Schur Class Operator Functions
and Automorphisms of Hardy Algebras
Paul S. Muhly
1and Baruch Solel
2Received: June 12, 2007
Communicated by Joachim Cuntz
Abstract.
LetE
beaW
∗
-orrespondeneoveravonNeumann alge-bra
M
andletH
∞
(
E
)
betheassoiatedHardyalgebra.If
σ
isa faith-ful normalrepresentationofM
onaHilbert spaeH
, thenonemay form thedual orrespondeneE
σ
andrepresentelementsin
H
∞
(
E
)
as
B
(
H
)
-valuedfuntionsontheunitballD
(
E
σ
)
∗
. Thefuntionsthat oneobtainsarealledShurlassfuntionsandmaybeharaterized in terms of ertain Pik-like kernels. We study these funtions and relate them to systemmatries and transfer funtions from systems theory. Weusetheinformationgainedtodesribetheautomorphism groupof
H
∞
(
E
)
intermsofspeialMöbiustransformationson
D
(
E
σ
)
. Partiularattentionisdevotedtothe
H
∞
-algebrasthatareassoiated tographs.
2000 Mathematis Subjet Classiation: 46E22, 46E50, 46G20, 46H15,46H25,46K50,46L08,46L89,
Keywords and Phrases: Hardy Algebras, Tensor Algebras, Shur lassfuntions,
W
∗
-orrespondene,nonommutativerealization the-ory,Möbiustransformations,freesemigroupalgebras,graphalgebras, Nevanlinna-Pikinterpolation
1
SupportedinpartbygrantsfromtheNationalSieneFoundationand fromthe U.S.-IsraelBinationalSieneFoundation.
2
1
Introduction
Let
M
beaW
∗
-algebraandlet
E
beaW
∗
-orrespondeneover
M
. In[31℄we builtanoperatoralgebrafromthisdatathatwealledtheHardyalgebraofE
and whih wedenotedH
∞
(
E
)
. If
M
=
E
=
C
-the omplexnumbers,thenH
∞
(
E
)
isthelassialHardyalgebraonsistingof allbounded analyti fun-tionsontheopen unit dis,
D
(seeExample 2.4below.) IfM
=
C
again,butE
=
C
n
, then
H
∞
(
E
)
is thefree semigroupalgebra
L
n
studied by Davidson andPitts[17℄,Popesu[32℄andothers(seeExample2.5.) Oneoftheprinipal disoveriesmadein [31℄,andthesoureofinspirationforthepresentpaper,is thatattahedtoeahfaithfulnormalrepresentationσ
ofM
thereisadual or-respondeneE
σ
, whihisa
W
∗
-orrespondeneovertheommutantof
σ
(
M
)
,σ
(
M
)
′
, and theelementsof
H
∞
(
E
)
denefuntions onthe openunit ballof
E
σ
,
D
(
E
σ
)
. Further,thevaluedistributiontheoryofthesefuntionsturnsout to be linked through our generalization of the Nevanlinna-Pik interpolation theorem [31, Theorem 5.3℄ with the positivity properties of ertain Pik-like kernelsofmappings betweenoperator spaes.
Inthesettingwhere
M
=
E
=
C
andσ
isthe1
-dimensional representationofC
onitself,thenE
σ
is
C
again. TherepresentationofH
∞
(
E
)
intermsof fun-tionson
D
(
E
σ
) =
D
isjusttheusualwaywethink ofH
∞
(
E
)
. Inthissetting, our Nevanlinna-Pik theorem is exatlythe lassialtheorem. If, however,
σ
is arepresentationofC
on aHilbert spaeH
,dim(
H
)
>
1
, thenE
σ
maybe identiedwith
B
(
H
)
andthenD
(
E
σ
)
beomesthespaeofstritontrations on
H
, i.e., all those operators of norm stritly less than1
. In this ase, the valueof anf
∈
H
∞
(
E
)
at a
T
∈
D
(
E
σ
)
is simplyf
(
T
)
,dened throughtheusual holomorphi funtional alulus. OurNevanlinna-Piktheorem givesa solutiontoproblems suh asthis: given
k
operatorsT
1
, T
2
, . . . , T
k
allofnorm lessthan1
andk
operators,A
1
, A
2
, . . . , A
k
,determinetheirumstanesunder whihoneanndaboundedanalytifuntionf
ontheopenunitdisofsup norm at most1
suh thatf
(
T
i) =
A
i
,i
= 1
,
2
, . . . , k
(See [31, Theorem6.1℄.) On the other hand, whenM
=
C,
E
=
C
n
, and
σ
is one dimensional, the spaeE
σ
is
C
n
and
D
(
E
σ
)
is the unit ball
B
n
. Elements in
H
∞
(
E
) =
L
n
arerealizedasholomorphifuntionson
B
n
thatlieinamultiplierspae stud-iedin detailbyArveson[5℄. Moreaurately,thefuntionalrepresentationof
H
∞
(
E
) =
L
n
in termsofthesefuntions expressesthisspaeasaquotient ofH
∞
(
E
) =
L
n
. The Nevanlinna-Pik theoremof [31℄ontainsthoseofDavid-son andPitts [18℄, Popesu[34℄, and Ariasand Popesu [4℄, whih dealwith interpolationproblemsforthesespaesoffuntions(possiblytensoredwiththe boundedoperatorsonanauxiliaryHilbertspae). Italsoontainssomeofthe results of Constaninesu and Johnson in [16℄ whih treatselementsof
L
n
as funtionsontheballofstritrowontrationswithvaluesintheoperatorson a Hilbert spae. (See their Theorem 3.4 in partiular.) This situation arises when onetakesM
=
C
andE
=
C
n
, but takes
σ
to besalar multipliation onanauxiliaryHilbert spae.those funtionson
D
(
E
σ
)
that arisefromevaluatingelementsofH
∞
(
E
)
. For this purpose, weintrodueafamilyof funtions on
D
(
E
σ
)
that weallShurlassoperatorfuntions(seeDenition3.1). Roughlyspeaking,thesefuntions aredened sothat aPik-likekernelthatonemayattahtoeahone is om-pletely positive denite in the sense of Barreto, Bhat, Liebsher and Skeide [14℄. In Theorem 3.3 weuse their Theorem 3.2.3 to give aKolmogorov-type representation of the kernel, from whih we derive an analogueof a unitary
systemmatrix
A
B
C
D
whosetransferfuntion
A
+
B
(
I
−
L
η
∗
D
)
−
1
L
∗
η
C
turns out to be the given Shur lass operator funtion. We then prove in Theorem 3.6that eahsuhtransferfuntion arises byevaluatinganelement in
H
∞
(
E
)
at pointsof
D
(
E
σ
)
andonversely, eah funtion in
H
∞
(
E
)
hasa representationintermsofatransferfuntion. Themeaningofthenotationwill bemadepreisebelow,but weuseitheretohighlighttheonnetionbetween our analysis and realization theory as it omes from mathematial systems theory. Thepointto keepinmindis thatfuntionson
D
(
E
σ
)
thatomefrom elementsof
H
∞
(
E
)
arenot, apriori, analytiin anyordinarysenseand itis not at alllear what analyti features theyhave. Our Theorems3.1 and3.6 together with[31, Theorem 5.3℄ showthat theShur lassoperator funtions are preisely the funtions one obtainswhen evaluating funtions in
H
∞
(
E
)
(ofnormatmost
1
)atpointsofD
(
E
σ
)
. Thefatthateahsuhfuntionmay berealizedasatransferfuntion exhibitsasurprisinglevelofanalytiitythat isnotevidentin thedenition of
H
∞
(
E
)
.
