This item is only available as the following downloads:

( PDF )

PAGE 1

forall x AnIntroductiontoFormalLogic P.D.Magnus UniversityatAlbany,StateUniversityofNewYork fecundity.com/logic,version1.26[090109] ThisbookisoeredunderaCreativeCommonslicense. Attribution-ShareAlike3.0

PAGE 2

Theauthorwouldliketothankthepeoplewhomadethisprojectpossible.Notable amongtheseareCristynMagnus,whoreadmanyearlydrafts;AaronSchiller,who wasanearlyadopterandprovidedconsiderable,helpfulfeedback;andBinKang, CraigErb,NathanCarter,WesMcMichael,andthestudentsofIntroductionto Logic,whodetectedvariouserrorsinpreviousversionsofthebook. c 2005{2009byP.D.Magnus.Somerightsreserved. Youarefreetocopythisbook,todistributeit,todisplayit,andtomakederivativeworks, underthefollowingconditions:aAttribution.Youmustgivetheoriginalauthorcredit.b ShareAlike.Ifyoualter,transform,orbuilduponthiswork,youmaydistributetheresulting workonlyunderalicenseidenticaltothisone.|Foranyreuseordistribution,youmust makecleartoothersthelicensetermsofthiswork.Anyoftheseconditionscanbewaivedif yougetpermissionfromthecopyrightholder.Yourfairuseandotherrightsareinnoway aectedbytheabove.|Thisisahuman-readablesummaryofthefulllicense,whichis availableon-lineat http://creativecommons.org/licenses/by-sa/3.0/ TypesettingwascarriedoutentirelyinL A T E X2 .Thestylefortypesettingproofs isbasedontch.styv0.4byPeterSelinger,UniversityofOttawa. Thiscopyof forall x iscurrentasofJanuary9,2009.Themostrecentversion isavailableon-lineat http://www.fecundity.com/logic

PAGE 3

Contents 1Whatislogic?5 1.1Arguments..............................6 1.2Sentences...............................6 1.3Twowaysthatargumentscangowrong..............7 1.4Deductivevalidity..........................8 1.5Otherlogicalnotions.........................10 1.6Formallanguages...........................12 PracticeExercises.............................15 2Sententiallogic17 2.1Sentenceletters............................17 2.2Connectives..............................19 2.3Othersymbolization.........................28 2.4SentencesofSL............................29 PracticeExercises.............................33 3Truthtables37 3.1Truth-functionalconnnectives....................37 3.2Completetruthtables........................38 3.3Usingtruthtables..........................41 3.4Partialtruthtables..........................42 PracticeExercises.............................44 4Quantiedlogic48 4.1Fromsentencestopredicates....................48 4.2BuildingblocksofQL........................50 4.3Quantiers..............................54 4.4TranslatingtoQL..........................57 4.5SentencesofQL............................68 4.6Identity................................71 PracticeExercises.............................76 5Formalsemantics83 5.1SemanticsforSL...........................84 3

PAGE 4

4 CONTENTS 5.2InterpretationsandmodelsinQL..................88 5.3Semanticsforidentity........................92 5.4Workingwithmodels.........................94 5.5TruthinQL..............................98 PracticeExercises.............................103 6Proofs107 6.1BasicrulesforSL...........................108 6.2Derivedrules.............................117 6.3Rulesofreplacement.........................119 6.4Rulesforquantiers.........................121 6.5Rulesforidentity...........................126 6.6Proofstrategy.............................128 6.7Proof-theoreticconcepts.......................129 6.8Proofsandmodels..........................131 6.9Soundnessandcompleteness.....................132 PracticeExercises.............................134 AOthersymbolicnotation140 BSolutionstoselectedexercises143 CQuickReference156

PAGE 5

Chapter1 Whatislogic? Logicisthebusinessofevaluatingarguments,sortinggoodonesfrombadones. Ineverydaylanguage,wesometimesusetheword`argument'torefertobelligerentshoutingmatches.Ifyouandafriendhaveanargumentinthissense, thingsarenotgoingwellbetweenthetwoofyou. Inlogic,wearenotinterestedintheteeth-gnashing,hair-pullingkindofargument.Alogicalargumentisstructuredtogivesomeoneareasontobelieve someconclusion.Hereisonesuchargument: Itisrainingheavily. Ifyoudonottakeanumbrella,youwillgetsoaked. : : Youshouldtakeanumbrella. Thethreedotsonthethirdlineoftheargumentmean`Therefore'andthey indicatethatthenalsentenceisthe conclusion oftheargument.Theother sentencesare premises oftheargument.Ifyoubelievethepremises,thenthe argumentprovidesyouwithareasontobelievetheconclusion. Thischapterdiscussessomebasiclogicalnotionsthatapplytoargumentsina naturallanguagelikeEnglish.Itisimportanttobeginwithaclearunderstandingofwhatargumentsareandofwhatitmeansforanargumenttobevalid. LaterwewilltranslateargumentsfromEnglishintoaformallanguage.We wantformalvalidity,asdenedintheformallanguage,tohaveatleastsomeof theimportantfeaturesofnatural-languagevalidity. 5

PAGE 6

6 forall x 1.1Arguments Whenpeoplemeantogivearguments,theytypicallyoftenusewordslike`therefore'and`because.'Whenanalyzinganargument,therstthingtodoisto separatethepremisesfromtheconclusion.Wordsliketheseareacluetowhat theargumentissupposedtobe,especiallyif|intheargumentasgiven|the conclusioncomesatthebeginningorinthemiddleoftheargument. premiseindicators: since,because,giventhat conclusionindicators: therefore,hence,thus,then,so Tobeperfectlygeneral,wecandenean argument asaseriesofsentences. Thesentencesatthebeginningoftheseriesarepremises.Thenalsentencein theseriesistheconclusion.Ifthepremisesaretrueandtheargumentisagood one,thenyouhaveareasontoaccepttheconclusion. Noticethatthisdenitionisquitegeneral.Considerthisexample: Thereiscoeeinthecoeepot. Thereisadragonplayingbassoononthearmoire. : : SalvadorDaliwasapokerplayer. Itmayseemoddtocallthisanargument,butthatisbecauseitwouldbe aterribleargument.Thetwopremiseshavenothingatalltodowiththe conclusion.Nevertheless,givenourdenition,itstillcountsasanargument| albeitabadone. 1.2Sentences Inlogic,weareonlyinterestedinsentencesthatcangureasapremiseor conclusionofanargument.Sowewillsaythata sentence issomethingthat canbetrueorfalse. Youshouldnotconfusetheideaofasentencethatcanbetrueorfalsewith thedierencebetweenfactandopinion.Often,sentencesinlogicwillexpress thingsthatwouldcountasfacts|suchas`Kierkegaardwasahunchback'or `Kierkegaardlikedalmonds.'Theycanalsoexpressthingsthatyoumightthink ofasmattersofopinion|suchas,`Almondsareyummy.' Also,therearethingsthatwouldcountas`sentences'inalinguisticsorgrammar coursethatwewillnotcountassentencesinlogic.

PAGE 7

ch.1whatislogic? 7 Questions Inagrammarclass,`Areyousleepyyet?'wouldcountasan interrogativesentence.Althoughyoumightbesleepyoryoumightbealert,the questionitselfisneithertruenorfalse.Forthisreason,questionswillnotcount assentencesinlogic.Supposeyouanswerthequestion:`Iamnotsleepy.'This iseithertrueorfalse,andsoitisasentenceinthelogicalsense.Generally, questions willnotcountassentences,but answers will. `Whatisthiscourseabout?'isnotasentence.`Nooneknowswhatthiscourse isabout'isasentence. Imperatives Commandsareoftenphrasedasimperativeslike`Wakeup!',`Sit upstraight',andsoon.Inagrammarclass,thesewouldcountasimperative sentences.Althoughitmightbegoodforyoutositupstraightoritmightnot, thecommandisneithertruenorfalse.Note,however,thatcommandsarenot alwaysphrasedasimperatives.`Youwillrespectmyauthority' is eithertrue orfalse|eitheryouwilloryouwillnot|andsoitcountsasasentenceinthe logicalsense. Exclamations `Ouch!'issometimescalledanexclamatorysentence,butit isneithertruenorfalse.Wewilltreat`Ouch,Ihurtmytoe!'asmeaningthe samethingas`Ihurtmytoe.'The`ouch'doesnotaddanythingthatcouldbe trueorfalse. 1.3Twowaysthatargumentscangowrong Considertheargumentthatyoushouldtakeanumbrellaonp.5,above.If premiseisfalse|ifitissunnyoutside|thentheargumentgivesyouno reasontocarryanumbrella.Evenifitisrainingoutside,youmightnotneedan umbrella.Youmightweararainpanchoorkeeptocoveredwalkways.Inthese cases,premisewouldbefalse,sinceyoucouldgooutwithoutanumbrella andstillavoidgettingsoaked. Supposeforamomentthatboththepremisesaretrue.Youdonotownarain pancho.Youneedtogoplaceswheretherearenocoveredwalkways.Nowdoes theargumentshowyouthatyoushouldtakeanumbrella?Notnecessarily. Perhapsyouenjoywalkingintherain,andyouwouldliketogetsoaked.In thatcase,eventhoughthepremisesweretrue,theconclusionwouldbefalse. Foranyargument,therearetwowaysthatitcouldbeweak.First,oneormore ofthepremisesmightbefalse.Anargumentgivesyouareasontobelieveits conclusiononlyifyoubelieveitspremises.Second,thepremisesmightfailto

PAGE 8

8 forall x supporttheconclusion.Evenifthepremisesweretrue,theformoftheargument mightbeweak.Theexamplewejustconsideredisweakinbothways. Whenanargumentisweakinthesecondway,thereissomethingwrongwith the logicalform oftheargument:Premisesofthekindgivendonotnecessarily leadtoaconclusionofthekindgiven.Wewillbeinterestedprimarilyinthe logicalformofarguments. Consideranotherexample: Youarereadingthisbook. Thisisalogicbook. : : Youarealogicstudent. Thisisnotaterribleargument.Mostpeoplewhoreadthisbookarelogic students.Yet,itispossibleforsomeonebesidesalogicstudenttoreadthis book.Ifyourroommatepickedupthebookandthumbedthroughit,theywould notimmediatelybecomealogicstudent.Sothepremisesofthisargument,even thoughtheyaretrue,donotguaranteethetruthoftheconclusion.Itslogical formislessthanperfect. Anargumentthathadnoweaknessofthesecondkindwouldhaveperfectlogical form.Ifitspremisesweretrue,thenitsconclusionwould necessarily betrue. Wecallsuchanargument`deductivelyvalid'orjust`valid.' Eventhoughwemightcounttheargumentaboveasagoodargumentinsome sense,itisnotvalid;thatis,itis`invalid.'Oneimportanttaskoflogicisto sortvalidargumentsfrominvalidarguments. 1.4Deductivevalidity Anargumentisdeductively valid ifandonlyifitisimpossibleforthepremises tobetrueandtheconclusionfalse. Thecrucialthingaboutavalidargumentisthatitisimpossibleforthepremises tobetrue atthesametime thattheconclusionisfalse.Considerthisexample: Orangesareeitherfruitsormusicalinstruments. Orangesarenotfruits. : : Orangesaremusicalinstruments. Theconclusionofthisargumentisridiculous.Nevertheless,itfollowsvalidly fromthepremises.Thisisavalidargument. If bothpremisesweretrue, then theconclusionwouldnecessarilybetrue.

PAGE 9

ch.1whatislogic? 9 Thisshowsthatadeductivelyvalidargumentdoesnotneedtohavetrue premisesoratrueconclusion.Conversely,havingtruepremisesandatrue conclusionisnotenoughtomakeanargumentvalid.Considerthisexample: LondonisinEngland. BeijingisinChina. : : ParisisinFrance. Thepremisesandconclusionofthisargumentare,asamatteroffact,alltrue. Thisisaterribleargument,however,becausethepremiseshavenothingtodo withtheconclusion.ImaginewhatwouldhappenifParisdeclaredindependence fromtherestofFrance.Thentheconclusionwouldbefalse,eventhoughthe premiseswouldbothstillbetrue.Thus,itis logicallypossible forthepremises ofthisargumenttobetrueandtheconclusionfalse.Theargumentisinvalid. Theimportantthingtorememberisthatvalidityisnotabouttheactualtruth orfalsityofthesentencesintheargument.Instead,itisabouttheformof theargument:Thetruthofthepremisesisincompatiblewiththefalsityofthe conclusion. Inductivearguments Therecanbegoodargumentswhichneverthelessfailtobedeductivelyvalid. Considerthisone: InJanuary1997,itrainedinSanDiego. InJanuary1998,itrainedinSanDiego. InJanuary1999,itrainedinSanDiego. : : ItrainseveryJanuaryinSanDiego. Thisisan inductive argument,becauseitgeneralizesfrommanycasestoa conclusionaboutallcases. Certainly,theargumentcouldbemadestrongerbyaddingadditionalpremises: InJanuary2000,itrainedinSanDiego.InJanuary2001 ::: andsoon.Regardlessofhowmanypremisesweadd,however,theargumentwillstillnotbe deductivelyvalid.Itispossible,althoughunlikely,thatitwillfailtorainnext JanuaryinSanDiego.Moreover,weknowthattheweathercanbeckle.No amountofevidenceshouldconvinceusthatitrainsthere every January.Who istosaythatsomeyearwillnotbeafreakishyearinwhichthereisnorain inJanuaryinSanDiego;evenasinglecounter-exampleisenoughtomakethe conclusionoftheargumentfalse.

PAGE 10

10 forall x Inductivearguments,evengoodinductivearguments,arenotdeductivelyvalid. Wewillnotbeinterestedininductiveargumentsinthisbook. 1.5Otherlogicalnotions Inadditiontodeductivevalidity,wewillbeinterestedinsomeotherlogical concepts. Truth-values Trueorfalseissaidtobethe truth-value ofsentence.Wedenedsentences asthingsthatcouldbetrueorfalse;wecouldhavesaidinsteadthatsentences arethingsthatcanhavetruth-values. Logicaltruth Inconsideringargumentsformally,wecareaboutwhatwouldbetrue if the premisesweretrue.Generally,wearenotconcernedwiththeactualtruthvalue ofanyparticularsentences|whethertheyare actually trueorfalse.Yetthere aresomesentencesthatmustbetrue,justasamatteroflogic. Considerthesesentences: 1.Itisraining. 2.Eitheritisraining,oritisnot. 3.Itisbothrainingandnotraining. Inordertoknowifsentence1istrue,youwouldneedtolookoutsideorcheckthe weatherchannel.Logicallyspeaking,itmightbeeithertrueorfalse.Sentences likethisarecalled contingent sentences. Sentence2isdierent.Youdonotneedtolookoutsidetoknowthatitistrue. Regardlessofwhattheweatherislike,itiseitherrainingornot.Thissentence is logicallytrue ;itistruemerelyasamatteroflogic,regardlessofwhatthe worldisactuallylike.Alogicallytruesentenceiscalleda tautology Youdonotneedtochecktheweathertoknowaboutsentence3,either.Itmust befalse,simplyasamatteroflogic.Itmightberaininghereandnotraining acrosstown,itmightberainingnowbutstoprainingevenasyoureadthis,but itisimpossibleforittobebothrainingandnotraininghereatthismoment.

PAGE 11

ch.1whatislogic? 11 Thethirdsentenceis logicallyfalse ;itisfalseregardlessofwhattheworldis like.Alogicallyfalsesentenceiscalleda contradiction Tobeprecise,wecandenea contingentsentence asasentencethatis neitheratautologynoracontradiction. Asentencemight always betrueandstillbecontingent.Forinstance,ifthere neverwereatimewhentheuniversecontainedfewerthanseventhings,then thesentence`Atleastseventhingsexist'wouldalwaysbetrue.Yetthesentence iscontingent;itstruthisnotamatteroflogic.Thereisnocontradictionin consideringapossibleworldinwhichtherearefewerthanseventhings.The importantquestioniswhetherthesentence must betrue,justonaccountof logic. Logicalequivalence Wecanalsoaskaboutthelogicalrelations between twosentences.Forexample: Johnwenttothestoreafterhewashedthedishes. Johnwashedthedishesbeforehewenttothestore. Thesetwosentencesarebothcontingent,sinceJohnmightnothavegoneto thestoreorwasheddishesatall.Yettheymusthavethesametruth-value.If eitherofthesentencesistrue,thentheybothare;ifeitherofthesentencesis false,thentheybothare.Whentwosentencesnecessarilyhavethesametruth value,wesaythattheyare logicallyequivalent Consistency Considerthesetwosentences: B1 MyonlybrotheristallerthanIam. B2 MyonlybrotherisshorterthanIam. Logicalonecannottelluswhich,ifeither,ofthesesentencesistrue.Yetwecan saythat if therstsentenceB1istrue, then thesecondsentenceB2must befalse.AndifB2istrue,thenB1mustbefalse.Itcannotbethecasethat bothofthesesentencesaretrue. Ifasetofsentencescouldnotallbetrueatthesametime,likeB1{B2,theyare saidtobe inconsistent .Otherwise,theyare consistent .

PAGE 12

12 forall x Wecanaskabouttheconsistencyofanynumberofsentences.Forexample, considerthefollowinglistofsentences: G1 Thereareatleastfourgiraesatthewildanimalpark. G2 Thereareexactlysevengorillasatthewildanimalpark. G3 Therearenotmorethantwomartiansatthewildanimalpark. G4 Everygiraeatthewildanimalparkisamartian. G1andG4togetherimplythatthereareatleastfourmartiangiraesatthe park.ThisconictswithG3,whichimpliesthattherearenomorethantwo martiangiraesthere.SothesetofsentencesG1{G4isinconsistent.Notice thattheinconsistencyhasnothingatalltodowithG2.G2justhappenstobe partofaninconsistentset. Sometimes,peoplewillsaythataninconsistentsetofsentences`containsa contradiction.'Bythis,theymeanthatitwouldbelogicallyimpossibleforall ofthesentencestobetrueatonce.Asetcanbeinconsistentevenwhenallof thesentencesinitareeithercontingentortautologous.Whenasinglesentence isacontradiction,thenthatsentencealonecannotbetrue. 1.6Formallanguages Hereisafamousvalidargument: Socratesisaman. Allmenaremortal. : : Socratesismortal. Thisisaniron-cladargument.Theonlywayyoucouldchallengetheconclusion isbydenyingoneofthepremises|thelogicalformisimpeccable.Whatabout thisnextargument? Socratesisaman. Allmenarecarrots. : : Socratesisacarrot. Thisargumentmightbelessinterestingthantherst,becausethesecond premiseisobviouslyfalse.Thereisnoclearsenseinwhichallmenarecarrots.Yettheargumentisvalid.Toseethis,noticethatbothargumentshave thisform:

PAGE 13

ch.1whatislogic? 13 S is M All M sare C s. : :S is C Inbotharguments S standsforSocratesand M standsforman.Intherst argument, C standsformortal;inthesecond, C standsforcarrot.Bothargumentshavethisform,andeveryargumentofthisformisvalid.Soboth argumentsarevalid. Whatwedidherewasreplacewordslike`man'or`carrot'withsymbolslike `M'or`C'soastomakethelogicalformexplicit.Thisisthecentralidea behindformallogic.Wewanttoremoveirrelevantordistractingfeaturesofthe argumenttomakethelogicalformmoreperspicuous. Startingwithanargumentina naturallanguage likeEnglish,wetranslatethe argumentintoa formallanguage .PartsoftheEnglishsentencesarereplaced withlettersandsymbols.Thegoalistorevealtheformalstructureofthe argument,aswedidwiththesetwo. Thereareformallanguagesthatworklikethesymbolizationwegaveforthese twoarguments.AlogiclikethiswasdevelopedbyAristotle,aphilosopherwho livedinGreeceduringthe4thcenturyBC.AristotlewasastudentofPlatoand thetutorofAlexandertheGreat.Aristotle'slogic,withsomerevisions,wasthe dominantlogicinthewesternworldformorethantwomillennia. InAristoteleanlogic,categoriesarereplacedwithcapitalletters.Everysentence ofanargumentisthenrepresentedashavingoneoffourforms,whichmedieval logicianslabeledinthisway:AAll A sare B s.ENo A sare B s.ISome A is B .OSome A isnot B Itisthenpossibletodescribevalid syllogisms ,three-lineargumentslikethe twoweconsideredabove.Medievallogiciansgavemnemonicnamestoallof thevalidargumentforms.Theformofourtwoarguments,forinstance,was called Barbara .Thevowelsinthename,allAs,representthefactthatthetwo premisesandtheconclusionareallAformsentences. TherearemanylimitationstoAristoteleanlogic.Oneisthatitmakesno distinctionbetweenkindsandindividuals.Sotherstpremisemightjustas wellbewritten`All S sare M s':AllSocratesesaremen.Despiteitshistorical importance,Aristoteleanlogichasbeensuperceded.Theremainderofthisbook willdeveloptwoformallanguages. TherstisSL,whichstandsfor sententiallogic .InSL,thesmallestunitsare sentencesthemselves.Simplesentencesarerepresentedaslettersandconnected withlogicalconnectiveslike`and'and`not'tomakemorecomplexsentences.

PAGE 14

14 forall x ThesecondisQL,whichstandsfor quantiedlogic .InQL,thebasicunitsare objects,propertiesofobjects,andrelationsbetweenobjects. Whenwetranslateanargumentintoaformallanguage,wehopetomakeits logicalstructureclearer.Wewanttoincludeenoughofthestructureofthe Englishlanguageargumentsothatwecanjudgewhethertheargumentisvalid orinvalid.IfweincludedeveryfeatureoftheEnglishlanguage,allofthe subtletyandnuance,thentherewouldbenoadvantageintranslatingtoa formallanguage.WemightaswellthinkabouttheargumentinEnglish. Atthesametime,wewouldlikeaformallanguagethatallowsustorepresent manykindsofEnglishlanguagearguments.ThisisonereasontopreferQLto Aristoteleanlogic;QLcanrepresenteveryvalidargumentofAristoteleanlogic andmore. Sowhendecidingonaformallanguage,thereisinevitablyatensionbetween wantingtocaptureasmuchstructureaspossibleandwantingasimpleformal language|simplerformallanguagesleaveoutmore.Thismeansthatthereis noperfectformallanguage.Somewilldoabetterjobthanothersintranslating particularEnglish-languagearguments. Inthisbook,wemaketheassumptionthat true and false aretheonlypossible truth-values.Logicallanguagesthatmakethisassumptionarecalled bivalent whichmeans two-valued .Aristoteleanlogic,SL,andQLareallbivalent,but therearelimitstothepowerofbivalentlogic.Forinstance,somephilosophers haveclaimedthatthefutureisnotyetdetermined.Iftheyareright,then sentencesabout whatwillbethecase arenotyettrueorfalse.Someformal languagesaccommodatethisbyallowingforsentencesthatareneithertruenor false,butsomethinginbetween.Otherformallanguages,so-calledparaconsistentlogics,allowforsentencesthatarebothtrue and false. Thelanguagespresentedinthisbookarenottheonlypossibleformallanguages. However,mostnonstandardlogicsextendonthebasicformalstructureofthe bivalentlogicsdiscussedinthisbook.Sothisisagoodplacetostart. Summaryoflogicalnotions Anargumentisdeductively valid ifitisimpossibleforthepremisesto betrueandtheconclusionfalse;itis invalid otherwise. A tautology isasentencethatmustbetrue,asamatteroflogic. A contradiction isasentencethatmustbefalse,asamatteroflogic. A contingentsentence isneitheratautologynoracontradiction.

PAGE 15

ch.1whatislogic? 15 Twosentencesare logicallyequivalent iftheynecessarilyhavethe sametruthvalue. Asetofsentencesis consistent ifitislogicallypossibleforallthemembersofthesettobetrueatthesametime;itis inconsistent otherwise. PracticeExercises Attheendofeachchapter,youwillndaseriesofpracticeproblemsthat reviewandexplorethematerialcoveredinthechapter.Thereisnosubstitute foractuallyworkingthroughsomeproblems,becauselogicismoreaboutaway ofthinkingthanitisaboutmemorizingfacts.Theanswerstosomeofthe problemsareprovidedattheendofthebookinappendixB;theproblemsthat aresolvedintheappendixaremarkedwitha ? PartA Whichofthefollowingare`sentences'inthelogicalsense? 1.EnglandissmallerthanChina. 2.GreenlandissouthofJerusalem. 3.IsNewJerseyeastofWisconsin? 4.Theatomicnumberofheliumis2. 5.Theatomicnumberofheliumis 6.Ihateovercookednoodles. 7.Blech!Overcookednoodles! 8.Overcookednoodlesaredisgusting. 9.Takeyourtime. 10.Thisisthelastquestion. PartB Foreachofthefollowing:Isitatautology,acontradiction,oracontingentsentence? 1.CaesarcrossedtheRubicon. 2.SomeoneoncecrossedtheRubicon. 3.NoonehasevercrossedtheRubicon. 4.IfCaesarcrossedtheRubicon,thensomeonehas. 5.EventhoughCaesarcrossedtheRubicon,noonehasevercrossedthe Rubicon. 6.IfanyonehasevercrossedtheRubicon,itwasCaesar. ? PartC LookbackatthesentencesG1{G4onp.12,andconsidereachofthe followingsetsofsentences.Whichareconsistent?Whichareinconsistent?

PAGE 16

16 forall x 1.G2,G3,andG4 2.G1,G3,andG4 3.G1,G2,andG4 4.G1,G2,andG3 ? PartD Whichofthefollowingispossible?Ifitispossible,giveanexample. Ifitisnotpossible,explainwhy. 1.Avalidargumentthathasonefalsepremiseandonetruepremise 2.Avalidargumentthathasafalseconclusion 3.Avalidargument,theconclusionofwhichisacontradiction 4.Aninvalidargument,theconclusionofwhichisatautology 5.Atautologythatiscontingent 6.Twologicallyequivalentsentences,bothofwhicharetautologies 7.Twologicallyequivalentsentences,oneofwhichisatautologyandoneof whichiscontingent 8.Twologicallyequivalentsentencesthattogetherareaninconsistentset 9.Aconsistentsetofsentencesthatcontainsacontradiction 10.Aninconsistentsetofsentencesthatcontainsatautology

PAGE 17

Chapter2 Sententiallogic ThischapterintroducesalogicallanguagecalledSL.Itisaversionof sentential logic ,becausethebasicunitsofthelanguagewillrepresententiresentences. 2.1Sentenceletters InSL,capitallettersareusedtorepresentbasicsentences.Consideredonlyasa symbolofSL,theletter A couldmeananysentence.Sowhentranslatingfrom EnglishintoSL,itisimportanttoprovidea symbolizationkey .Thekeyprovides anEnglishlanguagesentenceforeachsentenceletterusedinthesymbolization. Forexample,considerthisargument: Thereisanappleonthedesk. Ifthereisanappleonthedesk,thenJennymadeittoclass. : : Jennymadeittoclass. ThisisobviouslyavalidargumentinEnglish.Insymbolizingit,wewantto preservethestructureoftheargumentthatmakesitvalid.Whathappensif wereplaceeachsentencewithaletter?Oursymbolizationkeywouldlooklike this: A: Thereisanappleonthedesk. B: Ifthereisanappleonthedesk,thenJennymadeittoclass. C: Jennymadeittoclass. Wewouldthensymbolizetheargumentinthisway: 17

PAGE 18

18 forall x A B : :C Thereisnonecessaryconnectionbetweensomesentence A ,whichcouldbeany sentence,andsomeothersentences B and C ,whichcouldbeanysentences. Thestructureoftheargumenthasbeencompletelylostinthistranslation. Theimportantthingabouttheargumentisthatthesecondpremiseisnot merely any sentence,logicallydivorcedfromtheothersentencesintheargument.Thesecondpremisecontainstherstpremiseandtheconclusion asparts Oursymbolizationkeyfortheargumentonlyneedstoincludemeaningsfor A and C ,andwecanbuildthesecondpremisefromthosepieces.Sowesymbolize theargumentthisway: A If A ,then C : :C Thispreservesthestructureoftheargumentthatmakesitvalid,butitstill makesuseoftheEnglishexpression`If ::: then ::: .'Althoughweultimately wanttoreplacealloftheEnglishexpressionswithlogicalnotation,thisisa goodstart. Thesentencesthatcanbesymbolizedwithsentencelettersarecalled atomic sentences ,becausetheyarethebasicbuildingblocksoutofwhichmorecomplex sentencescanbebuilt.Whateverlogicalstructureasentencemighthaveislost whenitistranslatedasanatomicsentence.FromthepointofviewofSL,the sentenceisjustaletter.Itcanbeusedtobuildmorecomplexsentences,butit cannotbetakenapart. Thereareonlytwenty-sixlettersofthealphabet,butthereisnologicallimit tothenumberofatomicsentences.Wecanusethesamelettertosymbolize dierentatomicsentencesbyaddingasubscript,asmallnumberwrittenafter theletter.Wecouldhaveasymbolizationkeythatlookslikethis: A 1 : Theappleisunderthearmoire. A 2 : ArgumentsinSLalwayscontainatomicsentences. A 3 : AdamAntistakinganairplanefromAnchoragetoAlbany. . A 294 : Alliterationangersotherwiseaableastronauts. Keepinmindthateachoftheseisadierentsentenceletter.Whenthereare subscriptsinthesymbolizationkey,itisimportanttokeeptrackofthem.

PAGE 19

ch.2sententiallogic 19 2.2Connectives Logicalconnectivesareusedtobuildcomplexsentencesfromatomiccomponents.TherearevelogicalconnectivesinSL.Thistablesummarizesthem, andtheyareexplainedbelow. symbol whatitiscalled whatitmeans : negation `Itisnotthecasethat ::: & conjunction `Both ::: and ::: disjunction `Either ::: or ::: conditional `If ::: then ::: $ biconditional ` ::: ifandonlyif ::: Negation Considerhowwemightsymbolizethesesentences: 1.MaryisinBarcelona. 2.MaryisnotinBarcelona. 3.MaryissomewherebesidesBarcelona. Inordertosymbolizesentence1,wewillneedonesentenceletter.Wecan provideasymbolizationkey: B: MaryisinBarcelona. Notethatherewearegiving B adierentinterpretationthanwedidinthe previoussection.Thesymbolizationkeyonlyspecieswhat B means ina speciccontext .Itisvitalthatwecontinuetousethismeaningof B solong aswearetalkingaboutMaryandBarcelona.Later,whenwearesymbolizing dierentsentences,wecanwriteanewsymbolizationkeyanduse B tomean somethingelse. Now,sentence1issimply B Sincesentence2isobviouslyrelatedtothesentence1,wedonotwantto introduceadierentsentenceletter.ToputitpartlyinEnglish,thesentence means`Not B .'Inordertosymbolizethis,weneedasymbolforlogicalnegation. Wewilluse` : .'Nowwecantranslate`Not B 'to : B Sentence3isaboutwhetherornotMaryisinBarcelona,butitdoesnotcontain theword`not.'Nevertheless,itisobviouslylogicallyequivalenttosentence2.

PAGE 20

20 forall x Theybothmean:ItisnotthecasethatMaryisinBarcelona.Assuch,wecan translatebothsentence2andsentence3as : B Asentencecanbesymbolizedas : A ifitcanbeparaphrasedin Englishas`Itisnotthecasethat A .' Considerthesefurtherexamples: 4.Thewidgetcanbereplacedifitbreaks. 5.Thewidgetisirreplaceable. 6.Thewidgetisnotirreplaceable. Ifwelet R mean`Thewidgetisreplaceable',thensentence4canbetranslated as R Whataboutsentence5?Sayingthewidgetisirreplaceablemeansthatitis notthecasethatthewidgetisreplaceable.Soeventhoughsentence5isnot negativeinEnglish,wesymoblizeitusingnegationas : R Sentence6canbeparaphrasedas`Itisnotthecasethatthewidgetisirreplaceable.'Usingnegationtwice,wetranslatethisas :: R .Thetwonegationsina roweachworkasnegations,sothesentencemeans`Itisnotthecasethat ::: itisnotthecasethat :::R .'IfyouthinkaboutthesentenceinEnglish,itis logicallyequivalenttosentence4.SowhenwedenelogicalequivalenceisSL, wewillmakesurethat R and :: R arelogicallyequivalent. Moreexamples: 7.Elliottishappy. 8.Elliottisunhappy. Ifwelet H mean`Elliotishappy',thenwecansymbolizesentence7as H However,itwouldbeamistaketosymbolizesentence8as : H .IfElliottis unhappy,thenheisnothappy|butsentence8doesnotmeanthesamething as`ItisnotthecasethatElliottishappy.'Itcouldbethatheisnothappybut thatheisnotunhappyeither.Perhapsheissomewherebetweenthetwo.In ordertosymbolizesentence8,wewouldneedanewsentenceletter. Foranysentence A :If A istrue,then : A isfalse.If : A istrue,then A isfalse. Using`T'fortrueand`F'forfalse,wecansummarizethisina characteristic truthtable fornegation:

PAGE 21

ch.2sententiallogic 21 A : A T F F T Wewilldiscusstruthtablesatgreaterlengthinthenextchapter. Conjunction Considerthesesentences: 9.Adamisathletic. 10.Barbaraisathletic. 11.Adamisathletic,andBarbaraisalsoathletic. Wewillneedseparatesentencelettersfor9and10,sowedenethissymbolizationkey: A: Adamisathletic. B: Barbaraisathletic. Sentence9canbesymbolizedas A Sentence10canbesymbolizedas B Sentence11canbeparaphrasedas` A and B .'Inordertofullysymbolizethis sentence,weneedanothersymbol.Wewilluse`&.'Wetranslate` A and B as A & B .Thelogicalconnective`&'iscalled conjunction ,and A and B are eachcalled conjuncts Noticethatwemakenoattempttosymbolize`also'insentence11.Wordslike `both'and`also'functiontodrawourattentiontothefactthattwothingsare beingconjoined.Theyarenotdoinganyfurtherlogicalwork,sowedonotneed torepresenttheminSL. Somemoreexamples: 12.Barbaraisathleticandenergetic. 13.BarbaraandAdamarebothathletic. 14.AlthoughBarbaraisenergetic,sheisnotathletic. 15.Barbaraisathletic,butAdamismoreathleticthansheis. Sentence12isobviouslyaconjunction.Thesentencesaystwothingsabout Barbara,soinEnglishitispermissibletorefertoBarbaraonlyonce.Itmight

PAGE 22

22 forall x betemptingtotrythiswhentranslatingtheargument:Since B means`Barbara isathletic',onemightparaphrasethesentencesas` B andenergetic.'Thiswould beamistake.Oncewetranslatepartofasentenceas B ,anyfurtherstructureis lost. B isanatomicsentence;itisnothingmorethantrueorfalse.Conversely, `energetic'isnotasentence;onitsownitisneithertruenorfalse.Weshould insteadparaphrasethesentenceas` B andBarbaraisenergetic.'Nowweneed toaddasentencelettertothesymbolizationkey.Let E mean`Barbarais energetic.'Nowthesentencecanbetranslatedas B & E Asentencecanbesymbolizedas A & B ifitcanbeparaphrased inEnglishas`Both A ,and B .'Eachoftheconjunctsmustbea sentence. Sentence13saysonethingabouttwodierentsubjects.ItsaysofbothBarbara andAdamthattheyareathletic,andinEnglishweusetheword`athletic'only once.IntranslatingtoSL,itisimportanttorealizethatthesentencecanbe paraphrasedas,`Barbaraisathletic,andAdamisathletic.'Thistranslatesas B & A Sentence14isabitmorecomplicated.Theword`although'setsupacontrast betweentherstpartofthesentenceandthesecondpart.Nevertheless,the sentencesaysboththatBarbaraisenergeticandthatsheisnotathletic.In ordertomakeeachoftheconjunctsanatomicsentence,weneedtoreplace`she' with`Barbara.' Sowecanparaphrasesentence14as,` Both Barbaraisenergetic, and Barbara isnotathletic.'Thesecondconjunctcontainsanegation,soweparaphrasefurther:` Both Barbaraisenergetic anditisnotthecasethat Barbaraisathletic.' Thistranslatesas E & : B Sentence15containsasimilarcontrastivestructure.Itisirrelevantforthe purposeoftranslatingtoSL,sowecanparaphrasethesentenceas` Both Barbara isathletic, and AdamismoreathleticthanBarbara.'Noticethatweonceagain replacethepronoun`she'withhername.Howshouldwetranslatethesecond conjunct?Wealreadyhavethesentenceletter A whichisaboutAdam'sbeing athleticand B whichisaboutBarbara'sbeingathletic,butneitherisaboutone ofthembeingmoreathleticthantheother.Weneedanewsentenceletter.Let R mean`AdamismoreathleticthanBarbara.'Nowthesentencetranslatesas B & R Sentencesthatcanbeparaphrased` A ,but B 'or`Although A B arebestsymbolizedusingconjunction: A & B Itisimportanttokeepinmindthatthesentenceletters A B ,and R areatomic sentences.ConsideredassymbolsofSL,theyhavenomeaningbeyondbeing

PAGE 23

ch.2sententiallogic 23 trueorfalse.WehaveusedthemtosymbolizedierentEnglishlanguagesentencesthatareallaboutpeoplebeingathletic,butthissimilarityiscompletely lostwhenwetranslatetoSL.Noformallanguagecancaptureallthestructure oftheEnglishlanguage,butaslongasthisstructureisnotimportanttothe argumentthereisnothinglostbyleavingitout. Foranysentences A and B A & B istrueifandonlyifboth A and B aretrue. Wecansummarizethisinthecharacteristictruthtableforconjunction: A B A & B T T T T F F F T F F F F Conjunctionis symmetrical becausewecanswaptheconjunctswithoutchangingthetruth-valueofthesentence.Regardlessofwhat A and B are, A & B is logicallyequivalentto B & A Disjunction Considerthesesentences: 16.EitherDenisonwillplaygolfwithme,orhewillwatchmovies. 17.EitherDenisonorEllerywillplaygolfwithme. Forthesesentenceswecanusethissymbolizationkey: D: Denisonwillplaygolfwithme. E: Ellerywillplaygolfwithme. M: Denisonwillwatchmovies. Sentence16is`Either D or M .'Tofullysymbolizethis,weintroduceanewsymbol.Thesentencebecomes D M .The` 'connectiveiscalled disjunction and D and M arecalled disjuncts Sentence17isonlyslightlymorecomplicated.Therearetwosubjects,butthe Englishsentenceonlygivestheverbonce.Intranslating,wecanparaphrase itas.`EitherDenisonwillplaygolfwithme,orEllerywillplaygolfwithme.' Nowitobviouslytranslatesas D E .

