Projections in Normed Linear Spaces

Projections in Normed Linear Spaces
Author(s): Dwight B. Goodner

Source: Transactions of the American Mathematical Society, Vol. 69, No. 1 (Jul., 1950), pp. 89-
108
Published by: American Mathematical Society
Stable URL: http://www.jstor.org/stable/1990599 .
Accessed: 09/12/2014 19:44
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
American Mathematical Society is collaborating with JSTOR to digitize, preserve and extend access to
Transactions of the American Mathematical Society.
http://www.jstor.org
This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

All use subject to JSTOR Terms and Conditions
PROJECTIONS IN NORMED LINEAR SPACES(1)
BY
DWIGHT B. GOODNER
1. Introduction.The object of this paper is to study normedlinear spaces
which have what we shall call propertyP8 and to study in particular those
normedlinear spaces which have propertyP1. A normedlinear space X has
propertyP8, s > 1, if and only if forevery normedlinear space Y containing
X, thereexists a projection T, IIT|| < s, of Y onto X.
The terminologyused is forthe most part as given in Banach [3](2) and
Murray [16]. Thus a Banach space means a complete normedlinear space;
a linear transformationmeans a bounded and distributivetransformation;
and a projection means a linear transformationT with the propertythat
T2=T. Our notation is, we believe, for the most part standard and it is
chosen so as to be the same as that of our referenceswheneverpossible.
In ?2 we study some of the general properties of spaces which have
propertyP8, and we develop a necessaryand sufficient conditionfora space
to have property P8. Lemma 2.3 is of particular importance and is used
throughoutthe remainderof this paper.
In ?3 we discuss some problemswhich led to the formulationof Akilov's
theorem [2, p. 646] and also give a new proofof the theorem.This theorem
conditionforcertainnormedlinearlattices to have property
gives a sufficient
Ps.
?4 is a study of some of the conditionsimposed upon the unit sphere of
a space which has propertyP1. An importanttechnique,used in many later
proofs,is developed and used to discover several necessary conditionsfor a
space to have propertyP1.
In ?5 we characterizenormed linear spaces which have propertyP1 and
which have extremepoints on theirunit spheres.A partial orderingis intro-
duced into each space, and then a theoremof Clarkson's is applied to show
that the space is equivalent to the space of all continuous real-valued func-
tions over some compact Hausdorffspace H.
Kakutani [10, p. 998] has shown that everyabstract (M)-space has a con-
crete representation.In ?6 we find that an abstract (M)-space which has
propertyP1 has a concreterepresentationas the space of all real-valued con-
tinuous functionsover some compact Hausdorffspace H. The space H is
found to have the curious property of being extremallydisconnected. A
Presented to the Society, September 1, 1949; received by the editors August 23, 1949.
(1) This paper is a condensation of a doctoral dissertationwrittenat the Universityof
Illinois under the directionof ProfessorMahlon M. Day.
(2) Numbers in brackets referto the referencescited at the end of the paper.
89

90 D. B. GOODNER [July
necessaryand sufficient conditionforspaces having propertyPi to be reflexive

is given. Finally, the known examples of spaces having propertyPi are dis-
cussed.
2. Spaces which have propertyP8. In this section we shall study some of
the propertiesof normedlinearspaces whichhave what we shall call property
P8.
2.1 DEFINITION.A normed linear space X has propertyP8, s _ 1, if and
only ifforeverynormedlinear space Y containingX thereexistsa projection
T, TI|<5s,of Yonto X.
In parts of our work it will be necessary to specifythat Y is a Banach
space. This requirementinvolves no real loss of generality,since any normed
linearspace YDX is contained in a completespace Y' such that a projection
T of norms of Y' onto X is a projectionof norms of Y onto X.
Our firsttheoremshows that a normedlinearspace whichhas propertyPJ
is a Banach space.
2.2 THEOREM. If X is a normedlinear space whichhas propertyP8, thenX

is complete.
Proof. Let Y be a Banach space containingX. Since X has propertyP8,

there exists a projection T, ||T|| <s of Y onto X. Let {x, } be any Cauchy
sequence in X and let limn,OO x, =y. Then y is in Y since Y is complete.
Therefore limnO T(xn) = T(y) since T is continuous. But lim_OT(xn)
-lim"_0 xl,and hence T(y) =y. Thereforey is in X and X is complete.
Two normed linear spaces are equivalent [3, p. 180] if there exists be-
tween them a one-to-onetransformationU which is linear in both directions
and || Ull= 11
U-lI1= 1. This definitionleads us to the followingextremelyuse-
ful lemma.
2.3 LEMMA.If a normedlinear space X is equivalentto a normedlinear

space Y and if X has propertyP8, then Y has propertyP,.
Proof. Let W be a normedlinear space containing Y and let

Z = XU (WnGCY) = X'U W'
where CY is the complementof Y and W' = WnGCY.Then if V is the equiva-

lence between X and Y, we can extend V to U on Z by defining
U(z) = U(w') = w' if zEW'CZ
and
U(z) = U(x) = V(x) = y if zEXCZ.
Then U is a one-to-onetransformationbetween Z and W.
We definethe operation of addition between the elementsof Z to mean

1950] PROJECTIONS IN NORMED LINEAR SPACES 91
Z + z2 = X if U(zl) + U(z2) = wl + w2 = y = U(x)

and to mean
Zl + Z2 = WI if U(zl) + U(z2) = wl + w2 = w' = U(w').
If a is a real number,we definescalar multiplicationby
azl = 2= U-1[aU(z)]
Finally, we definea norm in Z by
jjzjj = 1U(z)| = llwll.
It is easily verifiedthat these definitionsmake Z a normedlinear space and
that U is an equivalence betweenZ and W.
Since X has propertyP8, there exists a projection T, T|I<5s,of Z onto
X. Let T'= VT U-1. T' is a projectionof W onto Y and T'JJ ?s. Hence Y
has propertyP8, which was to be shown.
The proofof our next lemma uses the concept of the adjoint transforma-
tion T* of a linear transformationT. If X is a linear subspace of the normed
linear space Y and T is a projection of Y onto X, the adjoint mapping T*
of T is definedby the relation T*(f) =g implies g(y) =f(Ty) for all y in Y,
where gC Y* and fCX*. ||T*| = ||T|; T* is one-to-oneon X* into Y*; and
if X and Y are complete, X is isomorphicto Y/T(0)-1 and T* is an iso-
morphismof X* and [T(O)-']' [6, Theorem 1.13, chap. 4].
2.4 LEMMA. If the normedlinear space X has propertyP8 and if Y is a
normedlinear space containingX, thereis a subspace G of Y* whichis iso-
morphicto X*, and thereis a projectionTA, ITAII s, of Y* ontoG.
Proof.Since X has propertyP, and Y is a normedlinear space containing
X, there is a projection T, 11T|<s, of Y onto X. Let T* be the adjoint
mapping of T and let G = T*(X*). Then
G = {hIh(y) = k(Ty) = T*f(y), h E Y*, f G X*}.
Define the transformationTA by
TAh= g implies g(y) = TAh(y) = h(Ty), g E G.
The reader may verify that TA is a projection, ||Tz| = ||T||, of Y* onto
G= T*(X*).
The finaltheoremin this chapter gives a necessaryand sufficientcondi-
tion fora space to have propertyP8.
2.5 THEOREM. A normedlinear space X 7haspropertyP8 if and onlyif for
each Banach space Y containingX thereexist a closed linear subspace G of
Y* and a projectionT- of Y* ontoG such that
a. |TAIl <s,

b. Y=G?iEX,
c. TAh=g implies TAh(gi+x)=g(g?+x)=h(x), hGY*, gCG, g-CG?,
and xCX.
Proof. Suppose TA and G exist. Then Y=G?iEX by hypothesis. (G.
= {ylg(y)=0, yC Y, gCG}.) If we defineT1 by the relationTi(y) = Ti(gl+x)
=x, T1 is a projectionof Y onto X [16, p. 139].
T 1(X*) {h| h(y) = Tlf(y) = f(Tly), f C X*}