Ourseondobjetiveisto onnettheusualholomorphipropertiesof
D
(
E
σ
)
with the automorphisms of
H
∞
(
E
)
. As aspae,
D
(
E
σ
)
is theunit ball of a
J
∗
-triplesystem. Consequently,everyholomorphiautomorphismofD
(
E
σ
)
is theomposition ofaMöbiustransformation andalinearisometry [20℄. Eah ofthese implementsanautomorphismofthealgebraofallbounded, omplex-valued analyti funtions on
D
(
E
σ
)
, but in our setting only ertain of them implement automorphisms of
H
∞
(
E
)
- those for whih the Möbius part is determinedbyaentralelementof
E
σ
(seeTheorem4.21).Ourproofrequires thefatthattheevaluationoffuntionsin
H
∞
(
E
)
(ofnormatmost
1
)atpoints ofD
(
E
σ
)
arepreiselytheShurlassoperatorfuntionson
D
(
E
σ
)
. Indeed,the whole analysisisanintriate point-ounterpoint interplayamongelements of
H
∞
(
E
)
, Shur lass funtions, transfer funtions and lassial funtion theory on
D
(
E
σ
)
. In the last setion, we apply our general analysis of the automorphisms of
H
∞
(
E
)
to the speial ase of
H
∞
-algebras oming from direted graphs.
also use (See Setion 3 and, in partiular, the proof of Theorem 3.3.) The analysis of Ball et al. makes additional ties between the theory of abstrat Hardyalgebrasthat wedevelophereandlassialfuntiontheoryontheunit dis.
2
Preliminaries
Westartbyintroduingthebasidenitionsandonstrutions. Weshallfollow Lane[24℄forthegeneraltheoryofHilbert
C
∗
-modulesthatweshalluse. Let
A
beaC
∗
-algebraand
E
bearightmoduleoverA
endowedwithabi-additive maph·
,
·i
:
E
×
E
→
A
(referred to asanA
-valuedinner produt) suh that, forξ, η
∈
E
anda
∈
A
,h
ξ, ηa
i
=
h
ξ, η
i
a
,h
ξ, η
i
∗
=
h
η, ξ
i
, and
h
ξ, ξ
i ≥
0
,withh
ξ, ξ
i
= 0
only whenξ
= 0
. Also,E
is assumed to be ompletein thenormk
ξ
k
:=
kh
ξ, ξ
ik
1
/
2
. We write
L
(
E
)
for the spae of ontinuous, adjointable,A
-modulemapsonE
. ItisknowntobeaC
∗
-algebra. If
M
isavonNeumann algebraandifE
isaHilbertC
∗
-moduleover
M
,thenE
issaidtobeself-dualin aseeveryontinuousM
-modulemapfromE
toM
isgivenbyaninnerprodut withanelementofE
. LetA
andB
beC
∗
-algebras. A
C
∗
-orrespondenefrom
A
toB
is aHilbertC
∗
-module
E
overB
endowed witha struture of aleft moduleoverA
viaanondegenerate∗
-homomorphismϕ
:
A
→ L
(
E
)
.Whendealingwithaspei
C
∗
-orrespondene,
E
,fromaC
∗
-algebra
A
toaC
∗
-algebra
B
, itwill be onvenientsometimes to suppress theϕ
in formulas involvingtheleft ation and simplywriteaξ
ora
·
ξ
forϕ
(
a
)
ξ
. This should ausenoonfusioninontext.If
E
is aC
∗
-orrespondene from
A
toB
and ifF
is a orrespondene fromB
toC
, then thebalanedtensor produt,E
⊗
B
F
is anA, C
-bimodule thatarriestheinner produtdenedbytheformula
h
ξ
1
⊗
η
1
, ξ
2
⊗
η
2
i
E⊗
B
F
:=
h
η
1
, ϕ
(
h
ξ
1
, ξ
2
i
E
)
η
2
i
F
TheHausdorompletion ofthis bimodule isagaindenotedby
E
⊗
B
F
. InthispaperwedealmostlywithorrespondenesovervonNeumannalgebras that satisfy some natural additional properties as indiated in the following denition. (Forexamplesand moredetailssee[31℄).Definition 2.1
LetM
andN
bevonNeumannalgebrasandletE
beaHilbertC
∗
-moduleover
N
. ThenE
isalleda HilbertW
∗
-moduleover
N
inaseE
is self-dual. ThemoduleE
isalledaW
∗
-orrespondenefrom
M
toN
inaseE
isaself-dualC
∗
-orrespondenefrom
M
toN
suhthatthe∗
-homomorphismϕ
:
M
→ L
(
E
)
, giving the left module struture onE
, is normal. IfM
=
N
weshallsay that
E
isaW
∗
-orrespondene over
M
.Wenote thatif
E
isaHilbertW
∗
-moduleoveravonNeumannalgebra,then
L
(
E
)
isnotonlyaC
∗
-algebra,butis alsoa
W
∗
Definition 2.2
An isomorphism of aW
∗
-orrespondene
E
1
overM
1
and aW
∗
-orrespondene
E
2
overM
2
is a pair(
σ,
Ψ)
whereσ
:
M
1
→
M
2
is an isomorphism of von Neumann algebras,Ψ :
E
1
→
E
2
is a vetor spae isomorphism preserving theσ
-topology and fore, f
∈
E
1
anda, b
∈
M
1
, we haveΨ(
aeb
) =
σ
(
a
)Ψ(
e
)
σ
(
b
)
andh
Ψ(
e
)
,
Ψ(
f
)
i
=
σ
(
h
e, f
i
)
.When onsidering the tensor produt
E
⊗
M
F
of twoW
∗
-orrespondenes, one needs to takethe losure of the
C
∗
-tensor produt in the
σ
-topologyof [6℄ in order to get aW
∗
-orrespondene. However, we will not distinguish notationallybetweenthe
C
∗
-tensorprodutandthe
W
∗
-tensorprodut. Note alsothatgivena
W
∗
-orrespondene
E
overM
andaHilbertspaeH
equipped with anormalrepresentationσ
ofM
,weanform theHilbert spaeE
⊗
σ
H
by deningh
ξ
1
⊗
h
1
, ξ
2
⊗
h
2
i
=
h
h
1
, σ
(
h
ξ
1
, ξ
2
i
)
h
2
i
. Thus,H
is viewed as a orrespondenefromM
toC
viaσ
andE
⊗
σ
H
isjust thetensor produtofE
andH
asW
∗
-orrespondenes.