PAGE 24

24 forall x Asentencecanbesymbolizedas A B ifitcanbeparaphrased inEnglishas`Either A ,or B .'Eachofthedisjunctsmustbea sentence. SometimesinEnglish,theword`or'excludesthepossibilitythatbothdisjuncts aretrue.Thisiscalledan exclusiveor .An exclusiveor isclearlyintended whenitsays,onarestaurantmenu,`Entreescomewitheithersouporsalad.' Youmayhavesoup;youmayhavesalad;but,ifyouwant both soup and salad, thenyouhavetopayextra. Atothertimes,theword`or'allowsforthepossibilitythatbothdisjunctsmight betrue.Thisisprobablythecasewithsentence17,above.Imightplaywith Denison,withEllery,orwithbothDenisonandEllery.Sentence17merelysays thatIwillplaywith atleast oneofthem.Thisiscalledan inclusiveor Thesymbol` 'representsan inclusiveor .So D E istrueif D istrue,if E istrue,orifboth D and E aretrue.Itisfalseonlyifboth D and E arefalse. Wecansummarizethiswiththecharacteristictruthtablefordisjunction: A B A B T T T T F T F T T F F F Likeconjunction,disjunctionissymmetrical. A B islogicallyequivalentto B A Thesesentencesaresomewhatmorecomplicated: 18.Eitheryouwillnothavesoup,oryouwillnothavesalad. 19.Youwillhaveneithersoupnorsalad. 20.Yougeteithersouporsalad,butnotboth. Welet S 1 meanthatyougetsoupand S 2 meanthatyougetsalad. Sentence18canbeparaphrasedinthisway:`Either itisnotthecasethat you getsoup,or itisnotthecasethat yougetsalad.'Translatingthisrequiresboth disjunctionandnegation.Itbecomes : S 1 _: S 2 Sentence19alsorequiresnegation.Itcanbeparaphrasedas,` Itisnotthecase that eitherthatyougetsouporthatyougetsalad.'Weneedsomewayof indicatingthatthenegationdoesnotjustnegatetherightorleftdisjunct,but rathernegatestheentiredisjunction.Inordertodothis,weputparentheses

PAGE 25

ch.2sententiallogic 25 aroundthedisjunction:`Itisnotthecasethat S 1 S 2 .'Thisbecomessimply : S 1 S 2 Noticethattheparenthesesaredoingimportantworkhere.Thesentence : S 1 S 2 wouldmean`Eitheryouwillnothavesoup,oryouwillhavesalad.' Sentence20isan exclusiveor .Wecanbreakthesentenceintotwoparts.The rstpartsaysthatyougetoneortheother.Wetranslatethisas S 1 S 2 Thesecondpartsaysthatyoudonotgetboth.Wecanparaphrasethisas, `Itisnotthecaseboththatyougetsoupandthatyougetsalad.'Usingboth negationandconjunction,wetranslatethisas : S 1 & S 2 .Nowwejustneedto putthetwopartstogether.Aswesawabove,`but'canusuallybetranslatedas aconjunction.Sentence20canthusbetranslatedas S 1 S 2 & : S 1 & S 2 Although` 'isan inclusiveor ,wecansymbolizean exclusiveor inSL.Wejust needmorethanoneconnectivetodoit. Conditional Forthefollowingsentences,let R mean`Youwillcuttheredwire'and B mean `Thebombwillexplode.' 21.Iftheyoucuttheredwire,thenthebombwillexplode. 22.Thebombwillexplodeonlyifyoucuttheredwire. Sentence21canbetranslatedpartiallyas`If R ,then B .'Wewillusethe symbol` 'torepresentlogicalentailment.Thesentencebecomes R B .The connectiveiscalleda conditional .Thesentenceontheleft-handsideofthe conditional R inthisexampleiscalledthe antecedent .Thesentenceonthe right-handside B iscalledthe consequent Sentence22isalsoaconditional.Sincetheword`if'appearsinthesecond halfofthesentence,itmightbetemptingtosymbolizethisinthesamewayas sentence21.Thatwouldbeamistake. Theconditional R B saysthat if R weretrue, then B wouldalsobetrue.It doesnotsaythatyourcuttingtheredwireisthe only waythatthebombcould explode.Someoneelsemightcutthewire,orthebombmightbeonatimer. Thesentence R B doesnotsayanythingaboutwhattoexpectif R isfalse. Sentence22isdierent.Itsaysthattheonlyconditionsunderwhichthebomb willexplodeinvolveyourhavingcuttheredwire;i.e.,ifthebombexplodes, thenyoumusthavecutthewire.Assuch,sentence22shouldbesymbolizedas B R .

PAGE 26

26 forall x Itisimportanttorememberthattheconnective` 'saysonlythat,ifthe antecedentistrue,thentheconsequentistrue.Itsaysnothingaboutthe causal connectionbetweenthetwoevents.Translatingsentence22as B R does notmeanthatthebombexplodingwouldsomehowhavecausedyourcutting thewire.Bothsentence21and22suggestthat,ifyoucuttheredwire,your cuttingtheredwirewouldbethecauseofthebombexploding.Theydieron the logical connection.Ifsentence22weretrue,thenanexplosionwouldtell us|thoseofussafelyawayfromthebomb|thatyouhadcuttheredwire. Withoutanexplosion,sentence22tellsusnothing. Theparaphrasedsentence` A onlyif B 'islogicallyequivalentto`If A ,then B .' `If A then B 'meansthatif A istruethensois B .Soweknowthatifthe antecedent A istruebuttheconsequent B isfalse,thentheconditional`If A then B 'isfalse.Whatisthetruthvalueof`If A then B 'underother circumstances?Suppose,forinstance,thattheantecedent A happenedtobe false.`If A then B 'wouldthennottellusanythingabouttheactualtruthvalue oftheconsequent B ,anditisunclearwhatthetruthvalueof`If A then B wouldbe. InEnglish,thetruthofconditionalsoftendependsonwhat would bethecase iftheantecedent weretrue |evenif,asamatteroffact,theantecedentis false.ThisposesaproblemfortranslatingconditionalsintoSL.Consideredas sentencesofSL, R and B intheaboveexampleshavenothingintrinsictodo witheachother.Inordertoconsiderwhattheworldwouldbelikeif R were true,wewouldneedtoanalyzewhat R saysabouttheworld.Since R isan atomicsymbolofSL,however,thereisnofurtherstructuretobeanalyzed. Whenwereplaceasentencewithasentenceletter,weconsideritmerelyas someatomicsentencethatmightbetrueorfalse. InordertotranslateconditionalsintoSL,wewillnottrytocaptureallthe subtletiesoftheEnglishlanguage`If ::: then ::: .'Instead,thesymbol` 'will bea materialconditional .Thismeansthatwhen A isfalse,theconditional A B isautomaticallytrue,regardlessofthetruthvalueof B .Ifboth A and B aretrue,thentheconditional A B istrue. Inshort, A B isfalseifandonlyif A istrueand B isfalse.Wecansummarize thiswithacharacteristictruthtablefortheconditional. A B A B T T T T F F F T T F F T

PAGE 27

ch.2sententiallogic 27 Theconditionalis asymmetrical .Youcannotswaptheantecedentandconsequentwithoutchangingthemeaningofthesentence,because A B and B A arenotlogicallyequivalent. Notallsentencesoftheform`If ::: then ::: 'areconditionals.Considerthis sentence: 23.Ifanyonewantstoseeme,thenIwillbeontheporch. IfIsaythis,itmeansthatIwillbeontheporch,regardlessofwhetheranyone wantstoseemeornot|butifsomeonedidwanttoseeme,thentheyshould lookformethere.Ifwelet P mean`Iwillbeontheporch,'thensentence23 canbetranslatedsimplyas P Biconditional Considerthesesentences: 24.Thegureontheboardisatriangleonlyifithasexactlythreesides. 25.Thegureontheboardisatriangleifithasexactlythreesides. 26.Thegureontheboardisatriangleifandonlyifithasexactlythree sides. Let T mean`Thegureisatriangle'and S mean`Thegurehasthreesides.' Sentence24,forreasonsdiscussedabove,canbetranslatedas T S Sentence25isimportantlydierent.Itcanbeparaphrasedas,`Ifthegurehas threesides,thenitisatriangle.'Soitcanbetranslatedas S T Sentence26saysthat T istrue ifandonlyif S istrue;wecaninfer S from T andwecaninfer T from S .Thisiscalleda biconditional ,becauseitentails thetwoconditionals S T and T S .Wewilluse` $ 'torepresentthe biconditional;sentence26canbetranslatedas S $ T Wecouldabidewithoutanewsymbolforthebiconditional.Sincesentence26 means` T S and S T ,'wecouldtranslateitas T S & S T .We wouldneedparenthesestoindicatethat T S and S T areseparate conjuncts;theexpression T S & S T wouldbeambiguous. Becausewecouldalwayswrite A B & B A insteadof A $ B ,we donotstrictlyspeaking need tointroduceanewsymbolforthebiconditional. Nevertheless,logicallanguagesusuallyhavesuchasymbol.SLwillhaveone, whichmakesiteasiertotranslatephraseslike`ifandonlyif.'

PAGE 28

28 forall x A $ B istrueifandonlyif A and B havethesametruthvalue.Thisisthe characteristictruthtableforthebiconditional: A B A $ B T T T T F F F T F F F T 2.3Othersymbolization WehavenowintroducedalloftheconnectivesofSL.Wecanusethemtogether totranslatemanykindsofsentences.Considertheseexamplesofsentencesthat usetheEnglish-languageconnective`unless': 27.Unlessyouwearajacket,youwillcatchcold. 28.Youwillcatchcoldunlessyouwearajacket. Let J mean`Youwillwearajacket'andlet D mean`Youwillcatchacold.' Wecanparaphrasesentence27as`Unless J D .'Thismeansthatifyoudonot wearajacket,thenyouwillcatchcold;withthisinmind,wemighttranslateit as : J D .Italsomeansthatifyoudonotcatchacold,thenyoumusthave wornajacket;withthisinmind,wemighttranslateitas : D J Whichoftheseisthecorrecttranslationofsentence27?Bothtranslationsare correct,becausethetwotranslationsarelogicallyequivalentinSL. Sentence28,inEnglish,islogicallyequivalenttosentence27.Itcanbetranslatedaseither : J D or : D J Whensymbolizingsentenceslikesentence27andsentence28,itiseasytoget turnedaround.Sincetheconditionalisnotsymmetric,itwouldbewrongto translateeithersentenceas J !: D .Fortunately,thereareotherlogically equivalentexpressions.Bothsentencesmeanthatyouwillwearajacketor| ifyoudonotwearajacket|thenyouwillcatchacold.Sowecantranslate themas J D .Youmightworrythatthe`or'hereshouldbean exclusiveor However,thesentencesdonotexcludethepossibilitythatyoumight both wear ajacket and catchacold;jacketsdonotprotectyoufromallthepossibleways thatyoumightcatchacold. Ifasentencecanbeparaphrasedas`Unless A B ,'thenitcanbe symbolizedas A B

PAGE 29

ch.2sententiallogic 29 Symbolizationofstandardsentencetypesissummarizedonp.156. 2.4SentencesofSL Thesentence`Applesarered,orberriesareblue'isasentenceofEnglish,and thesentence` A B 'isasentenceofSL.Althoughwecanidentifysentencesof Englishwhenweencounterthem,wedonothaveaformaldenitionof`sentence ofEnglish'.InSL,itispossibletoformallydenewhatcountsasasentence. ThisisonerespectinwhichaformallanguagelikeSLismoreprecisethana naturallanguagelikeEnglish. ItisimportanttodistinguishbetweenthelogicallanguageSL,whichweare developing,andthelanguagethatweusetotalkaboutSL.Whenwetalk aboutalanguage,thelanguagethatwearetalkingaboutiscalledthe object language .Thelanguagethatweusetotalkabouttheobjectlanguageis calledthe metalanguage TheobjectlanguageinthischapterisSL.ThemetalanguageisEnglish|not conversationalEnglish,butEnglishsupplementedwithsomelogicalandmathematicalvocabulary.Thesentence` A B 'isasentenceintheobjectlanguage, becauseitusesonlysymbolsofSL.Theword`sentence'isnotitselfpartofSL, however,sothesentence`ThisexpressionisasentenceofSL'isnotasentence ofSL.Itisasentenceinthemetalanguage,asentencethatweusetotalk about SL. Inthissection,wewillgiveaformaldenitionfor`sentenceofSL.'Thedenition itselfwillbegiveninmathematicalEnglish,themetalanguage. Expressions TherearethreekindsofsymbolsinSL: sentenceletters A;B;C;:::;Z withsubscripts,asneeded A 1 ;B 1 ;Z 1 ;A 2 ;A 25 ;J 375 ;::: connectives : ,&, $ parentheses Wedenean expressionofsl asanystringofsymbolsofSL.Takeanyofthe symbolsofSLandwritethemdown,inanyorder,andyouhaveanexpression.

PAGE 30

30 forall x Well-formedformulae Sinceanysequenceofsymbolsisanexpression,manyexpressionsofSLwillbe gobbledegook.Ameaningfulexpressioniscalleda well-formedformula .Itis commontousetheacronym w ;thepluralisws. Obviously,individualsentenceletterslike A and G 13 willbews.Wecan formfurtherwsoutofthesebyusingthevariousconnectives.Usingnegation, wecanget : A and : G 13 .Usingconjunction,wecanget A & G 13 G 13 & A A & A ,and G 13 & G 13 .Wecouldalsoapplynegationrepeatedlytogetwslike :: A orapplynegationalongwithconjunctiontogetwslike : A & G 13 and : G 13 & : G 13 .Thepossiblecombinationsareendless,evenstartingwithjust thesetwosentenceletters,andthereareinnitelymanysentenceletters.So thereisnopointintryingtolistallthews. Instead,wewilldescribetheprocessbywhichwscanbeconstructed.Consider negation:Givenanyw A ofSL, : A isawofSL.Itisimportantherethat A isnotthesentenceletter A .Rather,itisavariablethatstandsinforany watall.Noticethatthisvariable A isnotasymbolofSL,so : A isnotan expressionofSL.Instead,itisanexpressionofthemetalanguagethatallowsus totalkaboutinnitelymanyexpressionsofSL:alloftheexpressionsthatstart withthenegationsymbol.Because A ispartofthemetalanguage,itiscalleda metavariable Wecansaysimilarthingsforeachoftheotherconnectives.Forinstance,if A and B arewsofSL,then A & B isawofSL.Providingclauseslike thisforalloftheconnectives,wearriveatthefollowingformaldenitionfora well-formedformulaofSL: 1.Everyatomicsentenceisaw. 2.If A isaw,then : A isawofSL. 3.If A and B arews,then A & B isaw. 4.If A and B arews,then A B isaw. 5.If A and B arews,then A B isaw. 6.If A and B arews,then A $ B isaw. 7.AllandonlywsofSLcanbegeneratedbyapplicationsoftheserules. Noticethatwecannotimmediatelyapplythisdenitiontoseewhetheranarbitraryexpressionisaw.Supposewewanttoknowwhetherornot ::: D isawofSL.Lookingatthesecondclauseofthedenition,weknowthat ::: D isaw if :: D isaw.Sonowweneedtoaskwhetherornot :: D

PAGE 31

ch.2sententiallogic 31 isaw.Againlookingatthesecondclauseofthedenition, :: D isaw if : D is.Again, : D isaw if D isaw.Now D isasentenceletter,anatomic sentenceofSL,soweknowthat D isawbytherstclauseofthedenition. Soforacompoundformulalike ::: D ,wemustapplythedenitionrepeatedly. Eventuallywearriveattheatomicsentencesfromwhichthewisbuiltup. Denitionslikethisarecalled recursive .Recursivedenitionsbeginwithsome speciablebaseelementsanddenewaystoindenitelycompoundthebase elements.Justastherecursivedenitionallowscomplexsentencestobebuilt upfromsimpleparts,youcanuseittodecomposesentencesintotheirsimpler parts.Todeterminewhetherornotsomethingmeetsthedenition,youmay havetoreferbacktothedenitionmanytimes. Theconnectivethatyoulooktorstindecomposingasentenceiscalledthe mainlogicaloperator ofthatsentence.Forexample:Themainlogical operatorof : E F G isnegation, : .Themainlogicaloperatorof : E F G isdisjunction, Sentences Recallthatasentenceisameaningfulexpressionthatcanbetrueorfalse.Since themeaningfulexpressionsofSLarethewsandsinceeverywofSLiseither trueorfalse,thedenitionforasentenceofSListhesameasthedenitionfor aw.Noteveryformallanguagewillhavethisnicefeature.Inthelanguage QL,whichisdevelopedlaterinthebook,therearewswhicharenotsentences. TherecursivestructureofsentencesinSLwillbeimportantwhenweconsider thecircumstancesunderwhichaparticularsentencewouldbetrueorfalse. Thesentence ::: D istrueifandonlyifthesentence :: D isfalse,andsoon throughthestructureofthesentenceuntilwearriveattheatomiccomponents: ::: D istrueifandonlyiftheatomicsentence D isfalse.Wewillreturnto thispointinthenextchapter. Notationalconventions Awlike Q & R mustbesurroundedbyparentheses,becausewemightapply thedenitionagaintousethisaspartofamorecomplicatedsentence.Ifwe negate Q & R ,weget : Q & R .Ifwejusthad Q & R withouttheparentheses andputanegationinfrontofit,wewouldhave : Q & R .Itismostnatural toreadthisasmeaningthesamethingas : Q & R ,somethingverydierent than : Q & R .Thesentence : Q & R meansthatitisnotthecasethatboth Q and R aretrue; Q mightbefalseor R mightbefalse,butthesentencedoes nottelluswhich.Thesentence : Q & R meansspecicallythat Q isfalseand

PAGE 32

32 forall x that R istrue.Assuch,parenthesesarecrucialtothemeaningofthesentence. So,strictlyspeaking, Q & R withoutparenthesesis not asentenceofSL.When usingSL,however,wewilloftenbeabletorelaxtheprecisedenitionsoasto makethingseasierforourselves.Wewilldothisinseveralways. First,weunderstandthat Q & R meansthesamethingas Q & R .Asamatter ofconvention,wecanleaveoparenthesesthatoccur aroundtheentiresentence Second,itcansometimesbeconfusingtolookatlongsentenceswithmany, nestedpairsofparentheses.Weadopttheconventionofusingsquarebrackets `['and`]'inplaceofparenthesis.Thereisnologicaldierencebetween P Q and[ P Q ],forexample.Theunwieldysentence H I I H & J K couldbewritteninthisway: H I I H & J K Third,wewillsometimeswanttotranslatetheconjunctionofthreeormore sentences.Forthesentence`Alice,Bob,andCandiceallwenttotheparty', supposewelet A mean`Alicewent', B mean`Bobwent',and C mean`Candice went.'Thedenitiononlyallowsustoformaconjunctionoutoftwosentences, sowecantranslateitas A & B & C oras A & B & C .Thereisnoreason todistinguishbetweenthese,sincethetwotranslationsarelogicallyequivalent. Thereisnologicaldierencebetweentherst,inwhich A & B isconjoined with C ,andthesecond,inwhich A isconjoinedwith B & C .Sowemight aswelljustwrite A & B & C .Asamatterofconvention,wecanleaveout parentheseswhenweconjointhreeormoresentences. Fourth,asimilarsituationariseswithmultipledisjunctions.`EitherAlice,Bob, orCandicewenttotheparty'canbetranslatedas A B C oras A B C Sincethesetwotranslationsarelogicallyequivalent,wemaywrite A B C Theselattertwoconventionsonlyapplytomultipleconjunctionsormultipledisjunctions.Ifaseriesofconnectivesincludesbothdisjunctionsandconjunctions, thentheparenthesesareessential;aswith A & B C and A & B C .The parenthesesarealsorequiredifthereisaseriesofconditionalsorbiconditionals; aswith A B C and A $ B $ C Wehaveadoptedthesefourrulesas notationalconventions ,notaschangesto thedenitionofasentence.Strictlyspeaking, A B C isstillnotasentence. Instead,itisakindofshorthand.Wewriteitforthesakeofconvenience,but wereallymeanthesentence A B C Ifwehadgivenadierentdenitionforaw,thenthesecouldcountasws. Wemighthavewrittenrule3inthisway:If A B ::: Z arews,then

PAGE 33

ch.2sententiallogic 33 A & B & ::: & Z ,isaw."ThiswouldmakeiteasiertotranslatesomeEnglishsentences,butwouldhavethecostofmakingourformallanguagemore complicated.Wewouldhavetokeepthecomplexdenitioninmindwhenwe developtruthtablesandaproofsystem.Wewantalogicallanguagethatis expressivelysimple andallowsustotranslateeasilyfromEnglish,butwealsowant a formallysimple language.Adoptingnotationalconventionsisacompromise betweenthesetwodesires. PracticeExercises ? PartA Usingthesymbolizationkeygiven,translateeachEnglish-language sentenceintoSL. M: Thosecreaturesaremeninsuits. C: Thosecreaturesarechimpanzees. G: Thosecreaturesaregorillas. 1.Thosecreaturesarenotmeninsuits. 2.Thosecreaturesaremeninsuits,ortheyarenot. 3.Thosecreaturesareeithergorillasorchimpanzees. 4.Thosecreaturesareneithergorillasnorchimpanzees. 5.Ifthosecreaturesarechimpanzees,thentheyareneithergorillasnormen insuits. 6.Unlessthosecreaturesaremeninsuits,theyareeitherchimpanzeesor theyaregorillas. PartB Usingthesymbolizationkeygiven,translateeachEnglish-language sentenceintoSL. A: MisterAcewasmurdered. B: Thebutlerdidit. C: Thecookdidit. D: TheDuchessislying. E: MisterEdgewasmurdered. F: Themurderweaponwasafryingpan. 1.EitherMisterAceorMisterEdgewasmurdered. 2.IfMisterAcewasmurdered,thenthecookdidit. 3.IfMisterEdgewasmurdered,thenthecookdidnotdoit. 4.Eitherthebutlerdidit,ortheDuchessislying. 5.ThecookdiditonlyiftheDuchessislying.

PAGE 34

34 forall x 6.Ifthemurderweaponwasafryingpan,thentheculpritmusthavebeen thecook. 7.Ifthemurderweaponwasnotafryingpan,thentheculpritwaseither thecookorthebutler. 8.MisterAcewasmurderedifandonlyifMisterEdgewasnotmurdered. 9.TheDuchessislying,unlessitwasMisterEdgewhowasmurdered. 10.IfMisterAcewasmurdered,hewasdoneinwithafryingpan. 11.Thecookdidit,sothebutlerdidnot. 12.OfcoursetheDuchessislying! ? PartC Usingthesymbolizationkeygiven,translateeachEnglish-language sentenceintoSL. E 1 : Avaisanelectrician. E 2 : Harrisonisanelectrician. F 1 : Avaisareghter. F 2 : Harrisonisareghter. S 1 : Avaissatisedwithhercareer. S 2 : Harrisonissatisedwithhiscareer. 1.AvaandHarrisonarebothelectricians. 2.IfAvaisareghter,thensheissatisedwithhercareer. 3.Avaisareghter,unlesssheisanelectrician. 4.Harrisonisanunsatisedelectrician. 5.NeitherAvanorHarrisonisanelectrician. 6.BothAvaandHarrisonareelectricians,butneitherofthemnditsatisfying. 7.Harrisonissatisedonlyifheisareghter. 8.IfAvaisnotanelectrician,thenneitherisHarrison,butifsheis,thenhe istoo. 9.AvaissatisedwithhercareerifandonlyifHarrisonisnotsatisedwith his. 10.IfHarrisonisbothanelectricianandareghter,thenhemustbesatised withhiswork. 11.ItcannotbethatHarrisonisbothanelectricianandareghter. 12.HarrisonandAvaarebothreghtersifandonlyifneitherofthemisan electrician. ? PartD Giveasymbolizationkeyandsymbolizethefollowingsentencesin SL. 1.AliceandBobarebothspies. 2.IfeitherAliceorBobisaspy,thenthecodehasbeenbroken. 3.IfneitherAlicenorBobisaspy,thenthecoderemainsunbroken.

PAGE 35

ch.2sententiallogic 35 4.TheGermanembassywillbeinanuproar,unlesssomeonehasbrokenthe code. 5.Eitherthecodehasbeenbrokenorithasnot,buttheGermanembassy willbeinanuproarregardless. 6.EitherAliceorBobisaspy,butnotboth. PartE GiveasymbolizationkeyandsymbolizethefollowingsentencesinSL. 1.IfGregorplaysrstbase,thentheteamwilllose. 2.Theteamwillloseunlessthereisamiracle. 3.Theteamwilleitherloseoritwon't,butGregorwillplayrstbaseregardless. 4.Gregor'smomwillbakecookiesifandonlyifGregorplaysrstbase. 5.Ifthereisamiracle,thenGregor'smomwillnotbakecookies. PartF Foreachargument,writeasymbolizationkeyandtranslatetheargumentaswellaspossibleintoSL. 1.IfDorothyplaysthepianointhemorning,thenRogerwakesupcranky. Dorothyplayspianointhemorningunlesssheisdistracted.SoifRoger doesnotwakeupcranky,thenDorothymustbedistracted. 2.ItwilleitherrainorsnowonTuesday.Ifitrains,Nevillewillbesad.If itsnows,Nevillewillbecold.Therefore,Nevillewilleitherbesadorcold onTuesday. 3.IfZoogrememberedtodohischores,thenthingsarecleanbutnotneat. Ifheforgot,thenthingsareneatbutnotclean.Therefore,thingsare eitherneatorclean|butnotboth. ? PartG Foreachofthefollowing:aIsitawofSL?bIsitasentenceof SL,allowingfornotationalconventions? 1. A 2. J 374 _: J 374 3. :::: F 4. : & S 5. G & : G 6. A A 7. A A & : F D $ E 8.[ Z $ S W ]&[ J X ] 9. F $: D J C & D PartH

PAGE 36

36 forall x 1.ArethereanywsofSLthatcontainnosentenceletters?Whyorwhy not? 2.Inthechapter,wesymbolizedan exclusiveor using ,&,and : .How couldyoutranslatean exclusiveor usingonlytwoconnectives?Isthere anywaytotranslatean exclusiveor usingonlyoneconnective?

PAGE 37

Chapter3 Truthtables ThischapterintroducesawayofevaluatingsentencesandargumentsofSL. Althoughitcanbelaborious,thetruthtablemethodisapurelymechanical procedurethatrequiresnointuitionorspecialinsight. 3.1Truth-functionalconnnectives Anynon-atomicsentenceofSLiscomposedofatomicsentenceswithsentential connectives.Thetruth-valueofthecompoundsentencedependsonlyonthe truth-valueoftheatomicsentencesthatcompriseit.Inordertoknowthe truth-valueof D $ E ,forinstance,youonlyneedtoknowthetruth-value of D andthetruth-valueof E .Connectivesthatworkinthiswayarecalled truth-functional Inthischapter,wewillmakeuseofthefactthatallofthelogicaloperators inSLaretruth-functional|itmakesitpossibletoconstructtruthtablesto determinethelogicalfeaturesofsentences.Youshouldrealize,however,that thisisnotpossibleforalllanguages.InEnglish,itispossibletoformanew sentencefromanysimplersentence X bysaying`Itispossiblethat X .'The truth-valueofthisnewsentencedoesnotdependdirectlyonthetruth-valueof X .Evenif X isfalse,perhapsinsomesense X could havebeentrue|thenthe newsentencewouldbetrue.Someformallanguages,called modallogics ,have anoperatorforpossibility.Inamodallogic,wecouldtranslate`Itispossible that X 'as X .However,theabilitytotranslatesentenceslikethesecomeat acost:The operatorisnottruth-functional,andsomodallogicsarenot amenabletotruthtables. 37

PAGE 38

38 forall x A : A T F F T A B A & B A B A B A $ B T T T T T T T F F T F F F T F T T F F F F F T T Table3.1:ThecharacteristictruthtablesfortheconnectivesofSL. 3.2Completetruthtables Thetruth-valueofsentencesthatcontainonlyoneconnectiveisgivenbythe characteristictruthtableforthatconnective.Toputthemallinoneplace,the truthtablesfortheconnectivesofSLarerepeatedintable3.1. Thecharacteristictruthtableforconjunction,forexample,givesthetruthconditionsforanysentenceoftheform A & B .Eveniftheconjuncts A and B are long,complicatedsentences,theconjunctionistrueifandonlyifboth A and B aretrue.Considerthesentence H & I H .Weconsiderallthepossible combinationsoftrueandfalsefor H and I ,whichgivesusfourrows.Wethen copythetruth-valuesforthesentencelettersandwritethemunderneaththe lettersinthesentence. H I H & I H T T TTT T F TFT F T FTF F F FFF Nowconsiderthesubsentence H & I .Thisisaconjunction A & B with H as A andwith I as B H and I arebothtrueontherstrow.Sinceaconjunction istruewhenbothconjunctsaretrue,wewriteaTunderneaththeconjunction symbol.Wecontinuefortheotherthreerowsandgetthis: H I H & I H A & B T T T T TT T F T F FT F T F F TF F F F F FF Theentiresentenceisaconditional A B with H & I as A andwith H as B .Onthesecondrow,forexample, H & I isfalseand H istrue.Sincea conditionalistruewhentheantecedentisfalse,wewriteaTinthesecondrow

PAGE 39

ch.3truthtables 39 underneaththeconditionalsymbol.Wecontinuefortheotherthreerowsand getthis: H I H & I H A B T T T T T T F F T T F T F T F F F F T F ThecolumnofTsunderneaththeconditionaltellsusthatthesentence H & I I istrueregardlessofthetruth-valuesof H and I .Theycanbetrueorfalsein anycombination,andthecompoundsentencestillcomesouttrue.Itiscrucial thatwehaveconsideredallofthepossiblecombinations.Ifweonlyhadatwolinetruthtable,wecouldnotbesurethatthesentencewasnotfalseforsome othercombinationoftruth-values. Inthisexample,wehavenotrepeatedalloftheentriesineverysuccessivetable. Whenactuallywritingtruthtablesonpaper,however,itisimpracticaltoerase wholecolumnsorrewritethewholetableforeverystep.Althoughitismore crowded,thetruthtablecanbewritteninthisway: H I H & I H T T TTTTT T F TFFTT F T FFTTF F F FFFTF Mostofthecolumnsunderneaththesentenceareonlythereforbookkeeping purposes.Whenyoubecomemoreadeptwithtruthtables,youwillprobably nolongerneedtocopyoverthecolumnsforeachofthesentenceletters.Inany case,thetruth-valueofthesentenceoneachrowisjustthecolumnunderneath themainlogicaloperatorofthesentence;inthiscase,thecolumnunderneath theconditional. A completetruthtable hasarowforallthepossiblecombinationsofTand Fforallofthesentenceletters.Thesizeofthecompletetruthtabledependson thenumberofdierentsentencelettersinthetable.Asentencethatcontains onlyonesentenceletterrequiresonlytworows,asinthecharacteristictruth tablefornegation.Thisistrueevenifthesameletterisrepeatedmanytimes, asinthesentence[ C $ C C ]& : C C .Thecompletetruthtable requiresonlytwolinesbecausethereareonlytwopossibilities: C canbetrue oritcanbefalse.AsinglesentencelettercanneverbemarkedbothTandF onthesamerow.Thetruthtableforthissentencelookslikethis:

PAGE 40

40 forall x C [ C $ C C ]& : C C T TTTTT F FTTT F FTFFF F FFTF Lookingatthecolumnunderneaththemainconnective,weseethatthesentence isfalseonbothrowsofthetable;i.e.,itisfalseregardlessofwhether C istrue orfalse. Asentencethatcontainstwosentencelettersrequiresfourlinesforacomplete truthtable,asinthecharacteristictruthtablesandthetablefor H & I I Asentencethatcontainsthreesentencelettersrequireseightlines.Forexample: M N P M & N P T T T T T TTT T T F T T TTF T F T T T FTT T F F T F FFF F T T F F TTT F T F F F TTF F F T F F FTT F F F F F FFF Fromthistable,weknowthatthesentence M & N P mightbetrueorfalse, dependingonthetruth-valuesof M N ,and P Acompletetruthtableforasentencethatcontainsfourdierentsentenceletters requires16lines.Fiveletters,32lines.Sixletters,64lines.Andsoon.Tobe perfectlygeneral:Ifacompletetruthtablehas n dierentsentenceletters,then itmusthave2 n rows. Inordertollinthecolumnsofacompletetruthtable,beginwiththerightmostsentenceletterandalternateTsandFs.Inthenextcolumntotheleft, writetwoTs,writetwoFs,andrepeat.Forthethirdsentenceletter,writefour TsfollowedbyfourFs.Thisyieldsaneightlinetruthtableliketheoneabove. Fora16linetruthtable,thenextcolumnofsentencelettersshouldhaveeight TsfollowedbyeightFs.Fora32linetable,thenextcolumnwouldhave16Ts followedby16Fs.Andsoon.