= {h h(y) = h(g + x) = h(x)} =G.
Define TA by TA(h)=g implies g(y) = Tlh(y) =h(Tly) for all y in Y. Then
Tl(h) = TA(h) for all h, so IIT1A| = || TA| < s. But || TilI = || T||I <s. Therefore,
X has propertyP8.
If X has propertyP, and if Y is a Banach space containingX, thereis a
projection T, |T|| <s, of Y onto X. T*(X*) = [T(O)ft]', and by Lemma 2.4
there is a projection TA, ||TA| <s, of Y* onto [T(O)-']'. Let G= [T(O)-']'.
Then G? = [T(O)-']'l = T(0)-1. Hence Y= G ODXsince Y= T(0)-1 dX. Con-
dition c is satisfiedby our definitionof TA; thereforethe proofis complete.
3. Akilov's theorem. Akilov in a short paper [2, p. 643] considered two
problemson the extensionof linear operators fromone normedlinear space
to anotherand in a later paper [1, p. 417] showed that the propertiesstudied
in the two problemsare equivalent. Day [5, p. 241 ] noted that both problems
are equivalent to the studyof spaces withpropertyP,. We shall state Akilov's
problemstogetherwith a third problemand show that the three properties
involved are equivalent; that is, a space whichis a solution (has the property
described in the problem) to one of the problemsis a solution to the other
two. However, beforewe state the problemsa definitionis necessary.
3.1 DEFINITION. Given a linear subspace X of a normed linear space Z
and a linear transformationT on X into a normed linear space Y, we shall
call a linear transformationT1 on Z into Y an extensionof T if TI(x) = T(x)
forall x in X and |T1Ti =||T1.
PROBLEM 1. Given a normed linear space Y, does there exist for each
normedlinear space X, each linear transformationT of X into Y, and each
normedlinear space Z containingX, an extensionT1 of T on Z into Y?
The familiarHahn-Banach theorem [3, p. 55] provides an affirmative
answer to problem 1 if the normedlinear space Y is the real line.
PROBLEM 2. Given a normed linear space X, does there exist for each
normed linear space Y, each linear transformationT fromX into Y, and
each normedlinear space Z containingX, an extensionT1 of T on Z into Y?
PROBLEM 3. Given a normed linear space X, does there exist for every
normed linear space Z containingX, an extension T1 of the identitytrans-
formationI of X onto X such that T1 transformsZ onto X?
In this case T1, if it exists,is a projectionof norm one of Z onto X and

any space X which provides a solution to problem3 has propertyP1.

Proof that problem 2 implies problem 3. Suppose X is a space which
satisfiesthe conditionsof problem2 and let Y=X. Then the identitytrans-
formationI of X onto X is a linear transformationof X into Y. Hence, by
problem2 thereis an extensionT1 of I such that T1 transformsZ into X; that
is, thereis a projectionT1, || Till = 1, of Z onto X.
Proof that problem 1 implies problem 3. Suppose Y is a space which
satisfiesthe conditionsof problem 1 and let X= Y. Then the identitytrans-
formationI of X onto X is a linear transformationof X into Y. Hence by
problem 1 there exists an extension T1 of I such that T1 transformsZ into
X; that is, there is a projection T1, IT1! I
=1, of Z onto X.
Proof that problem 3 implies problem 2. Suppose X is a space which
satisfiesthe conditionsof problem3 and let T be a linear transformationof
X into Y. If Z contains X, there is a projection P, |P|| = 1, of Z onto X.
Then the linear transformationT1= TP is an extensionof T, and T1 trans-
formsZ into Y.
Proof that problem 3 implies problem 1. Suppose Y is a normed linear
space which satisfiesthe conditionof problem3 and let T be a linear trans-
formationof a normed linear space X into Y. If m(S) is the space of all
bounded real-valued functionson S, the unit sphere of Y*, we can imbed
Y in m(S) by an equivalence U definedby U(y) =y' implies s(y) =y'(s) for
all s in S. Suppose Z is a normedlinearspace containingX. Then by Phillips'
Corollary 7.2 [18, p. 538] the linear transformationUT of X into m(S) has
an extension T1 which transformsZ into m(S). Since Y has propertyP1,
Y'-U(Y) has propertyP1 by Lemma 2.3. Hence thereis a projectionP, IIPII
= 1, ofmr(S)onto Y'. Let T' = U-'PT1. T' is the requiredextensionof T which
transformsZ into Y.
The proof of our next theorem is similar to the proof that problem 3
implies problem 1.
3.2 THEOREM. If thenormedlinear space Y has propertyP8, if T is a pro-

jection of normt, and if T projectsY ontoa subspaceX of Y, thenX has prop-
ertyPst.
Proof. If m(S) is the space of all bounded real-valued functionson S, the

unit sphere of Y*, we can imbed Y in m(S) by an equivalence U definedby
U(y) =y' implies s(y) =y'(s) for all s in S. Suppose Z is a normed linear
space containingX. Then by Phillips' Corollary 7.2 [18, p. 538] the linear
transformationU of X into m(S) has an extension V which transformsZ
intom(S). Since Y has propertyP8, Y' = U( Y) has propertyP8 by Lemma 2.3.
Hence thereis a projectionP, IIPII< s, of mr(S)onto Y'. Let T'-= TP-P V.
Then T' is a projectionof norm less than or equal to st, and T' projects Z
onto X. Hence X has propertyP8t.
Akilov's study of problems 1 and 2 led him to the formulationof a suffi-