Note alsothat,given anoperator
X
∈ L
(
E
)
andan operatorS
∈
σ
(
M
)
′
, the map
ξ
⊗
h
7→
Xξ
⊗
Sh
denes a bounded operator onE
⊗
σ
H
denoted byX
⊗
S
. TherepresentationofL
(
E
)
that resultswhenone letsS
=
I
,isalledthe representation of
L
(
E
)
indued byσ
and is often denoted byσ
E
. The omposition,
σ
E
◦
ϕ
isarepresentationof
M
whihweshallalsosayisindued byσ
,but weshallusually denoteitbyϕ
(
·
)
⊗
I
.Observe that if
E
is aW
∗
-orrespondene over a von Neumann algebra
M
, then we may form the tensor powersE
⊗n
,
n
≥
0
, whereE
⊗
0
is simply
M
viewed as the identity orrespondene overM
, and we may form theW
∗
-diret sum ofthe tensor powers,
F
(
E
) :=
E
⊗
0
⊕
E
⊗
1
⊕
E
⊗
2
⊕ · · ·
to obtain a
W
∗
-orrespondeneover
M
alledthe(full)Fokspae overE
. Theations ofM
onthe left and right ofF
(
E
)
are the diagonal ations and, when it is onvenient to do so, we make expliit the left ation by writingϕ
∞
for it. That is,fora
∈
M
,ϕ
∞(
a
) :=
diag
{
a, ϕ
(
a
)
, ϕ
(2)
(
a
)
, ϕ
(3)
(
a
)
,
· · · }
, wherefor all
n
,ϕ
(
n
)
(
a
)(
ξ
1
⊗
ξ
2
⊗ · · ·
ξ
n) = (
ϕ
(
a
)
ξ
1
)
⊗
ξ
2
⊗ · · ·
ξ
n
,ξ
1
⊗
ξ
2
⊗ · · ·
ξ
n
∈
E
⊗n
. The tensor algebra over
E
, denotedT
+
(
E
)
, is dened to be thenorm-losed subalgebra ofL
(
F
(
E
))
generated byϕ
∞(
M
)
and the reation operatorsT
ξ
,ξ
∈
E
,dened bytheformulaT
ξ
η
=
ξ
⊗
η
,η
∈ F
(
E
)
. Wereferthereaderto[28℄forthebasifats about
T
+
(
E
)
.Definition 2.3
([31 ℄)GivenaW
∗
-orrespondene
E
overthe von Neumann algebraM
,theultraweaklosureofthetensoralgebraofE
,T
+
(
E
)
,inL
(
F
(
E
))
, isalledthe HardyAlgebraofE
,andisdenotedH
∞
(
E
)
.
Example 2.4
IfM
=
E
=
C,
thenF
(
E
)
an be identied withℓ
2
(
Z
+
)
or,through the Fouriertransform,
H
2
(
T
)
. Thetensor algebrathen isisomorphi to the dis algebra
A
(
D
)
viewed asmultipliation operatorsonH
2
(
T
)
andthe Hardy algebraisrealizedasthe lassial Hardyalgebra
H
∞
(
T
)
.
Example 2.5
IfM
=
C
andE
=
C
n
, then
F
(
E
)
an be identied with the spael
2
(
F
+
n
)
, whereF
+
algebrathen iswhat Popesu refersto as the non ommutative dis algebra
A
n
andtheHardyalgebraisitsw
∗
-losure. ItwasstudiedbyPopesu[32 ℄and by DavidsonandPittswho denoteditby
L
n
[17 ℄.Weneedtoreviewsomebasifatsabouttherepresentationtheoryof
H
∞
(
E
)
andof
T
+
(
E
)
. See[28,31℄formoredetails.Definition 2.6
LetE
beaW
∗
-orrespondeneover avon Neumannalgebra
M
. Then:1. Aompletelyontrativeovariantrepresentationof
E
onaHilbertspaeH
isapair(
T, σ
)
,where(a)
σ
isanormal∗
-representationofM
inB
(
H
)
.(b)
T
is a linear, ompletely ontrative map fromE
toB
(
H
)
that is ontinuousintheσ
-topologyof [6 ℄onE
andthe ultraweaktopology onB
(
H
)
.
()
T
is a bimodule map in the sense thatT
(
SξR
) =
σ
(
S
)
T
(
ξ
)
σ
(
R
)
,ξ
∈
E
,andS, R
∈
M
.2. Aompletelyontrativeovariantrepresentation
(
T, σ
)
ofE
inB
(
H
)
is alled isometri inaseT
(
ξ
)
∗
T
(
η
) =
σ
(
h
ξ, η
i
)
(1)
forall
ξ, η
∈
E
.Itshould benotedthat theoperatorspaestrutureon
E
towhihDenition 2.6 refers is that whihE
inherits when viewed as a subspae of its linking algebra.As weshowedin [28, Lemmas3.43.6℄andin [31℄, ifaompletely ontrative ovariant representation,
(
T, σ
)
, ofE
inB
(
H
)
is given, then it determines a ontrationT
˜
:
E
⊗
σ
H
→
H
dened by the formulaT
˜
(
η
⊗
h
) :=
T
(
η
)
h
,η
⊗
h
∈
E
⊗
σ
H
. TheoperatorT
˜
intertwines therepresentationσ
onH
andtheinduedrepresentation
σ
E
◦
ϕ
=
ϕ
(
·
)
⊗
I
H
onE
⊗
σ
H
;i.e.˜
T
(
ϕ
(
·
)
⊗
I
) =
σ
(
·
) ˜
T .
(2)Infatwehavethefollowinglemmafrom[31,Lemma 2.16℄.
The importane of theompletely ontrative ovariantrepresentationsof
E
(or, equivalently, theintertwining ontrationsT
˜
asabove)is that theyyield allompletelyontrativerepresentationsofthetensoralgebra. Morepreisely, wehavethefollowing.Theorem 2.8
LetE
beaW
∗
-orrespondeneoveravonNeumannalgebra
M
. Toevery ompletely ontrative ovariantrepresentation,(
T, σ
)
,ofE
there is a unique ompletely ontrative representationρ
of the tensor algebraT
+
(
E
)
that satisesρ
(
T
ξ) =
T
(
ξ
)
ξ
∈
E
and
ρ
(
ϕ
∞(
a
)) =
σ
(
a
)
a
∈
M.
The map
(
T, σ
)
7→
ρ
isabijetion between theset ofall ompletelyontrative ovariant representations ofE
and all ompletely ontrative (algebra) repre-sentationsofT
+
(
E
)
whose restritions toϕ
∞(
M
)
are ontinuouswith respet tothe ultraweak topologyonL
(
F
(
E
))
.Definition 2.9
If(
T, σ
)
is aompletely ontrativeovariant representation of aW
∗
-orrespondene
E
overavonNeumannalgebraM
,weall the repre-sentationρ
ofT
+
(
E
)
desribed in Theorem 2.8 the integrated form of(
T, σ
)
andwriteρ
=
σ
×
T
.Remark 2.10
Oneoftheprinipaldiulties onefaesindealingwithT
+
(
E
)
andH
∞
(
E
)
istodeidewhenthe integratedform,
σ
×
T
,ofaompletely on-trative ovariant representation(
T, σ
)
extends fromT
+
(
E
)
toH
∞
(
E
)
. This problemarises alreadyinthesimplestsituation,vis. when
M
=
C
=
E
. Inthis setting,T
is given by asingle ontration operator on aHilbert spae,T
+
(
E
)
is the dis algebra andH
∞
(
E
)
is the spae of boundedanalyti funtions onthe dis. Therepresentation
σ
×
T
extendsfromthe disalgebratoH
∞
(
E
)
preisely whenthere isnosingularparttothe spetral measureofthe minimal unitary dilation of
T
. We arenot aware of aomparableresult inourgeneral ontext but we have some suient onditions. One of them is given in the following lemma. Itisnotaneessaryonditioningeneral.Lemma 2.11
[31 , Corollary 2.14℄ Ifk
T
˜
k
<
1
thenσ
×
T
extends to a ultra-weakly ontinuousrepresentation ofH
∞
(
E
)
.
In[31℄weintroduedandstudiedtheoneptsofdualityandofpointevaluation (forelementsof
H
∞
(
E
)
). These playaentralroleinouranalysishere.