PAGE 41

ch.3truthtables 41 3.3Usingtruthtables Tautologies,contradictions,andcontingentsentences RecallthatanEnglishsentenceisatautologyifitmustbetrueasamatterof logic.Withacompletetruthtable,weconsiderallofthewaysthattheworld mightbe.Ifthesentenceistrueoneverylineofacompletetruthtable,thenit istrueasamatteroflogic,regardlessofwhattheworldislike. Soasentenceisa tautologyinsl ifthecolumnunderitsmainconnectiveis Toneveryrowofacompletetruthtable. Conversely,asentenceisa contradictioninsl ifthecolumnunderitsmain connectiveisFoneveryrowofacompletetruthtable. Asentenceis contingentinsl ifitisneitheratautologynoracontradiction; i.e.ifitisTonatleastonerowandFonatleastonerow. Fromthetruthtablesintheprevioussection,weknowthat H & I H is atautology,that[ C $ C C ]& : C C isacontradiction,andthat M & N P iscontingent. Logicalequivalence TwosentencesarelogicallyequivalentinEnglishiftheyhavethesametruth valueasamatterlogic.Onceagain,truthtablesallowustodeneananalogous conceptforSL:Twosentencesare logicallyequivalentinsl iftheyhave thesametruth-valueoneveryrowofacompletetruthtable. Considerthesentences : A B and : A & : B .Aretheylogicallyequivalent? Tondout,weconstructatruthtable. A B : A B : A & : B T T F TTT FT F FT T F F TTF FT F TF F T F FTT TF F FT F F T FFF TF T TF Lookatthecolumnsforthemainconnectives;negationfortherstsentence, conjunctionforthesecond.Ontherstthreerows,bothareF.Onthenal row,bothareT.Sincetheymatchoneveryrow,thetwosentencesarelogically equivalent.

PAGE 42

42 forall x Consistency AsetofsentencesinEnglishisconsistentifitislogicallypossibleforthemall tobetrueatonce.Asetofsentencesis logicallyconsistentinsl ifthere isatleastonelineofacompletetruthtableonwhichallofthesentencesare true.Itis inconsistent otherwise. Validity AnargumentinEnglishisvalidifitislogicallyimpossibleforthepremisesto betrueandfortheconclusiontobefalseatthesametime.Anargumentis validinsl ifthereisnorowofacompletetruthtableonwhichthepremises areallTandtheconclusionisF;anargumentis invalidinsl ifthereissuch arow. Considerthisargument: : L J L : L : :J Isitvalid?Tondout,weconstructatruthtable. J L : L J L : L J T T FT T TTT F T T T F TF T TTF T F T F T FT T FTT F T F F F TF F FFF T F F Yes,theargumentisvalid.TheonlyrowonwhichboththepremisesareTis thesecondrow,andonthatrowtheconclusionisalsoT. 3.4Partialtruthtables Inordertoshowthatasentenceisatautology,weneedtoshowthatitisTon everyrow.Soweneedacompletetruthtable.Toshowthatasentenceis not atautology,however,weonlyneedoneline:alineonwhichthesentenceisF. Therefore,inordertoshowthatsomethingisnotatautology,itisenoughto provideaone-line partialtruthtable |regardlessofhowmanysentenceletters thesentencemighthaveinit.

PAGE 43

ch.3truthtables 43 Consider,forexample,thesentence U & T S & W .Wewanttoshowthat itis not atautologybyprovidingapartialtruthtable.WellinFfortheentire sentence.Themainconnectiveofthesentenceisaconditional.Inorderforthe conditionaltobefalse,theantecedentmustbetrueTandtheconsequent mustbefalseF.Sowelltheseinonthetable: S T U W U & T S & W T F F Inorderforthe U & T tobetrue,both U and T mustbetrue. S T U W U & T S & W T T TTT F F Nowwejustneedtomake S & W false.Todothis,weneedtomakeatleast oneof S and W false.Wecanmakeboth S and W falseifwewant.All thatmattersisthatthewholesentenceturnsoutfalseonthisline.Makingan arbitrarydecision,wenishthetableinthisway: S T U W U & T S & W F T T F TTT F FFF Showingthatsomethingisacontradictionrequiresacompletetruthtable. Showingthatsomethingis not acontradictionrequiresonlyaone-linepartial truthtable,wherethesentenceistrueonthatoneline. Asentenceiscontingentifitisneitheratautologynoracontradiction.So showingthatasentenceiscontingentrequiresa two-line partialtruthtable: Thesentencemustbetrueononelineandfalseontheother.Forexample,we canshowthatthesentenceaboveiscontingentwiththistruthtable: S T U W U & T S & W F T T F TTT F FFF F T F F FFT T FFF Notethattherearemanycombinationsoftruthvaluesthatwouldhavemade thesentencetrue,sotherearemanywayswecouldhavewrittenthesecond line. Showingthatasentenceis not contingentrequiresprovidingacompletetruth table,becauseitrequiresshowingthatthesentenceisatautologyorthatitisa contradiction.Ifyoudonotknowwhetheraparticularsentenceiscontingent, thenyoudonotknowwhetheryouwillneedacompleteorpartialtruthtable.

PAGE 44

44 forall x YES NO tautology? completetruthtable one-linepartialtruthtable contradiction? completetruthtable one-linepartialtruthtable contingent? two-linepartialtruthtable completetruthtable equivalent? completetruthtable one-linepartialtruthtable consistent? one-linepartialtruthtable completetruthtable valid? completetruthtable one-linepartialtruthtable Table3.2:Doyouneedacompletetruthtableorapartialtruthtable?It dependsonwhatyouaretryingtoshow. Youcanalwaysstartworkingonacompletetruthtable.Ifyoucompleterows thatshowthesentenceiscontingent,thenyoucanstop.Ifnot,thencompletethetruthtable.Eventhoughtwocarefullyselectedrowswillshowthat acontingentsentenceiscontingent,thereisnothingwrongwithllinginmore rows. Showingthattwosentencesarelogicallyequivalentrequiresprovidingacompletetruthtable.Showingthattwosentencesare not logicallyequivalentrequiresonlyaone-linepartialtruthtable:Makethetablesothatonesentence istrueandtheotherfalse. Showingthatasetofsentencesisconsistentrequiresprovidingonerowofatruth tableonwhichallofthesentencesaretrue.Therestofthetableisirrelevant, soaone-linepartialtruthtablewilldo.Showingthatasetofsentencesis inconsistent,ontheotherhand,requiresacompletetruthtable:Youmust showthatoneveryrowofthetableatleastoneofthesentencesisfalse. Showingthatanargumentisvalidrequiresacompletetruthtable.Showing thatanargumentis invalid onlyrequiresprovidingaone-linetruthtable:If youcanproducealineonwhichthepremisesarealltrueandtheconclusionis false,thentheargumentisinvalid. Table3.2summarizeswhenacompletetruthtableisrequiredandwhenapartial truthtablewilldo. PracticeExercises Ifyouwantadditionalpractice,youcanconstructtruthtablesforanyofthe sentencesandargumentsintheexercisesforthepreviouschapter. ? PartA Determinewhethereachsentenceisatautology,acontradiction,ora contingentsentence.Justifyyouranswerwithacompleteorpartialtruthtable

PAGE 45

ch.3truthtables 45 whereappropriate. 1. A A 2. : B & B 3. C !: C 4. : D D 5. A $ B $: A $: B 6. A & B B & A 7. A B B A 8. : [ A B A ] 9. A & B B A 10. A $ [ A B & : B ] 11. : A B $ : A & : B 12. : A & B $ A 13. A & B & : A & B & C 14. A B C 15.[ A & B & C ] B 16. A & : A B C 17. : C A B 18. B & D $ [ A $ A C ] ? PartB Determinewhethereachpairofsentencesislogicallyequivalent. Justifyyouranswerwithacompleteorpartialtruthtablewhereappropriate. 1. A : A 2. A A A 3. A A A $ A 4. A _: B A B 5. A & : A : B $ B 6. : A & B : A _: B 7. : A B : A !: B 8. A B : B !: A 9.[ A B C ],[ A B C ] 10.[ A B & C ],[ A B & C ] ? PartC Determinewhethereachsetofsentencesisconsistentorinconsistent. Justifyyouranswerwithacompleteorpartialtruthtablewhereappropriate. 1. A A : A !: A A & A A A 2. A & B C !: B C 3. A B A C B C 4. A B B C A : C 5. B & C A A B : B C 6. A B B C C !: A

PAGE 46

46 forall x 7. A $ B C C !: A A !: B 8. A B C : D : E F ? PartD Determinewhethereachargumentisvalidorinvalid.Justifyyour answerwithacompleteorpartialtruthtablewhereappropriate. 1. A A : :A 2. A A A $ A : : A 3. A A & : A : : : A 4. A $: B $ A : :A 5. A B A : : : A !: B 6. A B B : :A 7. A B B C : A : :B & C 8. A B B C : B : :A & C 9. B & A C C & A B : : C & B A 10. A $ B B $ C : :A $ C ? PartE Answereachofthequestionsbelowandjustifyyouranswer. 1.Supposethat A and B arelogicallyequivalent.Whatcanyousayabout A $ B ? 2.Supposethat A & B C iscontingent.Whatcanyousayaboutthe argument A B : : C "? 3.Supposethat f A ; B ; C g isinconsistent.Whatcanyousayabout A & B & C ? 4.Supposethat A isacontradiction.Whatcanyousayabouttheargument A B : : C "? 5.Supposethat C isatautology.Whatcanyousayabouttheargument A B : : C "? 6.Supposethat A and B arelogicallyequivalent.Whatcanyousayabout A B ? 7.Supposethat A and B are not logicallyequivalent.Whatcanyousay about A B ? PartF Wecouldleavethebiconditional $ outofthelanguage.Ifwedid that,wecouldstillwrite` A $ B 'soastomakesentenceseasiertoread,but thatwouldbeshorthandfor A B & B A .Theresultinglanguage wouldbeformallyequivalenttoSL,since A $ B and A B & B A arelogicallyequivalentinSL.Ifwevaluedformalsimplicityoverexpressive richness,wecouldreplacemoreoftheconnectiveswithnotationalconventions andstillhavealanguageequivalenttoSL. Thereareanumberofequivalentlanguageswithonlytwoconnectives.Itwould beenoughtohaveonlynegationandthematerialconditional.Showthisby writingsentencesthatarelogicallyequivalenttoeachofthefollowingusingonly parentheses,sentenceletters,negation : ,andthematerialconditional .

PAGE 47

ch.3truthtables 47 1. ?A B 2. ?A & B 3. ?A $ B WecouldhavealanguagethatisequivalenttoSLwithonlynegationand disjunctionasconnectives.Showthis:Usingonlyparentheses,sentenceletters, negation : ,anddisjunction ,writesentencesthatarelogicallyequivalent toeachofthefollowing. 4. A & B 5. A B 6. A $ B The Sheerstroke isalogicalconnectivewiththefollowingcharacteristictruthtable: A B A j B T T F T F T F T T F F T 7.WriteasentenceusingtheconnectivesofSLthatislogicallyequivalent to A j B EverysentencewrittenusingaconnectiveofSLcanberewrittenasalogically equivalentsentenceusingoneormoreSheerstrokes.UsingonlytheSheer stroke,writesentencesthatareequivalenttoeachofthefollowing. 8. : A 9. A & B 10. A B 11. A B 12. A $ B

PAGE 48

Chapter4 Quantiedlogic ThischapterintroducesalogicallanguagecalledQL.Itisaversionof quantied logic ,becauseitallowsforquantierslike all and some .Quantiedlogicisalso sometimescalled predicatelogic ,becausethebasicunitsofthelanguageare predicatesandterms. 4.1Fromsentencestopredicates Considerthefollowingargument,whichisobviouslyvalidinEnglish: Ifeveryoneknowslogic,theneithernoonewillbeconfusedoreveryonewill.Everyonewillbeconfusedonlyifwetrytobelievea contradiction.Thisisalogicclass,soeveryoneknowslogic. : : Ifwedon'ttrytobelieveacontradiction,thennoonewillbe confused. InordertosymbolizethisinSL,wewillneedasymbolizationkey. L: Everyoneknowslogic. N: Noonewillbeconfused. E: Everyonewillbeconfused. B: Wetrytobelieveacontradiction. Noticethat N and E arebothaboutpeoplebeingconfused,buttheyaretwo separatesentenceletters.Wecouldnotreplace E with : N .Whynot? : N means`Itisnotthecasethatnoonewillbeconfused.'Thiswouldbethecase 48

PAGE 49

ch.4quantiedlogic 49 ifevenonepersonwereconfused,soitisalongwayfromsayingthat everyone willbeconfused. Oncewehaveseparatesentencelettersfor N and E ,however,weeraseany connectionbetweenthetwo.Theyarejusttwoatomicsentenceswhichmight betrueorfalseindependently.InEnglish,itcouldneverbethecasethat bothnooneandeveryonewasconfused.AssentencesofSL,however,thereisa truth-valueassignmentforwhich N and E arebothtrue. Expressionslike`noone',`everyone',and`anyone'arecalled quantiers .By translating N and E asseparateatomicsentences,weleaveoutthe quantier structure ofthesentences.Fortunately,thequantierstructureisnotwhat makesthisargumentvalid.Assuch,wecansafelyignoreit.Toseethis,we translatetheargumenttoSL: L N E E B L : : : B N ThisisavalidargumentinSL.Youcandoatruthtabletocheckthis. Nowconsideranotherargument.ThisoneisalsovalidinEnglish. Willardisalogician.Alllogicianswearfunnyhats. : : Willardwearsafunnyhat. TosymbolizeitinSL,wedeneasymbolizationkey: L: Willardisalogician. A: Alllogicianswearfunnyhats. F: Willardwearsafunnyhat. Nowwesymbolizetheargument: L A : :F Thisis invalid inSL.Again,youcanconrmthiswithatruthtable.Thereis somethingverywronghere,becausethisisclearlyavalidargumentinEnglish. ThesymbolizationinSLleavesoutalltheimportantstructure.Onceagain,

PAGE 50

50 forall x thetranslationtoSLoverlooksquantierstructure:Thesentence`Alllogicians wearfunnyhats'isaboutbothlogiciansandhat-wearing.Bynottranslating thisstructure,welosetheconnectionbetweenWillard'sbeingalogicianand Willard'swearingahat. SomeargumentswithquantierstructurecanbecapturedinSL,liketherst example,eventhoughSLignoresthequantierstructure.Otherargumentsare completelybotchedinSL,likethesecondexample.Noticethattheproblem isnotthatwehavemadeamistakewhilesymbolizingthesecondargument. Thesearethebestsymbolizationswecangiveforthesearguments inSL Generally,ifanargumentcontainingquantierscomesout validinSL ,then theEnglishlanguageargumentisvalid.Ifitcomesout invalidinSL ,thenwe cannotsaytheEnglishlanguageargumentisinvalid.Theargumentmightbe validbecauseofquantierstructurewhichthenaturallanguageargumenthas andwhichtheargumentinSLlacks. Similarly,ifasentencewithquantierscomesoutasa tautologyinSL ,thenthe Englishsentenceislogicallytrue.Ifcomesoutas contingentinSL ,thenthis mightbebecauseofthestructureofthequantiersthatgetsremovedwhenwe translateintotheformallanguage. Inordertosymbolizeargumentsthatrelyonquantierstructure,weneedto developadierentlogicallanguage.Wewillcallthislanguagequantiedlogic, QL. 4.2BuildingblocksofQL Justassentenceswerethebasicunitofsententiallogic,predicateswillbethe basicunitofquantiedlogic.Apredicateisanexpressionlike`isadog.'This isnotasentenceonitsown.Itisneithertruenorfalse.Inordertobetrueor false,weneedtospecifysomething:Whoorwhatisitthatisadog? Thedetailsofthiswillbeexplainedintherestofthechapter,buthereis thebasicidea:InQL,wewillrepresentpredicateswithcapitalletters.For instance,wemightlet D standfor` isadog.'Wewilluselower-case lettersasthenamesofspecicthings.Forinstance,wemightlet b standfor Bertie.Theexpression Db willbeasentenceinQL.Itisatranslationofthe sentence`Bertieisadog.' Inordertorepresentquantierstructure,wewillalsohavesymbolsthatrepresentquantiers.Forinstance,` 9 'willmean`Thereissome .'Sotosay thatthereisadog,wecanwrite 9 xDx ;thatis:Thereissome x suchthat x is adog.

PAGE 51

ch.4quantiedlogic 51 Thatwillcomelater.Westartbydeningsingulartermsandpredicates. SingularTerms InEnglish,a singularterm isawordorphrasethatreferstoa specic person, place,orthing.Theword`dog'isnotasingularterm,becausethereareagreat manydogs.Thephrase`Philip'sdogBertie'isasingularterm,becauseitrefers toaspeciclittleterrier. A propername isasingulartermthatpicksoutanindividualwithoutdescribingit.Thename`Emerson'isapropername,andthenamealonedoesnottell youanythingaboutEmerson.Ofcourse,somenamesaretraditionallygivento boysandotheraretraditionallygiventogirls.If`JackHathaway'isusedasa singularterm,youmightguessthatitreferstoaman.However,thenamedoes notnecessarilymeanthatthepersonreferredtoisaman|oreventhatthe creaturereferredtoisaperson.Jackmightbeagiraeforallyoucouldtelljust fromthename.Thereisagreatdealofphilosophicalactionsurroundingthis issue,buttheimportantpointhereisthatanameisasingulartermbecauseit picksoutasingle,specicindividual. Othersingulartermsmoreobviouslyconveyinformationaboutthethingto whichtheyrefer.Forinstance,youcantellwithoutbeingtoldanythingfurther that`Philip'sdogBertie'isasingulartermthatreferstoadog.A definite description picksoutanindividualbymeansofauniquedescription.In English,denitedescriptionsareoftenphrasesoftheform`thesuch-and-so.' Theyreferto the specicthingthatmatchesthegivendescription.Forexample, `thetallestmemberofMontyPython'and`therstemperorofChina'are denitedescriptions.Adescriptionthatdoesnotpickoutaspecicindividual isnotadenitedescription.`AmemberofMontyPython'and`anemperorof China'arenotdenitedescriptions. Wecanusepropernamesanddenitedescriptionstopickoutthesamething. Thepropername`MountRainier'namesthelocationpickedoutbythedenite description`thehighestpeakinWashingtonstate.'Theexpressionsrefertothe sameplaceindierentways.YoulearnnothingfrommysayingthatIamgoing toMountRainier,unlessyoualreadyknowsomegeography.Youcouldguess thatitisamountain,perhaps,buteventhisisnotasurething;forallyou knowitmightbeacollege,likeMountHolyoke.YetifIweretosaythatIwas goingtothehighestpeakinWashingtonstate,youwouldknowimmediately thatIwasgoingtoamountaininWashingtonstate. InEnglish,thespecicationofasingulartermmaydependoncontext;`Willard' meansaspecicpersonandnotjustsomeonenamedWillard;`P.D.Magnus' asalogicalsingulartermmeans me andnottheotherP.D.Magnus.Welive withthiskindofambiguityinEnglish,butitisimportanttokeepinmindthat

PAGE 52

52 forall x singulartermsinQLmustrefertojustonespecicthing. InQL,wewillsymbolizesingulartermswithlower-caseletters a through w Wecanaddsubscriptsifwewanttousesomelettermorethanonce.So a;b;c;:::w;a 1 ;f 32 ;j 390 ,and m 12 arealltermsinQL. Singulartermsarecalled constants becausetheypickoutspecicindividuals. Notethat x;y ,and z arenotconstantsinQL.Theywillbe variables ,letters whichdonotstandforanyspecicthing.Wewillneedthemwhenweintroduce quantiers. Predicates Thesimplestpredicatesarepropertiesofindividuals.Theyarethingsyoucan sayaboutanobject.` isadog'and` isamemberofMontyPython' arebothpredicates.IntranslatingEnglishsentences,thetermwillnotalways comeatthebeginningofthesentence:`Apianofellon 'isalsoapredicate. Predicateslikethesearecalled one-place or monadic ,becausethereisonly oneblanktollin.Aone-placepredicateandasingulartermcombinetomake asentence. Otherpredicatesareaboutthe relation betweentwothings.Forinstance,` isbiggerthan ',` istotheleftof ',and` owesmoneyto .'Theseare two-place or dyadic predicates,becausetheyneedtobe lledinwithtwotermsinordertomakeasentence. Ingeneral,youcanthinkaboutpredicatesasschematicsentencesthatneedtobe lledoutwithsomenumberofterms.Conversely,youcanstartwithsentences andmakepredicatesoutofthembyremovingterms.Considerthesentence, `VinnieborrowedthefamilycarfromNunzio.'Byremovingasingularterm,we canrecognizethissentenceasusinganyofthreedierentmonadicpredicates: borrowedthefamilycarfromNunzio. Vinnieborrowed fromNunzio. Vinnieborrowedthefamilycarfrom Byremovingtwosingularterms,wecanrecognizethreedierentdyadicpredicates: Vinnieborrowed from borrowedthefamilycarfrom borrowed fromNunzio. Byremovingallthreesingularterms,wecanrecognizeone three-place or

PAGE 53

ch.4quantiedlogic 53 triadic predicate: borrowed from IfwearetranslatingthissentenceintoQL,shouldwetranslateitwithaone-, two-,orthree-placepredicate?Itdependsonwhatwewanttobeabletosay. Iftheonlythingthatwewilldiscussbeingborrowedisthefamilycar,then thegeneralityofthethree-placepredicateisunnecessary.Iftheonlyborrowing weneedtosymbolizeisdierentpeopleborrowingthefamilycarfromNunzio, thenaone-placepredicatewillbeenough. Ingeneral,wecanhavepredicateswithasmanyplacesasweneed.Predicates withmorethanoneplacearecalled polyadic .Predicateswith n places,for somenumber n ,arecalled n-place or n-adic InQL,wesymbolizepredicateswithcapitalletters A through Z ,withorwithout subscripts.Whenwegiveasymbolizationkeyforpredicates,wewillnotuse blanks;instead,wewillusevariables.Byconvention,constantsarelistedatthe endofthekey.Sowemightwriteakeythatlookslikethis: Ax: x isangry. Hx: x ishappy. T 1 xy: x isastallortallerthan y T 2 xy: x isastoughortougherthan y Bxyz: y isbetween x and z d: Donald g: Gregor m: Marybeth Wecansymbolizesentencesthatuseanycombinationofthesepredicatesand terms.Forexample: 1.Donaldisangry. 2.IfDonaldisangry,thensoareGregorandMarybeth. 3.MarybethisatleastastallandastoughasGregor. 4.DonaldisshorterthanGregor. 5.GregorisbetweenDonaldandMarybeth. Sentence1isstraightforward: Ad .The` x 'inthekeyentry` Ax 'isjusta placeholder;wecanreplaceitwithothertermswhentranslating. Sentence2canbeparaphrasedas,`If Ad ,then Ag and Am .'QLhasallthe truth-functionalconnectivesofSL,sowetranslatethisas Ad Ag & Am Sentence3canbetranslatedas T 1 mg & T 2 mg .

PAGE 54

54 forall x Sentence4mightseemasifitrequiresanewpredicate.Ifweonlyneeded tosymbolizethissentence,wecoulddeneapredicatelike Sxy tomean` x isshorterthan y .'However,thiswouldignorethelogicalconnectionbetween `shorter'and`taller.'ConsideredonlyassymbolsofQL,thereisnoconnection between S and T 1 .Theymightmeananythingatall.Insteadofintroducinga newpredicate,weparaphrasesentence4usingpredicatesalreadyinourkey:`It isnotthecasethatDonaldisastallerortallerthanGregor.'Wecantranslate itas : T 1 dg Sentence5requiresthatwepaycarefulattentiontotheorderoftermsinthe key.Itbecomes Bdgm 4.3Quantiers Wearenowreadytointroducequantiers.Considerthesesentences: 6.Everyoneishappy. 7.EveryoneisatleastastoughasDonald. 8.Someoneisangry. Itmightbetemptingtotranslatesentence6as Hd & Hg & Hm .Yetthiswould onlysaythatDonald,Gregor,andMarybetharehappy.Wewanttosaythat everyone ishappy,evenifwehavenotdenedaconstanttonamethem.In ordertodothis,weintroducethe` 8 'symbol.Thisiscalledthe universal quantifier Aquantiermustalwaysbefollowedbyavariableandaformulathatincludes thatvariable.Wecantranslatesentence6as 8 xHx .ParaphrasedinEnglish, thismeans`Forall x x ishappy.'Wecall 8 x an x-quantier .Theformulathat followsthequantieriscalledthe scope ofthequantier.Wewillgiveaformal denitionofscopelater,butintuitivelyitisthepartofthesentencethatthe quantierquantiesover.In 8 xHx ,thescopeoftheuniversalquantieris Hx Sentence7canbeparaphrasedas,`Forall x x isatleastastoughasDonald.' Thistranslatesas 8 xT 2 xd Inthesequantiedsentences,thevariable x isservingasakindofplaceholder. Theexpression 8 x meansthatyoucanpickanyoneandputtheminas x .There isnospecialreasontouse x ratherthansomeothervariable.Thesentence 8 xHx meansexactlythesamethingas 8 yHy 8 zHz ,and 8 x 5 Hx 5 Totranslatesentence8,weintroduceanothernewsymbol:the existential quantifier 9 .Liketheuniversalquantier,theexistentialquantierrequires avariable.Sentence8canbetranslatedas 9 xAx .Thismeansthatthereis

PAGE 55

ch.4quantiedlogic 55 some x whichisangry.Moreprecisely,itmeansthatthereis atleastone angry person.Onceagain,thevariableisakindofplaceholder;wecouldjustaseasily havetranslatedsentence8as 9 zAz Considerthesefurthersentences: 9.Nooneisangry. 10.Thereissomeonewhoisnothappy. 11.Noteveryoneishappy. Sentence9canbeparaphrasedas,`Itisnotthecasethatsomeoneisangry.' Thiscanbetranslatedusingnegationandanexistentialquantier: :9 xAx Yetsentence9couldalsobeparaphrasedas,`Everyoneisnotangry.'Withthis inmind,itcanbetranslatedusingnegationandauniversalquantier: 8 x : Ax Bothoftheseareacceptabletranslations,becausetheyarelogicallyequivalent. Thecriticalthingiswhetherthenegationcomesbeforeorafterthequantier. Ingeneral, 8 x A islogicallyequivalentto :9 x : A .Thismeansthatanysentence whichcanbesymbolizedwithauniversalquantiercanbesymbolizedwithan existentialquantier,andviceversa.Onetranslationmightseemmorenatural thantheother,butthereisnologicaldierentintranslatingwithonequantier ratherthantheother.Forsomesentences,itwillsimplybeamatteroftaste. Sentence10ismostnaturallyparaphrasedas,`Thereissome x suchthat x is nothappy.'Thisbecomes 9 x : Hx .Equivalently,wecouldwrite :8 xHx Sentence11ismostnaturallytranslatedas :8 xHx .Thisislogicallyequivalent tosentence10andsocouldalsobetranslatedas 9 x : Hx AlthoughwehavetwoquantiersinQL,wecouldhaveanequivalentformal languagewithonlyonequantier.Wecouldproceedwithonlytheuniversal quantier,forinstance,andtreattheexistentialquantierasanotationalconvention.Weusesquarebrackets[]tomakesomesentencesmorereadable,but weknowthatthesearereallyjustparentheses.Inthesameway,wecould write` 9 x 'knowingthatthisisjustshorthandfor` :8 x : .'Thereisachoice betweenmakinglogicformallysimpleandmakingitexpressivelysimple.With QL,weoptforexpressivesimplicity.Both 8 and 9 willbesymbolsofQL. UniverseofDiscourse Giventhesymbolizationkeywehavebeenusing, 8 xHx means`Everyoneis happy.'Whoisincludedinthis everyone ?Whenweusesentenceslikethisin English,weusuallydonotmeaneveryonenowaliveontheEarth.Wecertainly donotmeaneveryonewhowaseveraliveorwhowilleverlive.Wemean

PAGE 56

56 forall x somethingmoremodest:everyoneinthebuilding,everyoneintheclass,or everyoneintheroom. Inordertoeliminatethisambiguity,wewillneedtospecifya universeof discourse |abbreviatedUD.TheUDisthesetofthingsthatwearetalking about.SoifwewanttotalkaboutpeopleinChicago,wedenetheUDtobe peopleinChicago.Wewritethisatthebeginningofthesymbolizationkey,like this: UD: peopleinChicago Thequantiers rangeover theuniverseofdiscourse.GiventhisUD, 8 x means `EveryoneinChicago'and 9 x means`SomeoneinChicago.'Eachconstant namessomememberoftheUD,sowecanonlyusethisUDwiththesymbolizationkeyaboveifDonald,Gregor,andMarybethareallinChicago.Ifwe wanttotalkaboutpeopleinplacesbesidesChicago,thenweneedtoinclude thosepeopleintheUD. InQL,theUDmustbe non-empty ;thatis,itmustincludeatleastonething. ItispossibletoconstructformallanguagesthatallowforemptyUDs,butthis introducescomplications. EvenallowingforaUDwithjustonemembercanproducesomestrangeresults. Supposewehavethisasasymbolizationkey: UD: theEielTower Px: x isinParis. Thesentence 8 xPx mightbeparaphrasedinEnglishas`EverythingisinParis.' Yetthatwouldbemisleading.Itmeansthateverything intheUD isinParis. ThisUDcontainsonlytheEielTower,sowiththissymbolizationkey 8 xPx justmeansthattheEielTowerisinParis. Non-referringterms InQL,eachconstantmustpickoutexactlyonememberoftheUD.Aconstant cannotrefertomorethanonething|itisa singular term.Eachconstantmust stillpickout something .Thisisconnectedtoaclassicphilosophicalproblem: theso-calledproblemofnon-referringterms. Medievalphilosopherstypicallyusedsentencesaboutthe chimera toexemplify thisproblem.Chimeraisamythologicalcreature;itdoesnotreallyexist. Considerthesetwosentences:

PAGE 57

ch.4quantiedlogic 57 12.Chimeraisangry. 13.Chimeraisnotangry. Itistemptingjusttodeneaconstanttomean`chimera.'Thesymbolization keywouldlooklikethis: UD: creaturesonEarth Ax: x isangry. c: chimera Wecouldthentranslatesentence12as Ac andsentence13as : Ac Problemswillarisewhenweaskwhetherthesesentencesaretrueorfalse. Oneoptionistosaythatsentence12isnottrue,becausethereisnochimera. Ifsentence12isfalsebecauseittalksaboutanon-existentthing,thensentence 13isfalseforthesamereason.Yetthiswouldmeanthat Ac and : Ac would bothbefalse.Giventhetruthconditionsfornegation,thiscannotbethecase. Sincewecannotsaythattheyarebothfalse,whatshouldwedo?Anotheroption istosaythatsentence12is meaningless becauseittalksaboutanon-existent thing.So Ac wouldbeameaningfulexpressioninQLforsomeinterpretations butnotforothers.Yetthiswouldmakeourformallanguagehostagetoparticularinterpretations.Sinceweareinterestedinlogicalform,wewanttoconsider thelogicalforceofasentencelike Ac apartfromanyparticularinterpretation. If Ac weresometimesmeaningfulandsometimesmeaningless,wecouldnotdo that. Thisisthe problemofnon-referringterms ,andwewillreturntoitlatersee p.74.TheimportantpointfornowisthateachconstantofQL must referto somethingintheUD,althoughtheUDcanbeanysetofthingsthatwelike. Ifwewanttosymbolizeargumentsaboutmythologicalcreatures,thenwemust deneaUDthatincludesthem.Thisoptionisimportantifwewanttoconsider thelogicofstories.Wecantranslateasentencelike`SherlockHolmeslivedat 221BBakerStreet'byincludingctionalcharacterslikeSherlockHolmesinour UD. 4.4TranslatingtoQL WenowhaveallofthepiecesofQL.Translatingmorecomplicatedsentences willonlybeamatterofknowingtherightwaytocombinepredicates,constants, quantiers,andconnectives.Considerthesesentences:

PAGE 58

58 forall x 14.Everycoininmypocketisaquarter. 15.Somecoinonthetableisadime. 16.Notallthecoinsonthetablearedimes. 17.Noneofthecoinsinmypocketaredimes. Inprovidingasymbolizationkey,weneedtospecifyaUD.Sincewearetalking aboutcoinsinmypocketandonthetable,theUDmustatleastcontainallof thosecoins.Sincewearenottalkingaboutanythingbesidescoins,weletthe UDbeallcoins.Sincewearenottalkingaboutanyspeciccoins,wedonot needtodeneanyconstants.Sowedenethiskey: UD: allcoins Px: x isinmypocket. Tx: x isonthetable. Qx: x isaquarter. Dx: x isadime. Sentence14ismostnaturallytranslatedwithauniversalquantier.TheuniversalquantiersayssomethingabouteverythingintheUD,notjustaboutthe coinsinmypocket.Sentence14meansthat,foranycoin, if thatcoinisinmy pocket then itisaquarter.Sowecantranslateitas 8 x Px Qx Sincesentence14isaboutcoinsthatarebothinmypocket and thatarequarters,itmightbetemptingtotranslateitusingaconjunction.However,the sentence 8 x Px & Qx wouldmeanthateverythingintheUDisbothinmy pocketandaquarter:Allthecoinsthatexistarequartersinmypocket.This iswouldbeacrazythingtosay,anditmeanssomethingverydierentthan sentence14. Sentence15ismostnaturallytranslatedwithanexistentialquantier.Itsays thatthereissomecoinwhichisbothonthetableandwhichisadime.Sowe cantranslateitas 9 x Tx & Dx Noticethatweneededtouseaconditionalwiththeuniversalquantier,but weusedaconjunctionwiththeexistentialquantier.Whatwoulditmeanto write 9 x Tx Dx ?Probablynotwhatyouthink.Itmeansthatthereissome memberoftheUDwhichwouldsatisfythesubformula;roughlyspeaking,there issome a suchthat Ta Da istrue.InSL, A B islogicallyequivalentto : A B ,andthiswillalsoholdinQL.So 9 x Tx Dx istrueifthereissome a suchthat : Ta Da ;i.e.,itistrueifsomecoinis either notonthetable or isadime.Ofcoursethereisacointhatisnotthetable|therearecoinslots ofotherplaces.So 9 x Tx Dx istriviallytrue.Aconditionalwillusually bethenaturalconnectivetousewithauniversalquantier,butaconditional withinthescopeofanexistentialquantiercandoverystrangethings.Asa generalrule,donotputconditionalsinthescopeofexistentialquantiersunless youaresurethatyouneedone.

PAGE 59

ch.4quantiedlogic 59 Sentence16canbeparaphrasedas,`Itisnotthecasethateverycoinonthe tableisadime.'Sowecantranslateitas :8 x Tx Dx .Youmightlook atsentence16andparaphraseitinsteadas,`Somecoinonthetableisnota dime.'Youwouldthentranslateitas 9 x Tx & : Dx .Althoughitisprobably notobvious,thesetwotranslationsarelogicallyequivalent.Thisisdueto thelogicalequivalencebetween :8 x A and 9 x : A ,alongwiththeequivalence between : A B and A & : B Sentence17canbeparaphrasedas,`Itisnotthecasethatthereissomedime inmypocket.'Thiscanbetranslatedas :9 x Px & Dx .Itmightalsobe paraphrasedas,`Everythinginmypocketisanon-dime,'andthencouldbe translatedas 8 x Px !: Dx .Againthetwotranslationsarelogicallyequivalent.Botharecorrecttranslationsofsentence17. Wecannowtranslatetheargumentfromp.49,theonethatmotivatedtheneed forquantiers: Willardisalogician.Alllogicianswearfunnyhats. : : Willardwearsafunnyhat. UD: people Lx: x isalogician. Fx: x wearsafunnyhat. w: Willard Translating,weget: Lw 8 x Lx Fx : :Fw ThiscapturesthestructurethatwasleftoutoftheSLtranslationofthisargument,andthisisavalidargumentinQL. Emptypredicates ApredicateneednotapplytoanythingintheUD.Apredicatethatappliesto nothingintheUDiscalledan empty predicate. Supposewewanttosymbolizethesetwosentences: 18.Everymonkeyknowssignlanguage.