cient condition (which we shall call Akilov's theorem) for a normed linear
space to have propertyP8. His proofof the conditionwas based on Kantoro-
vitch's extensiontheorem [13, p. 285]. The proofwe presenthere resembles
the proofof the Hahn-Banach theoremand does not make use of the Kan-
torovitchtheorem.
A space is of type B1+ [2, p. 643] if it is simultaneouslyof types K5 and B1
[11, pp. 210-212]. Using this definitionwe state and prove Akilov's theorem
[2, p. 646] in the followingform [5, p. 241].
3.3 THEOREM. If thenormedlinear space X is oftypeBi+,if theunitsphere
C of X has a least upper boundxo, and if IIxolI=s, thenX has propertyP..
Proof. Let Z be a normedlinear space containingX, let p be a sub-addi-
tive, positive-homogeneoustransformationon Z into X, and let f be a dis-
tributive transformationon X into X such that f(x) < p(x) for all x in X.
SubstitutingX forthe real numbersin a proofof Banach's [3, p. 28], we can
show there exists a maximal functionFo which is definedon all of Z and
which has the propertiesthat Fo(x) =f(x) forall x in X and Fo(z) ?p(z) forz
in Z.
Let p be the sub-additive positive-homogeneousfunctiondefinedby the
relation p(z) =xoIIzfl.If we take the identityfunctionI of X onto X as
1(x) ?p(x) forall x, and I can be extended to a distributivefunctionT such
that T(z) < P(z) = xoIIzlIand T(x) =x. T is a projection, T =s, of Z onto X.
Hence X has propertyP8 and the proofis complete.
Let I,n, be the space of ordered n-tuples of real numbers al
- {
(a,, a2, , an) }, 1 !n< oo, with the norm Ilall= (E f1 | ail P)l'P, 1 <p
< oo. If the usual partial ordering-that is, a'> a" if and only if a' >a" for
all i-is introduced,it can be easily verifiedthat Ip,n satisfiesthe conditions
of Akilov's theoremand has propertyP8, s =nlP.
In a later section we will considerother examples of spaces which satisfy
Akilov's theorem.However, forthe present,we shall content ourselves with
a briefexaminationof two spaces which do not have propertyP8 but which
satisfyall but one of the conditionsof the theorem.
Sobczyk [19, p. 945] proved there is no projectionof (m), the space of
bounded sequences, onto its subspace (co), the space of sequences convergent
to zero. Hence (co) does not have propertyP,. (co) is a space of type B1+but
its unit sphere has no upper bound in the space.
Fichtenholtzand Kantorovitch [7, p. 92] proved there is no projection
of M(O, 1), the space of all bounded functionson the unit interval,onto its
subspace C(O, 1), the space of continuous functionson the unit interval.
Hence C(O, 1) does not have propertyP8. C(O, 1) is a space of type B1, and
its unit sphere has a least upper bound of norm one. Hence C(O, 1) satisfies
all conditions of Akilov's theorem except the one requiring that all sets
bounded above have a least upper bound in the space.

4. The unitsphereofa space whichhas property

P1. Let X be a normed
linear space and let C be the unit sphere of X. The purpose of this section
is to study some of the conditionswhich are imposed on C if X has property
P1. In order to do this we make frequentuse of a technique which is of im-
portance in much of our later work. Brieflyit is this: our firstlemma tells us
that X has propertyP1 if and only if the unit sphere K of any normed linear
space Y containingX projectsonto C; we let Y be a linear (as yet unnormed)
space containingX and of dimensionone greater than the dimension of X
and findpoints in Y not in X which are symmetricabout the originand such
that theirconvex hull intersectsX in or on C; we then norm Y by lettingK,
the unit sphere of Y, be the closed convex hull of these points and C; then if
we have selected the points of K so that at least one of them must project
outside of C, we know that X does not have propertyP1.
4.1 LEMMA.The normedlinear space X withunit sphere C has property
P1 if and onlyiffor each normedlinear space Y containingX thereis 'projection
T of Y ontoX whichprojectstheunit sphereK of Y onto C.
Proof. If foreach Y containingX thereexistsa projectionT of Y onto X,
such that T(K) =C, then 11TIl IT(y)I=1ub,11,1 lIxII
=lubIIyII<1 =1. Hence X
has propertyP1.
Suppose that forsome Y containingX thereis no projectionof Y onto X
or that if a projection T of Y onto X exists, there is a yo in K such that
T(yo) is not in C. Then
ITI| |IIT(yo)II
> 1
and X does not have propertyPi.
4.2 THEOREM.If p and q#p, -p are extremepoints of theunit sphere C
of thenormedlinear space X whichhas propertyPi, theline segmentconnecting
p and q lies whollyon thesurfaceof C.
Proof. Assume the line segmentconnectingp and q does not lie whollyon
the surfaceof C. Then (p+q)/2 is inside C since C is convex. Also (p+q)/2 is
not equal to zero for q is not equal to -p. Hence there exists a point r
=t(p+q) such that 1/2<t<1 and r is in or on C. Let b=(2t-1)(p+q). Let
Y be a linear (as yet unnormed) space containingX and of dimensionone
greaterthan the dimensionof X; let yo be a fixedelementof Y not in X; let
K, the unit sphere in Y, be the smallest closed convex set containingC and
the eight points ? p ? yo, ? ? (yo+ b).
Our next step is to show that KnX =C. Let S be the convex hull of
?P?yo, ?q?(yo+b). If a point in C is of the formas +(1-a)c, s in S,
c in C, a 0 a real number,then s is in SQnX; so it sufficesforour purpose to
prove SnX is contained in C. S intersectsX in a polygon which can have
extremepoints only where a line segmentof S connects two extremepoints

of S which are not in C. Checkingall possibilitieswill show that the extreme

points of SnX are + p, ? q, and + r. But these points were chosen so that
the hexagon they determineis in C. Hence KQ,X= C.
Our finalstep is to show that everyprojectionof Y onto X carriesa point
of K outside of C. Let T be a projectionof Y onto X and let T(yo) = xo. Then
T carries the line segment fromP+yo to p-yo onto the line segment from
p+xo to p-xo. Since p is an extremepoint of C, eitherone end point of the
segmentis outside of C or both end points equal p and xo= 0. For T to be of
norm one, the latter condition must hold. Then T(q? (yo+b)) =q? b. But q
is an extremepoint so jjq+b||>1 or ||q-b||>1, and consequently IIT|I>1.
Hence our assumptionwas false and the theoremis proved.
4.3 THEOREM. If thenormedlinear space X has propertyP1 and if p is an

extreme pointoftheunitsphereC ofX, thenany pointq ofnormone,q p, -p,
to eitherp or -p bya line segmentlyingentirelyon thesurface
can be connected
of C.
Proof. Assume thereis a point q of normone such that the line segments
pq and -pq do not lie on the surface of C. Then (p+q)/2 and (-p+q)/2
are inside C and are not the zero element. Hence there is a point r1in C be-
tweenthepoints(p+q)/2 and (p+q)/IIp+qII and a pointr2in C betweenthe
points(-p+g)/2 and (-p+q)/II-p+qII. Let e,=1-IIrlIIand e2=1-11r211-
e1> 0 and e2> 0. Let e equal the smallerof el and e2and let k equal the larger
of IIp+qII and ||-p+qH|. Then ifbi=(e/k) (p+g) and b2=(e/k)(-p+q), ibillI
<e. Let Y be a linear (as yet unnormed) space which contains
<e and IIb2|I
X and which is of dimensionone greater than the dimensionof X, and let
yo be an elementof Y not in X. Let K, the unit sphere in Y, be the smallest
convex set containing C and the points + p ? yo, ? q ? (yo+ bi+ b2). Then
KnX=C.
Since X has propertyP1, there is a projection T, || T|| = 1, of Y onto X.
Let T(yo) =xo. T carries the line segmentfromp+yo to p-yo onto the line
segmentfromp+xo to p -xo. Since p is an extremepoint of C, eitherone end
of the line segment is outside of C or both ends equal p and xo= 0. For T
to be of norm one the latter condition must hold. Then T(q+yo+bl+b2)
=q+bl+b2. IIq+bl+b2II= 1+2e/kI >1. But jjq+yo+bi+b2jl =1; so, by
Lemma 4.1, X does not have propertyP1, and our assumption has led to a
contradiction.Hence the theoremis proved.
4.4 LEMMA. Let X be a normedlinear space withunit sphereC and let Y be

a linear space containingX and of dimensionone greaterthanthedimensionof
X. Let E be a subsetof C of diameterd <2 and let a=minimum (1, 2/d-1).
If yois fixedin Y and is notin E, and if K is theclosedconvexhull on C and
the set of all points of theformex+m(ax+yo), e- 1, m= ?1, x in E, then
KnX=C.