Definition 2.12
LetE
beaW
∗
-orrespondeneoveravonNeumannalgebra
M
andletσ
:
M
→
B
(
H
)
beafaithfulnormalrepresentationofM
onaHilbertspae
H
. Then theσ
-dual ofE
,denotedE
σ
,isdenedtobe
An important feature of the dual
E
σ
is that it is a
W
∗
-orrespondene, but overthe ommutant of
σ
(
M
)
,σ
(
M
)
′
.
Proposition 2.13
With respettothe ationofσ
(
M
)
′
andthe
σ
(
M
)
′
-valued innerprodutdenedasfollows,
E
σ
beomesa
W
∗
-orrespondeneover
σ
(
M
)
′
: For
Y
andX
inσ
(
M
)
′
,and
η
∈
E
σ
,
X
·
η
·
Y
:= (
I
⊗
X
)
ηY
,andforη
1
, η
2
∈
E
σ
,
h
η
1
, η
2
i
σ
(
M
)
′
:=
η
1
∗
η
2
.Inthefollowingremarkweexplainwhatwemeanbyevaluatinganelementof
H
∞
(
E
)
atapointintheopenunit ballofthedual.
Remark 2.14
The importaneof this dual spae,E
σ
,is that itislosely re-latedtotherepresentationsof
E
. Infat, theoperatorsinE
σ
whosenormdoes not exeed
1
arepreisely the adjoints of the operators ofthe formT
˜
for a o-variant pair(
T, σ
)
. In partiular, everyη
in the openunitball ofE
σ
(written
D
(
E
σ
)
) gives rise to a ovariant pair
(
T, σ
)
(withη
= ˜
T
∗
) suh that
σ
×
T
extendstoarepresentation ofH
∞
(
E
)
. Given
X
∈
H
∞
(
E
)
wean applythe representation assoiatedto
η
toit. The resultingoperator inB
(
H
)
will bedenotedbyX
b
(
η
∗
)
. Thus
b
X
(
η
∗
) = (
σ
×
η
∗
)(
X
)
.
In this way, we view every element in the Hardy algebra as a
B
(
H
)
-valued funtionb
X
:
D
(
E
σ
)
∗
→
B
(
H
)
onthe open unitballof
(
E
σ
)
∗
. Oneof ourprimaryobjetivesistounderstand the rangeofthe transform
X
→
X
b
,X
∈
H
∞
(
E
)
.
Example 2.15
SupposeM
=
E
=
C
andσ
the representation ofC
on some HilbertspaeH
. ThenitiseasytohekthatE
σ
isisomorphito
B
(
H
)
. FixanX
∈
H
∞
(
E
)
. Aswementionedabove,thisHardyalgebraisthelassial
H
∞
(
T
)
andweanidentify
X
withafuntionf
∈
H
∞
(
T
)
. Given
S
∈
D
(
E
σ
) =
B
(
H
)
, it is not hard to hek that
X
b
(
S
∗
)
, as dened above, is the operator
f
(
S
∗
)
denedthroughthe usualholomorphi funtionalalulus.
Example 2.16
In [17 ℄ Davidson and Pitts assoiate toevery element of the freesemigroupalgebraL
n
(seeExample2.5)afuntionontheopenunitballofC
n
. Thisisaspeialaseofouranalysiswhen
M
=
C,
E
=
C
n
and
σ
isaone dimensionalrepresentationofC.
Inthis aseσ
(
M
)
′
=
C
and
E
σ
=
C
n
. Note, however, thatourdenitionallowsustotake
σ
tobetherepresentationofC
on an arbitraryHilbert spaeH
. If we doso, thenE
σ
isisomorphi to
B
(
H
)
(
n
)
Example 2.17
Partof the reentwork ofPopesuin [35℄maybe astin our framework. We will follow his notation. Fix aHilbert spaeK
, and letE
be the olumn spaeB
(
K
)
n
. Take, also, aHilbert spae
H
and letσ
:
B
(
K
)
→
B
(
K
⊗
H
)
be the representation whih sendsa
∈
B
(
K
)
toa
⊗
I
H
. Then,sinetheommutantof
σ
(
B
(
K
))
isnaturallyisomorphitoB
(
H
)
,itiseasyto see thatE
σ
isthe olumnspae over
B
(
H
)
,B
(
H
)
n
. It follows that
D
(
E
σ
)
is the openunit ballin
B
(
H
)
n
. Afreeformalpowerseries withoeientsfrom
B
(
K
)
is aformal seriesF
=
P
α∈
F
+
n
A
α
⊗
Z
α
where
F
+
n
isthe freesemigroupon
n
generators,theA
α
areelements ofB
(
K
)
andwhereZ
α
is themonomial in nonommuting indeterminatesZ
1
,Z
2
,...,Z
n
determined byα
. IfF
has radiusofonvergeneequalto1
,thenonemay evaluateF
atpointsofD
(
E
σ
)
∗
togetafuntionon
D
(
E
σ
)
∗
withvaluesin
B
(
K
⊗
H
)
,vis.,F
((
S
1
, S
2
,
· · ·
S
n)) =
P
α∈
F
+
n
A
α
⊗
S
α
. See[35 , Theorem 1.1℄. Infat, under additional restritions ontheoeients
A
α
,F
maybeviewedasafuntionX
inH
∞
(
B
(
K
)
n
)
insuh awaythat
F
((
S
1
, S
2
,
· · ·
S
n)) =
X
b
(
S
1
, S
2
,
· · ·
S
n)
inthesensedenedin[31 ,p. 384℄ and disussed above in Remark 2.14. The spae that Popesu denotes byH
∞
(
B
(
X
)
n
1
)
arises whenK
=
C,
and isnaturallyisometrially isomorphi toL
n
[35 ,Theorem3.1℄. Wenotedinthe preeding examplethatL
n
isH
∞
(
C
n
)
. The point of [35 ℄, at least in part, is to study
H
∞
(
B
(
X
)
n
1
)
≃ L
n
=
H
∞
(
C
n
)
through all the representations
σ
ofC
on Hilbert spaesH
, that is, through evaluating funtions inH
∞
(
B
(
X
)
n
1
)
at points the unit ball ofB
(
H
)
n
for all possible
H
's. The spaeB
(
K
)
n
isMoritaequivalentto
C
n
inthesenseof [30 ℄, at leastwhen
dim(
K
)
<
∞
, and, in that asethe tensor algebrasT
+
(
B
(
K
)
n
)
and
T
+
(
C
n
)
are Morita equivalent in the sense desribed by [15℄. The tensor algebra
T
+
(
C
n
)
, in turn, is naturally isometrially isomorphi to Popesu's nonommutative dis algebra
A
n
[33 ℄. The analysis in [15 ℄ suggests a sense in whihC
n
and
B
(
K
)
n
are Morita equivalent even when
dim(
K
) =
∞
, and thattogetherwith[30℄suggeststhatH
∞
(
B
(
K
)
n
)
shouldbeMoritaequivalentto
H
∞
(
B
(
X
)
n
1
)
≃
H
∞
(
C
n
)
. Thiswouldsuggestanevenloseronnetionbetween Popesu's free power series, and all that goes with them, and the perspetive wehave takeninthis paper, whih,asweshall see, involvesgeneralizedShur funtionsandtransferfuntions. Theonnetionseemslikeapromisingavenue toexplore.In[31℄ weexploited theperspetiveofviewing elementsof theHardyalgebra as
B
(
H
)
-valued funtions on the open unit ball of the dual orrespondene to prove a Nevanlinna-Pik type interpolation theorem. In order to state it we introdue some notation: For operatorsB
1
andB
2
inB
(
H
)
, we writeAd
(
B
1
, B
2
)
forthemapfromB
(
H
)
toitselfthatsendsS
toB
1
SB
∗
2
. Also,givenelements
η
1
, η
2
inD
(
E
σ
)
, we let
θ
η
1
,η
2
denote the map, fromσ
(
M
)
′
to itself that sends
a
toh
η
1
, aη
2
i
. Thatis,θ
η
1
,η
2
(
a
) :=
h
η
1
, aη
2
i
=
η
∗
1
aη
2
,a
∈
σ
(
M
)
′
.