PAGE 60

60 forall x 19.Somemonkeyknowssignlanguage. Itispossibletowritethesymbolizationkeyforthesesentencesinthisway: UD: animals Mx: x isamonkey. Sx: x knowssignlanguage. Sentence18cannowbetranslatedas 8 x Mx Sx Sentence19becomes 9 x Mx & Sx Itistemptingtosaythatsentence18entailssentence19;thatis:ifevery monkeyknowssignlanguage,thenitmustbethatsomemonkeyknowssign language.ThisisavalidinferenceinAristoteleanlogic:All M sare S : : some M is S .However,theentailmentdoesnotholdinQL.Itispossibleforthe sentence 8 x Mx Sx tobetrueeventhoughthesentence 9 x Mx & Sx is false. Howcanthisbe?Theanswercomesfromconsideringwhetherthesesentences wouldbetrueorfalse iftherewerenomonkeys Wehavedened 8 and 9 insuchawaythat 8 A isequivalentto :9: A .As such,theuniversalquantierdoesn'tinvolvetheexistenceofanything|only non-existence.Ifsentence18istrue,thenthereare no monkeyswhodon'tknow signlanguage.Iftherewerenomonkeys,then 8 x Mx Sx wouldbetrue and 9 x Mx & Sx wouldbefalse. Weallowemptypredicatesbecausewewanttobeabletosaythingslike,`Ido notknowifthereareanymonkeys,butanymonkeysthatthereareknowsign language.'Thatis,wewanttobeabletohavepredicatesthatdonotormight notrefertoanything. Whathappensifweaddanemptypredicate R totheinterpretationabove?For example,wemightdene Rx tomean` x isarefrigerator.'Nowthesentence 8 x Rx Mx willbetrue.Thisiscounterintuitive,sincewedonotwantto saythatthereareawholebunchofrefrigeratormonkeys.Itisimportantto remember,though,that 8 x Rx Mx meansthatanymemberoftheUDthat isarefrigeratorisamonkey.SincetheUDisanimals,therearenorefrigerators intheUDandsothesetenceistriviallytrue. Ifyouwereactuallytranslatingthesentence`Allrefrigeratorsaremonkeys', thenyouwouldwanttoincludeappliancesintheUD.Thenthepredicate R wouldnotbeemptyandthesentence 8 x Rx Mx wouldbefalse.

PAGE 61

ch.4quantiedlogic 61 AUDmusthave atleast onemember. Apredicatemayapplytosome,all,ornomembersoftheUD. Aconstantmustpickout exactly onememberoftheUD. AmemberoftheUDmaybepickedoutbyoneconstant,many constants,ornoneatall. PickingaUniverseofDiscourse TheappropriatesymbolizationofanEnglishlanguagesentenceinQLwilldependonthesymbolizationkey.Insomeways,thisisobvious:Itmatterswhether Dx means` x isdainty'or` x isdangerous.'ThemeaningofsentencesinQL alsodependsontheUD. Let Rx mean` x isarose,'let Tx mean` x hasathorn,'andconsiderthis sentence: 20.Everyrosehasathorn. Itistemptingtosaythatsentence20shouldbetranslatedas 8 x Rx Tx .If theUDcontainsallroses,thatwouldbecorrect.YetiftheUDismerely things onmykitchentable ,then 8 x Rx Tx wouldonlymeanthateveryroseon mykitchentablehasathorn.Iftherearenorosesonmykitchentable,the sentencewouldbetriviallytrue. TheuniversalquantieronlyrangesovermembersoftheUD,soweneedto includeallrosesintheUDinordertotranslatesentence20.Wehavetwo options.First,wecanrestricttheUDtoincludeallrosesbut only roses.Then sentence20becomes 8 xTx .ThismeansthateverythingintheUDhasathorn; sincetheUDjustisthesetofroses,thismeansthateveryrosehasathorn. Thisoptioncansaveustroubleifeverysentencethatwewanttotranslateusing thesymbolizationkeyisaboutroses. Second,wecanlettheUDcontainthingsbesidesroses:rhododendrons,rats, ries,andwhatallelse.Thensentence20mustbe 8 x Rx Tx Ifwewantedtheuniversalquantiertomean every thing,withoutrestriction, thenwemighttrytospecifyaUDthatcontainseverything.Thiswouldlead toproblems.Does`everything'includethingsthathaveonlybeenimagined, likectionalcharacters?Ontheonehand,wewanttobeabletosymbolize argumentsaboutHamletorSherlockHolmes.Soweneedtohavetheoptionof includingctionalcharactersintheUD.Ontheotherhand,weneverneedto talkabouteverythingthatdoesnotexist.Thatmightnotevenmakesense.

PAGE 62

62 forall x Therearephilosophicalissuesherethatwewillnottrytoaddress.Wecan avoidthesedicultiesbyalwaysspecifyingtheUD.Forexample,ifwemeanto talkaboutplants,people,andcities,thentheUDmightbe`livingthingsand places.' Supposethatwewanttotranslatesentence20and,withthesamesymbolization key,translatethesesentences: 21.Esmereldahasaroseinherhair. 22.EveryoneiscrosswithEsmerelda. WeneedaUDthatincludesrosessothatwecansymbolizesentence20anda UDthatincludespeoplesowecantranslatesentence21{22.Hereisasuitable key: UD: peopleandplants Px: x isaperson. Rx: x isarose. Tx: x hasathorn. Cxy: x iscrosswith y Hxy: x has y intheirhair. e: Esmerelda Sincewedonothaveapredicatethatmeans` ::: hasaroseinherhair',translatingsentence21willrequireparaphrasing.Thesentencesaysthatthereisa roseinEsmerelda'shair;thatis,thereissomethingwhichisbotharoseandis inEsmerelda'shair.Soweget: 9 x Rx & Hex Itistemptingtotranslatesentence22as 8 xCxe .Unfortunately,thiswould meanthateverymemberoftheUDiscrosswithEsmerelda|bothpeopleand plants.Itwouldmean,forinstance,thattheroseinEsmerelda'shairiscross withher.Ofcourse,sentence22doesnotmeanthat. `Everyone'meanseveryperson,noteverymemberoftheUD.Sowecanparaphrasesentence22as,`EverypersoniscrosswithEsmerelda.'Weknowhowto translatesentenceslikethis: 8 x Px Cxe Ingeneral,theuniversalquantiercanbeusedtomean`everyone'iftheUD containsonlypeople.IftherearepeopleandotherthingsintheUD,then `everyone'mustbetreatedas`everyperson.' Translatingpronouns WhentranslatingtoQL,itisimportanttounderstandthestructureofthe sentencesyouwanttotranslate.WhatmattersisthenaltranslationinQL,

PAGE 63

ch.4quantiedlogic 63 andsometimesyouwillbeabletomovefromanEnglishlanguagesentence directlytoasentenceofQL.Othertimes,ithelpstoparaphrasethesentence oneormoretimes.Eachsuccessiveparaphraseshouldmovefromtheoriginal sentenceclosertosomethingthatyoucantranslatedirectlyintoQL. Forthenextseveralexamples,wewillusethissymbolizationkey: UD: people Gx: x canplayguitar. Rx: x isarockstar. l: Lemmy Nowconsiderthesesentences: 23.IfLemmycanplayguitar,thenheisarockstar. 24.Ifapersoncanplayguitar,thenheisarockstar. Sentence23andsentence24havethesameconsequent` ::: heisarockstar', buttheycannotbetranslatedinthesameway.Ithelpstoparaphrasethe originalsentences,replacingpronounswithexplicitreferences. Sentence23canbeparaphrasedas,`IfLemmycanplayguitar,then Lemmy is arockstar.'Thiscanobviouslybetranslatedas Gl Rl Sentence24mustbeparaphraseddierently:`Ifapersoncanplayguitar,then thatperson isarockstar.'Thissentenceisnotaboutanyparticularperson, soweneedavariable.Translatinghalfway,wecanparaphrasethesentenceas, `Foranyperson x ,if x canplayguitar,then x isarockstar.'Nowthiscanbe translatedas 8 x Gx Rx .Thisisthesameas,`Everyonewhocanplayguitar isarockstar.' Considerthesefurthersentences: 25.Ifanyonecanplayguitar,thenLemmycan. 26.Ifanyonecanplayguitar,thenheorsheisarockstar. Thesetwosentenceshavethesameantecedent`Ifanyonecanplayguitar ::: ', buttheyhavedierentlogicalstructures. Sentence25canbeparaphrased,`Ifsomeonecanplayguitar,thenLemmycan playguitar.'Theantecedentandconsequentareseparatesentences,soitcan besymbolizedwithaconditionalasthemainlogicaloperator: 9 xGx Gl .

PAGE 64

64 forall x Sentence26canbeparaphrased,`Foranyone,ifthatonecanplayguitar,then thatoneisarockstar.'Itwouldbeamistaketosymbolizethiswithan existentialquantier,becauseitistalkingabouteverybody.Thesentenceis equivalentto`Allguitarplayersarerockstars.'Itisbesttranslatedas 8 x Gx Rx TheEnglishwords`any'and`anyone'shouldtypicallybetranslatedusingquantiers.Asthesetwoexamplesshow,theysometimescallforanexistentialquantierasinsentence25andsometimesforauniversalquantierasinsentence 26.Ifyouhaveahardtimedeterminingwhichisrequired,paraphrasethe sentencewithanEnglishlanguagesentencethatuseswordsbesides`any'or `anyone.' Quantiersandscope Inthesentence 9 xGx Gl ,thescopeoftheexistentialquantieristheexpression Gx .Woulditmatterifthescopeofthequantierwerethewholesentence? Thatis,doesthesentence 9 x Gx Gl meansomethingdierent? Withthekeygivenabove, 9 xGx Gl meansthatifthereissomeguitarist, thenLemmyisaguitarist. 9 x Gx Gl wouldmeanthatthereissomeperson suchthatifthatpersonwereaguitarist,thenLemmywouldbeaguitarist. Recallthattheconditionalhereisamaterialconditional;theconditionalistrue iftheantecedentisfalse.Lettheconstant p denotetheauthorofthisbook, someonewhoiscertainlynotaguitarist.Thesentence Gp Gl istruebecause Gp isfalse.Sincesomeonenamely p satisesthesentence,then 9 x Gx Gl istrue.Thesentenceistruebecausethereisanon-guitarist,regardlessof Lemmy'sskillwiththeguitar. Somethingstrangehappenedwhenwechangedthescopeofthequantier,becausetheconditionalinQLisamaterialconditional.Inordertokeepthe meaningthesame,wewouldhavetochangethequantier: 9 xGx Gl means thesamethingas 8 x Gx Gl ,and 9 x Gx Gl meansthesamethingas 8 xGx Gl Thisodditydoesnotarisewithotherconnectivesorifthevariableisinthe consequentoftheconditional.Forexample, 9 xGx & Gl meansthesamething as 9 x Gx & Gl ,and Gl !9 xGx meansthesamethingsas 9 x Gl Gx Ambiguouspredicates Supposewejustwanttotranslatethissentence:

PAGE 65

ch.4quantiedlogic 65 27.Adinaisaskilledsurgeon. LettheUDbepeople,let Kx mean` x isaskilledsurgeon',andlet a mean Adina.Sentence27issimply Ka Supposeinsteadthatwewanttotranslatethisargument: Thehospitalwillonlyhireaskilledsurgeon.Allsurgeonsaregreedy. Billyisasurgeon,butisnotskilled.Therefore,Billyisgreedy,but thehospitalwillnothirehim. Weneedtodistinguishbeinga skilledsurgeon frommerelybeinga surgeon .So wedenethissymbolizationkey: UD: people Gx: x isgreedy. Hx: Thehospitalwillhire x Rx: x isasurgeon. Kx: x isskilled. b: Billy Nowtheargumentcanbetranslatedinthisway: 8 x : Rx & Kx !: Hx 8 x Rx Gx Rb & : Kb : :Gb & : Hb Nextsupposethatwewanttotranslatethisargument: Carolisaskilledsurgeonandatennisplayer.Therefore,Carolisa surgeonandaskilledtennisplayer. Ifwestartwiththesymbolizationkeyweusedforthepreviousargument,we couldaddapredicatelet Tx mean` x isatennisplayer'andaconstantlet c meanCarol.Thentheargumentbecomes: Rc & Kc & Tc : :Tc & Kc Thistranslationisadisaster!IttakeswhatinEnglishisaterribleargumentand translatesitasavalidargumentinQL.Theproblemisthatthereisadierence

PAGE 66

66 forall x betweenbeing skilledasasurgeon and skilledasatennisplayer .Translating thisargumentcorrectlyrequirestwoseparatepredicates,oneforeachtypeof skill.Ifwelet K 1 x mean` x isskilledasasurgeon'and K 2 x mean` x isskilled asatennisplayer,'thenwecansymbolizedtheargumentinthisway: Rc & K 1 c & Tc : :Tc & K 2 c LiketheEnglishlanguageargumentittranslates,thisisinvalid. Themoraloftheseexamplesisthatyouneedtobecarefulofsymbolizing predicatesinanambiguousway.Similarproblemscanarisewithpredicateslike good bad big ,and small .Justasskilledsurgeonsandskilledtennisplayershave dierentskills,bigdogs,bigmice,andbigproblemsarebigindierentways. Isitenoughtohaveapredicatethatmeans` x isaskilledsurgeon',ratherthan twopredicates` x isskilled'and` x isasurgeon'?Sometimes.Assentence27 shows,sometimeswedonotneedtodistinguishbetweenskilledsurgeonsand othersurgeons. Mustwealwaysdistinguishbetweendierentwaysofbeingskilled,good,bad, orbig?No.AstheargumentaboutBillyshows,sometimesweonlyneedtotalk aboutonekindofskill.Ifyouaretranslatinganargumentthatisjustabout dogs,itisnetodeneapredicatethatmeans` x isbig.'IftheUDincludes dogsandmice,however,itisprobablybesttomakethepredicatemean` x is bigforadog.' Multiplequantiers Considerthisfollowingsymbolizationkeyandthesentencesthatfollowit: UD:Peopleanddogs Dx: x isadog. Fxy: x isafriendof y Oxy: x owns y f:Fi g:Gerald 28.Fiisadog. 29.Geraldisadogowner. 30.Someoneisadogowner. 31.AllofGerald'sfriendsaredogowners. 32.Everydogowneristhefriendofadogowner.

PAGE 67

ch.4quantiedlogic 67 Sentence28iseasy: Df Sentence29canbeparaphrasedas,`ThereisadogthatGeraldowns.'Thiscan betranslatedas 9 x Dx & Ogx Sentence30canbeparaphrasedas,`Thereissome y suchthat y isadog owner.'Thesubsentence` y isadogowner'isjustlikesentence29,exceptthat itisabout y ratherthanbeingaboutGerald.Sowecantranslatesentence30 as 9 y 9 x Dx & Oyx Sentence31canbeparaphrasedas,`EveryfriendofGeraldisadogowner.' Translatingpartofthissentence,weget 8 x Fxg ` x isadogowner'.Again, itisimportanttorecognizethat` x isadogowner'isstructurallyjustlike sentence29.Sincewealreadyhaveanx-quantier,wewillneedadierent variablefortheexistentialquantier.Anyothervariablewilldo.Using z sentence31canbetranslatedas 8 x Fxg !9 z Dz & Oxz Sentence32canbeparaphrasedas`Forany x thatisadogowner,thereisa dogownerwhois x 'sfriend.'Partiallytranslated,thisbecomes 8 x x isadogowner !9 y y isadogowner& Fxy : Completingthetranslation,sentence32becomes 8 x 9 z Dz & Oxz !9 y 9 z Dz & Oyz & Fxy : Considerthissymbolizationkeyandthesesentences: UD: people Lxy: x likes y i: Imre. k: Karl. 33.ImrelikeseveryonethatKarllikes. 34.Thereissomeonewholikeseveryonewholikeseveryonethathelikes. Sentence33canbepartiallytranslatedas 8 x Karllikes x Imrelikes x .This becomes 8 x Lkx Lix Sentence34isalmostatongue-twister.Thereislittlehopeofwritingdownthe wholetranslationimmediately,butwecanproceedbysmallsteps.Aninitial, partialtranslationmightlooklikethis: 9 x everyonewholikeseveryonethat x likesislikedby x ThepartthatremainsinEnglishisauniversalsentence,sowetranslatefurther: 9 x 8 y y likeseveryonethat x likes x likes y :

PAGE 68

68 forall x Theantecedentoftheconditionalisstructurallyjustlikesentence33,with y and x inplaceofImreandKarl.Sosentence34canbecompletelytranslated inthisway 9 x 8 y 8 z Lxz Lyz Lxy Whensymbolizingsentenceswithmultiplequantiers,itisbesttoproceedby smallsteps.ParaphrasetheEnglishsentencesothatthelogicalstructureis readilysymbolizedinQL.Thentranslatepiecemeal,replacingthedaunting taskoftranslatingalongsentencewiththesimplertaskoftranslatingshorter formulae. 4.5SentencesofQL Inthissection,weprovideaformaldenitionfora well-formedformula w and sentence ofQL. Expressions TherearesixkindsofsymbolsinQL: predicates A;B;C;:::;Z withsubscripts,asneeded A 1 ;B 1 ;Z 1 ;A 2 ;A 25 ;J 375 ;::: constants a;b;c;:::;w withsubscripts,asneeded a 1 ;w 4 ;h 7 ;m 32 ;::: variables x;y;z withsubscripts,asneeded x 1 ;y 1 ;z 1 ;x 2 ;::: connectives : ,&, $ parentheses quantiers 8 ; 9 Wedenean expressionofql asanystringofsymbolsofSL.Takeanyofthe symbolsofQLandwritethemdown,inanyorder,andyouhaveanexpression. Well-formedformulae Bydenition,a termofql iseitheraconstantoravariable. An atomicformulaofql isann-placepredicatefollowedby n terms.

PAGE 69

ch.4quantiedlogic 69 JustaswedidforSL,wewillgivea recursive denitionforawofQL.In fact,mostofthedenitionwilllooklikethedenitionofforawofSL:Every atomicformulaisaw,andyoucanbuildnewwsbyapplyingthesentential connectives. Wecouldjustaddaruleforeachofthequantiersandbedonewithit.For instance:If A isaw,then 8 x A and 9 x A arews.However,thiswould allowforbizarresentenceslike 8 x 9 xDx and 8 xDw .Whatcouldthesepossibly mean?Wecouldadoptsomeinterpretationofsuchsentences,butinsteadwe willwritethedenitionofawsothatsuchabominationsdonotevencount aswell-formed. Inorderfor 8 x A tobeaw, A mustcontainthevariable x andmustnot alreadycontainanx-quantier. 8 xDw willnotcountasawbecause` x 'does notoccurin Dw ,and 8 x 9 xDx willnotcountasawbecause 9 xDx contains anx-quantier 1.Everyatomicformulaisaw. 2.If A isaw,then : A isaw. 3.If A and B arews,then A & B ,isaw. 4.If A and B arews, A B isaw. 5.If A and B arews,then A B isaw. 6.If A and B arews,then A $ B isaw. 7.If A isaw, x isavariable, A containsatleastoneoccurrenceof x ,and A containsno x -quantiers,then 8 xA isaw. 8.If A isaw, x isavariable, A containsatleastoneoccurrenceof x ,and A containsno x -quantiers,then 9 xA isaw. 9.AllandonlywsofQLcanbegeneratedbyapplicationsoftheserules. Noticethatthe` x 'thatappearsinthedenitionaboveisnotthevariable x .It isa meta-variable thatstandsinforanyvariableofQL.So 8 xAx isaw,but soare 8 yAy 8 zAz 8 x 4 Ax 4 ,and 8 z 9 Az 9 Wecannowgiveaformaldenitionforscope:The scope ofaquantieristhe subformulaforwhichthequantieristhemainlogicaloperator.

PAGE 70

70 forall x Sentences Asentenceissomethingthatcanbeeithertrueorfalse.InSL,everywwasa sentence.ThiswillnotbethecaseinQL.Considerthefollowingsymbolization key: UD: people Lxy: x loves y b: Boris Considertheexpression Lzz .Itisanatomicforumula:atwo-placepredicate followedbytwoterms.Allatomicformulaarews,so Lzz isaw.Doesit meananything?Youmightthinkthatitmeansthat z loveshimself,inthe samewaythat Lbb meansthatBorisloveshimself.Yet z isavariable;itdoes notnamesomepersonthewayaconstantwould.Thew Lzz doesnottell ushowtointerpret z .Doesitmeaneveryone?anyone?someone?Ifwehad az-quantier,itwouldtellushowtointerpret z .Forinstance, 9 zLzz would meanthatsomeonelovesthemself. Someformallanguagestreatawlike Lzz asimplicitlyhavingauniversal quantierinfront.WewillnotdothisforQL.Ifyoumeantosaythateveryone lovesthemself,thenyouneedtowritethequantier: 8 zLzz Inordertomakesenseofavariable,weneedaquantiertotellushowto interpretthatvariable.Thescopeofanx-quantier,forinstance,isthethe partoftheformulawherequantiertellshowtointerpret x Inordertobepreciseaboutthis,wedenea boundvariable tobeanoccurrenceofavariable x thatiswithinthescopeofan x -quantier.A free variable isanoccuranceofavariablethatisnotbound. Forexample,considerthew 8 x Ex Dy !9 z Ex Lzx .Thescopeof theuniversalquantier 8 x is Ex Dy ,sotherst x isboundbytheuniversal quantierbutthesecondandthird x sarefree.Thereisnoty-quantier,so the y isfree.Thescopeoftheexistentialquantier 9 z is Ex Lzx ,soboth occurrencesof z areboundbyit. Wedenea sentence ofQLasawofQLthatcontainsnofreevariables. Notationalconventions WewilladoptthesamenotationalconventionsthatwedidforSLp.31.First, wemayleaveotheoutermostparenthesesofaformula.Second,wewilluse squarebrackets`['and`]'inplaceofparenthesestoincreasethereadabilityof

PAGE 71

ch.4quantiedlogic 71 formulae.Third,wewillleaveoutparenthesesbetweeneachpairofconjuncts whenwritinglongseriesofconjunctions.Fourth,wewillleaveoutparentheses betweeneachpairofdisjunctswhenwritinglongseriesofdisjunctions. Substitutioninstance If A isaw, c aconstant,and x avariable,then A [ c j x ]isthewmadeby replacingeachoccuranceof x in A with c .Thisiscalleda substitution instance of 8 x A and 9 x A ; c iscalledthe instantiatingconstant Forexample: Aa Ba Af Bf ,and Ak Bk areallsubstitutioninstances of 8 x Ax Bx ;theinstantiatingconstantsare a f ,and k ,respectively. Raj Rdj ,and Rjj aresubstitutioninstancesof 9 zRzj ;theinstantiatingconstants are a d ,and j ,respectively. Thisdenitionwillbeusefullater,whenwedenetruthandderivabilityinQL. If 8 xPx istrue,theneverysubstitutioninstance Pa Pb Pc ...istrue.Toput thepointinformally,ifeverythingisa P ,then a isa P b isa P c isa P ,andso on.Conversely,ifsomesubstitutioninstanceof 9 xPx suchas Pa istrue,then 9 xPx mustbetrue.Informally,ifsomespecic a isa P ,thenthereissome P 4.6Identity Considerthissentence: 35.Pavelowesmoneytoeveryoneelse. LettheUDbepeople;thiswillallowustotranslate`everyone'asauniversal quantier.Let Oxy mean` x owesmoneyto y ',andlet p meanPavel.Nowwe cansymbolizesentence35as 8 xOpx .Unfortunately,thistranslationhassome oddconsequences.ItsaysthatPavelowesmoneytoeverymemberoftheUD, includingPavel;itentailsthatPavelowesmoneytohimself.However,sentence 35doesnotsaythatPavelowesmoneytohimself;heowesmoneytoeveryone else .Thisisaproblem,because 8 xOpx isthebesttranslationwecangiveof thissentenceintoQL. ThesolutionistoaddanothersymboltoQL.Thesymbol`='isatwo-place predicate.Sinceithasaspeciallogicalmeaning,wewriteitabitdierently: Fortwoterms t 1 and t 2 t 1 = t 2 isanatomicformula. Thepredicate x = y means` x isidenticalto y .'Thisdoesnotmeanmerely that x and y areindistinguishableorthatallofthesamepredicatesaretrueof

PAGE 72

72 forall x them.Rather,itmeansthat x and y aretheverysamething. Whenwewrite x 6 = y ,wemeanthat x and y arenotidentical.Thereisnoreason tointroducethisasanadditionalpredicate.Instead, x 6 = y isanabbreviation of : x = y Nowsupposewewanttosymbolizethissentence: 36.PavelisMisterCheckov. Lettheconstant c meanMisterCheckov.Sentence36canbesymbolizedas p = c .Thismeansthattheconstants p and c bothrefertothesameguy. Thisisallwellandgood,buthowdoesithelpwithsentence35?Thatsentence canbeparaphrasedas,`EveryonewhoisnotPavelisowedmoneybyPavel.' Thisisasentencestructurewealreadyknowhowtosymbolize:`Forall x ,if x isnotPavel,then x isowedmoneybyPavel.'InQLwithidentity,thisbecomes 8 x x 6 = p Opx Inadditiontosentencesthatusetheword`else',identitywillbehelpfulwhen symbolizingsomesentencesthatcontainthewords`besides'and`only.'Consider theseexamples: 37.NoonebesidesPavelowesmoneytoHikaru. 38.OnlyPavelowesHikarumoney. Weaddtheconstant h ,whichmeansHikaru. Sentence37canbeparaphrasedas,`NoonewhoisnotPavelowesmoneyto Hikaru.'Thiscanbetranslatedas :9 x x 6 = p & Oxh Sentence38canbeparaphrasedas,`PavelowesHikaru and noonebesidesPavel owesHikarumoney.'Wehavealreadytranslatedoneoftheconjuncts,andthe otherisstraightforward.Sentence38becomes Oph & :9 x x 6 = p & Oxh Expressionsofquantity Wecanalsouseidentitytosayhowmanythingsthereareofaparticularkind. Forexample,considerthesesentences: 39.Thereisatleastoneappleonthetable. 40.Thereareatleasttwoapplesonthetable. 41.Thereareatleastthreeapplesonthetable.

PAGE 73

ch.4quantiedlogic 73 LettheUDbe thingsonthetable ,andlet Ax mean` x isanapple.' Sentence39doesnotrequireidentity.Itcanbetranslatedadequatelyas 9 xAx : Thereissomeappleonthetable|perhapsmany,butatleastone. Itmightbetemptingtoalsotranslatesentence40withoutidentity.Yetconsider thesentence 9 x 9 y Ax & Ay .Itmeansthatthereissomeapple x intheUD andsomeapple y intheUD.Sincenothingprecludes x and y frompickingout thesamememberoftheUD,thiswouldbetrueeveniftherewereonlyone apple.Inordertomakesurethattherearetwo dierent apples,weneedan identitypredicate.Sentence40needstosaythatthetwoapplesthatexistare notidentical,soitcanbetranslatedas 9 x 9 y Ax & Ay & x 6 = y Sentence41requirestalkingaboutthreedierentapples.Itcanbetranslated as 9 x 9 y 9 z Ax & Ay & Az & x 6 = y & y 6 = z & x 6 = z Continuinginthisway,wecouldtranslate`Thereareatleast n applesonthe table.'Thereisasummaryofhowtosymbolizesentencesliketheseonp.157. Nowconsiderthesesentences: 42.Thereisatmostoneappleonthetable. 43.Thereareatmosttwoapplesonthetable. Sentence42canbeparaphrasedas,`Itisnotthecasethatthereareatleast two applesonthetable.'Thisisjustthenegationofsentence40: :9 x 9 y Ax & Ay & x 6 = y Sentence42canalsobeapproachedinanotherway.Itmeansthatanyapples thatthereareonthetablemustbetheselfsameapple,soitcanbetranslated as 8 x 8 y Ax & Ay x = y .Thetwotranslationsarelogicallyequivalent,so botharecorrect. Inasimilarway,sentence43canbetranslatedintwoequivalentways.Itcan beparaphrasedas,`Itisnotthecasethatthereare three ormoredistinct apples',soitcanbetranslatedasthenegationofsentence41.Usinguniversal quantiers,itcanalsobetranslatedas 8 x 8 y 8 z Ax & Ay & Az x = y x = z y = z : Seep.157forthegeneralcase. Theexamplesabovearesentencesaboutapples,butthelogicalstructureofthe sentencestranslatesmathematicalinequalitieslike a 3, a 2,andsoon.We alsowanttobeabletotranslatestatementsofequalitywhichsayexactlyhow manythingsthereare.Forexample:

PAGE 74

74 forall x 44.Thereisexactlyoneappleonthetable. 45.Thereareexactlytwoapplesonthetable. Sentence44canbeparaphrasedas,`Thereis atleast oneappleonthetable, andthereis atmost oneappleonthetable.'Thisisjusttheconjunction ofsentence39andsentence42: 9 xAx & 8 x 8 y Ax & Ay x = y .Thisis asomewhatcomplicatedwayofgoingaboutit.Itisperhapsmorestraightforwardtoparaphrasesentence44as,`Thereisathingwhichistheonly appleonthetable.'Thoughtofinthisway,thesentencecanbetranslated 9 x Ax & :9 y Ay & x 6 = y Similarly,sentence45maybeparaphrasedas,`Therearetwodierentapples onthetable,andthesearetheonlyapplesonthetable.'Thiscanbetranslated as 9 x 9 y Ax & Ay & x 6 = y & :9 z Az & x 6 = z & y 6 = z Finally,considerthissentence: 46.Thereareatmosttwothingsonthetable. Itmightbetemptingtoaddapredicatesothat Tx wouldmean` x isathing onthetable.'However,thisisunnecessary.SincetheUDisthesetofthings onthetable,allmembersoftheUDareonthetable.Ifwewanttotalk abouta thingonthetable ,weneedonlyuseaquantier.Sentence46canbe symbolizedlikesentence43whichsaidthattherewereatmosttwoapples, butleavingoutthepredicateentirely.Thatis,sentence46canbetranslatedas 8 x 8 y 8 z x = y x = z y = z Techniquesforsymbolizingexpressionsofquantity`atmost',`atleast',and `exactly'aresummarizedonp.157. Denitedescriptions RecallthataconstantofQLmustrefertosomememberoftheUD.Thisconstraintallowsustoavoidtheproblemofnon-referringterms.GivenaUDthat includedonlyactuallyexistingcreaturesbutaconstant c thatmeant`chimera' amythicalcreature,sentencescontaining c wouldbecomeimpossibletoevaluate. ThemostwidelyinuentialsolutiontothisproblemwasintroducedbyBertrand Russellin1905.Russellaskedhowweshouldunderstandthissentence: 47.ThepresentkingofFranceisbald.

PAGE 75

ch.4quantiedlogic 75 Thephrase`thepresentkingofFrance'issupposedtopickoutanindividualby meansofadenitedescription.However,therewasnokingofFrancein1905 andthereisnonenow.Sincethedescriptionisanon-referringterm,wecannot justdeneaconstanttomean`thepresentkingofFrance'andtranslatethe sentenceas Kf Russell'sideawasthatsentencesthatcontaindenitedescriptionshaveadifferentlogicalstructurethansentencesthatcontainpropernames,eventhough theysharethesamegrammaticalform.Whatdowemeanwhenweuseanunproblematic,referringdescription,like`thehighestpeakinWashingtonstate'? Wemeanthatthereissuchapeak,becausewecouldnottalkaboutitotherwise.Wealsomeanthatitistheonlysuchpeak.Iftherewasanotherpeak inWashingtonstateofexactlythesameheightasMountRainier,thenMount Rainierwouldnotbe the highestpeak. Accordingtothisanalysis,sentence47issayingthreethings.First,itmakes an existence claim:ThereissomepresentkingofFrance.Second,itmakesa uniqueness claim:ThisguyistheonlypresentkingofFrance.Third,itmakes aclaimof predication :Thisguyisbald. Inordertosymbolizedenitedescriptionsinthisway,weneedtheidentitypredicate.Withoutit,wecouldnottranslatetheuniquenessclaimwhichaccording toRussellisimplicitinthedenitedescription. LettheUDbe peopleactuallyliving ,let Fx mean` x isthepresentkingof France',andlet Bx mean` x isbald.'Sentence47canthenbetranslatedas 9 x Fx & :9 y Fy & x 6 = y & Bx .Thissaysthatthereissomeguywhoisthe presentkingofFrance,heistheonlypresentkingofFrance,andheisbald. Understoodinthisway,sentence47ismeaningfulbutfalse.Itsaysthatthis guyexists,buthedoesnot. Theproblemofnon-referringtermsismostvexingwhenwetrytotranslate negations.Soconsiderthissentence: 48.ThepresentkingofFranceisnotbald. AccordingtoRussell,thissentenceisambiguousinEnglish.Itcouldmean eitheroftwothings: 48a.ItisnotthecasethatthepresentkingofFranceisbald. 48b.ThepresentkingofFranceisnon-bald. Bothpossiblemeaningsnegatesentence47,buttheyputthenegationindierentplaces.

PAGE 76

76 forall x Sentence48aiscalleda wide-scopenegation ,becauseitnegatestheentire sentence.Itcanbetranslatedas :9 x Fx & :9 y Fy & x 6 = y & Bx .Thisdoes notsayanythingaboutthepresentkingofFrance,butrathersaysthatsome sentenceaboutthepresentkingofFranceisfalse.Sincesentence47iffalse, sentence48aistrue. Sentence48bsayssomethingaboutthepresentkingofFrance.Itsaysthathe lacksthepropertyofbaldness.Likesentence47,itmakesanexistenceclaim andauniquenessclaim;itjustdeniestheclaimofpredication.Thisiscalled narrow-scopenegation .Itcanbetranslatedas 9 x Fx & :9 y Fy & x 6 = y & : Bx .SincethereisnopresentkingofFrance,thissentenceisfalse. Russell'stheoryofdenitedescriptionsresolvestheproblemofnon-referring termsandalsoexplainswhyitseemedsoparadoxical.Beforewedistinguished betweenthewide-scopeandnarrow-scopenegations,itseemedthatsentences like48shouldbebothtrueandfalse.Byshowingthatsuchsentencesare ambiguous,Russellshowedthattheyaretrueunderstoodonewaybutfalse understoodanotherway. ForamoredetaileddiscussionofRussell'stheoryofdenitedescriptions,includingobjectionstoit,seethePeterLudlow'sentry`descriptions'in TheStanford EncyclopediaofPhilosophy :Summer2005edition,editedbyEdwardN.Zalta, http://plato.stanford.edu/archives/sum2005/entries/descriptions/ PracticeExercises PartA Identifywhichvariablesareboundandwhicharefree. 1. 9 xLxy & 8 yLyx 2. 8 xAx & Bx 3. 8 x Ax & Bx & 8 y Cx & Dy 4. 8 x 9 y [ Rxy Jz & Kx ] Ryx 5. 8 x 1 Mx 2 $ Lx 2 x 1 & 9 x 2 Lx 3 x 2 ? PartB 1.Identifywhichofthefollowingaresubstitutioninstancesof 8 xRcx : Rac Rca Raa Rcb Rbc Rcc Rcd Rcx 2.Identifywhichofthefollowingaresubstitutioninstancesof 9 x 8 yLxy : 8 yLby 8 xLbx Lab 9 xLxa ? PartC Usingthesymbolizationkeygiven,translateeachEnglish-language sentenceintoQL.