Proof. Let x1and X2be any two points of E and consider the convex hull
C' on exi+m(axl+yo) and e'x2+m'(ax2+yO) where e, e', m, and m' are
equal to ? 1 and a is positive. C'n(XXCC if w= (ex,+ax,+e'x2-ax2)/2 is in
C. If e=e', w-=((e+a)xl+(e-a)x2)/2 and a?< 1 implies I ? 1. If e = - e
w=(e+a)(xi-x2)/2 and thisis in C if |e+a| <2/1x1-x211. This is possible
fora>O only when IxI-x2l <2; a suitable choice is, in fact, 2/'| xl-X2| -1.
Then w lies in C if a=minimum (1, 2/d-1). Hence KQXX= C.
4.5 COROLLARY. If thenormedlinear space X has propertyP1, each extreme
pointof theunit sphereC is at a distance2 fromeveryotherextremepointof C.
Proof. Form K as described in Lemma 4.4 and let K be the unit sphere
of Y. Since X has propertyP1, thereis a projection T, jjT|| = 1, of Y onto X.
Hence exi+m(axi+yo) in K must project into ex1+m(axi+T(yo)) =ex1
+m(axl+u) in C. Then axi+-u is in (C-xi)fC\(C-+xi). If xi is an extreme
point, (C-xl)n(C+xi) =O, so ax1+u=O or u= -ax1. If both xi and X2 Xl
are extremepointsof C, u - ax1=-ax2, so a mtistbe zero. Then by Lemma
4.4 the set E consistingof x1 and x2 must be of diameter 2.
By a face of the unit sphere C of the normedlinearspace X we shall mean
a maximal convex subset of the surfaceof C.
4.6 LEMMA. If the normedlinear space X has propertyP1 and if E of
diameterless than or equal to one is a subsetof theunit sphereC, thenall faces
of C whichcontainpointsof E intersectin somepointof C.
Proof. Let Y be a linear (as yet unnormed) space containing X and of
dimension one greater than the dimension of X. Form K as described in
Lemma 4.4 and let K be the unit sphere of Y. There is a projection T,
|T| =1, of Y onto X since X has propertyP1. If ex,+rm(ax,+yo) is in K,
ex1+m(ax1+T(yo)) is in C. Then ax1+T(yo)=ax1+u is in (C-x1) n (C+xi).
Hence u is in (C-xi-ax1)Qn(C+x1-ax1). Since the diameter of the set E
which contains xi is less than or equal to one, a 1, u is in (C - 2xi) \C,,
and -u is in CQ (C+ 2x1). If IIxlI =fi(x1) = If,=I1 forfi in X*, then Ht
- {x fi(x) = 1 } separates C fromC + 2x1, so u lies in H1inC. Suppose x1 and
x2 are in E and x| xjjI I x21I=1; then there exist fi and f2 in X* such that
IIfilI=If2| =1 and f1(x1)=f2(x2)==1. Let Fi-={x xEC andfi(x) =1, i=1, 2J
-u is in CQn(C+2x2) as well as CC(C+2x,). Then F&r)F2D -u. Hence all
faces of C which contain points of E intersectin a common point.
The proofof our firstcorollarymakes use of Mazur's theorem [14, p. 731
that if K is a convex body in a normedlinear space X and if E is a flatsub-
set of X which contains no interiorpoints of K, then there exists a hyper-
plane H such that HIDE and K lies on one side of H.
4.7 COROLLARY. If X, Y, E, yo T, and u are as in the lemma, if x4i
= X211= 1, if xlEE, and if thelinesegment
fromxl to X2 is on thesurfaceofC,
thentheline segment
from-u to X2 is on thesurfaceof C.

Proof. By Mazur's theoremthere is a hyperplaneH containingthe line

throughxi and x2 ancl such that C lies on one side of -I. Then by the lemma
-u belongs to HnC. Since HnC is a convex set on the surface of C, the
from- u to x2must lie on the surfaceof C.
line segmnent
The next corollaryis fromAkilov [1, p. 417].
4.8 COROLLARY. No n-dimensionalnormedlinear space (1 <n < oo) such
thatthetangenthyperplaneat each pointoftheunitsphereis unique has property
P1.
Proof. Let X be an n-dimensionalnormed linear space (1 <n < oo) with
propertyP1. Since X is finite-dimensional its unit sphere C has an extreme
point e. Let E be the set of all pointson the surfaceof C at a distance less than
or equal to 1/2 frome. Let Y, T, yo, and u be as in the lemma. Then all
tangent hyperplanes to C which contain points of E must pass throughe.
Hence a tangenthyperplaneat e is not unique.
4.9 THEOREM. A normedlinear space X does not havepropertyP1 if there
exists in X a hyperplaneH= { xIf(x) = =f
I fH}I such thatHnCC (C the unit
sphereofX) is an n-dimensionalsimplex,3 < n.
Proof. Let Y be a linearspace containingX and of dimensionone greater
than the dimension of X. Suppose H exists and let Pi, P2, * * *, p,Pbe the
verticesof F=HnC. The centersof the faces of F are qi= (1/(n-1)) ,joi P
wherei,j-1,2, ,n.
flqi- qklf _ IlPk - Pill 1.

n-i
__
Hence E== {qi}, = 1, 2, * * *, n, is a set ofdiameterless than or equal to one.

Let yo be a fixed point in Y which is not in X and construct the points
?qi?(qi+yo), i=1, 2, * * *, n. Then if K is the closed convex hull of these
points aind C, KnX= C by Lemma 4.4. Let K be the unit sphere of Y. If
X has propertyP1 there exists a projection T, || T| = 1, of Y onto X. T( ? qi
? (qi+yo)) = ?qi? (qi+u) is in C and by our previous work -u must be
on everyface of F. Since this is impossible,X cannot have propertyP1.
Our nextcorollaryis fromAkilov [1, p. 417].
4.10 COROLLARY. No abstract(L)-space (AL) of dimensiongreaterthan
or equal to 3 has propertyP1.
Proof. Realize (AL) as a space L' of measurable functionswhich are
integrable with respect to a completely additive measure on some space
S [9, p. 533]; take three disjoint subsets Si, i=1, 2, 3, of S such that
each Si is of positive finite measure, and let L be the subspace of L'
spanned by the three characteristicfunctionsof these sets. Then let Tf=g
mean that g(s) =0 if s is not in one of the threesets and let g(s) be equal to