Theorem 2.18
([31 ,Theorem5.3℄)LetE
beaW
∗
-orrespondeneoveravon Neumannalgebra
M
andletσ
:
M
→
B
(
H
)
beafaithfulnormalrepresentation ofM
onaHilbertspaeH
. Fixk
pointsη
1
, . . . η
k
inthediskD
(
E
σ
)
2
k
operatorsB
1
, . . . B
k
, C
1
, . . . C
k
inB
(
H
)
. Then thereexistsanX
inH
∞
(
E
)
suhthat
k
X
k ≤
1
andB
i
X
b
(
η
i
∗
) =
C
i
for
i
= 1
,
2
, . . . , k,
if and only if the map fromM
k(
σ
(
M
)
′
)
into
M
k
(
B
(
H
))
denedbythek
×
k
matrix(
Ad
(
B
i
, B
j
)
−
Ad
(
C
i
, C
j
))
◦
(
id
−
θ
η
i
,η
j
)
−
1
(3)
isompletely positive.
That is,themap
T
,say,givenbythematrix(3) isomputedbytheformulaT
((
a
ij
)) = (
b
ij
)
,
where
b
ij
=
B
i((
id
−
θ
η
i
,η
j
)
−
1
(
a
ij)
B
∗
j
−
C
i((
id
−
θ
η
i
,η
j
)
−
1
(
a
ij
)
C
j
∗
and
(
id
−
θ
η
i
,η
j
)
−
1
(
a
ij
) =
a
ij
+
θ
η
i
,η
j
(
a
ij) +
θ
η
i
,η
j
(
θ
η
i
,η
j
(
a
ij)) +
· · ·
We lose this setion with two tehnial lemmas that will be needed in our analysis. Let
M
andN
beW
∗
-algebrasand let
E
be aW
∗
-orrespondene from
M
toN
. Given aσ
-losed suborrespondeneE
0
ofE
we know that the orthogonal projetionP
ofE
ontoE
0
is a right module map. (See [6, Consequenes1.8(ii)℄). InthefollowinglemmaweshowthatP
alsopreserves theleftation.Lemma 2.19
LetE
be aW
∗
-orrespondene from the von Neumann algebra
M
to the von Neumann algebraN
, and letE
0
be a subW
∗
-orrespondene
E
0
ofE
that is losed in theσ
-topology of [6 , Consequenes 1.8 (ii)℄. IfP
is the orthogonal projetion from
E
ontoE
0
, thenP
is abimodule map; i.e.,P
(
aξb
) =
aP
(
ξ
)
b
for alla
∈
M
andb
∈
N
.Proof.
ItsuestohekthatP
(
eξ
) =
eP
(
ξ
)
forallξ
∈
E
andprojetionse
∈
M
. Forξ, η
∈
E
andaprojetione
∈
M
,wehavek
eξ
+
f η
k
2
=
kh
eξ, eξ
i
+
h
f η, f η
ik ≤ kh
eξ, eξ
ik
+
kh
f η, f η
ik
=
k
eξ
k
2
+
k
f η
k
2
,
where
f
= 1
−
e
. So,foreveryλ
∈
R
wehave(
λ
+ 1)
2
k
f P
(
eξ
)
k
2
=
k
f P
(
eξ
+
λf P
(
eξ
))
k
2
≤ k
eξ
+
λf P
(
eξ
)
k
2
≤ k
eξ
k
2
+
λ
2
k
f P
(
eξ
)
k
2
.
Hene,forevery
λ
∈
R,
and,thus,
(
I
−
e
)
P
(
eξ
) =
f P
(
eξ
) = 0
.
Replaing
e
byf
=
I
−
e
wegeteP
((
I
−
e
)
ξ
) = 0
and,therefore,P
(
eξ
) =
eP
(
eξ
) =
eP
(
ξ
)
.
Sine
M
isspanned byitsprojetions,wearedone.Lemma 2.20
LetE
beaW
∗
-orrespondeneover
M
,letσ
beafaithfulnormal representationofM
ontheHilbertspaeE
,andletE
σ
bethe
σ
-dual orrespon-deneoverN
:=
σ
(
M
)
′
. Then
(i) The left ation of
N
onE
σ
is faithful if and only if
E
is full (i.e. if and only if the ultraweakly losed ideal generated by the inner produtsh
ξ
1
, ξ
2
i
,ξ
1
, ξ
2
∈
E
,isall ofM
).(ii) Theleft ationof
M
onE
isfaithfulifand onlyifE
σ
isfull.
Proof.
We shall prove (i). Part (ii) then follows by duality (using [31, Theorem 3.6℄). GivenS
∈
N
,Sη
= 0
for everyη
∈
E
σ
if and only if for all
η
∈
E
σ
and
g
∈ E
,(
I
⊗
S
)
η
(
g
) = 0
. Sine the losed subspaespanned by the rangesof allη
∈
E
σ
is allof
E
⊗
M
E
([31℄), this is equivalent to the equationξ
⊗
Sg
= 0
holdingforallg
∈ E
andξ
∈
E
. Sineh
ξ
⊗
Sg, ξ
⊗
Sg
i
=
h
g, S
∗
h
ξ, ξ
i
Sg
i
, we nd that
SE
σ
= 0
if and only if
σ
(
h
E, E
i
)
S
= 0
, whereh
E, E
i
isthe ultraweakly losedidealgenerated byall innerproduts. Ifthisidealisallof
M
wendthattheequationSE
σ
= 0
impliesthat
S
= 0
. Inthe other diretion,ifthisisnotthease, thenthisideal isoftheform(
I
−
q
)
M
forsomeentralnonzeroprojetionq
andthenS
=
σ
(
q
)
isdierentfrom0
but vanishesonE
σ
.
3
Schur class operator functions and realization
Throughout this setion,
E
will be a xedW
∗
-orrespondene over the von Neumannalgebra
M
andσ
willbeafaithfulrepresentationofM
onaHilbert spaeE
. We then form theσ
-dual ofE
,E
σ
, whih is aorrespondeneover
N
:=
σ
(
M
)
′
, and we write
D
(
E
σ
)
for its open unit ball. Further, we write
D
(
E
σ
)
∗
for
{
η
∗
|
η
∈
D
(
E
σ
)
}
.
The following denition is learly motivated by the ondition appearing in Theorem2.18andShur'stheorem fromlassialfuntion theory.
