) for the group generated by g m G.

I n this paper we shall prove the following theorem.

T~OREM. Let K be a totally disconnected Kronecker subset of G and D a countable comloact Ha-subset of G such that

Gp (K) N Gp (D) ---- {0}.

Then K U D is an Ha-set.

(2)

194 ARKIV :FOR NATEMATIK. YO1. 9 :No. 2

_~emarks. Varopoulos [3] has p r o v e d t h a t if K is a n y t o t a l l y disconnected c o m p a c t Hi-subset of G a n d D is a n y H~-subset of G t h e n K 0 D is Ha(~) w i t h fi(a) > O.

The interest of our t h e o r e m lies in t h e f a c t t h a t in this special case we h a v e

=

l~owever we m u s t p o i n t o u t t h a t the conclusion o f o u r t h e o r e m fails if we replace t h e condition ))Kronecker~) b y ~)HI~ for t h e set K . To see this observe t h a t t h e set { ( x , 0 ) , ( - - x , 0 ) , ( 0 , ~ y ) , ( 0 , y ) } c R ~ ( x , y C R ) is n o t a n Hi-set of R 2 despite t h e f a c t tha~ it is t h e u n i o n o f t w o Hi-sets

E 1 ~--- {(x, 0 ) , (-- x , 0)}, E 2 = {(0, - - y ) , (0, y)}

t h a t satisfy Gp (E~) f] Gp (E2) = {0}.

W e shall prove first:

L ] ~ M A . Let K and D be as in the theorem with D = { x 1 , x 2 . . . . , x , } a f i n i t e set. T h e n for every Z E G there exists

Z, "-> 1 uniformly on K

r~---> o o

for all j = 1 , 2 . . . r.

{X.},,~, Z,, e G such that and X,(xi) --> Z(xi)

n---> o o

Proof. L e t L be t h e set of points t E T" for which there exists a sequence {Z, e }~=1 such t h a t Z,[~ - ~ 1 u n i f o r m l y , a n d (Z~(xl), Z,(x2), . . . . g~(x~)) --~ t

in T . " - ~ ~

L e t us also d e n o t e b y H ~ {Z(Xl), g(x2) . . . Z(x,)}. W e observe t h a t b o t h L a n d H are closed subgroups of T" a n d t h a t L c H . W e shall prove t h a t L = H .

T o w a r d s t h a t let us suppose b y contradiction t h a t L ~= H . There exists t h e n a character O C T~ o f T ~ such t h a t

O ( L ) = I a n d O(H) % 1. (1)

t h a t

O(lim ( g n ( x l ) , . . . , gn(x~))) = lim O(z,(Xl) . . . g,(x,)) -~ lim X-(Y) = 1.

n n n

c O

B u t this implies t h a t i f {~vn E G},=I is a sequence o f characters such t h a t Since O E T " a n d O(H) 4= 1, there exist n 1 , n 2 . . . n, E Z such O(•(Xl) , Z(x~) . . . Z(x,)) = z ( n l x l + n~x 2 + . . . + n~x~) = Z(Y) V g e G where y = nix 1 + n~x~ t - . . . . + n,x, :/: O.

L e t n o w {xn}~=l be a sequence of characters such t h a t t h e sequence of points {(gn(xl) , . . . , g,(xr)) E T'}n~=l converges t o w a r d s a p o i n t o f L, (1) implies t h e n t h a t we have:

(3)

A P R O B L E M O N T H E U N I O N O F H E L S O N S E T S 195

~n [K ---> 1 uniformly on K t h e n ~n(y)--> 1

It--> ~3

and from t h a t using Corollary i of [2] we deduce the required result t h a t y E Gp (K).

Proof of the theorem.

Case 1. D

is finite.

Let

/ x E M ( K U D ) ;

w e can write / x = / ~ l + / x 2 where / 4 E M ( K ) and

I~ E M(D).

Then

ll~xll~

=

ll~l[

a n d

I]/~=ll~ => ~l[~I[.

We first observe t h a t there exists ~ E R a n d Z l , Z2 E G such t h a t

Re e'~,(zl) >_ 11/611 - ~- (3)

Indeed choose q E R and Ze E G such t h a t (2) holds. L e t us show t h a t we can find h E ~ satisfying

(3).

We can choose ~ E

C(K), etep

[~1 ~ 1 such t h a t

f ~ d ~ >_ I I ~ l l - 8/2

a n d then since K is Kronecker we can approximate ~ uniformly b y h E G as close as we like. This shows t h a t Zl can be chosen to satisfy (3).

Now from the lemma we can find a sequence ~n E G such t h a t ~v~ --> 1 uniformly on K a n d

We have t h e n

and Therefore

YJ" [D --+ Z ~-1z2 [D"

iq~ ^

R e e ~2(XxY-) - + R e e ~ ~ ( Z 2 ) 9

I1~11oo >~ sut) I~(x~w.)l ~ sup Re {e'~p(X~.)} ~ II/~I1 -4- ~1I/~.11 -- 2~

n r t

>_

~(11~11 + I1/~II) - 2~ ___ ~1>I1 - 2e.

S i n c e e is arbitrary the conclusion of the theorem follows.

Case 2. D

is countable.

L e t # E M ( K U D ) we can write

P = / 6 +/~2 +/~3, /11 q

M(K), /as,/~3

E

M(D)

(4)

196 A R E / V F O R I~TEMATI]~. r e 1 . 9 N o . 2

with the support of /~2 Tying in a finite subset of D and 11/~311 ~ s (s ~ 0).

Now b y case 1

Therefore

1[Pll

And since e is arbitrary this completes the proof of the theorem.

I t is m y pleasure to express m y gratitude to Dr. N. Th. Varopoulos for his advice and critisisms and the Mittag-Leffler institute for its stimulating hospitality.

References

1. I=~UDIN, W., Fourier analysis on groups. Interseience, No 12, 1962.

2. V ~ o P o u L o s , N. TH., Sur les ensembles de Kroneeker. C. R. Acad. Sci. Paris 268 (1969), 954-- 957.

3. --))-- Groups of continuous functions in harmonic analysis. Acta Math. 125 (1970), 109-- 154.

Received May, 1970 E. Galanis

D e p a r t m e n t of P u r e M a t h e m a t i c s 16, Mill L a n e

Cambridge E n g l a n d

Références

Documents relatifs

In particular, it was shown that right strongly non-singular, right semi- hereditary rings are left semi-hereditary too, so that we shall call such rings right strongly

Leffler modules is a Mittag-Leffler module has the property that any intersection of finitely generated left ideals is a finitely generated left.. ideal and all its

customers simultaneously present in the system during a busy cycle the generating function of the distribution is known. For every one of these variables the

- We can assume without loss of generality that there exists a discontinuous character (p on some commutative, unital Frechet algebra A. The advantage of the point of view

Si l'on met maintenant chacune des fonctions F- k (x', c^) sous la forme B (3), on obtient une nouvelle expression générale d'une fonction analytique uniforme ayant un nombre fini

Toute utilisation commerciale ou impression systématique est constitutive d’une infrac- tion pénale. Toute copie ou impression de ce fichier doit contenir la présente mention

POONEN, Using elliptic curves of rank one towards the undecidability of Hilbert’s tenth problem over rings of algebraic integers. Algorithmic Number Theory

Par des bourses accordées, à des étudiants en Mathé- matiques des deux sexes, appartenant aux pays men- tionnés ci-dessus, pour la poursuite de leurs études, dans leur pays ou