the average value off over Si, i = 1, 2, 3, if s belongs to Si. It may be easily
verifiedthat T is a projection,IIT|| = 1, of L' onto L. Then ifL' has property
P1, L has propertyP1 by Theorem 3.2. But L does not have propertyPI by
Theorem 4.9. Hence L' does not have propertyP1. Lemma 2.3 then tells us
that (AL) does not have propertyPi.
5. A characterizationof certain normed linear spaces which have prop-
erty P1. The purpose of this section is to characterize those normed linear
spaces which have propertyP1 and which have extremepoints on theirunit
spheres.Since an equivalence preservesextremepoints,such a characteriza-
tion includesspaces with propertyP1, any one of whichhas an extremepoint
on its unit sphere; hence it includes finite-dimensionalspaces, conjugate
spaces, and reflexivespaces, which have propertyP1. The extremepoint is
used to introducea partial orderingin X. This partial orderingenables us
to show that a space with propertyP1 is equivalent to the space of all real-
valued continuous functionsover some compact Hausdorffspace H. From
this the characterizationfollowseasily.
Let C, the unit sphere of the normedlinear space X which has property
P1, have an extremepoint u. Let
Cu = { x I x = tx', x' E C + u, t real,t > 0}.
C. is closed [4, p. 849]. We definexl>x2, x1and x2in X, if and only ifX1-X2
is in C,. This definitiongives X a partial orderingsuch that the unit sphere
C of X consists of just those elementsx with -u_x_u [4, p. 850].
5.1 LEMMA.Let thenormedlinear space Y withunit sphereK containthe
normedlinear space X in whichthe unit sphere C has an extremepoint u. A
projectionT of Y ontoX maps K ontoC if and onlyif it maps Ku-{ yIy = ty',
y'EK+u, t_O} ontoCu.
Proof. Suppose T maps K onto C and let y=ty' be in K,, where t is a
real numbergreaterthan or equal to zero and y' is in K+u. Then y = t(y"+u)
wherey" is in K.
T(y) = T(t(y" + u)) = t(T(y") + T(u)) = t(x" + u)
wherex " ? C. Hence
T(y) = t(x" + u) = t(x') C C. where x' C + u.
Hence T maps K. onto C.
Suppose T does not map K onto C; that is, suppose thereis a y in K such
that T(y) =x is not in C. Then l|x|I> 1. We assume x is in C.-u, forother-
wise we are done. Then x+uFC.; that is, x> -u. -yFK since yEK and
hence T(-y)=-xEC since xEC. Suppose -x+u is in C.. Then u>x.
Hence -u<x?u and x is in C, a contradictionsince ||x|| > 1. Then -x+u is
not in C, as we assumed above. Thus u-y is in K. but T(u-y)= u-x is

not in Cu. Hence if T(K) $ C, T(K,) S C,; that is, T(Ku) = Cu implies T(K)
=C.
5.2 LEMMA. If thenormedlinear space X has propertyP1, if theunitsphere

C of X has an extremepoint u, and if X1_ X2 means X1- X2 is in C,, thenevery
setin X boundedabovehas a least upper bound.
Proof. Suppose there is a set A in X such that A is bounded above but
has no least upper bound. Let B be the set of all upper bounds of A and ex-
pand A to include all lower bounds of B. Then a<x<b, a in A and b in B,
is impossiblesince x would be a least upper bound of the originalset A. Let
Y be a linear (as yet unnormed) space of dimension one greater than the
dimensionof X, let yo be a fixedelementof Y not in X, and let K ' be the
smallest half-conewith vertex at the origincontaining Cl, yo+B, and -yo
-A. Then K'QX = Cu.
If thereexists a projection T of K ' onto Cu, forb in B, T(yo+b) = T(yo)
+ bE Cu and b? -T(yo). Likewise fora in A, T(-yo-a) -T(yo) -a Cu
and - T(yo)> a. Then - T(yo) would have to lie between A and B, an im-
possibility.Hence theredoes not exist a projectionof K ' onto C,.
Let a be a real numbergreaterthan zero and let bECB. If x=aab, x+ayo
=a(b+yo) FK'. If x+ayoCK', xeK' -ayo and x =a(b+yo) -ayo =ab
forsome b in B. Hence fora>O, x+oFK' if and only if x==ab with b in
B. Similarly,fora<O, x+ayoFK ' if and only if x=aa with a in A, and for
t=0, x+ayoFK ' if and only ifxECu.
We are now ready to show that K ' is closed. Suppose a is a real number
greaterthan zero. Let aon(xn+yo)-*x+ayo whereanXn+anYOSK'. an>O for
n large enoughsince an->a. a?xn,->xand by a previousresultxnCB foran > 0,
so forlarge enough n, a?nXn> ana foreach a in A. For E>0 thereexistsan n(e)
such that (e/2)u>(a -an)a if n>n(e). Also there exists m(e) such that if
n>m(e), (E/2)u_anXn-x; that is, x>anxn-(e/2)u. Combiningthese results
gives x_?aa-Eu. But since Cu is closed, x> aa forall a in A. Then x&CaB by
the definitionof B and x+ayo =a(b+yo)CK'. By similar arguments we
see that the same conclusionholds fora < 0 and a =0. Hence K ' is closed.
Let
K = {x+ayoll al 1, x+yo+X uE K',
u-x-ayoy K '}.
K may be used as the unit sphere of Y. Then
K + u = {x + ayol lal 1, x + ayo C Ku, 2u - x - ayo C Ku'.
KuEKU' since K+uCK'. Let zCK '; then

z = x + a(-a - yo) + 3(b+ yo)
wherex Cu, aCA, b :B, and a and 3 are real numbersgreaterthan or equal
to zero. For all real numbersX >O, XzCK '. To get X - a+0 <?1 we need only

small 2u -XzWKEK'.
take X small. It remainsto be shown that forX sufficiently
In this connectionit sufficesto show that fork sufficiently
large,ku-zeKU',
forby settingX= 2/k> 0, we have 2u -XzCXK ' = K '. Take ki= max (lIx|I, 1).
Then kiu-xC4C.. Take k2so that k2u+aaCaaB, that is, so that k2u+aa=ab'
forsome b' in B. Take k3so that k3u-/3bE-f3A, that is, k3u-/3b= -3a' for
some a' in A. Then ku-z=(k1+k2+k3)u-x-a(-a-yo)-13(b+yo) =kiu
-x+a(yoj+-b') +3(-yo -a') EKu'. Hence zCKu and KudK'. But we have
previouslyshown that KuCK' so K =K'.
There is a projection T, ||T|| = 1, of Y onto X since X has propertyP1.
Then T(K) = by b Lemma 4.1 andT(K. ) == by Lemma 5.1. But we have
shown above that T(K ') = T(Ku) - Cu. Hence our assumptionis false and A
must have a least upper bound. This completes the proofof the lemma.
We now quote a theorem of Clarkson's [4, p. 8471 with the notation
modifiedslightlyso as to be consistentwith the notation of our previous
work.
5.3 THEOREM. Let X be a real Banach space. In orderthatthereshall exist

a compactHausdorftspace H such thatX is equivalentto C(H), thespace of all
real-valuedcontinuousfunctionsoverH, thefollowingtwoconditionsare neces-
sary and sufficient:
A. Thereexistsan elementu of C, theunitsphereofX, suchthatany element
of S, thesurfaceof C, is co-facialeitherwithu or -u.
B. The half-coneC has thepropertythattheintersectionof twoofits trans-
lates is itselfa translate.
The followinglemma is a direct result of this theoremand our previous
work.
5.4 LEMMA. If the normedlinear space X has propertyP1 and if the unit
sphereC ofX has an extremepointu, thenX is equivalentto a space C(H) ofall
real-valuedcontinuousfunctionson some compactHausdorffspace H.
Proof. X satisfiesClarkson's conditionA by Theorem 4.3. To show that
X satisfiesClarkson's conditionB, we let xi and x2 be any two elementsin X
and prove that Cu+xi and CG+x2 intersectin a translateof C4.
C.+xi= {xIx-xi>0}. C.+x2= {X lx-X2o0}.
Hence (Cu+X1)n(Cu+X2) =X{X-X_ 0, x-x2_0}. Let k=max (jjxuj|,
1). ku?xi and ku>x2 sinceu>x ifl|xii<?. Then xi and x2are bounded
11x211,
above. Thereforeby Lemma 5.2 thereexists in X an elementx3which is the
least upper bound of xi and x2. Hence (Cu+xi)Gn (Cu+X2) = C,+x3 is a trans-
late of Cu. Then X satisfiesClarkson's conditionsand is thereforeequivalent
to a space C(H) of continuous functionsover some compact Hausdorff
space H.
Lemma 5.4 enables us to give anotherproofof Corollary4.5. If the space