Definition 3.1
LetΩ
be asubset ofD
(
E
σ
)
and let
Ω
∗
=
{
ω
∗
|
ω
∈
Ω
}
. A funtion
Z
: Ω
∗
→
B
(
E
)
will be alled a Shur lass operator funtion (with valuesin
B
(
E
)
)if,foreveryk
andeveryhoieofelementsη
1
, η
2
, . . . , η
k
inΩ
, the map fromM
k(
N
)
toM
k(
B
(
E
))
denedbythek
×
k
matrixof maps,((
id
−
Ad
(
Z
(
η
i
∗
)
, Z
(
η
∗
j
)))
◦
(
id
−
θ
η
i
,η
j
)
−
1
)
,
Note that,when
M
=
E
=
B
(
E
)
andσ
isthe identityrepresentationofB
(
E
)
onE
,σ
(
M
)
′
is
C
I
E
,E
σ
isisomorphito
C
andD
(
E
σ
)
∗
anbeidentiedwith the open unit dis
D
ofC.
In this ase our denition reovers the lassial Shur lass funtions. More preisely, these funtions are usually dened as analyti funtionsZ
from anopensubsetΩ
ofD
into the losed unit ballofB
(
E
)
butitisknownthatsuhfuntionsarepreiselythoseforwhihthePikkernel
k
Z
(
z, w
) = (
I
−
Z
(
z
)
Z
(
w
)
∗
)(1
−
z
w
¯
)
−
1
is positivesemi-denite on
Ω
. The argument of [31, Remark 5.4℄ shows that the positivity of this kernelis equivalent, in ourase, to the onditionof Denition 3.1. This ondition, in turn, isthesameasassertingthatthekernelk
Z(
ζ
∗
, ω
∗
) := (
id
−
Ad
(
Z
(
ζ
∗
)
, Z
(
ω
∗
))
◦
(
id
−
θ
ζ,ω
)
−
1
(4)isaompletelypositivedenitekernelon
Ω
∗
in thesenseofDenition3.2.2of [14℄.
Forthesakeofompleteness,wereordthefatthat everyelementof
H
∞
(
E
)
ofnormat mostonegivesrisetoaShurlassoperatorfuntion.
Theorem 3.2
LetE
beaW
∗
-orrespondeneoveravonNeumannalgebra
M
and letσ
be afaithful normal representation ofM
inB
(
H
)
for some Hilbert spaeH
. IfX
isanelementofH
∞
(
E
)
ofnormatmostone,thenthefuntion
η
∗
→
X
b
(
η
∗
)
dened in Remark 2.14 is a Shur lass operator funtion on
D
((
E
σ
))
∗
with valuesin
B
(
H
)
.Proof.
OnesimplytakesB
i
=
I
foralli
andC
i
=
X
b
(
η
∗
i
)
in Theorem2.18.Theorem 3.3
LetE
beaW
∗
-orrespondeneoveravonNeumannalgebra
M
. Suppose also thatσ
a faithful normal representation ofM
on a Hilbert spaeE
and thatq
1
andq
2
are projetions inσ
(
M
)
. Finally, suppose thatΩ
is asubset of
D
((
E
σ
))
and that
Z
is a Shur lass operator funtion onΩ
∗
with values in
q
2
B
(
E
)
q
1
. Then thereisaHilbert spaeH
,anormalrepresentationτ
ofN
:=
σ
(
M
)
′
on
H
and operatorsA, B, C
andD
fullling the following onditions:(i) Theoperator
A
lies inq
2
σ
(
M
)
q
1
.(ii) Theoperators
C
,B
,andD
,areinthespaesB
(
E
1
, E
σ
⊗
τ
H
)
,B
(
H,
E
2
)
, andB
(
H, E
σ
⊗
τ
H
)
,respetively,andeahintertwinestherepresentationsof
N
=
σ
(
M
)
′
on the relevant spaes (i.e. , for every
S
∈
N
,CS
=
(
S
⊗
I
H
)
C
,Bτ
(
S
) =
SB
andDτ
(
S
) = (
S
⊗
I
H)
D
).(iii) Theoperator matrix
V
=
A
B
C
D
,
(5)viewedas anoperator from
E
1
⊕
H
toE
2
⊕
(
E
σ
⊗
τ
H
)
, isaoisometry,(iv) For every
η
∗
in
Ω
∗
,
Z
(
η
∗
) =
A
+
B
(
I
−
L
∗
η
D
)
−
1
L
∗
η
C
(6)where
L
η
:
H
→
E
σ
⊗
H
is dened by the formula
L
η
h
=
η
⊗
h
(soL
∗
η
(
θ
⊗
h
) =
τ
(
h
η, θ
i
)
h
).Remark 3.4
BeforegivingtheproofofTheorem3.3,wewanttonotethatthe resultbears astrong resemblane tostandardresultsin the literature. Weall speial attention to [1 , 2, 7, 9 , 10 , 11 , 12 , 13℄. Indeed, we reommend [7 ℄, whih is a survey that explains the general strategy for proving the theorem. What isnovel inourapproahisthe adaptation of the resultsin the literature toaommodateompletelypositive denitekernels.Sinethe matrixin equation(5) andthefuntion inequation (6)arefamiliar onstrutsinmathematialsystemstheory,morepartiularlyfrom
H
∞
-ontrol theory(see,e.g.,[38℄),weadoptthefollowingterminology.
Definition 3.5
LetE
beaW
∗
-orrespondeneover avon Neumannalgebra
M
. Supposethatσ
isafaithfulnormalrepresentation ofM
onaHilbertspaeE
and thatq
1
andq
2
are projetions inσ
(
M
)
. Then an operator matrixV
=
A
B
C
D
,where the entries
A
,B
,C
,andD
,satisfyonditions(
i
)
and(
ii
)
of Theorem 3.3for somenormal representation
τ
ofσ
(
M
)
′
on aHilbert spae
H
,isalleda systemmatrixprovidedV
isaoisometry (that isunitary,ifE
is full). If
V
is a system matrix, then the funtionA
+
B
(
I
−
L
∗
η
D
)
−
1
L
∗
η
C
,η
∗
∈
D
(
E
σ
)
∗
isalledthe transferfuntion determinedby
V
.Proof.
Aswejustremarked,thehypothesisthatZ
isaShurlassfuntion onΩ
∗
meansthat thekernel
k
Z
in equation(4)isompletelypositivedenite inthesenseof[14℄. Consequently,wemayapplyTheorem3.2.3of[14℄,whihis alovelyextensionofKolmogorov'srepresentationtheoremforpositivedenite kernels,to ndanN
-B
(
E
)
W
∗
-orrespondene
F
andafuntionι
fromΩ
∗
to
F
suhthatF
isspannedbyN ι
(Ω
∗
)
B
(
E
)
andsuhthatforevery
η
1
andη
2
inΩ
∗
andevery
a
∈
N
,(
id
−
Ad
(
Z
(
η
∗
1
)
, Z
(
η
2
∗
)))
◦
(
id
−
θ
η
1
,η
2
)
−
1
(
a
) =
h
ι
(
η
1
)
, aι
(
η
2
)
i
.
Itfollowsthatforevery
b
∈
N
and everyη
1
, η
2
inΩ
∗
,
b
−
Z
(
η
∗
1
)
bZ
(
η
∗
2
)
∗
=
h
ι
(
η
1
)
, bι
(
η
2
)
i − h
ι
(
η
1
)
,
h
η
1
, bη
2
i
ι
(
η
2
)
i
=
h
ι
(
η
1
)
, bι
(
η
2
)
i − h
η
1
⊗
ι
(
η
1
)
, bη
2
⊗
ι
(
η
2
)
i
.