PAGE 77

ch.4quantiedlogic 77 UD: allanimals Ax: x isanalligator. Mx: x isamonkey. Rx: x isareptile. Zx: x livesatthezoo. Lxy: x loves y a: Amos b: Bouncer c: Cleo 1.Amos,Bouncer,andCleoallliveatthezoo. 2.Bouncerisareptile,butnotanalligator. 3.IfCleolovesBouncer,thenBouncerisamonkey. 4.IfbothBouncerandCleoarealligators,thenAmoslovesthemboth. 5.Somereptilelivesatthezoo. 6.Everyalligatorisareptile. 7.Anyanimalthatlivesatthezooiseitheramonkeyoranalligator. 8.Therearereptileswhicharenotalligators. 9.Cleolovesareptile. 10.Bouncerlovesallthemonkeysthatliveatthezoo. 11.AllthemonkeysthatAmosloveslovehimback. 12.Ifanyanimalisanreptile,thenAmosis. 13.Ifanyanimalisanalligator,thenitisareptile. 14.EverymonkeythatCleolovesisalsolovedbyAmos. 15.ThereisamonkeythatlovesBouncer,butsadlyBouncerdoesnotreciprocatethislove. PartD ThesearesyllogisticguresidentiedbyAristotleandhissuccessors, alongwiththeirmedievalnames.TranslateeachargumentintoQL. Barbara All B sare C s.All A sare B s. : : All A sare C s. Baroco All C sare B s.Some A isnot B : : Some A isnot C Bocardo Some B isnot C .All A sare B s. : : Some A isnot C Celantes No B sare C s.All A sare B s. : : No C sare A s. Celarent No B sare C s.All A sare B s. : : No A sare C s. Cemestres No C sare B s.No A sare B s. : : No A sare C s. Cesare No C sare B s.All A sare B s. : : No A sare C s. Dabitis All B sare C s.Some A is B : : Some C is A Darii All B sare C s.Some A is B : : Some A is C .

PAGE 78

78 forall x Datisi All B sare C s.Some A is B : : Some A is C Disamis Some B is C .All A sare B s. : : Some A is C Ferison No B sare C s.Some A is B : : Some A isnot C Ferio No B sare C s.Some A is B : : Some A isnot C Festino No C sare B s.Some A is B : : Some A isnot C Baralipton All B sare C s.All A sare B s. : : Some C is A Frisesomorum Some B is C .No A sare B s. : : Some C isnot A PartE Usingthesymbolizationkeygiven,translateeachEnglish-language sentenceintoQL. UD: allanimals Dx: x isadog. Sx: x likessamuraimovies. Lxy: x islargerthan y b: Bertie e: Emerson f: Fergis 1.Bertieisadogwholikessamuraimovies. 2.Bertie,Emerson,andFergisarealldogs. 3.EmersonislargerthanBertie,andFergisislargerthanEmerson. 4.Alldogslikesamuraimovies. 5.Onlydogslikesamuraimovies. 6.ThereisadogthatislargerthanEmerson. 7.IfthereisadoglargerthanFergis,thenthereisadoglargerthanEmerson. 8.NoanimalthatlikessamuraimoviesislargerthanEmerson. 9.NodogislargerthanFergis. 10.AnyanimalthatdislikessamuraimoviesislargerthanBertie. 11.ThereisananimalthatisbetweenBertieandEmersoninsize. 12.ThereisnodogthatisbetweenBertieandEmersoninsize. 13.Nodogislargerthanitself. 14.Everydogislargerthansomedog. 15.Thereisananimalthatissmallerthaneverydog. 16.Ifthereisananimalthatislargerthananydog,thenthatanimaldoes notlikesamuraimovies. PartF Foreachargument,writeasymbolizationkeyandtranslatetheargumentintoQL.

PAGE 79

ch.4quantiedlogic 79 1.Nothingonmydeskescapesmyattention.Thereisacomputeronmy desk.Assuch,thereisacomputerthatdoesnotescapemyattention. 2.Allmydreamsareblackandwhite.OldTVshowsareinblackandwhite. Therefore,someofmydreamsareoldTVshows. 3.NeitherHolmesnorWatsonhasbeentoAustralia.Apersoncouldseea kangarooonlyiftheyhadbeentoAustraliaortoazoo.AlthoughWatson hasnotseenakangaroo,Holmeshas.Therefore,Holmeshasbeentoa zoo. 4.NooneexpectstheSpanishInquisition.NooneknowsthetroublesI've seen.Therefore,anyonewhoexpectstheSpanishInquisitionknowsthe troublesI'veseen. 5.Anantelopeisbiggerthanabreadbox.Iamthinkingofsomethingthat isnobiggerthanabreadbox,anditiseitheranantelopeoracantaloupe. Assuch,Iamthinkingofacantaloupe. 6.Allbabiesareillogical.Nobodywhoisillogicalcanmanageacrocodile. Bertholdisababy.Therefore,Bertholdisunabletomanageacrocodile. ? PartG Usingthesymbolizationkeygiven,translateeachEnglish-language sentenceintoQL. UD: candies Cx: x haschocolateinit. Mx: x hasmarzipaninit. Sx: x hassugarinit. Tx: Borishastried x Bxy: x isbetterthan y 1.Borishasnevertriedanycandy. 2.Marzipanisalwaysmadewithsugar. 3.Somecandyissugar-free. 4.Theverybestcandyischocolate. 5.Nocandyisbetterthanitself. 6.Borishasnevertriedsugar-freechocolate. 7.Borishastriedmarzipanandchocolate,butnevertogether. 8.Anycandywithchocolateisbetterthananycandywithoutit. 9.Anycandywithchocolateandmarzipanisbetterthananycandythat lacksboth. PartH Usingthesymbolizationkeygiven,translateeachEnglish-language sentenceintoQL. UD: peopleanddishesatapotluck Rx: x hasrunout. Tx: x isonthetable.

PAGE 80

80 forall x Fx: x isfood. Px: x isaperson. Lxy: x likes y e: Eli f: Francesca g: theguacamole 1.Allthefoodisonthetable. 2.Iftheguacamolehasnotrunout,thenitisonthetable. 3.Everyonelikestheguacamole. 4.Ifanyonelikestheguacamole,thenElidoes. 5.Francescaonlylikesthedishesthathaverunout. 6.Francescalikesnoone,andnoonelikesFrancesca. 7.Elilikesanyonewholikestheguacamole. 8.Elilikesanyonewholikesthepeoplethathelikes. 9.Ifthereisapersononthetablealready,thenallofthefoodmusthave runout. ? PartI Usingthesymbolizationkeygiven,translateeachEnglish-language sentenceintoQL. UD: people Dx: x dancesballet. Fx: x isfemale. Mx: x ismale. Cxy: x isachildof y Sxy: x isasiblingof y e: Elmer j: Jane p: Patrick 1.AllofPatrick'schildrenareballetdancers. 2.JaneisPatrick'sdaughter. 3.Patrickhasadaughter. 4.Janeisanonlychild. 5.AllofPatrick'sdaughtersdanceballet. 6.Patrickhasnosons. 7.JaneisElmer'sniece. 8.PatrickisElmer'sbrother. 9.Patrick'sbrothershavenochildren. 10.Janeisanaunt. 11.Everyonewhodancesballethasasisterwhoalsodancesballet. 12.Everymanwhodancesballetisthechildofsomeonewhodancesballet.

PAGE 81

ch.4quantiedlogic 81 PartJ Usingthesymbolizationkeygiven,translateeachEnglish-language sentenceintoQLwithidentity.Thelastsentenceisambiguousandcanbe translatedtwoways;youshouldprovidebothtranslations.Hint:Identityis onlyrequiredforthelastfoursentences. UD: people Kx: x knowsthecombinationtothesafe. Sx: x isaspy. Vx: x isavegetarian. Txy: x trusts y h: Hofthor i: Ingmar 1.Hofthorisaspy,butnovegetarianisaspy. 2.NooneknowsthecombinationtothesafeunlessIngmardoes. 3.Nospyknowsthecombinationtothesafe. 4.NeitherHofthornorIngmarisavegetarian. 5.Hofthortrustsavegetarian. 6.EveryonewhotrustsIngmartrustsavegetarian. 7.EveryonewhotrustsIngmartrustssomeonewhotrustsavegetarian. 8.OnlyIngmarknowsthecombinationtothesafe. 9.IngmartrustsHofthor,butnooneelse. 10.Thepersonwhoknowsthecombinationtothesafeisavegetarian. 11.Thepersonwhoknowsthecombinationtothesafeisnotaspy. ? PartK Usingthesymbolizationkeygiven,translateeachEnglish-language sentenceintoQLwithidentity.Thelasttwosentencesareambiguousandcan betranslatedtwoways;youshouldprovidebothtranslationsforeach. UD: cardsinastandarddeck Bx: x isblack. Cx: x isaclub. Dx: x isadeuce. Jx: x isajack. Mx: x isamanwithanaxe. Ox: x isone-eyed. Wx: x iswild. 1.Allclubsareblackcards. 2.Therearenowildcards. 3.Thereareatleasttwoclubs. 4.Thereismorethanoneone-eyedjack. 5.Thereareatmosttwoone-eyedjacks. 6.Therearetwoblackjacks.

PAGE 82

82 forall x 7.Therearefourdeuces. 8.Thedeuceofclubsisablackcard. 9.One-eyedjacksandthemanwiththeaxearewild. 10.Ifthedeuceofclubsiswild,thenthereisexactlyonewildcard. 11.Themanwiththeaxeisnotajack. 12.Thedeuceofclubsisnotthemanwiththeaxe. PartL Usingthesymbolizationkeygiven,translateeachEnglish-language sentenceintoQLwithidentity.Thelasttwosentencesareambiguousandcan betranslatedtwoways;youshouldprovidebothtranslationsforeach. UD: animalsintheworld Bx: x isinFarmerBrown'seld. Hx: x isahorse. Px: x isaPegasus. Wx: x haswings. 1.Thereareatleastthreehorsesintheworld. 2.Thereareatleastthreeanimalsintheworld. 3.ThereismorethanonehorseinFarmerBrown'seld. 4.TherearethreehorsesinFarmerBrown'seld. 5.ThereisasinglewingedcreatureinFarmerBrown'seld;anyothercreaturesintheeldmustbewingless. 6.ThePegasusisawingedhorse. 7.TheanimalinFarmerBrown'seldisnotahorse. 8.ThehorseinFarmerBrown'selddoesnothavewings.

PAGE 83

Chapter5 Formalsemantics Inthischapter,wedescribea formalsemantics forSLandforQL.Theword `semantics'comesfromthegreekwordfor`mark'andmeans`relatedtomeaning.'Soaformalsemanticswillbeamathematicalaccountofmeaninginthe formallanguage. Aformal,logicallanguageisbuiltfromtwokindsofelements:logicalsymbols andnon-logicalsymbols.Connectiveslike`&'andquantierslike` 8 'are logicalsymbols,becausetheirmeaningisspeciedwithintheformallanguage. Whenwritingasymbolizationkey,youarenotallowedtochangethemeaning ofthelogicalsymbols.Youcannotsay,forinstance,thatthe` : 'symbolwill mean`not'inoneargumentand`perhaps'inanother.The` : 'symbolalways meanslogicalnegation.ItisusedtotranslatetheEnglishlanguageword`not', butitisasymbolofaformallanguageandisdenedbyitstruthconditions. ThesentenceslettersinSLarenon-logicalsymbols,becausetheirmeaningis notdenedbythelogicalstructureofSL.Whenwetranslateanargumentfrom EnglishtoSL,forexample,thesentenceletter M doesnothaveitsmeaning xedinadvance;instead,weprovideasymbolizationkeythatsayshow M shouldbeinterpretedinthatargument.InQL,thepredicatesandconstants arenon-logicalsymbols. IntranslatingfromEnglishtoaformallanguage,weprovidedsymbolization keyswhichwereinterpretationsofallthenon-logicalsymbolsweusedinthe translation.An interpretation givesameaningtoallthenon-logicalelements ofthelanguage. Itispossibletoprovidedierentinterpretationsthatmakenoformaldierence. InSL,forexample,wemightsaythat D means`TodayisTuesday';wemightsay insteadthat D means`TodayisthedayafterMonday.'Thesearetwodierent 83

PAGE 84

84 forall x interpretations,becausetheyusedierentEnglishsentencesforthemeaningof D .Yet,formally,thereisnodierencebetweenthem.Allthatmattersoncewe havesymbolizedthesesentencesiswhethertheyaretrueorfalse.Inorderto characterizewhatmakesadierenceintheformallanguage,weneedtoknow whatmakessentencestrueorfalse.Forthis,weneedaformalcharacterization of truth WhenwegavedenitionsforasentenceofSLandforasentenceofQL,we distinguishedbetweenthe objectlanguage andthe metalanguage .The objectlanguageisthelanguagethatweare talkingabout :eitherSLorQL.The metalanguageisthelanguagethatweusetotalkabouttheobjectlanguage: English,supplementedwithsomemathematicaljargon.Itwillbeimportantto keepthisdistinctioninmind. 5.1SemanticsforSL Thissectionprovidesarigorous,formalcharacterizationof truthinSL which buildsonwhatwealreadyknowfromdoingtruthtables.Wewereabletouse truthtablestoreliablytestwhetherasentencewasatautologyinSL,whether twosentenceswereequivalent,whetheranargumentwasvalid,andsoon.For instance: A isatautologyinSLifitisToneverylineofacompletetruthtable. Thisworkedbecauseeachlineofatruthtablecorrespondstoawaytheworld mightbe.WeconsideredallthepossiblecombinationsofTandFforthe sentencelettersthatmadeadierencetothesentenceswecaredabout.The truthtableallowedustodeterminewhatwouldhappengiventhesedierent combinations. Onceweconstructatruthtable,thesymbols`T'and`F'aredivorcedfromtheir metalinguisticmeaningof`true'and`false'.We interpret `T'asmeaning`true', buttheformalpropertiesofTaredenedbythecharacteristictruthtables forthevariousconnectives.Thetableswouldbethesameifwehadusedthe symbols`1'and`0',andcomputerscanbeprogrammedtollouttruthtables withouthavinganysensethat1meanstrueand0meansfalse. Formally,whatwewantisafunctionthatassignsa1or0toeachofthesentences ofSL.WecaninterpretthisfunctionasadenitionoftruthforSLifitassigns 1toallofthetruesentencesofSLand0toallofthefalsesentencesofSL.Call thisfunction` v 'for`valuation'.Wewant v toabeafunctionsuchthatfor anysentence A v A =1if A istrueand v A =0if A isfalse. RecallthattherecursivedenitionofawforSLhadtwostages:Therststep saidthatatomicsentencessolitarysentencelettersarews.Thesecondstage allowedforwstobeconstructedoutofmorebasicws.Therewereclausesof

PAGE 85

ch.5formalsemantics 85 thedenitionforallofthesententialconnectives.Forexample,if A isaw, then : A isaw. Ourstrategyfordeningthetruthfunction, v ,willalsobeintwosteps.The rststepwillhandletruthforatomicsentences;thesecondstepwillhandle truthforcompoundsentences. TruthinSL HowcanwedenetruthforanatomicsentenceofSL?Consider,forexample, thesentence M .Withoutaninterpretation,wecannotsaywhether M istrue orfalse.Itmightmeananything.Ifweuse M tosymbolize`Themoonorbits theEarth',then M istrue.Ifuse M tosymbolize`Themoonisagiantturnip', then M isfalse. Moreover,thewayyouwoulddiscoverwhetherornot M istruedependson what M means.If M means`Itismonday,'thenyouwouldneedtochecka calendar.If M means`Jupiter'smoonIohassignicantvolcanicactivity,'then youwouldneedtocheckanastronomytext|andastronomersknowbecause theysentsatellitestoobserveIo. WhenwegiveasymbolizationkeyforSL,weprovideaninterpretationofthe sentencelettersthatweuse.ThekeygivesanEnglishlanguagesentenceforeach sentenceletterthatweuse.Inthisway,theinterpretationspecieswhateachof thesentenceletters means .However,thisnotenoughtodeterminewhetheror notthatsentenceistrue.Thesentencesaboutthemoon,forinstance,require thatyouknowsomerudimentaryastronomy.Imagineasmallchildwhobecame convincedthatthemoonisagiantturnip.Shecouldunderstandwhatthe sentence`Themoonisagiantturnip'means,butmistakenlythinkthatitwas true. Consideranotherexample:If M means`Itismorningnow',thenwhetheritis trueornotdependsonwhenyouarereadingthis.Iknowwhatthesentence means,but|sinceIdonotknowwhenyouwillbereadingthis|Idonotknow whetheritistrueorfalse. Soaninterpretationalonedoesnotdeterminewhetherasentenceistrueor false.Truthorfalsitydependsalsoonwhattheworldislike.If M meant`The moonisagiantturnip'andtherealmoonwereagiantturnip,then M would betrue.Toputthepointinageneralway,truthorfalsityisdeterminedbyan interpretation plus awaythattheworldis. INTERPRETATION+STATEOFTHEWORLD= TRUTH/FALSITY

PAGE 86

86 forall x Inprovidingalogicaldenitionoftruth,wewillnotbeabletogiveanaccount ofhowanatomicsentenceismadetrueorfalsebytheworld.Instead,we willintroducea truthvalueassignment .Formally,thiswillbeafunctionthat tellsusthetruthvalueofalltheatomicsentences.Callthisfunction` a 'for `assignment'.Wedene a forallsentenceletters P ,suchthat a P = 1if P istrue ; 0otherwise. Thismeansthat a takesanysentenceofSLandassignsiteitheraoneora zero;oneifthesentenceistrue,zeroifthesentenceisfalse.Thedetailsofthe function a aredeterminedbythemeaningofthesentenceletterstogetherwith thestateoftheworld.If D means`Itisdarkoutside',then a D =1atnight orduringaheavystorm,while a D =0onaclearday. Youcanthinkof a asbeinglikearowofatruthtable.Whereasatruthtable rowassignsatruthvaluetoafewatomicsentences,thetruthvalueassignment assignsavaluetoeveryatomicsentenceofSL.Thereareinnitelymanysentence letters,andthetruthvalueassignmentgivesavaluetoeachofthem.When constructingatruthtable,weonlycareaboutsentencelettersthataectthe truthvalueofsentencesthatinterestus.Assuch,weignoretherest.Strictly speaking,everyrowofatruthtablegivesa partial truthvalueassignment. Itisimportanttonotethatthetruthvalueassignment, a ,isnotpartofthe languageSL.Rather,itispartofthemathematicalmachinerythatweareusing todescribeSL.Itencodeswhichatomicsentencesaretrueandwhicharefalse. Wenowdenethetruthfunction, v ,usingthesamerecursivestructurethatwe usedtodeneawofSL. 1.If A isasentenceletter,then v A = a A 2.If A is : B forsomesentence B ,then v A = 1if v B =0 ; 0otherwise. 3.If A is B & C forsomesentences B ; C ,then v A = 1if v B =1and v C =1, 0otherwise. Itmightseemasifthisdenitioniscircular,becauseitusestheword`and' intryingtodene`and.'Notice,however,thatthisisnotadenitionofthe Englishword`and';itisadenitionoftruthforsentencesofSLcontainingthe logicalsymbol`&.'Wedenetruthforobjectlanguagesentencescontaining thesymbol`&'usingthemetalanguageword`and.'Thereisnothingcircular aboutthat.

PAGE 87

ch.5formalsemantics 87 4.If A is B C forsomesentences B ; C ,then v A = 0if v B =0and v C =0, 1otherwise. 5.If A is B C forsomesentences B ; C ,then v A = 0if v B =1and v C =0, 1otherwise. 6.If A is B $ C forsomesentences B ; C ,then v A = 1if v B = v C ; 0otherwise. Sincethedenitionof v hasthesamestructureasthedenitionofaw,we knowthat v assignsavalueto every wofSL.SincethesentencesofSLand thewsofSLarethesame,thismeansthat v returnsthetruthvalueofevery sentenceofSL. TruthinSLisalwaystruth relativeto sometruthvalueassignment,because thedenitionoftruthforSLdoesnotsaywhetheragivensentenceistrueor false.Rather,itsayshowthetruthofthatsentencerelatestoatruthvalue assignment. OtherconceptsinSL WorkingwithSLsofar,wehavedonewithoutaprecisedenitionof`tautology', `contradiction',andsoon.Truthtablesprovidedawayto checkif asentence wasatautologyinSL,buttheydidnot dene whatitmeanstobeatautology inSL.WewillgivedenitionsoftheseconceptsforSLintermsofentailment. Therelationofsemanticentailment,` A entails B ',meansthatthereisnotruth valueassignmentforwhich A istrueand B isfalse.Putdierently,itmeans that B istrueforanyandalltruthvalueassignmentsforwhich A istrue. Weabbreviatethiswithasymbolcalledthe doubleturnstile : A j = B means` A semanticallyentails B .' Wecantalkaboutentailmentbetweenmorethantwosentences: f A 1 ; A 2 ; A 3 ; gj = B meansthatthereisnotruthvalueassignmentforwhichallofthesentencesin theset f A 1 ; A 2 ; A 3 ; g aretrueand B isfalse.

PAGE 88

88 forall x Wecanalsousethesymbolwithjustonesentence: j = C meansthat C istruefor alltruthvalueassignments.Thisisequivalenttosayingthatthatthesentence isentailedbyanything. Thedoubleturnstilesymbolallowsustogiveconcisedenitionsforvarious conceptsofSL: A tautologyinsl isasentence A suchthat j = A A contradictioninsl isasentence A suchthat j = : A Asentenceis contingentinsl ifandonlyifitisneitheratautologynoracontradiction. Anargument P 1 ; P 2 ; : : C "is validinsl ifandonlyif f P 1 ; P 2 ; gj = C Twosentences A and B are logicallyequivalentinsl ifand onlyifboth A j = B and B j = A Logicalconsistencyissomewhathardertodeneintermsofsemanticentailment.Instead,wewilldeneitinthisway: Theset f A 1 ; A 2 ; A 3 ; g is consistentinsl ifandonlyifthereis atleastonetruthvalueassignmentforwhichallofthesentencesare true.Thesetis inconsistentinsl ifandifonlythereisnosuch assignment. 5.2InterpretationsandmodelsinQL InSL,aninterpretationorsymbolizationkeyspecieswhateachofthesentence lettersmeans.Theinterpretationofasentenceletteralongwiththestateofthe worlddetermineswhetherthesentenceletteristrueorfalse.Sincethebasic unitsaresentenceletters,aninterpretationonlymattersinsofarasitmakes sentenceletterstrueorfalse.Formally,thesemanticsforSLisstrictlyinterms oftruthvalueassignments.Twointerpretationsarethesame,formally,ifthey makeforthesametruthvalueassignment. WhatisaninterpretationinQL?LikeasymbolizationkeyforQL,aninterpretationrequiresaUD,aschematicmeaningforeachofthepredicates,andan objectthatispickedoutbyeachconstant.Forexample: UD: comicbookcharacters Fx: x ghtscrime. b: theBatman

PAGE 89

ch.5formalsemantics 89 w: BruceWayne Considerthesentence Fb .Thesentenceistrueonthisinterpretation,but| justasinSL|thesentenceisnottrue justbecause oftheinterpretation.Most peopleinourcultureknowthatBatmanghtscrime,butthisrequiresamodicumofknowledgeaboutcomicbooks.Thesentence Fb istruebecauseofthe interpretation plus somefactsaboutcomicbooks.Thisisespeciallyobvious whenweconsider Fw .BruceWayneisthesecretidentityoftheBatmanin thecomicbooks|theidentityclaim b = w istrue|so Fw istrue.Sinceitis a secret identity,however,othercharactersdonotknowthat Fw istrueeven thoughtheyknowthat Fb istrue. Wecouldtrytocharacterizethisasatruthvalueassignment,aswedidforSL. Thetruthvalueassignmentwouldassign0or1toeachatomicw: Fb Fw andsoon.Ifweweretodothat,however,wemightjustaswelltranslatethe sentencesfromQLtoSLbyreplacing Fb and Fw withsentenceletters.We couldthenrelyonthedenitionoftruthforSL,butatthecostofignoringall thelogicalstructureofpredicatesandterms.Inwritingasymbolizationkey forQL,wedonotgiveseparatedenitionsfor Fb and Fw .Instead,wegive meaningsto F b ,and w .Thisisessentialbecausewewanttobeabletouse quantiers.Thereisnoadequatewaytotranslate 8 xFx intoSL. Sowewantaformalcounterparttoaninterpretationforpredicatesandconstants,notjustforsentences.Wecannotuseatruthvalueassignmentforthis, becauseapredicateisneithertruenorfalse.Intheinterpretationgivenabove, F istrue of theBatmani.e., Fb istrue,butitmakesnosenseatalltoask whether F onitsownistrue.ItwouldbelikeaskingwhethertheEnglish languagefragment` ::: ghtscrime'istrue. Whatdoesaninterpretationdoforapredicate,ifitdoesnotmakeittrueor false?Aninterpretationhelpstopickouttheobjectstowhichthepredicate applies.Interpreting Fx tomean` x ghtscrime'picksoutBatman,Superman, Spiderman,andotherheroesasthethingsthatare F s.Formally,thisisa setofmembersoftheUDtowhichthepredicateapplies;thissetiscalledthe extension ofthepredicate. Manypredicateshaveindenitelylargeextensions.Itwouldbeimpractical totryandwritedownallofthecomicbookcrimeghtersindividually,so insteadweuseanEnglishlanguageexpressiontointerpretthepredicate.This issomewhatimprecise,becausetheinterpretationalonedoesnottellyouwhich membersoftheUDareintheextensionofthepredicate.Inordertogure outwhetheraparticularmemberoftheUDisintheextensionofthepredicate togureoutwhetherBlackLightningghtscrime,forinstance,youneedto knowaboutcomicbooks.Ingeneral,theextensionofapredicateistheresult ofaninterpretation alongwith somefacts.

PAGE 90

90 forall x Sometimesitispossibletolistallofthethingsthatareintheextensionofa predicate.InsteadofwritingaschematicEnglishsentence,wecanwritedown theextensionasasetofthings.Supposewewantedtoaddaone-placepredicate M tothekeyabove.Wewant Mx tomean` x livesinWayneManor',sowe writetheextensionasasetofcharacters: extension M = f BruceWayne,Alfredthebutler,DickGrayson g Youdonotneedtoknowanythingaboutcomicbookstobeabletodetermine that,onthisinterpretation, Mw istrue:BruceWayneisjustspeciedtobeone ofthethingsthatis M .Similarly, 9 xMx isobviouslytrueonthisinterpretation: ThereisatleastonememberoftheUDthatisan M |infact,therearethree ofthem. Whataboutthesentence 8 xMx ?Thesentenceisfalse,becauseitisnottrue thatallmembersoftheUDare M .Itrequiresthebarestminimumofknowledge aboutcomicbookstoknowthatthereareothercharactersbesidesjustthese three.Althoughwespeciedtheextensionof M inaformallypreciseway,we stillspeciedtheUDwithanEnglishlanguagedescription.Formallyspeaking, aUDisjustasetofmembers. Theformalsignicanceofapredicateisdeterminedbyitsextension,butwhat shouldwesayaboutconstantslike b and w ?ThemeaningofaconstantdetermineswhichmemberoftheUDispickedoutbytheconstant.Theindividual thattheconstantpicksoutiscalledthe referent oftheconstant.Both b and w havethesamereferent,sincetheybothrefertothesamecomicbook character.Youcanthinkofaconstantletterasanameandthereferentasthe thingnamed.InEnglish,wecanusethedierentnames`Batman'and`Bruce Wayne'torefertothesamecomicbookcharacter.Inthisinterpretation,we canusethedierentconstants` b 'and` w 'torefertothesamememberofthe UD. Sets Weusecurlybrackets` f 'and` g 'todenotesets.Themembersofthesetcanbe listedinanyorder,separatedbycommas.Thefactthatsetscanbeinanyorder isimportant,becauseitmeansthat f foo,bar g and f bar,foo g arethesameset. Itispossibletohaveasetwithnomembersinit.Thisiscalledthe empty set .Theemptysetissometimeswrittenas fg ,butusuallyitiswrittenasthe singlesymbol ; .

PAGE 91

ch.5formalsemantics 91 Models Aswehaveseen,aninterpretationinQLisonlyformallysignicantinsofaras itdeterminesaUD,anextensionforeachpredicate,andareferentforeach constant.Wecallthisformalstructurea model forQL. Toseehowthisworks,considerthissymbolizationkey: UD: PeoplewhoplayedaspartoftheThreeStooges Hx: x hadheadhair. f: MisterFine IfyoudonotknowanythingabouttheThreeStooges,youwillnotbeableto saywhichsentencesofQLaretrueonthisinterpretation.Perhapsyoujust rememberLarry,Curly,andMoe.Isthesentence Hf trueorfalse?Itdepends onwhichofthestoogesisMisterFine. Whatisthemodelthatcorrespondstothisinterpretation?Thereweresix peoplewhoplayedaspartoftheThreeStoogesovertheyears,sotheUD willhavesixmembers:LarryFine,MoeHoward,andCurlyHoward,Shemp Howard,JoeBesser,andCurlyJoeDeRita.Curly,Joe,andCurlyJoewerethe onlycompletelybaldstooges.Theresultisthismodel: UD= f Larry,Curly,Moe,Shemp,Joe,CurlyJoe g extension H = f Larry,Moe,Shemp g referent f =Larry YoudonotneedtoknowanythingabouttheThreeStoogesinordertoevaluate whethersentencesaretrueorfalseinthis model Hf istrue,sincethereferent of f Larryisintheextensionof H .Both 9 xHx and 9 x : Hx aretrue,since thereisatleastonememberoftheUDthatisintheextensionof H andatleast onememberthatisnotintheextensionof H .Inthisway,themodelcaptures alloftheformalsignicanceoftheinterpretation. Nowconsiderthisinterpretation: UD:wholenumberslessthan10 Ex: x iseven. Nx: x isnegative. Lxy: x islessthan y Txyz: x times y equals z Whatisthemodelthatgoeswiththisinterpretation?TheUDistheset f 1 ; 2 ; 3 ; 4 ; 5 ; 6 ; 7 ; 8 ; 9 g .

PAGE 92

92 forall x Theextensionofaone-placepredicatelike E or N isjustthesubsetoftheUD ofwhichthepredicateistrue.Roughlyspeaking,theextensionofthepredicate E isthesetof E sintheUD.Theextensionof E isthesubset f 2 ; 4 ; 6 ; 8 g .There aremanyevennumbersbesidesthesefour,butthesearetheonlymembersof theUDthatareeven.TherearenonegativenumbersintheUD,so N hasan emptyextension;i.e.extension N = ; Theextensionofatwo-placepredicatelike L issomewhatvexing.Itseemsasif theextensionof L oughttocontain1,since1islessthanalltheothernumbers; itoughttocontain2,since2islessthanalloftheothernumbersbesides1;and soon.EverymemberoftheUDbesides9islessthansomememberoftheUD. Whatwouldhappenifwejustwroteextension L = f 1 ; 2 ; 3 ; 4 ; 5 ; 6 ; 7 ; 8 g ? Theproblemisthatsetscanbewritteninanyorder,sothiswouldbethesame aswritingextension L = f 8 ; 7 ; 6 ; 5 ; 4 ; 3 ; 2 ; 1 g .Thisdoesnottelluswhichof themembersofthesetarelessthanwhichothermembers. Weneedsomewayofshowingthat1islessthan8butthat8isnotlessthan 1.Thesolutionistohavetheextensionof L consistofpairsofnumbers. An orderedpair islikeasetwithtwomembers,exceptthattheorder does matter.Wewriteorderedpairswithanglebrackets` < 'and` > '.Theordered pair < foo,bar > isdierentthantheorderedpair < bar,foo > .Theextension of L isacollectionoforderedpairs,allofthepairsofnumbersintheUDsuch thattherstnumberislessthanthesecond.Writingthisoutcompletely: extension L = f < 1,2 > < 1,3 > < 1,4 > < 1,5 > < 1,6 > < 1,7 > < 1,8 > < 1,9 > < 2,3 > < 2,4 > < 2,5 > < 2,6 > < 2,7 > < 2,8 > < 2,9 > < 3,4 > < 3,5 > < 3,6 > < 3,7 > < 3,8 > < 3,9 > < 4,5 > < 4,6 > < 4,7 > < 4,8 > < 4,9 > < 5,6 > < 5,7 > < 5,8 > < 5,9 > < 6,7 > < 6,8 > < 6,9 > < 7,8 > < 7,9 > < 8,9 > g Three-placepredicateswillworksimilarly;theextensionofathree-placepredicateisasetoforderedtripleswherethepredicateistrueofthosethreethings inthatorder .Sotheextensionof T inthismodelwillcontainorderedtriples like < 2,4,8 > ,because2 4=8. Generally,theextensionofann-placepredicateisasetofallorderedn-tuples suchthat a 1 { a n aremembersoftheUDandthepredicateis trueof a 1 { a n inthatorder. 5.3Semanticsforidentity IdentityisaspecialpredicateofQL.Wewriteitabitdierentlythanother two-placepredicates: x = y insteadof Ixy .Wealsodonotneedtoincludeit

PAGE 93

ch.5formalsemantics 93 inasymbolizationkey.Thesentence x = y alwaysmeans` x isidenticalto y ,' anditcannotbeinterpretedtomeananythingelse.Inthesameway,whenyou constructamodel,youdonotgettopickandchoosewhichorderedpairsgo intotheextensionoftheidentitypredicate.Italwayscontainsjusttheordered pairofeachobjectintheUDwithitself. Thesentence 8 xIxx ,whichcontainsanordinarytwo-placepredicate,iscontingent.Whetheritistrueforaninterpretationdependsonhowyouinterpret I andwhetheritistrueinamodeldependsontheextensionof I Thesentence 8 xx = x isatautology.Theextensionofidentitywillalways makeittrue. Noticethatalthoughidentityalwayshasthesameinterpretation,itdoesnot alwayshavethesameextension.TheextensionofidentitydependsontheUD. IftheUDinamodelistheset f Doug g ,thenextension=inthatmodelis f < Doug,Doug > g .IftheUDistheset f Doug,Omar g ,thenextension=in thatmodelis f < Doug,Doug > < Omar,Omar > g .Andsoon. Ifthereferentoftwoconstantsisthesame,thenanythingwhichistrueofone istrueoftheother.Forexample,ifreferent a =referent b ,then Aa $ Ab Ba $ Bb Ca $ Cb Rca $ Rcb 8 xRxa $8 xRxb ,andsoonforanytwo sentencescontaining a and b .However,thereverseisnottrue. Itispossiblethatanythingwhichistrueof a isalsotrueof b ,yetfor a and b still tohavedierentreferents.Thismayseempuzzling,butitiseasytoconstruct amodelthatshowsthis.Considerthismodel: UD= f Rosencrantz,Guildenstern g referent a =Rosencrantz referent b =Guildenstern forallpredicates P ,extension P = ; extension== f < Rosencrantz,Rosencrantz > < Guildenstern,Guildenstern > g ThisspeciesanextensionforeverypredicateofQL:Alltheinnitely-many predicatesareempty.Thismeansthatboth Aa and Ab arefalse,andtheyare equivalent;both Ba and Bb arefalse;andsoonforanytwosentencesthatcontain a and b .Yet a and b refertodierentthings.Wehavewrittenouttheextensionofidentitytomakethisclear:Theorderedpair < referent a ; referent b > isnotinit.Inthismodel, a = b isfalseand a 6 = b istrue.