X has propertyP1 and an extremepoint on its unit sphere,X is equivalent

to a space C(H1) by Lemma 5.4. The extremepoints of the unit sphere of
C(H) are all at a distance two fromeach other.Since an equivalence preserves
distances and extremepoints,the extremepoints of the unit sphere of X are
at a distance two fromeach other.
5.5 THEOREM. A normedlinear space X withan extremepointu on its unit
sphereC has propertyPi if and only if X is equivalentto a space Y of typeBj+
whoseunitsphereS has a leastupperboundv ofnormone.
Proof. Suppose X is equivalent to a space Y of type B1+whose unit sphere
S has a least upper bound v of normone. Then Y has propertyP1 by Akilov's
theoremn. ThereforeX has propertyPi by Lemma 2.3.
If X has an extremepoint on its unit sphere C and if X has propertyP1,
then X is equivalent to a space C(H) of all real-valued continuousfunctions
over some compact Hausdorffspace H, by Lemma 5.4. C(H) is a space of
type B1 when given its usual normand partial ordering.But C(H1) has prop-
ertyPi and hence is of type B1+by Lemma 5.2. Let v in C(H) be the element
such that v(t) = 1 forall t in H. ||v|| = 1 and v is the least upper bound of S.
Hence C(H) is a space of typeB1+whose unit spherehas a least upper bound v
of normone.
P1.
6. Abstract(M)-spaces whichhave property
6.1 THEOREM. An abstract(M)-space X whichhas propertyPi has a unit
element.
Proof. X is isometric and lattice isomorphic to a space C(H; ta, ta;
Xa; C9Y) by Kakutani's theorem [10, p. 998]. Let Q be the transformation
between the two spaces. Q preserves property Pi by Lemma 2.3, so
C(H; ta, ta; Xa; azTI) has property P1. Hence there is a projection T,
|T=1, of C(H), the space of all continuous functionson H, onto its sub-
space C(H; ta. t;Xat; aEz-TC).
We now show that if y is in C(H) and y>0, then T(y) =x_0. If y=0,
T(y) =0 and it is true forthis case. Suppose y>0 but T(y) =x is not greater
than or equal to 0. Then Ilyll>0 and IxII>0. Let yo=y/||yII and xo=x/IIyII.
Then |Iyoll=1 and |IxoI>0. Let x'=xoAO. x' is in C(H; ta, t' ; Xa; E9)
since the space is a lattice; x' <0; |X'|I=k >0. Since H is compact, thereis a
t in H such that x'(t) =-k. Let x"=x'/||x'| =x'/k. |x"|| =1, x"(t) -1,
and x"<0. Then ||yo+x"||?1 and T(yo+x") =xo+x". But I (xo+x")(t)I
= -k-1 | > 1,whichis a contradiction sinceIITJ =-1.Therefore ourassump-
tion was false and T(y) 0 when y ?0.
Let v in C(H) be the element such that v(s) = 1 forall s in H. If y is in
C(H) and IIYII< 1, thenv-y _ 0; in particularif x is in C(H; ta,tat Xa; EzC9)
and ||x||<1, then v-x_O. Hence T(v-x)=e-x_O since T is a positive
transformation. Then the element u = Q-(e) in X is the unit element of X.

6.2 COROLLARY. An abstract (M)-space X which has propertyP1 has a

of theform C(QIJ),whereC(H) is the space of all real-
concreterepresentation
valued continuousfunctionsoversome compactIHausdorffspace H.
Proof. By Theorem 6.1, X has a concrete representationof the form

C(H; ta, tc; Xa; aC9J) which has a unit element.Then by a proofof Kaku-
tani's [10, p. 1005], C(II; ta, t '; Xa; a CM) is actually the entirespace C(H).
In Lemma 5.4 and again in Corollary 6.2 we found that certain spaces
which have propertyP1 are each equivalent to a space of all continuousreal-
valued functionsover some compact Hausdorffspace H. We shall now turn
our attentionto the compact Hausdorffspace H and findthat it possesses a
very curious property known as extremal disconnectivity.A Hausdorff
space H is said to be extremallydisconnectedif every pair of disjoint open
subsets of H have disjoint closures.
6.3 THEOREM. The space C(H) of all real-valuedcontinuousfunctionson a

compactHausdorffspace H has propertyP1 if and onlyif H is extremallydis-
connected.
Proof. Suppose C(H) has propertyP1. Then by Lemma 5.2 every set of
elements of C(H) which is bounded above has a least upper bound. Stone
has shown [20, p. 186] that if C(H) has this property,then everyopen set in
H has a closure which is open. Then by a theoremof Hewitt's [8, p. 326],
the closures of disjoint open sets in H are disjoint. From this we conclude
that H is extremallydisconnected.
C(fI) is a space of type B1 and its unit sphere has a least upper bound e
of normone wheree is the elementidenticallyequal to one on H. If H is ex-
tremallydisconnected,every set in C(H) that is bounded above has a least
upper bound [20, p. 186]. Hence C(H) is a space of typeB1+whose unitsphere
has a least upper bound e of normone. Then by Akilov's theorem,C(H) has
propertyP1.
6.4 THEOREM. The unitsphereS of C(QH),thespace ofall boundedcontinuous
real-valuedfunctionson an extremallydisconnectedHausdorffspace H, is the
closedconvexhull of its extremepoints.
Proof. Let K be the convex hull of the extremepoints of S; let F be any
elementof S; let e be a real numbergreaterthan zero; let N be the smallest
integergreater than 2/e; let the elementsof H be denoted by x's. Our first
step is to show that thereis an elementof K withine of F, that is, thereis an
elementg in K such that ||F-gj <?e. Let
El= I{xIF(x) > 1 -e.
E1 is open [15, p. 40]. Then E' =cEl, the closure of E1, is open and closed
since H is extremallydisconnected [8, p. 326]. Hence CE1', the complement

of E', is open and closed. The characteristicfunctionCE1of E' is upper semi-

continuous since E' is closed and lower semi-continuoussince E1' is open.
Hence CE; is continuous. Let
E2= {xIF(x) > 1 - 2.
Since E2 is open, cE2 is open and closed and hence E' = CE1' qcE2 iS open and
closed. Theii CE2is continuous. In generalif
En= {xIF(x) > 1 - nE,n < NJ,
then
Cn-1
Et CU E! n ~(cEn)
is open and closed and has a continuous characteristicfunctionCE'. Also

