ABSTRACT
Wehaveseenthat,inmanycases,thequantiflabilityofembed-<b:lwh-phrasesispragmaticallydetenninedinthatitiscorrelatedwith the(largelypragmatic)variabilityofmanypredicateswithrespectwhethertheypresupposetheircomplement,i.e.,whethertheyhavea factiveinterpretation.Iftheydo(orcan),then,inthelogicalrepresentation,thecomplementisaccommodatedintotherestrictivetenn,thus sanctioningquantification.Thecrucialroleofcontextraisesproblems formorestrictlysemanticaccountssuchasGroenendijkandStokhofs, accordingtowhichitisinherentpropertiesofthematrixpredicatethat makesthedifference,orBoer's,accordingtowhichwh-clausesare inherentlyfactive.Nevertheless,whileIthinkBoer'sgeneralapproach iswrong,hisstrategyoffactivitynullificationdoes,inanumberof cases,havesomeplausibility,asinBaker'sinterrogativestructures.