PAGE 94

94 forall x 5.4Workingwithmodels WewillusethedoubleturnstilesymbolforQLmuchaswedidforSL.` A j = B meansthat` A entails B ':When A and B aretwosentencesofQL, A j = B meansthatthereisnomodelinwhich A istrueand B isfalse. j = A means that A istrueineverymodel. ThisallowsustogivedenitionsforvariousconceptsinQL.Becauseweare usingthesamesymbol,thesedenitionswilllooksimilartothedenitionsin SL.Remember,however,thatthedenitionsinQLareintermsof models rather thanintermsoftruthvalueassignments. A tautologyinql isasentence A thatistrueineverymodel; i.e., j = A A contradictioninql isasentence A thatisfalseineverymodel; i.e., j = : A Asentenceis contingentinql ifandonlyifitisneitheratautologynoracontradiction. Anargument P 1 ; P 2 ; : : C "is validinql ifandonlyifthere isnomodelinwhichallofthepremisesaretrueandtheconclusion isfalse;i.e., f P 1 ; P 2 ; gj = C .Itis invalidinql otherwise. Twosentences A and B are logicallyequivalentinsl ifand onlyifboth A j = B and B j = A Theset f A 1 ; A 2 ; A 3 ; g is consistentinql ifandonlyifthereis atleastonemodelinwhichallofthesentencesaretrue.Thesetis inconsistentinql ifandifonlythereisnosuchmodel. Constructingmodels Supposewewanttoshowthat 8 xAxx Bd is not atautology.Thisrequires showingthatthesentenceisnottrueineverymodel;i.e.,thatitisfalseinsome model.Ifwecanprovidejustonemodelinwhichthesentencefalse,thenwe willhaveshownthatthesentenceisnotatautology. Whatwouldsuchamodellooklike?Inorderfor 8 xAxx Bd tobefalse,the antecedent 8 xAxx mustbetrue,andtheconsequent Bd mustbefalse. Toconstructsuchamodel,westartwithaUD.Itwillbeeasiertospecify extensionsforpredicatesifwehaveasmallUD,sostartwithaUDthathas justonemember.Formally,thissinglemembermightbeanything.Let'ssayit isthecityofParis.

PAGE 95

ch.5formalsemantics 95 Wewant 8 xAxx tobetrue,sowewantallmembersoftheUDtobepaired withthemselfintheextensionof A ;thismeansthattheextensionof A must be f < Paris,Paris > g Wewant Bd tobefalse,sothereferentof d mustnotbeintheextensionof B Wegive B anemptyextension. SinceParisistheonlymemberoftheUD,itmustbethereferentof d .The modelwehaveconstructedlookslikethis: UD= f Paris g extension A = f < Paris,Paris > g extension B = ; referent d =Paris Strictlyspeaking,amodelspeciesanextensionfor every predicateofQLand areferentfor every constant.Assuch,itisgenerallyimpossibletowritedown acompletemodel.Thatwouldrequirewritingdowninnitelymanyextensions andinnitelymanyreferents.However,wedonotneedtoconsidereverypredicateinordertoshowthattherearemodelsinwhich 8 xAxx Bd isfalse. Predicateslike H andconstantslike f 13 makenodierencetothetruthor falsityofthissentence.Itisenoughtospecifyextensionsfor A and B anda referentfor d ,aswehavedone.Thisprovidesa partialmodel inwhichthe sentenceisfalse. Perhapsyouarewondering:Whatdoesthepredicate A meaninEnglish?The partialmodelcouldcorrespondtoaninterpretationlikethisone: UD: Paris Axy : x isinthesamecountryas y Bx : x wasfoundedinthe20thcentury. d : theCityofLights However,allthatthepartialmodeltellsusisthat A isapredicatewhichistrue ofParisandParis.ThereareindenitelymanypredicatesinEnglishthathave thisextension. Axy mightinsteadtranslate` x isthesamesizeas y 'or` x and y arebothcities.'Similarly, Bx issomepredicatethatdoesnotapplytoParis;it mightinsteadtranslate` x isonanisland'or` x isasubcompactcar.'Whenwe specifytheextensionsof A and B ,wedonotspecifywhatEnglishpredicates A and B shouldbeusedtotranslate.Weareconcernedwithwhetherthe 8 xAxx Bd comesouttrueorfalse,andallthatmattersfortruthandfalsity inQListheinformationinthemodel:theUD,theextensionsofpredicates, andthereferentsofconstants. Wecanjustaseasilyshowthat 8 xAxx Bd isnotacontradiction.Weneed onlyspecifyamodelinwhich 8 xAxx Bd istrue;i.e.,amodelinwhicheither 8 xAxx isfalseor Bd istrue.Hereisonesuchpartialmodel:

PAGE 96

96 forall x UD= f Paris g extension A = f < Paris,Paris > g extension B = f Paris g referent d =Paris Wehavenowshownthat 8 xAxx Bd isneitheratautologynoracontradiction.Bythedenitionof`contingentinQL,'thismeansthat 8 xAxx Bd is contingent.Ingeneral,showingthatasentenceiscontingentwillrequiretwo models:oneinwhichthesentenceistrueandanotherinwhichthesentenceis false. Supposewewanttoshowthat 8 xSx and 9 xSx arenotlogicallyequivalent. Weneedtoconstructamodelinwhichthetwosentenceshavedierenttruth values;wewantoneofthemtobetrueandtheothertobefalse.Westartby specifyingaUD.Again,wemaketheUDsmallsothatwecanspecifyextensions easily.Wewillneedatleasttwomembers.LettheUDbe f Duke,Miles g .If wechoseaUDwithonlyonemember,thetwosentenceswouldendupwith thesametruthvalue.Inordertoseewhy,tryconstructingsomepartialmodels withone-memberUDs. Wecanmake 9 xSx truebyincludingsomethingintheextensionof S ,andwe canmake 8 xSx falsebyleavingsomethingoutoftheextensionof S .Itdoes notmatterwhichoneweincludeandwhichoneweleaveout.MakingDukethe only S ,wegetapartialmodelthatlookslikethis: UD= f Duke,Miles g extension S = f Duke g Thispartialmodelshowsthatthetwosentencesare not logicallyequivalent. Backonp.66,wesaidthatthisargumentwouldbeinvalidinQL: Rc & K 1 c & Tc : :Tc & K 2 c Inordertoshowthatitisinvalid,weneedtoshowthatthereissomemodelin whichthepremisesaretrueandtheconclusionisfalse.Wecanconstructsuch amodeldeliberately.Hereisonewaytodoit: UD= f Bjork g extension T = f Bjork g extension K 1 = f Bjork g extension K 2 = ; extension R = f Bjork g referent c =Bjork Similarly,wecanshowthatasetofsentencesisconsistentbyconstructinga

PAGE 97

ch.5formalsemantics 97 Table5.1:Itisrelativelyeasytoansweraquestionifyoucandoitbyconstructingamodelortwo.Itismuchharderifyouneedtoreasonaboutallpossible models.Thistableshowswhenconstructingmodelsisenough. YESNO Is A atautology?showthat A mustbe trueinanymodel constructamodel in which A isfalse Is A acontradiction?showthat A mustbe falseinanymodel constructamodel in which A istrue Is A contingent? constructtwomodels oneinwhich A istrue andanotherinwhich A isfalse eithershowthat A isa tautologyorshowthat A isacontradiction Are A and B equivalent? showthat A and B musthavethesame truthvalueinany model constructamodel in which A and B have dierenttruthvalues Istheset A consistent? constructamodel in whichallthesentences in A aretrue showthatthesentencescouldnotallbe trueinanymodel Istheargument ` P : : C 'valid? showthatanymodelin which P istruemust beamodelinwhich C istrue constructamodel in which P istrueand C isfalse modelinwhichallofthesentencesaretrue. Reasoningaboutallmodels Wecanshowthatasentenceis not atautologyjustbyprovidingonecarefully speciedmodel:amodelinwhichthesentenceisfalse.Toshowthatsomething isatautology,ontheotherhand,itwouldnotbeenoughtoconstructten,one hundred,orevenathousandmodelsinwhichthesentenceistrue.Itisonly atautologyifitistruein every model,andthereareinnitelymanymodels. Thiscannotbeavoidedjustbyconstructingpartialmodels,becausethereare innitelymanypartialmodels. Consider,forexample,thesentence Raa $ Raa .Therearetwologicallydistinct partialmodelsofthissentencethathavea1-memberUD.Thereare32distinct partialmodelsthathavea2-memberUD.Thereare1526distinctpartialmodels thathavea3-memberUD.Thereare262,144distinctpartialmodelsthathave

PAGE 98

98 forall x a4-memberUD.Andsoontoinnity.Inordertoshowthatthissentenceis atautology,weneedtoshowsomethingaboutallofthesemodels.Thereisno hopeofdoingsobydealingwiththemoneatatime. Nevertheless, Raa $ Raa isobviouslyatautology.Wecanproveitwitha simpleargument: Therearetwokindsofmodels:thoseinwhich < referent a ; referent a > isintheextensionof R andthoseinwhichitisnot.Intherst kindofmodel, Raa istrue;bythetruthtableforthebiconditional, Raa $ Raa isalsotrue.Inthesecondkindofmodel, Raa isfalse; thismakes Raa $ Raa true.Sincethesentenceistrueinbothkinds ofmodel,andsinceeverymodelisoneofthetwokinds, Raa $ Raa istrueineverymodel.Therefore,itisatautology. Thisargumentisvalid,ofcourse,anditsconclusionistrue.However,itisnot anargumentinQL.Rather,itisanargumentinEnglish about QL;itisan argumentinthemetalanguage.Thereisnoformalprocedureforevaluating orconstructingnaturallanguageargumentslikethisone.Theimprecisionof naturallanguageistheveryreasonwebeganthinkingaboutformallanguages. Therearefurtherdicultieswiththisapproach. Considerthesentence 8 x Rxx Rxx ,anotherobvioustautology.Itmight betemptingtoreasoninthisway:` Rxx Rxx istrueineverymodel,so 8 x Rxx Rxx mustbetrue.'Theproblemisthat Rxx Rxx is not truein everymodel.Itisnotasentence,andsoitis neither true nor false.Wedonot yethavethevocabularytosaythatwewanttosayabout Rxx Rxx .Inthe nextsection,weintroducetheconceptof satisfaction ;afterdoingso,wewillbe betterabletoprovideanargumentthat 8 x Rxx Rxx isatautology. Itisnecessarytoreasonaboutaninnityofmodelstoshowthatasentence isatautology.Similarly,itisnecessarytoreasonaboutaninnityofmodels toshowthatasentenceisacontradition,thattwosentencesareequivalent, thatasetofsentencesisinconsistent,orthatanargumentisvalid.Thereare otherthingswecanshowbycarefullyconstructingamodelortwo.Table5.1 summarizeswhichthingsarewhich. 5.5TruthinQL ForSL,wesplitthedenitionoftruthintotwoparts:atruthvalueassignment a forsentencelettersandatruthfunction v forallsentences.Thetruth functioncoveredthewaythatcomplexsentencescouldbebuiltoutofsentence lettersandconnectives.

PAGE 99

ch.5formalsemantics 99 InthesamewaythattruthforSLisalways truthgivenatruthvalueassignment truthforQLis truthinamodel .ThesimplestatomicsentenceofQLconsists ofaone-placepredicatefollowedbyaconstant,like Pj .Itistrueinamodel M ifandonlyifthereferentof j isintheextensionof P in M Wecouldgooninthiswaytodenetruthforallatomicsentencesthatcontain onlypredicatesandconstants:Consideranysentenceoftheform Rc 1 ::: c n where R isann-placepredicateandthe c sareconstants.Itistruein M ifand onlyif < referent c 1 ;:::; referent c n > isinextension R in M Wecouldthendenetruthforsentencesbuiltupwithsententialconnectivesin thesamewaywedidforSL.Forexample,thesentence Pj Mda istruein M ifeither Pj isfalsein M or Mda istruein M Unfortunately,thisapproachwillfailwhenweconsidersentencescontaining quantiers.Consider 8 xPx .Whenisittrueinamodel M ?Theanswercannot dependonwhether Px istrueorfalsein M ,becausethe x in Px isafree variable. Px isnotasentence.Itisneithertruenorfalse. WewereabletogivearecursivedenitionoftruthforSLbecauseeverywellformedformulaofSLhasatruthvalue.ThisisnottrueinQL,sowecannot denetruthbystartingwiththetruthofatomicsentencesandbuildingup.We alsoneedtoconsidertheatomicformulaewhicharenotsentences.Inorder todothiswewilldene satisfaction ;everywell-formedformulaofQLwillbe satisedornotsatised,evenifitdoesnothaveatruthvalue.Wewillthenbe abletodene truth forsentencesofQLintermsofsatisfaction. Satisfaction Theformula Px says,roughly,that x isoneofthe P s.Thiscannotbequite right,however,because x isavariableandnotaconstant.Itdoesnotnameany particularmemberoftheUD.Instead,itsmeaninginasentenceisdetermined bythequantierthatbindsit.Thevariable x muststand-inforeverymember oftheUDinthesentence 8 xPx ,butitonlyneedstostand-inforonemember in 9 xPx .Sincewewantthedenitionofsatisfactiontocover Px withoutany quantierwhatsoever,wewillstartbysayinghowtointerpretafreevariable likethe x in Px Wedothisbyintroducinga variableassignment .Formally,thisisafunction thatmatchesupeachvariablewithamemberoftheUD.Callthisfunction `a.'The`a'isfor`assignment',butthisisnotthesameasthetruthvalue assignmentthatweusedindeningtruthforSL. Theformula Px issatisedinamodel M byavariableassignment a ifandonly if a x ,theobjectthat a assignsto x ,isinthetheextensionofPin M .

PAGE 100

100 forall x Whenis 8 xPx satised?Itisnotenoughif Px issatisedin M by a ,because thatjustmeansthat a x isinextension P 8 xPx requiresthateveryother memberoftheUDbeinextension P aswell. Soweneedanotherbitoftechnicalnotation:ForanymemberoftheUDand anyvariable x ,let a [ j x ]bethevariableassignmentthatassignsto x but agreeswith a inallotherrespects.Wehaveused,theGreekletterOmega, tounderscorethefactthatitissomememberoftheUDandnotsomesymbol ofQL.Suppose,forexample,thattheUDispresidentsoftheUnitedStates. Thefunction a [GroverCleveland j x ]assignsGroverClevelandtothevariable x regardlessofwhat a assignsto x ;foranyothervariable, a [GroverCleveland j x ] agreeswith a Wecannowsayconciselythat 8 xPx issatisedinamodel M byavariable assignment a ifandonlyif,foreveryobjectintheUDof M Px issatised in M by a [ j x ]. Youmayworrythatthisiscircular,becauseitgivesthesatisfactionconditions forthesentence 8 xPx usingthephrase`foreveryobject.'However,itisimportanttorememberthedierencebetweenalogicalsymbollike` 8 'andanEnglish languagewordlike`every.'Thewordispartofthemetalanguagethatweusein deningsatisfactionconditionsforobjectlanguagesentencesthatcontainthe symbol. Wecannowgiveageneraldenitionofsatisfaction,extendingfromthecaseswe havealreadydiscussed.Wedeneafunction s for`satisfaction'inamodel M suchthatforanyw A andvariableassignment a s A ;a =1if A issatised in M by a ;otherwise s A ;a =0. 1.If A isanatomicwoftheform Pt 1 ::: t n and i istheobjectpickedout by t i ,then s A ;a = 1if < 1 ::: n > isinextension P in M ; 0otherwise. Foreachterm t i :If t i isaconstant,then i =referent t i .If t i isa variable,then i = a t i 2.If A is : B forsomew B ,then s A ;a = 1if s B ;a =0 ; 0otherwise. 3.If A is B & C forsomews B ; C ,then s A ;a = 1if s B ;a =1and s C ;a =1, 0otherwise.

PAGE 101

ch.5formalsemantics 101 4.If A is B C forsomews B ; C ,then s A ;a = 0if s B ;a =0and s C ;a =0, 1otherwise. 5.If A is B C forsomews B ; C ,then s A ;a = 0if s B ;a =1and s C ;a =0, 1otherwise. 6.If A is B $ C forsomesentences B ; C ,then s A ;a = 1if s B ;a = s C ;a ; 0otherwise. 7.If A is 8 xB forsomew B andsomevariable x ,then s A ;a = 1if s B ;a [ j x ]=1foreverymemberoftheUD ; 0otherwise. 8.If A is 9 xB forsomew B andsomevariable x ,then s A ;a = 1if s B ;a [ j x ]=1foratleastonememberoftheUD ; 0otherwise. ThisdenitionfollowsthesamestructureasthedenitionofawforQL,so weknowthateverywofQLwillbecoveredbythisdenition.Foramodel M andavariableassignment a ,anywwilleitherbesatisedornot.Nowsare leftoutorassignedconictingvalues. Truth Considerasimplesentencelike 8 xPx .Bypart7inthedenitionofsatisfaction, thissentenceissatisedif a [ j x ]satises Px in M foreveryintheUD.By part1ofthedenition,thiswillbethecaseifeveryisintheextension of P .Whether 8 xPx issatiseddoesnotdependontheparticularvariable assignment a .Ifthissentenceissatised,thenitistrue.Thisisaformalization ofwhatwehavesaidallalong: 8 xPx istrueifeverythingintheUDisinthe extensionof P ThesamethingholdsforanysentenceofQL.Becauseallofthevariablesare bound,asentenceissatisedornotregardlessofthedetailsofthevariable assignment.Sowecandenetruthinthisway:Asentence A is truein M ifandonlyifsomevariableassignmentsatises A in M ; A is falsein M otherwise.

PAGE 102

102 forall x TruthinQLis truthinamodel .SentencesofQLarenotat-footedlytrue orfalseasmeresymbols,butonlyrelativetoamodel.Amodelprovidesthe meaningofthesymbols,insofarasitmakesanydierencetotruthandfalsity. Reasoningaboutallmodelsreprise Attheendofsection5.4,wewerestymiedwhenwetriedtoshowthat 8 x Rxx Rxx isatautology.Havingdenedsatisfaction,wecannowreasoninthisway: Considersomearbitrarymodel M .NowconsideranarbitrarymemberoftheUD;forthesakeofconvenience,callit.Itmustbe thecaseeitherthat < ; > isintheextensionof R orthatitis not.If < ; > isintheextensionof R ,then Rxx issatisedbya variableassignmentthatassignsto x bypart1ofthedenition ofsatisfaction;sincetheconsequentof Rxx Rxx issatised, theconditionalissatisedbypart5.If < ; > isnotinthe extensionof R ,then Rxx isnotsatisedbyavariableassignment thatassignsto x bypart1;sinceantecedentof Rxx Rxx is notsatised,theconditionalissatisedbypart5.Ineithercase, Rxx Rxx issatised.ThisistrueforanymemberoftheUD,so 8 x Rxx Rxx issatisedbyanytruthvalueassignmentbypart 7.So 8 x Rxx Rxx istruein M bythedenitionoftruth. ThisargumentholdsregardlessoftheexactUDandregardlessof theexactextensionof R ,so 8 x Rxx Rxx istrueinanymodel. Therefore,itisatautology. Givingargumentsaboutallpossiblemodelstypicallyrequiresclevercombinationoftwostrategies: 1.Dividecasesbetweentwopossiblekinds,suchthateverycasemustbeone kindortheother.Intheargumentonp.98,forexample,wedistinguished twokindsofmodelsbasedonwhetherornotaspecicorderedpairwasin extension R .Intheargumentabove,wedistinguishedcasesinwhichanorderedpairwasinextension R andcasesinwhichitwasnot. 2.Consideranarbitraryobjectasawayofshowingsomethingmoregeneral.In theargumentabove,itwascrucialthatwasjustsomearbitrarymemberof theUD.Wedidnotassumeanythingspecialabout.Assuch,whateverwecould showtoholdofmustholdofeverymemberoftheUD|ifwecouldshow itfor,wecouldshowitforanything.Inthesameway,wedidnotassume anythingspecialabout M ,andsowhateverwecouldshowabout M musthold forallmodels.

PAGE 103

ch.5formalsemantics 103 Consideronemoreexample.Theargument 8 x Hx & Jx : : 8 xHx isobviously valid.Wecanonlyshowthattheargumentisvalidbyconsideringwhatmust betrueineverymodelinwhichthepremiseistrue. Consideranarbitrarymodel M inwhichthepremise 8 x Hx & Jx istrue.Theconjunction Hx & Jx issatisedregardlessofwhatis assignedto x ,so Hx mustbealsobypart3ofthedenitionof satisfaction.Assuch, 8 x Hx issatisedbyanyvariableassignmentbypart7ofthedenitionofsatisfactionandtruein M by thedenitionoftruth.Sincewedidnotassumeanythingabout M besides 8 x Hx & Jx beingtrue, 8 x Hx mustbetrueinanymodel inwhich 8 x Hx & Jx istrue.So 8 x Hx & Jx j = 8 xHx Evenforasimpleargumentlikethisone,thereasoningissomewhatcomplicated. Forlongerarguments,thereasoningcanbeinsuerable.Theproblemarises becausetalkingaboutaninnityofmodelsrequiresreasoningthingsoutin English.Whatarewetodo? Wemighttrytoformalizeourreasoningaboutmodels,codifyingthedivide-andconquerstrategiesthatweusedabove.Thisapproach,originallycalled semantic tableaux ,wasdevelopedinthe1950sbyEvertBethandJaakkoHintikka.Their tableauxarenowmorecommonlycalled truthtrees Amoretraditionalapproachistoconsiderdeductiveargumentsasproofs.A proofsystem consistsofrulesthatformallydistinguishbetweenlegitimateand illegitimatearguments|withoutconsideringmodelsorthemeaningsofthe symbols.Inthenextchapter,wedevelopproofsystemsforSLandQL. PracticeExercises ? PartA Determinewhethereachsentenceistrueorfalseinthemodelgiven. UD= f Corwin,Benedict g extension A = f Corwin,Benedict g extension B = f Benedict g extension N = ; referent c =Corwin 1. Bc 2. Ac $: Nc 3. Nc Ac Bc 4. 8 xAx 5. 8 x : Bx

PAGE 104

104 forall x 6. 9 x Ax & Bx 7. 9 x Ax Nx 8. 8 x Nx _: Nx 9. 9 xBx !8 xAx ? PartB Determinewhethereachsentenceistrueorfalseinthemodelgiven. UD= f Waylan,Willy,Johnny g extension H = f Waylan,Willy,Johnny g extension W = f Waylan,Willy g extension R = f < Waylan,Willy > < Willy,Johnny > < Johnny,Waylan > g referent m =Johnny 1. 9 x Rxm & Rmx 2. 8 x Rxm Rmx 3. 8 x Hx $ Wx 4. 8 x Rxm Wx 5. 8 x Wx Hx & Wx 6. 9 xRxx 7. 9 x 9 yRxy 8. 8 x 8 yRxy 9. 8 x 8 y Rxy Ryx 10. 8 x 8 y 8 z Rxy & Ryz Rxz PartC Determinewhethereachsentenceistrueorfalseinthemodelgiven. UD= f Lemmy,Courtney,Eddy g extension G = f Lemmy,Courtney,Eddy g extension H = f Courtney g extension M = f Lemmy,Eddy g referent c =Courtney referent e =Eddy 1. Hc 2. He 3. Mc Me 4. Gc _: Gc 5. Mc Gc 6. 9 xHx 7. 8 xHx 8. 9 x : Mx 9. 9 x Hx & Gx 10. 9 x Mx & Gx 11. 8 x Hx Mx

PAGE 105

ch.5formalsemantics 105 12. 9 xHx & 9 xMx 13. 8 x Hx $: Mx 14. 9 xGx & 9 x : Gx 15. 8 x 9 y Gx & Hy PartD Writeoutthemodelthatcorrespondstotheinterpretationgiven. UD:naturalnumbersfrom10to13 Ox: x isodd. Sx: x islessthan7. Tx: x isatwo-digitnumber. Ux: x isthoughttobeunlucky. Nxy: x isthenextnumberafter y PartE Showthateachofthefollowingiscontingent. 1. ?Da & Db 2. ? 9 xTxh 3. ?Pm & :8 xPx 4. 8 zJz $9 yJy 5. 8 x Wxmn _9 yLxy 6. 9 x Gx !8 yMy ? PartF Showthatthefollowingpairsofsentencesarenotlogicallyequivalent. 1. Ja Ka 2. 9 xJx Jm 3. 8 xRxx 9 xRxx 4. 9 xPx Qc 9 x Px Qc 5. 8 x Px !: Qx 9 x Px & : Qx 6. 9 x Px & Qx 9 x Px Qx 7. 8 x Px Qx 8 x Px & Qx 8. 8 x 9 yRxy 9 x 8 yRxy 9. 8 x 9 yRxy 8 x 9 yRyx PartG Showthatthefollowingsetsofsentencesareconsistent. 1. f Ma, : Na,Pa, : Qa g 2. f Lee Lef : Lfe : Lff g 3. f: Ma & 9 xAx Ma Fa 8 x Fx Ax g 4. f Ma Mb Ma !8 x : Mx g 5. f8 yGy 8 x Gx Hx 9 y : Iy g 6. f9 x Bx Ax 8 x : Cx 8 x Ax & Bx Cx g

PAGE 106

106 forall x 7. f9 xXx 9 xYx 8 x Xx $: Yx g 8. f8 x Px Qx 9 x : Qx & Px g 9. f9 z Nz & Ozz 8 x 8 y Oxy Oyx g 10. f:9 x 8 yRxy 8 x 9 yRxy g PartH Constructmodelstoshowthatthefollowingargumentsareinvalid. 1. 8 x Ax Bx : : 9 xBx 2. 8 x Rx Dx 8 x Rx Fx : : 9 x Dx & Fx 3. 9 x Px Qx : : 9 xPx 4. Na & Nb & Nc : : 8 xNx 5. Rde 9 xRxd : :Red 6. 9 x Ex & Fx 9 xFx !9 xGx : : 9 x Ex & Gx 7. 8 xOxc 8 xOcx : : 8 xOxx 8. 9 x Jx & Kx 9 x : Kx 9 x : Jx : : 9 x : Jx & : Kx 9. Lab !8 xLxb 9 xLxb : :Lbb PartI 1. ? Showthat f: Raa; 8 x x = a Rxa g isconsistent. 2. ? Showthat f8 x 8 y 8 z x = y y = z x = z ; 9 x 9 yx 6 = y g isconsistent. 3. ? Showthat f8 x 8 yx = y; 9 xx 6 = a g isinconsistent. 4.Showthat 9 x x = h & x = i iscontingent. 5.Showthat f9 x 9 y Zx & Zy & x = y : Zd d = s g isconsistent. 6.Showthat` 8 x Dx !9 yTyx : : 9 y 9 zy 6 = z 'isinvalid. PartJ 1.Manylogicbooksdeneconsistencyandinconsistencyinthisway:Aset f A 1 ; A 2 ; A 3 ; g isinconsistentifandonlyif f A 1 ; A 2 ; A 3 ; gj = B & : B forsomesentence B .Asetisconsistentifitisnotinconsistent." Doesthisdenitionleadtoanydierentsetsbeingconsistentthanthe denitiononp.88?Explainyouranswer. 2. ? Ourdenitionoftruthsaysthatasentence A is truein M ifandonlyif somevariableassignmentsatises A in M .Woulditmakeanydierence ifwesaidinsteadthat A is truein M ifandonlyif every variable assignmentsatises A in M ?Explainyouranswer.

PAGE 107

Chapter6 Proofs ConsidertwoargumentsinSL: ArgumentA P Q : P : : Q ArgumentB P Q P : : Q Clearly,thesearevalidarguments.Youcanconrmthattheyarevalidby constructingfour-linetruthtables.ArgumentAmakesuseofaninferenceform thatisalwaysvalid:Givenadisjunctionandthenegationofoneofthedisjuncts, theotherdisjunctfollowsasavalidconsequence.Thisruleiscalled disjunctive syllogism ArgumentBmakesuseofadierentvalidform:Givenaconditionalandits antecedent,theconsequentfollowsasavalidconsequence.Thisiscalled modus ponens Whenweconstructtruthtables,wedonotneedtogivenamestodierentinferenceforms.Thereisnoreasontodistinguishmodusponensfromadisjunctive syllogism.Forthissamereason,however,themethodoftruthtablesdoesnot clearlyshow why anargumentisvalid.Ifyouweretodoa1028-linetruthtable foranargumentthatcontainstensentenceletters,thenyoucouldchecktosee iftherewereanylinesonwhichthepremiseswerealltrueandtheconclusion werefalse.Ifyoudidnotseesuchalineandprovidedyoumadenomistakes inconstructingthetable,thenyouwouldknowthattheargumentwasvalid. Yetyouwouldnotbeabletosayanythingfurtheraboutwhythisparticular argumentwasavalidargumentform. 107

PAGE 108

108 forall x Theaimofa proofsystem istoshowthatparticularargumentsarevalidina waythatallowsustounderstandthereasoninginvolvedintheargument.We beginwithbasicargumentforms,likedisjunctivesyllogismandmodusponens. Theseformscanthenbecombinedtomakemorecomplicatedarguments,like thisone: : L J L : L : :J Bymodusponens,andentail J L .Thisisan intermediateconclusion Itfollowslogicallyfromthepremises,butitisnottheconclusionwewant.Now J L andentail J ,bydisjunctivesyllogism.Wedonotneedanewrule forthisargument.Theproofoftheargumentshowsthatitisreallyjusta combinationofruleswehavealreadyintroduced. Formally,a proof isasequenceofsentences.Therstsentencesofthesequence areassumptions;thesearethepremisesoftheargument.Everysentencelater inthesequencefollowsfromearliersentencesbyoneoftherulesofproof.The nalsentenceofthesequenceistheconclusionoftheargument. ThischapterbeginswithaproofsystemforSL,whichisthenextendedtocover QLandQLplusidentity. 6.1BasicrulesforSL Indesigningaproofsystem,wecouldjuststartwithdisjunctivesyllogismand modusponens.Wheneverwediscoveredavalidargumentwhichcouldnotbe provenwithruleswealreadyhad,wecouldintroducenewrules.Proceedingin thisway,wewouldhaveanunsystematicgrabbagofrules.Wemightaccidently addsomestrangerules,andwewouldsurelyendupwithmorerulesthanwe need. Instead,wewilldevelopwhatiscalleda naturaldeduction system.Ina naturaldeductionsystem,therewillbetworulesforeachlogicaloperator:an introduction rulethatallowsustoproveasentencethathasitasthemain logicaloperatorandan elimination rulethatallowsustoprovesomething givenasentencethathasitasthemainlogicaloperator. Inadditiontotherulesforeachlogicaloperator,wewillalsohaveareiterationrule.Ifyoualreadyhaveshownsomethinginthecourseofaproof,the reiterationruleallowsyoutorepeatitonanewline.Forinstance:

PAGE 109

ch.6proofs 109 1 A 2 A R1 Whenweaddalinetoaproof,wewritetherulethatjustiesthatline.Wealso writethenumbersofthelinestowhichtherulewasapplied.Thereiteration ruleaboveisjustiedbyoneline,thelinethatyouarereiterating.Sothe`R1' online2oftheproofmeansthatthelineisjustiedbythereiterationruleR appliedtoline1. Obviously,thereiterationrulewillnotallowustoshowanything new .Forthat, wewillneedmorerules.Theremainderofthissectionwillgiveintroduction andeliminationrulesforallofthesententialconnectives.Thiswillgiveus acompleteproofsystemforSL.Laterinthechapter,weintroducerulesfor quantiersandidentity. Alloftherulesintroducedinthischapteraresummarizedstartingonp.158. Conjunction Thinkforamoment:Whatwouldyouneedtoshowinordertoprove E & F ? Ofcourse,youcouldshow E & F byproving E andseparatelyproving F .This holdsevenifthetwoconjunctsarenotatomicsentences.Ifyoucanprove [ A J V ]and[ V L $ F N ],thenyouhaveeectivelyproven [ A J V ]&[ V L $ F N ] : Sothiswillbeourconjunctionintroductionrule,whichweabbreviate&I: m A n B A & B &I m n Alineofproofmustbejustiedbysomerule,andherewehave`&Im,n.' Thismeans:Conjunctionintroductionappliedtoline m andline n .Theseare variables,notreallinenumbers; m issomelineand n issomeotherline.In anactualproof,thelinesarenumbered1 ; 2 ; 3 ;::: andrulesmustbeapplied tospeciclinenumbers.Whenwedenetherule,however,weusevariables tounderscorethepointthattherulemaybeappliedtoanytwolinesthatare alreadyintheproof.Ifyouhave K online8and L online15,youcanprove K & L atsomelaterpointintheproofwiththejustication`&I8,15.' Now,considertheeliminationruleforconjunction.Whatareyouentitledto concludefromasentencelike E & F ?Surely,youareentitledtoconclude E ;if

PAGE 110

110 forall x E & F weretrue,then E wouldbetrue.Similarly,youareentitledtoconclude F .Thiswillbeourconjunctioneliminationrule,whichweabbreviate&E: m A & B A &E m B &E m Whenyouhaveaconjunctiononsomelineofaproof,youcanuse&Etoderive eitheroftheconjuncts.The&Erulerequiresonlyonesentence,sowewrite onelinenumberasthejusticationforapplyingit. Evenwithjustthesetworules,wecanprovidesomeproofs.Considerthis argument. [ A B C D ]&[ E F G H ] : : [ E F G H ]&[ A B C D ] Themainlogicaloperatorinboththepremiseandconclusionisconjunction. Sinceconjunctionissymmetric,theargumentisobviouslyvalid.Inorderto provideaproof,webeginbywritingdownthepremise.Afterthepremises,we drawahorizontalline|everythingbelowthislinemustbejustiedbyarule ofproof.Sothebeginningoftheprooflookslikethis: 1 [ A B C D ]&[ E F G H ] Fromthepremise,wecangeteachoftheconjunctsby&E.Theproofnow lookslikethis: 1 [ A B C D ]&[ E F G H ] 2 [ A B C D ]&E1 3 [ E F G H ]&E1 Therule&Irequiresthatwehaveeachoftheconjunctsavailablesomewhere intheproof.Theycanbeseparatedfromoneanother,andtheycanappearin anyorder.Sobyapplyingthe&Iruletolines3and2,wearriveatthedesired conclusion.Thenishedprooflookslikethis:

PAGE 111

ch.6proofs 111 1 [ A B C D ]&[ E F G H ] 2 [ A B C D ]&E1 3 [ E F G H ]&E1 4 [ E F G H ]&[ A B C D ]&I3,2 Thisproofistrivial,butitshowshowwecanuserulesofprooftogetherto demonstratethevalidityofanargumentform.Also:Usingatruthtableto showthatthisargumentisvalidwouldhaverequiredastaggering256lines, sincethereareeightsentencelettersintheargument. Disjunction If M weretrue,then M N wouldalsobetrue.Sothedisjunctionintroduction rule Iallowsustoderiveadisjunctionifwehaveoneofthetwodisjuncts: m A A B I m B A I m Noticethat B canbe any sentencewhatsoever.Sothefollowingisalegitimate proof: 1 M 2 M [ A $ B C & D ] $ [ E & F ] I1 Itmayseemoddthatjustbyknowing M wecanderiveaconclusionthat includessentenceslike A B ,andtherest|sentencesthathavenothingtodo with M .Yettheconclusionfollowsimmediatelyby I.Thisisasitshouldbe: Thetruthconditionsforthedisjunctionmeanthat,if A istrue,then A B is trueregardlessofwhat B is.Sotheconclusioncouldnotbefalseifthepremise weretrue;theargumentisvalid. Nowconsiderthedisjunctioneliminationrule.Whatcanyouconcludefrom M N ?Youcannotconclude M .Itmightbe M 'struththatmakes M N true,asintheexampleabove,butitmightnot.From M N alone,youcannot concludeanythingabouteither M or N specically.Ifyoualsoknewthat N wasfalse,however,thenyouwouldbeabletoconclude M Thisisjustdisjunctivesyllogism,itwillbethedisjunctioneliminationrule E.

