ABSTRACT
Strictlyspeaking,thisisnotquitecorrect.Proposi-
tionalfunctionsinIL*arealsotypedwithrespectto
astowhethereachargumentisann-placepropositional
functionoranindividual.So,eventhoughthespaceof
logicaltypesassociatedwithIL*iscertainlymore
limitedthantheoneassociatedwithIL,stillstrictly
infinitesetofpossibletypeassignmentstobasic
categories.Ofcourse,wemighteasilycutdowntoa
reasonablesizethenumberofpossibletypeassignmentsby
verysmalladicity(possibly~3)ofverbs.Theproblem iswhetherwecanfindaprincipledwayofcarryingthis
Oneofthecentralideasbuiltintotheversionof
combinatorialcapacitiesofphrasesareareflexof
ofconstrainingandmakingclearthenotionofbeinga
reflexofemployedinthelatterstatementwouldbefor
function(i.e.belongstoacategoryoftheformA/B)iff itissemanticallyafunction.Itiseasytoconceive
instance,Bennett's(1976)modificationofMontague'stype assignmentdoesnot,sinceCNistakenasaprimitive
(i.e.nonfunctional)categoryandyetmappedintoa functionaltype(namely<e,t>).Wecanstatetherefore thefollowinggeneralprinciple:
(12)Functionalcorrespondence:afunctionaltype mustcorrespondtoafunctionalcategoryra functionalcategorymustcorrespondtoa functionaltype.