Thus,
Set
G
1
:=
span
{
bZ
(
η
∗
)
∗
q
2
T
⊕
bι
(
η
)
q
2
T
|
b
∈
N, η
∈
Ω
∗
, T
∈
B
(
E
)
}
and
G
2
:=
span
{
bq
2
T
⊕
(
bη
⊗
ι
(
η
)
q
2
T
)
|
b
∈
N, η
∈
Ω
∗
, T
∈
B
(
E
)
}
.
Then
G
1
is a subN
-B
(
E
)
W
∗
-orrespondene of
B
(
E
)
⊕
F
(where we use the assumption thatq
2
Z
(
η
∗
) =
q
2
Z
(
η
∗
)
q
1
) andG
2
is a subN
-B
(
E
)
W
∗
-orrespondeneof
B
(
E
)
⊕
(
E
σ
⊗
N
F
)
. (ThelosureinthedenitionsofG
1
, G
2
isinthe
σ
-topologyof[6℄. ItthenfollowsthatG
1
andG
2
areW
∗
-orrespondenes [6,Consequenes1.8(i)℄). Dene
v
:
G
1
→
G
2
bytheequationv
(
bZ
(
η
∗
)
∗
q
2
T
⊕
bι
(
η
)
q
2
T
) =
bq
2
T
⊕
(
bη
⊗
ι
(
η
)
q
2
T
)
.
It followsfrom (7)that
v
is anisometry. Itis alsolearthat itisabimodule map. WewriteP
i
fortheorthogonalprojetionontoG
i
,i
= 1
,
2
andV
˜
forthe map˜
V
:=
P
2
vP
1
:
q
1
B
(
E
)
⊕
F
→
q
2
B
(
E
)
⊕
(
E
σ
⊗
N
F
)
.
Then
V
˜
is apartial isometry and, sineP
1
, v
andP
2
are all bimodule maps (seeLemma 2.19),soisV
˜
. WewriteV
˜
matriially:˜
V
=
α β
γ
δ
,
where
α
:
q
1
B
(
E
)
→
q
2
B
(
E
)
,β
:
F
→
q
2
B
(
E
)
,γ
:
q
1
B
(
E
)
→
E
σ
⊗
F
and
δ
:
F
→
E
σ
⊗
F
and all these maps are bimodule maps. Let
H
0
be the Hilbert spaeF
⊗
B
(
E
)
E
and note thatB
(
E
)
⊗
B
(
E
)
E
is isomorphitoE
(and theisomorphismpreservestheleftN
-ation). TensoringontherightbyE
(overB
(
E
)
)weobtainapartialisometryV
0
:=
A
0
B
0
C
0
D
0
:
E
1
H
0
→
E
2
E
σ
⊗
H
0
.
Here
A
0
=
α
⊗
I
E
,B
0
=
β
⊗
I
E
,C
0
=
γ
⊗
I
E
andD
0
=
δ
⊗
I
E
. Thesemaps are well dened beausethe mapsα, β, γ
andδ
arerightB
(
E
)
-module maps. Sinethesemaps arealso leftN
-modulemaps,soareA
0
, B
0
, C
0
andD
0
. BythedenitionofV
0
,itsinitialspaeisG
1
⊗ E
anditsnalspaeisG
2
⊗ E
. Infat,V
0
induesanequivalene ofthe representationsofN
onG
1
⊗ E
and onG
2
⊗ E
.Itwillbeonvenienttousethenotation
K
1
N
K
2
iftheHilbertspaesK
1
andK
2
arebothleftN
-modulesandtherepresentationofN
onK
1
isequivalenttoaUsingthisnotation,weanwrite
G
1
⊗ E ≃
N
G
2
⊗ E
. FormM
2
:= (
E
2
⊕
(
E
σ
⊗
H
0
))
⊖
(
G
2
⊗ E
)
, whih isaleftN
-module, andnote thatL
:=
F
(
E
σ
)
⊗ M
2
also is a left
N
-module, where the representation ofN
onL
is the indued representation. SineL
=
F
(
E
σ
)
⊗ M
2
=
L
∞
n
=0
((
E
σ
)
⊗n
⊗
(
M
2
))
,itisevident that(
E
σ
⊗
L
)
⊕ M
2
≃
N
L
. Indeed,theisomorphismsarejustthenaturalones that givetheassoiativityof thetensorprodutsinvolved. Thus,E
2
⊕
(
E
σ
⊗
(
H
0
⊕
L
)) =
E
2
⊕
(
E
σ
⊗
H
0
)
⊕
(
E
σ
⊗
L
) =
G
2
⊗E ⊕M
2
⊕
E
σ
⊗
L
≃
N
G
2
⊗E ⊕
L
≃
N
G
1
⊗ E ⊕
L
N
E
1
⊕
(
H
0
⊕
L
)
. Consequently,weobtainaoisometrioperatorV
:
E
1
⊕
(
H
0
⊕
L
)
→ E
2
⊕
E
σ
⊗
(
H
0
⊕
L
)
thatintertwinestherepresentationsof
N
and extendsV
0
. Note that,ifV
0
were knownto beanisometry (sothatG
2
⊗ E ≃
N
G
1
⊗ E
=
E
1
⊕
H
0
),thenwewouldhaveequivaleneaboveandV
anbehosentobeunitary.
Assume that
E
isfull. We alsowriteM
1
for(
E
1
⊕
H
0
)
⊖
G
1
⊗ E
. SineE
is full,therepresentationρ
ofN
onE
σ
⊗ E
isfaithful(Lemma2.20)anditfollows that everyrepresentationof
N
isquasiequivalenttoasubrepresentationofρ
. WriteE
∞
forthediretsumofinnitelymanyopiesofE
. ThenE
σ
⊗E
∞
isthediretsumofinnitelymanyopiesof
E
σ
⊗E
and,thus,everyrepresentationof
N
isequivalenttoasubrepresentationoftherepresentationofN
onE
σ
⊗ E
∞
.Inpartiular, weanwrite
M
1
⊕ E
∞
N
E
σ
⊗ E
∞
. ThusE
1
⊕
(
H
0
⊕ E
∞) =
(
G
1
⊗ E
)
⊕ M
1
⊕ E
∞
N
E
2
⊕
(
E
σ
⊗
H
0
)
⊕
(
E
σ
⊗ E
∞) =
E
2
⊕
(
E
σ
⊗
(
H
0
⊕ E
∞))
.So,replaing
H
0
byH
0
⊕ E
∞
,weanreplaeV
0
byanisometryand,usingthe argumentjustpresented,weonludethattheresultingV
isaunitaryoperator intertwiningtherepresentationsofN
andextendingV
0
.Sowelet
V
betheoisometryjust onstruted(andtreatitasunitarywhenE
isfull). WritingH
:=
H
0
⊕
L
,weanexpressV
inthematriialform asinpart(iii) ofthestatementof thetheorem. Conditions (i)and(ii) thenfollow from thefat that
V
intertwinestheindiatedrepresentationsofN
. It isleft to prove(iv).Setting
b
=
T
=
I
inthedenitionofv
aboveandwritingv
inamatriialform weseethatα β
γ
δ
Z
(
η
∗
)
∗
q
2
ι
(
η
)
q
2
=
q
2
η
⊗
ι
(
η
)
q
2
.