PAGE 112

112 forall x m A B n : B A E m n m A B n : A B E m n Conditional Considerthisargument: R F : : : R F Theargumentiscertainlyavalidone.Whatshouldtheconditionalintroduction rulebe,suchthatwecandrawthisconclusion? Webegintheproofbywritingdownthepremiseoftheargumentanddrawing ahorizontalline,likethis: 1 R F Ifwehad : R asafurtherpremise,wecouldderive F bythe Erule.Wedonot have : R asapremiseofthisargument,norcanwederiveitdirectlyfromthe premisewedohave|sowecannotsimplyprove F .Whatwewilldoinsteadis starta subproof ,aproofwithinthemainproof.Whenwestartasubproof,we drawanotherverticallinetoindicatethatwearenolongerinthemainproof. Thenwewriteinanassumptionforthesubproof.Thiscanbeanythingwe want.Here,itwillbehelpfultoassume : R .Ourproofnowlookslikethis: 1 R F 2 : R Itisimportanttonoticethatwearenotclaimingtohaveproven : R .Wedo notneedtowriteinanyjusticationfortheassumptionlineofasubproof.You canthinkofthesubproofasposingthequestion:Whatcouldweshow if : R weretrue?Foronething,wecanderive F .Sowedo: 1 R F 2 : R 3 F E1,2

PAGE 113

ch.6proofs 113 Thishasshownthat if wehad : R asapremise, then wecouldprove F .In eect,wehaveproven : R F .Sotheconditionalintroductionrule Iwill allowustoclosethesubproofandderive : R F inthemainproof.Ournal prooflookslikethis: 1 R F 2 : R 3 F E1,2 4 : R F I2{3 Noticethatthejusticationforapplyingthe Iruleistheentiresubproof. Usuallythatwillbemorethanjusttwolines. Itmayseemasiftheabilitytoassumeanythingatallinasubproofwouldlead tochaos:Doesitallowyoutoproveanyconclusionfromanypremises?The answerisno,itdoesnot.Considerthisproof: 1 A 2 B 3 B R2 Itmayseemasifthisisaproofthatyoucanderiveanyconclusions B from anypremise A .Whentheverticallineforthesubproofends,thesubproofis closed .Inordertocompleteaproof,youmustcloseallofthesubproofs.And youcannotclosethesubproofandusetheRruleagainonline4toderive B in themainproof.Onceyoucloseasubproof,youcannotreferbacktoindividual linesinsideit. Closingasubproofiscalled discharging theassumptionsofthatsubproof.So wecanputthepointthisway:Youcannotcompleteaproofuntilyouhave dischargedalloftheassumptionsbesidestheoriginalpremisesoftheargument. Ofcourse,itislegitimatetodothis: 1 A 2 B 3 B R2 4 B B I2{3

PAGE 114

114 forall x Thisshouldnotseemsostrange,though.Since B B isatautology,noparticularpremisesshouldberequiredtovalidlyderiveit.Indeed,aswewillsee,a tautologyfollowsfromanypremises. Putinageneralform,the Irulelookslikethis: m A want B n B A B I m { n Whenweintroduceasubproof,wetypicallywritewhatwewanttoderivein thecolumn.Thisisjustsothatwedonotforgetwhywestartedthesubproofif itgoesonforveortenlines.Thereisno`want'rule.Itisanotetoourselves andnotformallypartoftheproof. Althoughitisalwayspermissibletoopenasubproofwithanyassumptionyou please,thereissomestrategyinvolvedinpickingausefulassumption.Starting asubproofwithanarbitrary,wackyassumptionwouldjustwastelinesofthe proof.Inordertoderiveaconditionalbythe I,forinstance,youmustassume theantecedentoftheconditionalinasubproof. The Irulealsorequiresthattheconsequentoftheconditionalbethelastline ofthesubproof.Itisalwayspermissibletocloseasubproofanddischargeits assumptions,butitwillnotbehelpfultodosountilyougetwhatyouwant. Nowconsidertheconditionaleliminationrule.Nothingfollowsfrom M N alone,butifwehaveboth M N and M ,thenwecanconclude N .Thisrule, modusponens,willbetheconditionaleliminationrule E. m A B n A B E m n Nowthatwehaverulesfortheconditional,considerthisargument: P Q Q R : :P R Webegintheproofbywritingthetwopremisesasassumptions.Sincethemain logicaloperatorintheconclusionisaconditional,wecanexpecttousethe Irule.Forthat,weneedasubproof|sowewriteintheantecedentofthe conditionalasassumptionofasubproof:

PAGE 115

ch.6proofs 115 1 P Q 2 Q R 3 P Wemade P availablebyassumingitinasubproof,allowingustouse Eon therstpremise.Thisgivesus Q ,whichallowsustouse Eonthesecond premise.Havingderived R ,weclosethesubproof.Byassuming P wewereable toprove R ,soweapplythe Iruleandnishtheproof. 1 P Q 2 Q R 3 P want R 4 Q E1,3 5 R E2,4 6 P R I3{5 Biconditional Therulesforthebiconditionalwillbelikedouble-barreledversionsoftherules fortheconditional. Inordertoderive W $ X ,forinstance,youmustbeabletoprove X by assuming W and prove W byassuming X .Thebiconditionalintroduction rule $ Irequirestwosubproofs.Thesubproofscancomeinanyorder,and thesecondsubproofdoesnotneedtocomeimmediatelyaftertherst|but schematically,theruleworkslikethis: m A want B n B p B want A q A A $ B $ I m { n p { q Thebiconditionaleliminationrule $ Eletsyoudoabitmorethantheconditionalrule.Ifyouhavetheleft-handsubsentenceofthebiconditional,youcan derivetheright-handsubsentence.Ifyouhavetheright-handsubsentence,you

PAGE 116

116 forall x canderivetheleft-handsubsentence.Thisistherule: m A $ B n A B $ E m n m A $ B n B A $ E m n Negation HereisasimplemathematicalargumentinEnglish: Assumethereissomegreatestnaturalnumber.Callit A Thatnumberplusoneisalsoanaturalnumber. Obviously, A +1 >A Sothereisanaturalnumbergreaterthan A Thisisimpossible,since A isassumedtobethegreatestnaturalnumber. : : Thereisnogreatestnaturalnumber. Thisargumentformistraditionallycalleda reductio .ItsfullLatinnameis reductioadabsurdum ,whichmeans`reductiontoabsurdity.'Inareductio, weassumesomethingforthesakeofargument|forexample,thatthereis agreatestnaturalnumber.Thenweshowthattheassumptionleadstotwo contradictorysentences|forexample,that A isthegreatestnaturalnumber andthatitisnot.Inthisway,weshowthattheoriginalassumptionmusthave beenfalse. Thebasicrulesfornegationwillallowforargumentslikethis.Ifweassume somethingandshowthatitleadstocontradictorysentences,thenwehave proventhenegationoftheassumption.Thisisthenegationintroduction : I rule: m A forreductio n B n +1 : B n +2 : A : I m { n +1 Fortheruletoapply,thelasttwolinesofthesubproofmustbeanexplicit contradiction:somesentencefollowedonthenextlinebyitsnegation.We write`forreductio'asanotetoourselves,areminderofwhywestartedthe subproof.Itisnotformallypartoftheproof,andyoucanleaveitoutifyou nditdistracting.

PAGE 117

ch.6proofs 117 Toseehowtheruleworks,supposewewanttoprovethelawofnon-contradiction: : G & : G .Wecanprovethiswithoutanypremisesbyimmediatelystartinga subproof.Wewanttoapply : Itothesubproof,soweassume G & : G .We thengetanexplicitcontradictionby&E.Theprooflookslikethis: 1 G & : G forreductio 2 G &E1 3 : G &E1 4 : G & : G : I1{3 The : Erulewillworkinmuchthesameway.Ifweassume : A andshowthat itleadstoacontradiction,wehaveeectivelyproven A .Sotherulelookslike this: m : A forreductio n B n +1 : B n +2 A : E m { n +1 6.2Derivedrules Therulesofthenaturaldeductionsystemaremeanttobesystematic.Thereis anintroductionandaneliminationruleforeachlogicaloperator,butwhythese basicrulesratherthansomeothers?Manynaturaldeductionsystemshavea disjunctioneliminationrulethatworkslikethis: m A B n A C o B C C m n o Itmightseemasiftherewillbesomeproofsthatwecannotdowithourproof system,becausewedonothavethisrule.Yetthisisnotthecase.Ifyoucan doaproofwiththisrule,youcandoaproofwiththebasicrulesofthenatural deductionsystem.Considerthisproof:

PAGE 118

118 forall x 1 A B 2 A C 3 B C want C 4 : C forreductio 5 A forreductio 6 C I2,5 7 : C R4 8 : A : I5{7 9 B forreductio 10 C I3,9 11 : C R4 12 B E1,8 13 : B : I9{11 14 C : E4{13 A B ,and C aremeta-variables.TheyarenotsymbolsofSL,butstand-insfor arbitrarysentencesofSL.Sothisisnot,strictlyspeaking,aproofinSL.Itis morelikearecipe.Itprovidesapatternthatcanproveanythingthatthe rulecanprove,usingonlythebasicrulesofSL.Thismeansthatthe isnot reallynecessary.Addingittothelistofbasicruleswouldnotallowustoderive anythingthatwecouldnotderivewithoutit. Nevertheless,the rulewouldbeconvenient.Itwouldallowustodoinone linewhatrequireselevenlinesandseveralnestedsubproofswiththebasicrules. Sowewilladd totheproofsystemasaderivedrule. A derivedrule isaruleofproofthatdoesnotmakeanynewproofspossible. Anythingthatcanbeprovenwithaderivedrulecanbeprovenwithoutit.You canthinkofashortproofusingaderivedruleasshorthandforalongerproof thatusesonlythebasicrules.Anytimeyouusethe rule,youcouldalways taketenextralinesandprovethesamethingwithoutit. Forthesakeofconvenience,wewilladdseveralotherderivedrules.Oneis modustollens MT.

PAGE 119

ch.6proofs 119 m A B n : B : A MT m n Weleavetheproofofthisruleasanexercise.Notethatifwehadalreadyproven theMTrule,thentheproofofthe rulecouldhavebeendoneinonlyve lines. WealsoaddhypotheticalsyllogismHSasaderivedrule.Wehavealready givenaproofofitonp.115. m A B n B C A C HS m n 6.3Rulesofreplacement Considerhowyouwouldprovethisargument: F G & H : :F G Perhapsitistemptingtowritedownthepremiseandapplythe&Eruletothe conjunction G & H .Thisisimpermissible,however,becausethebasicrulesof proofcanonlybeappliedtowholesentences.Weneedtoget G & H onaline byitself.Wecanprovetheargumentinthisway: 1 F G & H 2 F want G 3 G & H E1,2 4 G &E3 5 F G I2{4 Wewillnowintroducesomederivedrulesthatmaybeappliedtopartofa sentence.Thesearecalled rulesofreplacement ,becausetheycanbeused toreplacepartofasentencewithalogicallyequivalentexpression.Onesimple ruleofreplacementiscommutivityabbreviatedComm,whichsaysthatwe canswaptheorderofconjunctsinaconjunctionortheorderofdisjunctsina disjunction.Wedenetherulethisway:

PAGE 120

120 forall x A & B B & A A B B A A $ B B $ A Comm Theboldarrowmeansthatyoucantakeasubformulaononesideofthearrow andreplaceitwiththesubformulaontheotherside.Thearrowisdouble-headed becauserulesofreplacementworkinbothdirections. Considerthisargument: M P P & M : : P M M & P Itispossibletogiveaproofofthisusingonlythebasicrules,butitwillbelong andinconvenient.WiththeCommrule,wecanprovideaproofeasily: 1 M P P & M 2 P M P & M Comm1 3 P M M & P Comm2 AnotherruleofreplacementisdoublenegationDN.WiththeDNrule,you canremoveorinsertapairofnegationsanywhereinasentence.Thisisthe rule: :: A A DN TwomorereplacementrulesarecalledDeMorgan'sLaws,namedforthe19thcenturyBritishlogicianAugustDeMorgan.AlthoughDeMorgandiddiscover theselaws,hewasnotthersttodoso.Therulescaptureusefulrelations betweennegation,conjunction,anddisjunction.Herearetherules,whichwe abbreviateDeM: : A B : A & : B : A & B : A _: B DeM Because A B isa materialconditional ,itisequivalentto : A B .Afurther replacementrulecapturesthisequivalence.WeabbreviatetheruleMC,for `materialconditional.'Ittakestwoforms: A B : A B A B : A B MC Nowconsiderthisargument: : P Q : :P & : Q

PAGE 121

ch.6proofs 121 Asalways,wecouldprovethisargumentusingonlythebasicrules.Withrules ofreplacement,though,theproofismuchsimpler: 1 : P Q 2 : : P Q MC1 3 :: P & : Q DeM2 4 P & : Q DN3 Analreplacementrulecapturestherelationbetweenconditionalsandbiconditionals.Wewillcallthisrulebiconditionalexchangeandabbreviateit $ ex. [ A B & B A ] A $ B $ ex 6.4Rulesforquantiers ForproofsinQL,weuseallofthebasicrulesofSLplusfournewbasicrules: bothintroductionandeliminationrulesforeachofthequantiers. SinceallofthederivedrulesofSLarederivedfromthebasicrules,theywill alsoholdinQL.Wewilladdanotherderivedrule,areplacementrulecalled quantiernegation. Universalelimination Ifyouhave 8 xAx ,itislegitimatetoinferthatanythingisan A .Youcaninfer Aa Ab Az Ad 3 |inshort,youcaninfer A c foranyconstant c .Thisisthe generalformoftheuniversaleliminationrule 8 E: m 8 xA A [ c j x ] 8 E m A [ c j x ]isasubstitutioninstanceof 8 xA .Thesymbolsforasubstitutioninstance arenotsymbolsofQL,soyoucannotwritetheminaproof.Instead,you writethesubsitutedsentencewiththeconstant c replacingalloccurancesof thevariable x in A .Forexample:

PAGE 122

122 forall x 1 8 x Mx Rxd 2 Ma Rad 8 E1 3 Md Rdd 8 E1 Existentialintroduction Whenisitlegitimatetoinfer 9 xAx ?Ifyouknowthatsomethingisan A |for instance,ifyouhave Aa availableintheproof. Thisistheexistentialintroductionrule 9 I: m A 9 xA [ x jj c ] 9 I m Itisimportanttonoticethat A [ x jj c ]isnotthesameasasubstitutioninstance. Wewriteitwithtwobarstoshowthatthevariable x doesnotneedtoreplace alloccurrencesoftheconstant c .Youcandecidewhichoccurrencestoreplace andwhichtoleaveinplace.Forexample: 1 Ma Rad 2 9 x Ma Rax 9 I1 3 9 x Mx Rxd 9 I1 4 9 x Mx Rad 9 I1 5 9 y 9 x Mx Ryd 9 I4 6 9 z 9 y 9 x Mx Ryz 9 I5 Universalintroduction Auniversalclaimlike 8 xPx wouldbeprovenifeverysubstitutioninstanceofit hadbeenproven,ifeverysentence Pa Pb ::: wereavailableinaproof.Alas, thereisnohopeofproving every substitutioninstance.Thatwouldrequire proving Pa Pb ::: Pj 2 ::: Ps 7 ::: ,andsoontoinnity.Thereareinnitely manyconstantsinQL,andsothisprocesswouldnevercometoanend. Considerasimpleargument: 8 xMx : : 8 yMy Itmakesnodierencetothemeaningofthesentencewhetherweusethevariable

PAGE 123

ch.6proofs 123 x orthevariable y ,sothisargumentisobviouslyvalid.Supposewebeginin thisway: 1 8 xMx want 8 yMy 2 Ma 8 E1 Wehavederived Ma .Nothingstopsusfromusingthesamejusticationto derive Mb ::: Mj 2 ::: Ms 7 ::: ,andsoonuntilwerunoutofspaceor patience.Wehaveeectivelyshownthewaytoprove M c foranyconstant c Fromthis, 8 xMx follows. 1 8 xMx 2 Ma 8 E1 3 8 yMy 8 I2 Itisimportantherethat a wasjustsomearbitraryconstant.Wehadnotmade anyspecialassumptionsaboutit.If Ma wereapremiseoftheargument,then thiswouldnotshowanythingabout all y .Forexample: 1 8 xRxa 2 Raa 8 E1 3 8 yRyy notallowed! Thisistheschematicformoftheuniversalintroductionrule 8 I: m A 8 xA [ x j c ] 8 I m c mustnotoccurinanyundischargedassumptions. Notethatwecandothisforanyconstantthatdoesnotoccurinanundischarged assumptionandforanyvariable. Notealsothattheconstantmaynotoccurinany undischarged assumption,but itmayoccurastheassumptionofasubproofthatwehavealreadyclosed.For example,wecanprove 8 z Dz Dz withoutanypremises.

PAGE 124

124 forall x 1 Df want Df 2 Df R1 3 Df Df I1{2 4 8 z Dz Dz 8 I3 Existentialelimination Asentencewithanexistentialquantiertellsusthatthereis some member oftheUDthatsatisesaformula.Forexample, 9 xSx tellsusroughlythat thereisatleastone S .Itdoesnottellus which memberoftheUDsatises S however.Wecannotimmediatelyconclude Sa Sf 23 ,oranyothersubstitution instanceofthesentence.Whatcanwedo? Supposethatweknewboth 9 xSx and 8 x Sx Tx .Wecouldreasoninthis way: Since 9 xSx ,thereissomethingthatisan S .Wedonotknowwhich constantsrefertothisthing,ifanydo,socallthisthing.From 8 x Sx Tx ,itfollowsthatifisan S ,thenitisa T .Therefore isa T .Becauseisa T ,weknowthat 9 xTx Inthisparagraph,weintroducedanameforthethingthatisan S .Wecalled it,sothatwecouldreasonaboutitandderivesomeconsequencesfromthere beingan S .Sinceisjustabogusnameintroducedforthepurposeofthe proofandnotagenuineconstant,wecouldnotmentionitintheconclusion. Yetwecouldderiveasentencethatdoesnotmention;namely, 9 xTx .This sentencedoesfollowfromthetwopremises. Wewanttheexistentialeliminationruletoworkinasimilarway.Yetsince GreekletterslikearenotsymbolsofQL,wecannotusetheminformalproofs. Instead,wewilluseconstantsofQLwhichdonototherwiseappearintheproof. Aconstantthatisusedtostandinforwhateveritisthatsatisesanexistential claimiscalleda proxy .Reasoningwiththeproxymustalloccurinsidea subproof,andtheproxycannotbeaconstantthatisdoingworkelsewherein theproof. Thisistheschematicformoftheexistentialeliminationrule 9 E:

PAGE 125

ch.6proofs 125 m 9 xA n A [ c j x ] p B B 9 E m n { p Theconstant c mustnotappearin 9 xA ,in B ,orinanyundischargedassumption. Withthisrule,wecangiveaformalproofthat 9 xSx and 8 x Sx Tx together entail 9 xTx .ThestructureoftheproofiseectivelythesameastheEnglishlanguageargumentwithwhichwebegan,exceptthatthesubproofusesthe constant` a 'ratherthanthebogusname. 1 9 xSx 2 8 x Sx Tx want 9 xTx 3 Sa 4 Sa Ta 8 E2 5 Ta E3,4 6 9 xTx 9 I5 7 9 xTx 9 E1,3{6 Quantiernegation WhentranslatingfromEnglishtoQL,wenotedthat :9 x : A islogicallyequivalentto 8 x A .InQL,theyareprovablyequivalent.Wecanproveonehalfof theequivalencewitharathergruesomeproof:

PAGE 126

126 forall x 1 8 xAx want :9 x : Ax 2 9: Ax forreductio 3 : Ac for 9 E 4 8 xAx forreductio 5 Ac 8 E1 6 : Ac R3 7 :8 xAx : I4{6 8 8 xAx R1 9 :8 xAx 9 E3{7 10 :9: Ax : I2{9 Inordertoshowthatthetwosentencesaregenuinelyequivalent,weneeda secondproofthatassumes :9 x : A andderives 8 x A .Weleavethatproofasan exerciseforthereader. Itwilloftenbeusefultotranslatebetweenquantiersbyaddingorsubtracting negationsinthisway,soweaddtwoderivedrulesforthispurpose.Theserules arecalledquantiernegationQN: :8 xA 9 x : A :9 xA 8 x : A QN SinceQNisareplacementrule,itcanbeusedonwholesentencesoronsubformulae. 6.5Rulesforidentity TheidentitypredicateisnotpartofQL,butweadditwhenweneedtosymbolizecertainsentences.Forproofsinvolvingidentity,weaddtworulesofproof. Supposeyouknowthatmanythingsthataretrueof a arealsotrueof b .For example: Aa & Ab Ba & Bb : Ca & : Cb Da & Db : Ea & : Eb ,andsoon. Thiswouldnotbeenoughtojustifytheconclusion a = b .Seep.93.Ingeneral, therearenosentencesthatdonotalreadycontaintheidentitypredicatethat couldjustifytheconclusion a = b .Thismeansthattheidentityintroduction rulewillnotjustify a = b oranyotheridentityclaimcontainingtwodierent constants.

PAGE 127

ch.6proofs 127 However,itisalwaystruethat a = a .Ingeneral,nopremisesarerequired inordertoconcludethatsomethingisidenticaltoitself.Sothiswillbethe identityintroductionrule,abbreviated=I: c = c =I Noticethatthe=Iruledoesnotrequirereferringtoanypriorlinesoftheproof. Foranyconstant c ,youcanwrite c = c onanypointwithonlythe=Iruleas justication. Ifyouhaveshownthat a = b ,thenanythingthatistrueof a mustalsobetrueof b .Foranysentencewith a init,youcanreplacesomeoralloftheoccurrencesof a with b andproduceanequivalentsentence.Forexample,ifyoualreadyknow Raa ,thenyouarejustiedinconcluding Rab Rba Rbb .Recallthat A [ a jj b ]is thesentenceproducedbyreplacing a in A with b .Thisisnotthesameasa substitutioninstance,because b mayreplacesomeoralloccurrencesof a .The identityeliminationrule=Ejustiesreplacingtermswithothertermsthat areidenticaltoit: m a = b n A A [ a jj b ]=E m n A [ b jj a ]=E m n Toseetherulesinaction,considerthisproof:

PAGE 128

128 forall x 1 8 x 8 yx = y 2 9 xBx 3 8 x Bx !: Cx want :9 xCx 4 Be 5 8 ye = y 8 E1 6 e = f 8 E5 7 Bf =E6,4 8 Bf !: Cf 8 E3 9 : Cf E8,7 10 : Cf 9 E2,4{9 11 8 x : Cx 8 I10 12 :9 xCx QN11 6.6Proofstrategy Thereisnosimplerecipeforproofs,andthereisnosubstituteforpractice. Here,though,aresomerulesofthumbandstrategiestokeepinmind. Workbackwardsfromwhatyouwant. Theultimategoalistoderivethe conclusion.Lookattheconclusionandaskwhattheintroductionruleisforits mainlogicaloperator.Thisgivesyouanideaofwhatshouldhappen justbefore thelastlineoftheproof.Thenyoucantreatthislineasifitwereyourgoal. Askwhatyoucoulddotoderivethisnewgoal. Forexample:Ifyourconclusionisaconditional A B ,plantousethe I rule.Thisrequiresstartingasubproofinwhichyouassume A .Inthesubproof, youwanttoderive B Workforwardsfromwhatyouhave. Whenyouarestartingaproof,look atthepremises;later,lookatthesentencesthatyouhavederivedsofar.Think abouttheeliminationrulesforthemainoperatorsofthesesentences.These willtellyouwhatyouroptionsare. Forexample:Ifyouhave 8 x A ,thinkaboutinstantiatingitforanyconstant thatmightbehelpful.Ifyouhave 9 x A andintendtousethe 9 Erule,thenyou

PAGE 129

ch.6proofs 129 shouldassume A [ c j x ]forsome c thatisnotinuseandthenderiveaconclusion thatdoesnotcontain c Forashortproof,youmightbeabletoeliminatethepremisesandintroduce theconclusion.Alongproofisformallyjustanumberofshortproofslinked together,soyoucanllthegapbyalternatelyworkingbackfromtheconclusion andforwardfromthepremises. Changewhatyouarelookingat. Replacementrulescanoftenmakeyour lifeeasier.Ifaproofseemsimpossible,tryoutsomedierentsubstitutions. Forexample:Itisoftendiculttoproveadisjunctionusingthebasicrules.If youwanttoshow A B ,itisofteneasiertoshow : A B andusetheMC rule. Showing :9 x A canalsobehard,anditisofteneasiertoshow 8 x : A anduse theQNrule. Somereplacementrulesshouldbecomesecondnature.Ifyouseeanegated disjunction,forinstance,youshouldimmediatelythinkofDeMorgan'srule. Donotforgetindirectproof. Ifyoucannotndawaytoshowsomething directly,tryassumingitsnegation. Rememberthatmostproofscanbedoneeitherindirectlyordirectly.Oneway mightbeeasier|orperhapsonesparksyourimaginationmorethantheother| buteitheroneisformallylegitimate. Repeatasnecessary. Onceyouhavedecidedhowyoumightbeabletoget totheconclusion,askwhatyoumightbeabletodowiththepremises.Then considerthetargetsentencesagainandaskhowyoumightreachthem. Persist. Trydierentthings.Ifoneapproachfails,thentrysomethingelse. 6.7Proof-theoreticconcepts Wewillusethesymbol` ` 'toindicatethataproofispossible.Thissymbolis calledthe turnstile .Sometimesitiscalleda singleturnstile ,tounderscorethe factthatthisisnotthedoubleturnstilesymbol j =thatweusedtorepresent semanticentailmentinch.5.

PAGE 130

130 forall x Whenwewrite f A 1 ; A 2 ;::: g` B ,thismeansthatitispossibletogiveaproof of B with A 1 A 2 ::: aspremises.Withjustonepremise,weleaveoutthe curlybraces,so A ` B meansthatthereisaproofof B with A asapremise. Naturally, ` C meansthatthereisaproofof C thathasnopremises. Often,logicalproofsarecalled derivations .So A ` B canbereadas` B is derivablefrom A .' A theorem isasentencethatisderivablewithoutanypremises;i.e., T isa theoremifandonlyif ` T Itisnottoohardtoshowthatsomethingisatheorem|youjusthavetogive aproofofit.Howcouldyoushowthatsomethingis not atheorem?Ifits negationisatheorem,thenyoucouldprovideaproof.Forexample,itiseasy toprove : Pa & : Pa ,whichshowsthat Pa & : Pa cannotbeatheorem.For asentencethatisneitheratheoremnorthenegationofatheorem,however, thereisnoeasywaytoshowthis.Youwouldhavetodemonstratenotjustthat certainproofstrategiesfail,butthatnoproofispossible.Evenifyoufailin tryingtoproveasentenceinathousanddierentways,perhapstheproofis justtoolongandcomplexforyoutomakeout. Twosentences A and B are provablyequivalent ifandonlyifeachcanbe derivedfromtheother;i.e., A ` B and B ` A Itisrelativelyeasytoshowthattwosentencesareprovablyequivalent|itjust requiresapairofproofs.Showingthatsentencesare not provablyequivalent wouldbemuchharder.Itwouldbejustashardasshowingthatasentence isnotatheorem.Infact,theseproblemsareinterchangeable.Canyouthink ofasentencethatwouldbeatheoremifandonlyif A and B wereprovably equivalent? Thesetofsentences f A 1 ; A 2 ;::: g is provablyinconsistent ifandonlyifa contradictionisderivablefromit;i.e.,forsomesentence B f A 1 ; A 2 ;::: g` B and f A 1 ; A 2 ;::: g`: B Itiseasytoshowthatasetisprovablyinconsistent:Youjustneedtoassume thesentencesinthesetandproveacontradiction.Showingthatasetis not provablyinconsistentwillbemuchharder.Itwouldrequiremorethanjust providingaproofortwo;itwouldrequireshowingthatproofsofacertainkind are impossible .

PAGE 131

ch.6proofs 131 6.8Proofsandmodels Asyoumightalreadysuspect,thereisaconnectionbetween theorems and tautologies Thereisaformalwayofshowingthatasentenceisatheorem:Proveit.For eachline,wecanchecktoseeifthatlinefollowsbythecitedrule.Itmaybe hardtoproduceatwentylineproof,butitisnotsohardtocheckeachline oftheproofandconrmthatitislegitimate|andifeachlineoftheproof individuallyislegitimate,thenthewholeproofislegitimate.Showingthata sentenceisatautology,though,requiresreasoninginEnglishaboutallpossible models.Thereisnoformalwayofcheckingtoseeifthereasoningissound. Givenachoicebetweenshowingthatasentenceisatheoremandshowingthat itisatautology,itwouldbeeasiertoshowthatitisatheorem. Contrawise,thereisnoformalwayofshowingthatasentenceis not atheorem. WewouldneedtoreasoninEnglishaboutallpossibleproofs.Yetthereisa formalmethodforshowingthatasentenceisnotatautology.Weneedonly constructamodelinwhichthesentenceisfalse.Givenachoicebetweenshowing thatasentenceisnotatheoremandshowingthatitisnotatautology,itwould beeasiertoshowthatitisnotatautology. Fortunately,asentenceisatheoremifandonlyifitisatautology.Ifwe provideaproofof ` A andthusshowthatitisatheorem,itfollowsthat A isa tautology;i.e., j = A .Similarly,ifweconstructamodelinwhich A isfalseand thusshowthatitisnotatautology,iffollowsthat A isnotatheorem. Ingeneral, A ` B ifandonlyif A j = B .Assuch: Anargumentis valid ifandonlyif theconclusionisderivablefromthe premises . Twosentencesare logicallyequivalent ifandonlyiftheyare provably equivalent . Asetofsentencesis consistent ifandonlyifitis notprovablyinconsistent Youcanpickandchoosewhentothinkintermsofproofsandwhentothinkin termsofmodels,doingwhicheveriseasierforagiventask.Table6.1summarizes whenitisbesttogiveproofsandwhenitisbesttogivemodels. Inthisway,proofsandmodelsgiveusaversatiletoolkitforworkingwith arguments.IfwecantranslateanargumentintoQL,thenwecanmeasureits logicalweightinapurelyformalway.Ifitisdeductivelyvalid,wecangivea formalproof;ifitisinvalid,wecanprovideaformalcounterexample.

PAGE 132

132 forall x YES NO Is A atautology? prove ` A giveamodelinwhich A isfalse Is A acontradiction? prove `: A giveamodelinwhich A istrue Is A contingent? giveamodelinwhich A istrueandanother inwhich A isfalse prove ` A or `: A Are A and B equivalent? prove A ` B and B ` A giveamodelinwhich A and B havedierent truthvalues Istheset A consistent? giveamodelinwhich allthesentencesin A aretrue takingthesentencesin A ,prove B and : B Istheargument ` P : : C 'valid? prove P ` C giveamodelinwhich P istrueand C isfalse Table6.1:Sometimesitiseasiertoshowsomethingbyprovidingproofsthan itisbyprovidingmodels.Sometimesitistheotherwayround.Itdependson whatyouaretryingtoshow. 6.9Soundnessandcompleteness Thistoolkitisincrediblyconvenient.Itisalsointuitive,becauseitseemsnatural thatprovabilityandsemanticentailmentshouldagree.Yet,donotbefooled bythesimilarityofthesymbols` j ='and` ` .'Thefactthatthesetwoarereally interchangeableisnotasimplethingtoprove. Whyshouldwethinkthatanargumentthat canbeproven isnecessarilya valid argument?Thatis,whythinkthat A ` B implies A j = B ? Thisistheproblemof soundness .Aproofsystemis sound ifthereareno proofsofinvalidarguments.Demonstratingthattheproofsystemissound wouldrequireshowingthat any possibleproofistheproofofavalidargument. Itwouldnotbeenoughsimplytosucceedwhentryingtoprovemanyvalid argumentsandtofailwhentryingtoproveinvalidones. Fortunately,thereisawayofapproachingthisinastep-wisefashion.Ifusing the&Eruleonthelastlineofaproofcouldneverchangeavalidargument intoaninvalidone,thenusingtherulemanytimescouldnotmakeanargument invalid.Similarly,ifusingthe&Eand Erulesindividuallyonthelastline ofaproofcouldneverchangeavalidargumentintoaninvalidone,thenusing themincombinationcouldnoteither.

PAGE 133

ch.6proofs 133 Thestrategyistoshowforeveryruleofinferencethatitalonecouldnotmakea validargumentintoaninvalidone.Itfollowsthattherulesusedincombination wouldnotmakeavalidargumentinvalid.Sinceaproofisjustaseriesof lines,eachjustiedbyaruleofinference,thiswouldshowthateveryprovable argumentisvalid. Consider,forexample,the&Erule.Supposeweuseittoadd A & B toavalid argument.Inorderfortheruletoapply, A and B mustalreadybeavailablein theproof.Sincetheargumentsofarisvalid, A and B areeitherpremisesofthe argumentorvalidconsequencesofthepremises.Assuch,anymodelinwhich thepremisesaretruemustbeamodelinwhich A and B aretrue.According tothedenitionof truthinql ,thismeansthat A & B isalsotrueinsuch amodel.Therefore, A & B validlyfollowsfromthepremises.Thismeansthat usingthe&Eruletoextendavalidproofproducesanothervalidproof. Inordertoshowthattheproofsystemissound,wewouldneedtoshowthisfor theotherinferencerules.Sincethederivedrulesareconsequencesofthebasic rules,itwouldsucetoprovidesimilarargumentsforthe16otherbasicrules. Thistediousexercisefallsbeyondthescopeofthisbook. Givenaproofthattheproofsystemissound,itfollowsthateverytheoremisa tautology. Itisstillpossibletoask:Whythinkthat every validargumentisanargument thatcanbeproven?Thatis,whythinkthat A j = B implies A ` B ? Thisistheproblemof completeness .Aproofsystemis complete ifthereis aproofofeveryvalidargument.CompletenessforalanguagelikeQLwasrst provenbyKurtGodelin1929.Theproofisbeyondthescopeofthisbook. Theimportantpointisthat,happily,theproofsystemforQLisbothsoundand complete.Thisisnotthecaseforallproofsystemsandallformallanguages. BecauseitistrueofQL,wecanchoosetogiveproofsorconstructmodels| whicheveriseasierforthetaskathand. Summaryofdenitions Asentence A isa theorem ifandonlyif ` A . Twosentences A and B are provablyequivalent ifandonlyif A ` B and B ` A . f A 1 ; A 2 ;::: g is provablyinconsistent ifandonlyif,forsomesentence B f A 1 ; A 2 ;::: g` B & : B .