U 1EX = H and E ' nEJ =0 wheni Pj. It is easilyverified that EN 1 miCE
are extremepoints of S when mi= + 1.
Let kn=ne/2 for n <N and kN=1. Form the functions
~~Kc~L
= ki (-CEI N
gi + E CE? + (1 - k) (CE + E CE)
N
= (1 - E)CE1 + ? CE'
j=2
. . . . . . . . . . . . . . . . . . . . . . . . . .
n N
gn= E(1- E)CE+ E CE1 for n<N,
i=1 i=n+l
* . . . . . . . . . . . . . . . . . . . . . . . .
N-I
gN = E (1 - tE)CE' - CEN.
i=1
An element g is in K if and only if there exist an integern, elements

gj !,g2' * g' in K, and positive numbersa,, a2, Ian with j=1 ai= 1
such that g= >i=l aig! [6,Theorem 2.9, Chap. 2].
In the above set of equations each knis positiveor zero and is less than or
equal to 1. Hence O<kn 1-kn<1 for n=l, . , N. kn+(1-kn) =1
gn=knkn_l+(1-kn) gn-1 n=2, 3, * , N where gn- is equal to gn-l with the
sign of CEB changed. gn ICK when gnlCK; so by the theoremjust quoted
gnCK when gn_- CK. Then since giCK, all the gnare in K; in particular gN
is in K.
For anv x in H there is an i <N such that E' contains x. Then IF(x)
-gN(x) ?| e. Hence ||F-gN|I <e; that is, forany F in S and e>0 thereis an
elementg in K withine of F.

If we now choose a sequence { e(n) } of E's, such that e(n) = 1/n, we can
finda sequence of functionsge(n) in K such that lim g
g) = F. Hence F
is in the closureof K; that is, S is the closed convex hull of its extremepoints.
6.5 COROLLARY. If the normedlinear space X has propertyP1 and if the
unit sphereC of X has an extremepoint,thenC is theclosedconvexhull of its
extremepoints.
Proof. X is equivalent to C(H), the space of all real-valued continuous
functionsover some compact Hausdorffspace H, by Lemma 5.4. Then C(H)
has propertyP1 and hence H is extremallydisconnected.Theorem 6.4 tells
us the unit sphere of C(H) is the closed convex hull of its extreme points.
Then C is the closed convex hull of its extremepoints since an equivalence
preservesextreme points, line segments,and distances.
6.6 COROLLARY. If the normedlinear space X has propertyP1, if X is
infinite-dimensional,
and if theunit sphereC of X has an extremepoint,then
C has an infinitenumberof extremepoints.
Proof. If C has a finitenumber of extreme points, it is not the closed
convex hull of its extremepoints. Hence it does not have propertyP1 by the
corollaryabove.
Our next theoremgives a necessaryand sufficient conditionfora normed
linear space X to be equivalent to a space of type Bj+ which has a unit ele-
ment.A space of type Bj+ whichhas a unit elementis in turnequivalent to an
abstract (M)-space C(H).
6.7 THEOREM. A normedlinear space X whichhas propertyP1 is equivalent
to a space Y oftypeBj+ whichhas a unit elementif and onlyif theunitsphereC
of X has an extremepoint.
Proof. Suppose the unit sphere C of X has an extreme point u. Let
x1>x2 mean x1-x2 CCu.Then X is equivalent to a space C(H) of all continu-
ous real-valuedfunctionson some compact Hausdorffspace H, by Lemma 5.4.
The element e which is identicallyequal to one on H is the unit element in
C(H). Since C(H) has propertyP1,everyset in C(H) which is bounded above
has a least upper bound, by a previouslemma. Hence C(H) is a space of type
Bj+ and has a unit element e.
Suppose X is equivalent to a space Y of type Bj+ which has a unit element
e. e is an extreme point. Let U be the equivalence between X and Y and
considerthe point u = U-1(e) in X. lull= 1 so u is in C. Suppose u = (x] +x2)/2
where x1 and x2 are in C.
U(Xl) + U(x2) yl + Y2
=
U(u) = e-
2
-=____ 2_
2
where yi and Y2 are in S the unit sphere of Y. Then Yi = Y2 = e since e is an

extremepoint. Hence x1=X2= u and u is an extremepoint.

6.8 THEOREM. Let X be a normedlinear space whichhas propertyP1. Then
X is reflexive
if and onlyif it is finite-dimensional.
Proof. Let X be reflexiveand definex > y to mean x - y C C.. Then by
Lemma 5.4, X is equivalent to the space C(H) of all real-valued continuous
functionsover some compact Hausdorffspace H. Since a space is reflexive
ifand only ifeveryspace isomorphicto it is reflexive,C(H) must be reflexive.
C(H) is an abstract (M)-space. But an abstract (M)-space is reflexiveonly
when it is finite-dimensional[10, p. 1022]. Hence C(H) is finite-dimensional
and it followsthat X is finite-dimensional.
A theoremof Myers' [17, p. 404] gives us a bit of additional information
about H. Since H is a compact Hausdorffspace, it is completely regular.
Then Myers' theoremtells us that C(H) is reflexiveifand only ifH consists
of a finitenumberof points.
6.9 THEOREM. The conjugatespace (AL)* of an abstract(L)-space (AL)
has propertyP1.
Proof. (AL)* is an abstract (M)-space with a unit element [10, p. 1021]
and consequentlyis a space of type B1 whose unit sphere has a least upper
bound of norm one. Hence it sufficesto prove that if a set E in (AL)* is
bounded above by an elementgo in (AL)*, then E has a least upper bound,
for then (AL)* has propertyP1 by Akilov's theorem.
Since f g in (AL)* means f(x) >g(x) for all x in the positive half-cone
K of (AL), let G= {glgo>g>E, g in (AL)*}. Then GC{g|go?>g?>fo} forfo
fixedin E. Hence G is contained in a sphere in (AL)*. If K*= {flf(x)?0 for
x in K }, thenG = [nfEE(K* -f) ] n (go- K*). Since K* is weak starclosed,so is
G. Then G is weak star closed and bounded and hence is weak star comnpact.
If x is in K and 4(x)=glb0EGg(x), then by compactness Gx=G I{gjg(x)
=sb(x) } cannot be empty. If xl, X2, * * *, x. are in K, let gi, g2, **a gnbe
such that giCGGi and defineg=glAg2A . . Agn. Then gi(xj) >?4O(xj)forall
i, j since gi G, so g(xj) ?q$(xj) for all j. Also g(xj) <gj(xj) =45(xj) for all j
so g(xj) =q$(xj) and gefln1Gxi. Then by coiipactness nfxEKGX.0
If g'and g" are in AXEKGX,theng'Ag" is in G. (g'- (g'Ag"))(x) _0 forall
x in K and ifg' > g'Ag" thereis an elementxoin K such that (g'- (g'Ag")) (x0)
>0; that is, g'(xo) > (g'Ag")(xo). But this is impossiblesince g' is in G,(.
Hence g'==g'Ag". Likewise g"=g'Ag" so gl = g2 and the intersectioncon-
tains only one element,which is a least upper bound of E.
6.10 COROLLARY. The secondconjugatespace (AM)** of an abstract(M)-
space (AM) has propertyP,.
Proof. (AMI)* is an abstract (L)-space (AL) [10, p. 1021], and hence
(AM)**=(AL)* has propertyP,.