Tensoringby
I
E
ontherightandidentifyingB
(
E
)
⊗
B
(
E
)
E
withE
asabove,we ndthatA
0
B
0
C
0
D
0
Z
(
η
∗
)
∗
g
ι
(
η
)
⊗
g
=
g
η
⊗
(
ι
(
η
)
⊗
g
)
,
for
g
∈ E
2
. SineA, B, C
andD
extendA
0
, B
0
, C
0
andD
0
respetively, we andropthesubsript0
. Wealsousethefat that thematrixweobtainisa oisometry, andthus its adjointequals itsinverse onits range. We onludethat
A
∗
C
∗
B
∗
D
∗
g
η
⊗
(
ι
(
η
)
⊗
g
)
=
Z
(
η
∗
)
∗
g
ι
(
η
)
⊗
g
.
(8)Thus
ι
(
η
)
⊗
g
=
B
∗
g
+
D
∗
(
η
⊗
(
ι
(
η
)
⊗
g
)) =
B
∗
g
+
D
∗
L
η(
ι
(
η
)
⊗
g
)
and
Combiningthisequalitywiththeotherequationthatwegetfrom(8),wehave
Z
(
η
∗
)
∗
g
=
A
∗
g
+
C
∗
L
η
(
I
−
D
∗
L
η)
−
1
B
∗
g , g
∈ E
.
Takingadjointsyields(iv).
Thus,Theorem3.3assertsthateveryShurlassfuntiondeterminesasystem matrixwhose transferfuntion representsthefuntion. The systemmatrixis notuniqueingeneral,butastheproofofTheorem3.3shows,itarisesthrough aseries of naturalhoies. Of ourse,equation (6) suggeststhat everyShur lassfuntionrepresentsanelementin
H
∞
(
E
)
. Thisisindeedthease,asthe followingonverseshows.
Theorem 3.6
LetE
be aW
∗
-orrespondene over a
W
∗
-algebra
M
, and letσ
be a faithful normal representation ofM
on a Hilbert spaeE
. IfV
=
A
B
C
D
is a system matrix determined by a normal representation
τ
ofN
:=
σ
(
M
)
′
on a Hilbert spae
H
, then there is anX
∈
H
∞
(
E
)
,
k
X
k ≤
1
, suhthatb
X
(
η
∗
) =
A
+
B
(
I
−
L
∗
η
D
)
−
1
L
∗
η
C
,for all
η
∗
∈
D
(
E
σ
)
∗
and, onversely, every
X
∈
H
∞
(
E
)
,
k
X
k ≤
1
, may be representedinthis fashion for asuitablesystemmatrixV
=
A
B
C
D
.
Proof.
Foreveryn
≥
0
wedeneanoperatorK
n
fromE
to(
E
σ
)
⊗n
⊗ E
as follows. For
n
= 0
,wesetK
0
=
A
-anoperatorinB
(
E
)
. Forn
= 1
,wedeneK
1
,mappingE
toE
σ
⊗ E
,tobe
(
I
1
⊗
B
)
C
,whereforallk
≥
1
,I
k
denotesthe identityoperatoron(
E
σ
)
⊗k
. For
n
≥
2
,wesetK
n
:= (
I
n
⊗
B
)(
I
n−
1
⊗
D
)
· · ·
(
I
1
⊗
D
)
C.
Note,rst,thatitfollowsfrom thepropertiesof
A, B, C
andD
that,foreveryn
≥
0
and everya
∈
N
,K
n
a
= (
ϕ
n(
a
)
⊗
I
E
)
K
n
whereϕ
n
denes the leftmultipliationon
(
E
σ
)
⊗n
. Thus,writing
ι
fortheidentityrepresentationofN
onE
,K
n
lies intheι
-dual of(
E
σ
)
⊗n
whih, byTheorem 3.6and Lemma3.7 of [31℄, is isomorphi to
E
⊗n
. Hene, for every
n
≥
0
,K
n
denes a unique elementξ
n
inE
⊗n
.
For every
n
≥
0
andη
∈
E
σ
we shall write
L
n(
η
)
for the operator from(
E
σ
)
⊗n
⊗ E
to
(
E
σ
)
⊗
(
n
+1)
⊗ E
given by tensoring on the left by
η
. Also note that, fork
≥
1
andn
≥
0
,I
k
⊗
K
n
is an operator from(
E
σ
)
⊗k
⊗ E
to
(
E
σ
)
⊗
(
k
+
n
)
⊗ E
. With thisnotation,itiseasytoseethat,forall
k
≥
1
andn
≥
0
,(
I
k
+1
⊗
K
n)
L
k(
η
) =
L
k
+
n
(
η
)(
I
k
⊗
K
n)
.
(9)Note, too,thatweanwrite
andeveryoperatoron
F
(
E
σ
)
⊗E
anbewritteninamatriialformwithrespet to this deomposition(with indiesstarting at
0
). Foreverym
,0
≤
m
≤ ∞
, weletS
m
betheoperatordenedbythematrixwhosei, j
entryisI
j
⊗
K
i−j
, if0
≤
j
≤
i
≤
m
,andis0
otherwise. (Form
=
∞
,itisnotlearyetthat the matrixsoonstrutedrepresentsabounded operator,but thiswill beveried later).Sofarwehavenotusedtheassumptionthat
V
isaoisometry. Butifwetake this into aount, form the produtV V
∗
, and set it equalto
I
E⊕
(
E
σ
⊗H
)
, we ndthatI
E
−
AA
∗
=
BB
∗
(10)CC
∗
=
I
E
σ
⊗
τ
H
−
DD
∗
(11)AC
∗
=
−
BD
∗
(12)
Welaimthat,for
1
≤
j
≤
i
≤
m
,thefollowingequationshold,(
I
−
S
m
S
m
∗
)i,j
= (
I
i
⊗
B
)(
I
i−
1
⊗
D
)
· · ·
DD
∗
· · ·
(
I
j−
1
⊗
D
∗
)(
I
j
⊗
B
∗
);
(13)that for
0
< i
≤
m
,(
I
−
S
m
S
∗
m
)i,
0
= (
I
i
⊗
B
)(
I
i−
1
⊗
D
)
· · ·
DB
∗
,
(14)andthat for
i
=
j
= 0
,(
I
−
S
m
S
m
∗
)
0
,
0
=
BB
∗
.
(15)Equation(15)followsimmediatelyfrom(10)sine
(
S
m)
0
,
0
=
A
. For0
< i
≤
m
weompute(
I
−
S
m
S
∗
m
)i,
0
=
−
(
S
m)i,
0
(
S
m)
∗
0
,
0
=
−
(
I
i
⊗
B
)(
I
i−
1
⊗
D
)
· · ·
(
I
1
⊗
D
)
CA
∗
= (
I
i
⊗
B
)(
I
i−
1
⊗
D
)
· · ·
(
I
1
⊗
D
)
DB
∗
where, in the last equalitywe used (12). It is left to prove(13). Letus writeR
i,j
for theleft hand sideof (13). (Forj
= 0
< i
we haveR
i,
0
= (
I
i
⊗
B
)(
I
i−
1
⊗
D
)
· · ·
DB
∗
and when bothare
0
,R
0
,
0
=
BB
∗
). Wehave
K
0
K
∗
0
=
AA
∗
=
I
−
BB
∗
=
I
−
R
0
,
0
R
∗
0
,
0
. For0 =
j < i
≤
m
wehaveK
i
K
∗
0
= (
I
i
⊗
B
)(
I
i−
1
⊗
D
)
· · ·
(
I
1
⊗
D
)
CA
∗
=
−
(
I
i
⊗
B
)(
I
i−
1
⊗
D
)
· · ·