PAGE 134

134 forall x PracticeExercises ? PartA Provideajusticationruleandlinenumbersforeachlineofproof thatrequiresone. 1 W !: B 2 A & W 3 B J & K 4 W 5 : B 6 J & K 7 K 1 L $: O 2 L _: O 3 : L 4 : O 5 L 6 : L 7 L 1 Z C & : N 2 : Z N & : C 3 : N C 4 : N & : C 5 Z 6 C & : N 7 C 8 : C 9 : Z 10 N & : C 11 N 12 : N 13 N C ? PartB GiveaproofforeachargumentinSL. 1. K & L : :K $ L 2. A B C : : A & B C 3. P & Q R P !: R : :Q E 4. C & D E : :E D 5. : F G F H : :G H 6. X & Y X & Z : X & D D M: :M PartC GiveaproofforeachargumentinSL. 1. Q Q & : Q : : : Q 2. J !: J : : : J 3. E F F G : F : :E & G

PAGE 135

ch.6proofs 135 4. A $ B B $ C : :A $ C 5. M N M : : : M !: N 6. S $ T : :S $ T S 7. M N & O P N P : P : :M & O 8. Z & K K & M K D : :D PartD ShowthateachofthefollowingsentencesisatheoreminSL. 1. O O 2. N _: N 3. : P & : P 4. : A !: C A C 5. J $ [ J L & : L ] PartE ShowthateachofthefollowingpairsofsentencesareprovablyequivalentinSL. 1. :::: G G 2. T S : S !: T 3. R $ E E $ R 4. : G $ H : G $ H 5. U I : U & : I PartF Provideproofstoshoweachofthefollowing. 1. M & : N !: M ` N & M _: M 2. f C E & G : C G g` G 3. f Z & K $ Y & M D & D M g` Y Z 4. f W X Y Z X Y : Z g` W Y PartG Forthefollowing,provideproofsusingonlythebasicrules.Theproofs willbelongerthanproofsofthesameclaimswouldbeusingthederivedrules. 1.ShowthatMTisalegitimatederivedrule.Usingonlythebasicrules, provethefollowing: A B : B : : : A 2.ShowthatCommisalegitimateruleforthebiconditional.Usingonlythe basicrules,provethat A $ B and B $ A areequivalent. 3.Usingonlythebasicrules,provethefollowinginstanceofDeMorgan's Laws: : A & : B : : : A B 4.WithoutusingtheQNrule,prove :9 x : A `8 x A 5.Showthat $ exisalegitimatederivedrule.Usingonlythebasicrules, provethat D $ E and D E & E D areequivalent.

PAGE 136

136 forall x ? PartH Provideajusticationruleandlinenumbersforeachlineofproof thatrequiresone. 1 8 x 9 y Rxy Ryx 2 8 x : Rmx 3 9 y Rmy Rym 4 Rma Ram 5 : Rma 6 Ram 7 9 xRxm 8 9 xRxm 1 8 x 9 yLxy !8 zLzx 2 Lab 3 9 yLay !8 zLza 4 9 yLay 5 8 zLza 6 Lca 7 9 yLcy !8 zLzc 8 9 yLcy 9 8 zLzc 10 Lcc 11 8 xLxx 1 8 x Jx Kx 2 9 x 8 yLxy 3 8 xJx 4 Ja 5 Ja Ka 6 Ka 7 8 yLay 8 Laa 9 Ka & Laa 10 9 x Kx & Lxx 11 9 x Kx & Lxx 1 : 9 xMx _8 x : Mx 2 :9 xMx & :8 x : Mx 3 :9 xMx 4 8 x : Mx 5 :8 x : Mx 6 9 xMx _8 x : Mx ? PartI Provideaproofofeachclaim. 1. `8 xFx _:8 xFx 2. f8 x Mx $ Nx ;Ma & 9 xRxa g`9 xNx 3. f8 x : Mx Ljx ; 8 x Bx Ljx ; 8 x Mx Bx g`8 xLjx 4. 8 x Cx & Dt `8 xCx & Dt 5. 9 x Cx Dt `9 xCx Dt PartJ ProvideaproofoftheargumentaboutBillyonp.65.

PAGE 137

ch.6proofs 137 PartK LookbackatPartDonp.77.Provideproofstoshowthateachofthe argumentformsisvalidinQL. PartL Aristotleandhissuccessorsidentiedothersyllogisticforms.Symbolize eachofthefollowingargumentformsinQLandaddtheadditionalassumptions `Thereisan A 'and`Thereisa B .'Thenprovethatthesupplementedarguments formsarevalidinQL. Darapti: All A sare B s.All A sare C s. : : Some B is C .1.IfeveryMisL andeveryMisS,thensomeSisLDarapti. Felapton: No B sare C s.All A sare B s. : : Some A isnot C Barbari: All B sare C s.All A sare B s. : : Some A is C Camestros: All C sare B s.No A sare B s. : : Some A isnot C Celaront: No B sare C s.All A sare B s. : : Some A isnot C Cesaro: No C sare B s.All A sare B s. : : Some A isnot C Fapesmo: All B sare C s.No A sare B s. : : Some C isnot A PartM Provideaproofofeachclaim. 1. 8 x 8 yGxy `9 xGxx 2. 8 x 8 y Gxy Gyx `8 x 8 y Gxy $ Gyx 3. f8 x Ax Bx ; 9 xAx g`9 xBx 4. f Na !8 x Mx $ Ma ;Ma; : Mb g`: Na 5. `8 z Pz _: Pz 6. `8 xRxx !9 x 9 yRxy 7. `8 y 9 x Qy Qx PartN Showthateachpairofsentencesisprovablyequivalent. 1. 8 x Ax !: Bx :9 x Ax & Bx 2. 8 x : Ax Bd 8 xAx Bd 3. 9 xPx Qc 8 x Px Qc 4. Rca $8 xRxa 8 x Rca $ Rxa PartO Showthateachofthefollowingisprovablyinconsistent. 1. f Sa Tm Tm Sa Tm & : Sa g 2. f9 xRxa 8 x 8 yRyx g 3. f:9 x 9 yLxy Laa g

PAGE 138

138 forall x 4. f8 x Px Qx 8 z Pz Rz 8 yPy : Qa & : Rb g ? PartP Writeasymbolizationkeyforthefollowingargument,translateit, andproveit: Thereissomeonewholikeseveryonewholikeseveryonethathelikes. Therefore,thereissomeonewholikeshimself. PartQ Provideaproofofeachclaim. 1. f Pa Qb;Qb b = c; : Pa g` Qc 2. f m = n n = o;An g` Am Ao 3. f8 xx = m;Rma g`9 xRxx 4. :9 xx 6 = m `8 x 8 y Px Py 5. 8 x 8 y Rxy x = y ` Rab Rba 6. f9 xJx; 9 x : Jx g`9 x 9 yx 6 = y 7. f8 x x = n $ Mx ; 8 x Ox & Mx g` On 8. f9 xDx; 8 x x = p $ Dx g` Dp 9. f9 x Kx & 8 y Ky x = y & Bx ;Kd g` Bd 10. ` Pa !8 x Px x 6 = a PartR LookbackatPartFonp.78.Foreachargument:IfitisvalidinQL, giveaproof.Ifitisinvalid,constructamodeltoshowthatitisinvalid. ? PartS Foreachofthefollowingpairsofsentences:Iftheyarelogically equivalentinQL,giveproofstoshowthis.Iftheyarenot,constructamodel toshowthis. 1. 8 xPx Qc 8 x Px Qc 2. 8 xPx & Qc 8 x Px & Qc 3. Qc _9 xQx 9 x Qc Qx 4. 8 x 8 y 8 zBxyz 8 xBxxx 5. 8 x 8 yDxy 8 y 8 xDxy 6. 9 x 8 yDxy 8 y 9 xDxy ? PartT Foreachofthefollowingarguments:IfitisvalidinQL,giveaproof. Ifitisinvalid,constructamodeltoshowthatitisinvalid. 1. 8 x 9 yRxy : : 9 y 8 xRxy 2. 9 y 8 xRxy : : 8 x 9 yRxy 3. 9 x Px & : Qx : : 8 x Px !: Qx 4. 8 x Sx Ta Sd : :Ta

PAGE 139

ch.6proofs 139 5. 8 x Ax Bx 8 x Bx Cx : : 8 x Ax Cx 6. 9 x Dx Ex 8 x Dx Fx : : 9 x Dx & Fx 7. 8 x 8 y Rxy Ryx : :Rjj 8. 9 x 9 y Rxy Ryx : :Rjj 9. 8 xPx !8 xQx 9 x : Px : : 9 x : Qx 10. 9 xMx !9 xNx :9 xNx : : 8 x : Mx PartU 1.Ifyouknowthat A ` B ,whatcanyousayabout A & C ` B ?Explain youranswer. 2.Ifyouknowthat A ` B ,whatcanyousayabout A C ` B ?Explain youranswer.

PAGE 140

AppendixA Symbolicnotation Inthehistoryofformallogic,dierentsymbolshavebeenusedatdierenttimes andbydierentauthors.Often,authorswereforcedtousenotationthattheir printerscouldtypeset. Inonesense,thesymbolsusedforvariouslogicalconstantsisarbitrary.There isnothingwritteninheaventhatsaysthat` : 'mustbethesymbolfortruthfunctionalnegation.Wemighthavespeciedadierentsymboltoplaythat part.Oncewehavegivendenitionsforwell-formedformulaewandfor truthinourlogiclanguages,however,using` : 'isnolongerarbitrary.Thatis thesymbolfornegationinthistextbook,andsoitisthesymbolfornegation whenwritingsentencesinourlanguagesSLorQL. Thisappendixpresentssomecommonsymbols,sothatyoucanrecognizethem ifyouencountertheminanarticleorinanotherbook. summaryofsymbols negation : conjunction&, ^ disjunction conditional biconditional $ Negation Twocommonlyusedsymbolsarethe hoe ,` : ',andthe swungdash ` .'Insomemoreadvancedformalsystemsitisnecessarytodistinguishbetweentwokindsofnegation;thedistinctionissometimesrepresentedbyusing both` : 'and` .' Disjunction Thesymbol` 'istypicallyusedtosymbolizeinclusivedisjunction. Conjunction Conjunctionisoftensymbolizedwiththe ampersand ,`&.'The ampersandisactuallyadecorativeformoftheLatinword`et'whichmeans `and';itiscommonlyusedinEnglishwriting.Asasymbolinaformalsys140

PAGE 141

appendix:symbolicnotation 141 tem,theampersandisnottheword`and';itsmeaningisgivenbytheformal semanticsforthelanguage.Perhapstoavoidthisconfusion,somesystemsuse adierentsymbolforconjunction.Forexample,` ^ 'isacounterparttothe symbolusedfordisjunction.Sometimesasingledot,` ',isused.Insomeolder texts,thereisnosymbolforconjunctionatall;` A and B 'issimplywritten ` AB .' MaterialConditional Therearetwocommonsymbolsforthematerialconditional:the arrow ,` ',andthe hook ,` .' MaterialBiconditional The double-headedarrow ,` $ ',isusedinsystems thatusethearrowtorepresentthematerialconditional.Systemsthatusethe hookfortheconditionaltypicallyusethe triplebar ,` ',forthebiconditional. Quantiers TheuniversalquantieristypicallysymbolizedasanupsidedownA,` 8 ',andtheexistentialquantierasabackwardsE,` 9 .'Insometexts, thereisnoseparatesymbolfortheuniversalquantier.Instead,thevariableis justwritteninparenthesesinfrontoftheformulathatitbinds.Forexample, `all x are P 'iswritten x Px Insomesystems,thequantiersaresymbolizedwithlargerversionsofthesymbolsusedforconjunctionanddisjunction.Althoughquantiedexpressionscannotbetranslatedintoexpressionswithoutquantiers,thereisaconceptual connectionbetweentheuniversalquantierandconjunctionandbetweenthe existentialquantieranddisjunction.Considerthesentence 9 xPx ,forexample. Itmeansthat either therstmemberoftheUDisa P or thesecondoneis, or thethirdoneis,....Suchasystemusesthesymbol` W 'insteadof` 9 .' Polishnotation ThissectionbrieydiscussessententiallogicinPolishnotation,asystemof notationintroducedinthelate1920sbythePolishlogicianJanLukasiewicz. Lowercaselettersareusedassentenceletters.Thecapitalletter N isused fornegation. A isusedfordisjunction, K forconjunction, C fortheconditional, E forthebiconditional.`A'isforalternation,anothernameforlogical disjunction.`E'isforequivalence. notationPolish ofSLnotation : N & K A C $ E InPolishnotation,abinaryconnectiveiswritten before thetwosentencesthat itconnects.Forexample,thesentence A & B ofSLwouldbewritten Kab in Polishnotation.

PAGE 142

142 forall x Thesentences : A B and : A B areverydierent;themainlogical operatoroftherstistheconditional,butthemainconnectiveofthesecond isnegation.InSL,weshowthisbyputtingparenthesesaroundtheconditional inthesecondsentence.InPolishnotation,parenthesesareneverrequired.The left-mostconnectiveisalwaysthemainconnective.Therstsentencewould simplybewritten CNab andthesecond NCab ThisfeatureofPolishnotationmeansthatitispossibletoevaluatesentences simplybyworkingthroughthesymbolsfromrighttoleft.Ifyouwereconstructingatruthtablefor NKab ,forexample,youwouldrstconsiderthe truth-valuesassignedto b and a ,thenconsidertheirconjunction,andthen negatetheresult.ThegeneralruleforwhattoevaluatenextinSLisnotnearly sosimple.InSL,thetruthtablefor : A & B requireslookingat A and B thenlookinginthemiddleofthesentenceattheconjunction,andthenatthe beginningofthesentenceatthenegation.Becausetheorderofoperationscan bespeciedmoremechanicallyinPolishnotation,variantsofPolishnotation areusedastheinternalstructureformanycomputerprogramminglanguages.

PAGE 143

AppendixB Solutionstoselected exercises Manyoftheexercisesmaybeansweredcorrectlyindierentways.Wherethat isthecase,thesolutionhererepresentsonepossiblecorrectanswer. Chapter1PartC 1.consistent 2.inconsistent 3.consistent 4.consistent Chapter1PartD 1,2,3,6,8,and10arepossible. Chapter2PartA 1. : M 2. M _: M 3. G C 4. : C & : G 5. C : G & : M 6. M C G Chapter2PartC 1. E 1 & E 2 143

PAGE 144

144 forall x 2. F 1 S 1 3. F 1 E 1 4. E 2 & : S 2 5. : E 1 & : E 2 6. E 1 & E 2 & : S 1 S 2 7. S 2 F 2 8. : E 1 !: E 2 & E 1 E 2 9. S 1 $: S 2 10. E 2 & F 2 S 2 11. : E 2 & F 2 12. F 1 & F 2 $ : E 1 & : E 2 Chapter2PartD A: Aliceisaspy. B: Bobisaspy. C: Thecodehasbeenbroken. G: TheGermanembassywillbeinanuproar. 1. A & B 2. A B C 3. : A B !: C 4. G C 5. C _: C & G 6. A B & : A & B Chapter2PartG 1.anobno 2.anobyes 3.ayesbyes 4.anobno 5.ayesbyes 6.anobno 7.anobyes 8.anobyes 9.anobno Chapter3PartA 1.tautology 2.contradiction 3.contingent

PAGE 145

solutionsforch.3 145 4.tautology 5.tautology 6.contingent 7.tautology 8.contradiction 9.tautology 10.contradiction 11.tautology 12.contingent 13.contradiction 14.contingent 15.tautology 16.tautology 17.contingent 18.contingent Chapter3PartB 2,3,5,6,8,and9arelogicallyequivalent. Chapter3PartC 1,3,6,7,and8areconsistent. Chapter3PartD 3,5,8,and10arevalid. Chapter3PartE 1. A and B havethesametruthvalueoneverylineofacompletetruthtable, so A $ B istrueoneveryline.Itisatautology. 2.Thesentenceisfalseonsomelineofacompletetruthtable.Onthatline, A and B aretrueand C isfalse.Sotheargumentisinvalid. 3.Sincethereisnolineofacompletetruthtableonwhichallthreesentences aretrue,theconjunctionisfalseoneveryline.Soitisacontradiction. 4.Since A isfalseoneverylineofacompletetruthtable,thereisnolineon which A and B aretrueand C isfalse.Sotheargumentisvalid. 5.Since C istrueoneverylineofacompletetruthtable,thereisnolineon which A and B aretrueand C isfalse.Sotheargumentisvalid. 6.Notmuch. A B isatautologyif A and B aretautologies;itisacontradictioniftheyarecontradictions;itiscontingentiftheyarecontingent. 7. A and B havedierenttruthvaluesonatleastonelineofacomplete truthtable,and A B willbetrueonthatline.Onotherlines,itmight betrueorfalse.So A B iseitheratautologyoritiscontingent;itis not acontradiction. Chapter3PartF 1. : A B 2. : A !: B

PAGE 146

146 forall x 3. : [ A B !: B A ] Chapter4PartB 1. Rca Rcb Rcc ,and Rcd aresubstitutioninstancesof 8 xRcx 2.Oftheexpressionslisted,only 8 yLby isasubstitutioninstanceof 9 x 8 yLxy Chapter4PartC 1. Za & Zb & Zc 2. Rb & : Ab 3. Lcb Mb 4. Ab & Ac Lab & Lac 5. 9 x Rx & Zx 6. 8 x Ax Rx 7. 8 x Zx Mx Ax 8. 9 x Rx & : Ax 9. 9 x Rx & Lcx 10. 8 x Mx & Zx Lbx 11. 8 x Mx & Lax Lxa 12. 9 xRx Ra 13. 8 x Ax Rx 14. 8 x Mx & Lcx Lax 15. 9 x Mx & Lxb & : Lbx Chapter4PartG 1. :9 xTx 2. 8 x Mx Sx 3. 9 x : Sx 4. 9 x [ Cx & :9 yByx ] 5. :9 xBxx 6. :9 x Cx & : Sx & Tx 7. 9 x Cx & Tx & 9 x Mx & Tx & :9 x Cx & Mx & Tx 8. 8 x [ Cx !8 y : Cy Bxy ] 9. 8 x )]TJ/F8 9.9626 Tf 4.566 -8.07 Td [( Cx & Mx !8 y [ : Cy & : My Bxy ] Chapter4PartI 1. 8 x Cxp Dx 2. Cjp & Fj 3. 9 x Cxp & Fx

PAGE 147

solutionsforch.4 147 4. :9 xSxj 5. 8 x Cxp & Fx Dx 6. :9 x Cxp & Mx 7. 9 x Cjx & Sxe & Fj 8. Spe & Mp 9. 8 x Sxp & Mx !:9 yCyx 10. 9 x Sxj & 9 yCyx & Fj 11. 8 x Dx !9 y Sxy & Fy & Dy 12. 8 x Mx & Dx !9 y Cxy & Dy Chapter4PartK 1. 8 x Cx Bx 2. :9 xWx 3. 9 x 9 y Cx & Cy & x 6 = y 4. 9 x 9 y Jx & Ox & Jy & Oy & x 6 = y 5. 8 x 8 y 8 z Jx & Ox & Jy & Oy & Jz & Oz x = y x = z y = z 6. 9 x 9 y )]TJ/F11 9.9626 Tf 4.566 -8.07 Td [(Jx & Bx & Jy & By & 8 z [ Jz & Bz x = z y = z ] 7. 9 x 1 9 x 2 9 x 3 9 x 4 Dx 1 & Dx 2 & Dx 3 & Dx 4 & x 1 6 = x 2 & x 1 6 = x 3 & x 1 6 = x 4 & x 2 6 = x 3 & x 2 6 = x 4 & x 3 6 = x 4 & :9 y Dy & y 6 = x 1 & y 6 = x 2 & y 6 = x 3 & y 6 = x 4 8. 9 x )]TJ/F11 9.9626 Tf 4.566 -8.07 Td [(Dx & Cx & 8 y [ Dy & Cy x = y ]& Bx 9. 8 x Ox & Jx Wx & 9 x Mx & 8 y My x = y & Wx 10. 9 x )]TJ/F11 9.9626 Tf 4.566 -8.07 Td [(Dx & Cx & 8 y [ Dy & Cy x = y ]& Wx !9 x 8 y Wx $ x = y 11.widescope: :9 x Mx & 8 y My x = y & Jx narrowscope: 9 x Mx & 8 y My x = y & : Jx 12.widescope: :9 x 9 z )]TJ/F11 9.9626 Tf 4.566 -8.07 Td [(Dx & Cx & Mz & 8 y [ Dy & Cy x = y ]& 8 y [ My z = y & x = z ] narrowscope: 9 x 9 z )]TJ/F11 9.9626 Tf 4.566 -8.07 Td [(Dx & Cx & Mz & 8 y [ Dy & Cy x = y ]& 8 y [ My z = y & x 6 = z ] Chapter5PartA 2,3,4,6,8,and9aretrueinthemodel. Chapter5PartB 2,4,5,and7aretrueinthemodel. Chapter5PartD UD= f 10,11,12,13 g extension O = f 11,13 g extension S = ; extension T = f 10,11,12,13 g extension U = f 13 g extension N = f < 11,10 > < 12,11 > < 13,12 > g referent m =Johnny Chapter5PartE

PAGE 148

148 forall x 1.Thesentenceistrueinthismodel: UD= f Stan g extension D = f Stan g referent a =Stan referent b =Stan Anditisfalseinthismodel: UD= f Stan g extension D = ; referent a =Stan referent b =Stan 2.Thesentenceistrueinthismodel: UD= f Stan g extension T = f < Stan,Stan > g referent h =Stan Anditisfalseinthismodel: UD= f Stan g extension T = ; referent h =Stan 3.Thesentenceistrueinthismodel: UD= f Stan,Ollie g extension P = f Stan g referent m =Stan Anditisfalseinthismodel: UD= f Stan g extension P = ; referent m =Stan Chapter5PartF Therearemanypossiblecorrectanswers.Herearesome: 1.Makingtherstsentencetrueandthesecondfalse: UD= f alpha g extension J = f alpha g extension K = ; referent a =alpha 2.Makingtherstsentencetrueandthesecondfalse: UD= f alpha,omega g extension J = f alpha g referent m =omega 3.Makingtherstsentencefalseandthesecondtrue: UD= f alpha,omega g extension R = f < alpha,alpha > g

PAGE 149

solutionsforch.5 149 4.Makingtherstsentencefalseandthesecondtrue: UD= f alpha,omega g extension P = f alpha g extension Q = ; referent c =alpha 5.Makingtherstsentencetrueandthesecondfalse: UD= f iota g extension P = ; extension Q = ; 6.Makingtherstsentencefalseandthesecondtrue: UD= f iota g extension P = ; extension Q = f iota g 7.Makingtherstsentencetrueandthesecondfalse: UD= f iota g extension P = ; extension Q = f iota g 8.Makingtherstsentencetrueandthesecondfalse: UD= f alpha,omega g extension R = f < alpha,omega > < omega,alpha > g 9.Makingtherstsentencefalseandthesecondtrue: UD= f alpha,omega g extension R = f < alpha,alpha > < alpha,omega > g Chapter5PartI 1.Therearemanypossibleanswers.Hereisone: UD= f Harry,Sally g extension R = f < Sally,Harry > g referent a =Harry 2.Therearenopredicatesorconstants,soweonlyneedtogiveaUD.Any UDwith2memberswilldo. 3.Weneedtoshowthatitisimpossibletoconstructamodelinwhichthese arebothtrue.Suppose 9 xx 6 = a istrueinamodel.Thereissomethingin theuniverseofdiscoursethatis not thereferentof a .Sothereareatleast twothingsintheuniverseofdiscourse:referent a andthisotherthing. Callthisotherthing |weknow a 6 = .Butif a 6 = ,then 8 x 8 yx = y isfalse.Sotherstsentencemustbefalseifthesecondsentenceistrue. Assuch,thereisnomodelinwhichtheyarebothtrue.Therefore,they areinconsistent. Chapter5PartJ

PAGE 150

150 forall x 2.No,itwouldnotmakeanydierence.Thesatisfactionofasentencedoes notdependonthevariableassignment.Soasentencethatissatisedby some variableassignmentissatisedby every othervariableassignment aswell. Chapter6PartA 1 W !: B 2 A & W 3 B J & K 4 W &E2 5 : B E1,4 6 J & K E3,5 7 K &E6 1 L $: O 2 L _: O 3 : L 4 : O E2,3 5 L $ E1,4 6 : L R3 7 L : E3{6 1 Z C & : N 2 : Z N & : C 3 : N C 4 : N & : C DeM3 5 Z 6 C & : N E1,5 7 C &E6 8 : C &E4 9 : Z : I5{8 10 N & : C E2,9 11 N &E10 12 : N &E4 13 N C : E3{12 Chapter6PartB 1. 1 K & L want K $ L 2 K want L 3 L &E1 4 L want K 5 K &E1 6 K $ L $ I2{3,4{5

PAGE 151

solutionsforch.6 151 2. 1 A B C want A & B C 2 A & B want C 3 A &E2 4 B C E1,3 5 B &E2 6 C E4,5 7 A & B C I2{6 3. 1 P & Q R 2 P !: R want Q E 3 P &E1 4 : R E2,3 5 Q R &E1 6 Q E5,4 7 Q E I6 4. 1 C & D E want E D 2 : E want D 3 C & D E1,2 4 D &E3 5 : E D I2{4 6 E D MC5 5. 1 : F G 2 F H want G H 3 : G want H 4 :: F MT1,3 5 F DN4 6 H E2,5 7 : G H I3{6 8 G H MC7

PAGE 152

152 forall x 6. 1 X & Y X & Z 2 : X & D 3 D M want M 4 : X forreductio 5 : X _: Y I4 6 : X & Y DeM5 7 X & Z E1,6 8 X &E7 9 : X R4 10 X : E4{9 11 : M forreductio 12 D E3,11 13 X & D &I10,12 14 : X & D R2 15 M : E11{14 Chapter6PartH 1 8 x 9 y Rxy Ryx 2 8 x : Rmx 3 9 y Rmy Rym 8 E1 4 Rma Ram 5 : Rma 8 E2 6 Ram E4,5 7 9 xRxm 9 I6 8 9 xRxm 9 E3,4{7 1 8 x 9 yLxy !8 zLzx 2 Lab 3 9 yLay !8 zLza 8 E1 4 9 yLay 9 I2 5 8 zLza E3,4 6 Lca 8 E5 7 9 yLcy !8 zLzc 8 E1 8 9 yLcy 9 I5 9 8 zLzc E7,8 10 Lcc 8 E9 11 8 xLxx 8 I10

PAGE 153

solutionsforch.6 153 1 8 x Jx Kx 2 9 x 8 yLxy 3 8 xJx 4 Ja 8 E3 5 Ja Ka 8 E1 6 Ka E5,4 7 8 yLay 8 Laa 8 E7 9 Ka & Laa &I6,8 10 9 x Kx & Lxx 9 I9 11 9 x Kx & Lxx 9 E2,7{10 1 : 9 xMx _8 x : Mx 2 :9 xMx & :8 x : Mx DeM1 3 :9 xMx &E2 4 8 x : Mx QN3 5 :8 x : Mx &E2 6 9 xMx _8 x : Mx : E1{5 Chapter6PartI 1. 1 : 8 xFx _:8 xFx forreductio 2 :8 xFx & ::8 xFx DeM1 3 :8 xFx &E2 4 ::8 xFx &E2 5 8 xFx _:8 xFx : E1{4 2. 1 8 x Mx $ Nx 2 Ma & 9 xRxa want 9 xNx 3 Ma $ Na 8 E1 4 Ma &E2 5 Na $ E3,4 6 9 xNx 9 I5

PAGE 154

154 forall x 3. 1 8 x : Mx Ljx 2 8 x Bx Ljx 3 8 x Mx Bx want 8 xLjx 4 : Ma Lja 8 E1 5 Ma Lja _! 4 6 Ba Lja 8 E2 7 Ma Ba 8 E3 8 Lja 7,5,6 9 8 xLjx 8 I8 4. 1 8 x Cx & Dt want 8 xCx & Dt 2 Ca & Dt 8 E1 3 Ca &E2 4 8 xCx 8 I3 5 Dt &E2 6 8 xCx & Dt &I4,5 5. 1 9 x Cx Dt want 9 xCx Dt 2 Ca Dt for 9 E 3 : 9 xCx Dt forreductio 4 :9 xCx & : Dt DeM3 5 : Dt &E4 6 Ca E2,5 7 9 xCx 9 I6 8 :9 xCx &E4 9 9 xCx Dt : E3{8 10 9 xCx Dt 9 E1,2{9 Chapter6PartP Regardingthetranslationofthisargument,seep.67.

PAGE 155

solutionsforch.6 155 1 9 x 8 y [ 8 z Lxz Lyz Lxy ] 2 8 y [ 8 z Laz Lyz Lay ] 3 8 z Laz Laz Laa 8 E2 4 :9 xLxx forreductio 5 8 x : Lxx QN4 6 : Laa 8 E5 7 :8 z Laz Laz MT5,6 8 Lab 9 Lab R8 10 Lab Lab I8{{9 11 8 z Laz Laz 8 I10 12 :8 z Laz Laz R7 13 9 xLxx : E4{{12 14 9 xLxx 9 E1,2{{13 Chapter6PartS 2,3,and5arelogicallyequivalent. Chapter6PartT 2,4,5,7,and10arevalid.Herearecompleteanswersfor someofthem: 1. UD= f mocha,freddo g extension R = f < mocha,freddo > < freddo,mocha > g 2. 1 9 y 8 xRxy want 8 x 9 yRxy 2 8 xRxa 3 Rba 8 E2 4 9 yRby 9 I3 5 8 x 9 yRxy 8 I4 6 8 x 9 yRxy 9 E1,2{5

PAGE 156

QuickReference CharacteristicTruthTables A : A T F F T A B A & B A B A B A $ B T T T T T T T F F T F F F T F T T F F F F F T T Symbolization SententialConnectives chapter2 Itisnotthecasethat P : P Either P ,or Q P Q Neither P ,nor Q : P Q or : P & : Q Both P ,and Q P & Q If P ,then Q P Q P onlyif Q P Q P ifandonlyif Q P $ Q Unless P Q P unless Q P Q Predicates chapter4 All F sare G s. 8 x Fx Gx Some F sare G s. 9 x Fx & Gx Notall F sare G s. :8 x Fx Gx or 9 x Fx & : Gx No F sare G s. 8 x Fx !: Gx or :9 x Fx & Gx Identity section4.6 Only j is G 8 x Gx $ x = j Everythingbesides j is G 8 x x 6 = j Gx The F is G 9 x Fx & 8 y Fy x = y & Gx `TheFisnotG'canbetranslatedtwoways: ItisnotthecasethattheFisG.wide :9 x Fx & 8 y Fy x = y & Gx The F isnonG .narrow 9 x Fx & 8 y Fy x = y & : Gx 156

PAGE 157

Usingidentitytosymbolizequantities Thereareatleast F s. one 9 xFx two 9 x 1 9 x 2 Fx 1 & Fx 2 & x 1 6 = x 2 three 9 x 1 9 x 2 9 x 3 Fx 1 & Fx 2 & Fx 3 & x 1 6 = x 2 & x 1 6 = x 3 & x 2 6 = x 3 four 9 x 1 9 x 2 9 x 3 9 x 4 Fx 1 & Fx 2 & Fx 3 & Fx 4 & x 1 6 = x 2 & x 1 6 = x 3 & x 1 6 = x 4 & x 2 6 = x 3 & x 2 6 = x 4 & x 3 6 = x 4 n 9 x 1 9 x n Fx 1 & & Fx n & x 1 6 = x 2 & & x n )]TJ/F7 6.9738 Tf 6.226 0 Td [(1 6 = x n Thereareatmost F s. Onewaytosay`atmost n thingsare F 'istoputanegationsigninfrontofone ofthesymbolizationsaboveandsay : `atleast n +1thingsare F .'Equivalently: one 8 x 1 8 x 2 Fx 1 & Fx 2 x 1 = x 2 two 8 x 1 8 x 2 8 x 3 Fx 1 & Fx 2 & Fx 3 x 1 = x 2 x 1 = x 3 x 2 = x 3 three 8 x 1 8 x 2 8 x 3 8 x 4 Fx 1 & Fx 2 & Fx 3 & Fx 4 x 1 = x 2 x 1 = x 3 x 1 = x 4 x 2 = x 3 x 2 = x 4 x 3 = x 4 n 8 x 1 8 x n +1 Fx 1 & & Fx n +1 x 1 = x 2 __ x n = x n +1 Thereareexactly F s. Onewaytosay`exactly n thingsare F 'istoconjointwoofthesymbolizations aboveandsay`atleast n thingsare F '&`atmost n thingsare F .'Thefollowing equivalentformulaeareshorter: zero 8 x : Fx one 9 x Fx & :9 y Fy & x 6 = y two 9 x 1 9 x 2 Fx 1 & Fx 2 & x 1 6 = x 2 & :9 y )]TJ/F11 9.9626 Tf 4.567 -8.07 Td [(Fy & y 6 = x 1 & y 6 = x 2 three 9 x 1 9 x 2 9 x 3 Fx 1 & Fx 2 & Fx 3 & x 1 6 = x 2 & x 1 6 = x 3 & x 2 6 = x 3 & :9 y Fy & y 6 = x 1 & y 6 = x 2 & y 6 = x 3 n 9 x 1 9 x n Fx 1 & & Fx n & x 1 6 = x 2 & & x n )]TJ/F7 6.9738 Tf 6.226 0 Td [(1 6 = x n & :9 y Fy & y 6 = x 1 & & y 6 = x n SpecifyingthesizeoftheUD Removing F fromthesymbolizationsaboveproducessentencesthattalkabout thesizeoftheUD.Forinstance,`thereareatleast2thingsintheUD'may besymbolizedas 9 x 9 y x 6 = y 157

PAGE 158

BasicRulesofProof Reiteration m A A R m ConjunctionIntroduction m A n B A & B &I m n ConjunctionElimination m A & B A &E m B &E m DisjunctionIntroduction m A A B I m B A I m DisjunctionElimination m A B n : B A E m n m A B n : A B E m n ConditionalIntroduction m A want B n B A B I m { n ConditionalElimination m A B n A B E m n BiconditionalIntroduction m A want B n B p B want A q A A $ B $ I m { n p { q BiconditionalElimination m A $ B n B A $ E m n m A $ B n A B $ E m n NegationIntroduction m A forreductio n )]TJ/F8 9.9626 Tf 9.963 0 Td [(1 B n : B : A : I m { n NegationElimination m : A forreductio n )]TJ/F8 9.9626 Tf 9.963 0 Td [(1 B n : B A : E m { n

PAGE 159

QuantierRules ExistentialIntroduction m A 9 xA [ x jj c ] 9 I m x mayreplacesomeoralloccurrencesof c in A ExistentialElimination m 9 xA n A [ c j x ] p B B 9 E m n { p Theconstant c mustnotappearin 9 xA ,in B ,or inanyundischargedassumption. UniversalIntroduction m A 8 xA [ x j c ] 8 I m c mustnotoccurinanyundischargedassumptions. UniversalElimination m 8 xA A [ c j x ] 8 E m IdentityRules c = c =I m c = d n A A [ c jj d ]=E m n A [ d jj c ]=E m n Oneconstantmayreplacesomeoralloccurrences oftheother. DerivedRules Dilemma m A B n A C p B C C m n p ModusTollens m A B n : B : A MT m n HypotheticalSyllogism m A B n B C A C HS m n ReplacementRules Commutivity Comm A & B B & A A B B A A $ B B $ A DeMorgan DeM : A B : A & : B : A & B : A _: B DoubleNegation DN :: A A MaterialConditional MC A B : A B A B : A B BiconditionalExchange $ ex [ A B & B A ] A $ B QuantifierNegation QN :8 xA 9 x : A :9 xA 8 x : A

PAGE 160

IntheIntroductiontohisvolume SymbolicLogic CharlesLutwidgeDodsonadvised:Whenyou cometoanypassageyoudon'tunderstand, read itagain :ifyou still don'tunderstandit, readit again :ifyoufail,evenafter three readings,very likelyyourbrainisgettingalittletired.Inthat case,putthebookaway,andtaketootheroccupations,andnextday,whenyoucometoitfresh, youwillverylikelyndthatitis quite easy." Thesamemightbesaidforthisvolume,although readersareforgiveniftheytakeabreakforsnacks after two readings. abouttheauthor: P.D.MagnusisanassistantprofessorofphilosophyinAlbany,NewYork.Hisprimaryresearchis inthephilosophyofscience,concernedespecially withtheunderdeterminationoftheorybydata.