19501 PROJECTIONS IN NORMED LINEAR SPACES 107
6.11 COROLLARY.An abstract(M)-space (AM) has propertyP1 if and only

if thereis a projectionT, IITj I = 1, of (A M)** ontoQ(A M), whereQ is thena-
turalmappingof (A M) into (A M) **.
Proof. If there is a projection T, ||T= 1, of (AM)** onto (AM), then
(AM) has propertyP1 by Theorem 3.2.
If thereis no projectionof normone of (A M) ** onto Q(A M), then Q(A M),
and hence (A M), does not have propertyP1.
The real line R is the simplestexample of a space which has propertyP1.
That it has this propertywas firstshown by applying the Hahn-Banach
theorem [3, p. 55] to the identitytransformation of R onto R, since by the
theorem there is an extension of any linear functionalon R to a normed
linear space containingR. It is easily seen that R satisfiesthe conditionsof
Akilov's theorem. A result of Phillips' [18, p. 358] showed that (m), the
space of all bounded sequences, has propertyP1. It is easily verifiedthat (m)
also satisfiesthe conditionsof Akilov's theorem.Finally, the space M of all
essentially bounded measurable functionson the unit interval satisfiesthe
conditionsof Akilov's theoremand hence has propertyP1.
Each of the knownexamples of spaces having propertyP1 is the conjugate
space of an abstract (L)-space. Hence each of the examples has a concrete
representationas the space of all continuousreal-valued functionsover some
extremallydisconnectedcompact Hausdorffspace H. It is an open question
as to whetheror not a necessary condition fora space X to have property
P1 is that X is equivalent to the conjugate space of an abstract (L)-space.
BIBLIOGRAPHY
1. G. P. Akilov, Necessaryconditionsfor theextensionoflinear operations,C. R. (Doklady)
Acad. Sci. URSS. N. S. vol. 59 (1948) pp. 417-418. (Russian)
2. , On theextensionof linear operations,C. R. (Doklady) Acad. Sci. URSS. N. S.
vol. 57 (1947) pp. 643-646. (Russian)
3. S. Banach, Theoriedes operationslineaires,Warsaw, 1932.
4. J. A. Clarkson,A characterizationof C-spaces,Ann. of Math. vol. 48 (1947) pp. 845-850.
5. M. M. Day, A reviewof A kilov's"On theextensionof linear operations,"Mathematical
Reviews vol. 9 (1948) p. 241.
6. , Lecture notesfor theoryof linear operations,University of Illinois, 1947, un-
published.
7. G. Fichtenholtz, and L. Kantorovitch, Sur les operationslineaires dans l'espace des
functionsbornees,Studia Mathematica vol. 5 (1934) pp. 69-98.
8. E. Hewitt, A problemofset-theoretictopology,Duke Math. J. vol. 10 (1943) pp. 309-333.
9. S. Kakutani, Concreterepresentation of abstract(L)-spaces and themean ergodictheorem,
Ann. of Math. (2) vol. 42 (1941) pp. 523-537.
10. ofabstract(M)-spaces, Ann. of Math. (2) vol. 42 (1941)
, Concreterepresentation
pp. 994-1024.
11. L. Kantorovitch,Linear operationsin semi-ordered spaces, I, Rec. Math. (Mat. Sbornik)
N. S. vol. 7 (1940) pp. 210-279.
12. Rdume, Rec. Math. (Mat. Sbornik) N. S. vol. 2 (1937)
, Lineare halbgeordnete
pp. 121-168.

108 D. B. GOODNER
13. ---, Sur la theoriegenerale des operations dans les espaces semi-ordonnes,C. R.
(Doklady) Acad. Sci. URSS. N. S. vol. 1 (1936) pp. 283-286.
14. S. Mazur, Ober konvexeMengen in linearennormierten Rdumen,Studia Mathematica
vol. 4 (1933) pp. 70-84.
15. E. J. McShane, Integration,PrincetonUniversityPress, 1944.
16. F. J. Murray, On complementary manifoldsand projectionsin spaces Lp and lp,Trans.
Amer. Math. Soc. vol. 41 (1937) pp. 138-152.
17. S. B. Myers, Spaces ofcontinuousfunctions,Bull. Amer. Math. Soc. vol. 55 (1949) pp.
402-407.
18. R. A. Phillips, On linear transformations,
Trans. Amer. Math. Soc. vol. 48 (1940) pp.
516-541.
19. A. Sobczyk, Projectionofthespace (m) on its subspace (co), Bull. Amer. Math. Soc. vol.
47 (1941) pp. 938-947.
20. M. H. Stone, Boundednesspropertiesin functionlattices,Canadian Journal of Mathe-
matics vol. 1 (1949) pp. 176-186.
UNIVERSITY OF ILLINOIS,
URBANA, ILL.
FLORIDA STATE UNIVERSITY,
TALLAHASSEE, FLA.


Projections in Normed Linear Spaces

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Projections in Normed Linear Spaces

Uploaded by

Copyright:

Available Formats

Projections in Normed Linear Spaces

Author(s): Dwight B. Goodner

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

necessaryand sufficient conditionforspaces having propertyPi to be reflexive

2.2 THEOREM. If X is a normedlinear space whichhas propertyP8, thenX

Proof. Let Y be a Banach space containingX. Since X has propertyP8,

2.3 LEMMA.If a normedlinear space X is equivalentto a normedlinear

Proof. Let W be a normedlinear space containing Y and let

where CY is the complementof Y and W' = WnGCY.Then if V is the equiva-

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

Z + z2 = X if U(zl) + U(z2) = wl + w2 = y = U(x)

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

T 1(X*) {h| h(y) = Tlf(y) = f(Tly), f C X*}

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

any space X which provides a solution to problem3 has propertyP1.

3.2 THEOREM. If thenormedlinear space Y has propertyP8, if T is a pro-

Proof. If m(S) is the space of all bounded real-valued functionson S, the

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

4. The unitsphereofa space whichhas property

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

of S which are not in C. Checkingall possibilitieswill show that the extreme

4.3 THEOREM. If thenormedlinear space X has propertyP1 and if p is an

4.4 LEMMA. Let X be a normedlinear space withunit sphereC and let Y be

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

Proof. By Mazur's theoremthere is a hyperplaneH containingthe line

flqi- qklf _ IlPk - Pill 1.

Hence E== {qi}, = 1, 2, * * *, n, is a set ofdiameterless than or equal to one.

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

5.2 LEMMA. If thenormedlinear space X has propertyP1, if theunitsphere

KuEKU' since K+uCK'. Let zCK '; then

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

5.3 THEOREM. Let X be a real Banach space. In orderthatthereshall exist

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

X has propertyP1 and an extremepoint on its unit sphere,X is equivalent

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

6.2 COROLLARY. An abstract (M)-space X which has propertyP1 has a

Proof. By Theorem 6.1, X has a concrete representationof the form

6.3 THEOREM. The space C(H) of all real-valuedcontinuousfunctionson a

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

of E', is open and closed. The characteristicfunctionCE1of E' is upper semi-

is open and closed and has a continuous characteristicfunctionCE'. Also

An element g is in K if and only if there exist an integern, elements

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

extremepoint. Hence x1=X2= u and u is an extremepoint.

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

6.11 COROLLARY.An abstract(M)-space (AM) has propertyP1 if and only

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

This content downloaded from 128.235.251.160 on Tue, 9 Dec 2014 19:44:34 PM

You might also like

T 1(X) {h| h(y) = Tlf(y) = f(Tly), f C X}