turnsouttobeeitherhopelessormeaningless。Milldoesnotdirectlyassertthatall’ontological’speculationsaremerelycobwebsofthebrain。Buthetriestoshowthat，whatevertheymaybe，theyarestrictlyirrelevantinreasoning。Allmetaphysiciansareexpectedtogranthimcertainpostulates。
Theseoncegranted，hewillbeabletoaccountforthewholestructureofknowledge。’Intuitions，’transcendentalspeculations，andontologywillthenbedeprivedofthewholeconditionsunderwhichtheythrive。Idonotnowassert，hevirtuallysays，thatyourdoctrineiswrong，butIshallshowthatitisthrownaway。Itisapretenceofexplainingsomethingwhichliesaltogetherbeyondthelimitsofrealknowledge，andthereforeadmitsofnoexplanation。
Millstartsfromtheclassificationgiveninoldlogicaltextbooks，towhich，differentasarehisconclusions，heattachedaveryhighvalue。10*Theschoolmenhadbytheirelaborateacutenessestablishedawholesystemoflogicaldistinctionsanddefinitionswhicharebothimportantandaccurate，howeversteriletheinquiriesinwhichtheywereused。
Themachinerywasexcellent，thoughitscontriversforgotthatamillcannotgrindoutflourifyouputinnograin。Millbeginsaccordinglybyclassifyingthevariouskindsofwordsinthelightaffordedbypreviouslogicalsystems。
Heistogiveatheoryofproof。Thatwhichistobeprovedisaproposition；andapropositiondealswithnames，andmoreoverwiththenamesof’things，’notmerelywiththenames’ofourideasofthings。’11*That，insomesense，reasoninghastodowiththingsisofcoursehisessentialprinciple；andtheproblemconsequentlyarises，Whatareempirical’things’？12*
Thoughwecannotaskwhatare’ultimatethings，’thelogicianmustenumeratethevariouskindsofthingstowhichreferencemaybemadeinpredication。Millmakesoutaclassificationwhichheproposestosubstitute，provisionallyatanyrate，fortheAristoteliancategories。13*Thefirstandsimplestclassofnameablethingscorrespondstothings’inthemind，’thatis，’feelings，’or’statesofconsciousness，’sensations，emotions，thoughts，andvolitions。Thesecondclasscorrespondstothings’externaltothemind’。14*andtheseareeither’substances’or’attributes。’Hereourtaskislightenedbyawelcomediscovery。
Allphilosophers，itappears，arenowagreedupononepoint。SirW。Hamilton，Cousin，Kant，nay，accordingtoHamilton——thoughthatistoogoodtobetrue——nearlyallpreviousphilosophersadmitonetruth。15*Weknow，astheyagree，nothingabout’objects’exceptthesensationswhichtheygiveusandtheorderofthosesensations。Hencethetwo’substances，’bodyandmind，remainunknowable’inthemselves。’Bodyisthe’hiddenexternalcause’towhichwereferoursensations；16*andasbodyisthe’mysterioussomethingwhichexcitesthemindtofeel，somindisthemysterioussomethingwhichfeelsandthinks。’Themindis，ashesaysinlanguagequotedfromhisfather，’athreadofconsciousness，’aseriesof’feelings’：itisthe’myself’whichisconceivedasdistinctfromthefeelingsbutofwhichIcanyetknownothingexceptthatithasthefeelings。
Thus，althoughweknownothingofmindsandofbodies’inthemselves，’wedoknowtheirexistence。Thatisessentialtohisposition。The’threadofconsciousness’isa’finalinexplicability’withhim，butitcorrespondstosomerealentity。And，ontheotherside，wemustbelieve，insomesense，inthings。Thething，thoughknownonlythroughthesensationswhichitexcites，mustbesomethingmorethanameresensation，forthewholeofhislogicdefendsthethesisthatinsomewayorotherthoughthastoconformtofactsortotherelationsbetween’things。’Knowledge，however，isconfinedentirelytothesensationsandtheattributes；andthetwoareatbottomone。The’verbal’distinctionbetweenapropertyofthingsandthesensationwhichwereceivefromit，ismade，hesays，forconvenienceofdiscourseratherthanfromanydifferenceinthenatureofthethingdenoted。17*Thisbringsustoacriticalpoint。Attributes，hesays，followingtheolddistribution，areofQuality，Quantity，andRelation。NowQualityandQuantitymeansimplythesensationsexcitedbybodies。Tosaythatsnowiswhite，orthatthereisagallonofwater，meanssimplythatcertainsensationsofcolourorsizeareexcitedinusbysnoworavolumeofwater。Theattributecalled’Relation’introducesadifferentorderoffeelings。A’relation’supposesthattwothingsareinvolvedinsomeonefactorseriesoffacts。18*Butitisstillan’attribute’ora’stateofconsciousness。’Itisafeelingdifferentfromotherfeelingsbythecircumstancethattwo’things’insteadofoneareinvolved。Thisistheexplanationwhich，aswehaveseen，hepraisessowarmlyinhisfather’sAnalysis，andnowadoptsforhisownpurposes。Itenableshimtoclassifypredications。Allpredicationiseitheranassertionofsimpleexistenceoranassertionof’relations。’Byclassifyingthepossiblerelations，therefore，weobtainthepossibleformsofpredication。Itturnsoutaccordinglythatwecanmakefivepossiblepredications：wecanpredicate，first，simpleexistence；
orsecondly，’coexistence’；orthirdly，’sequence’thesetwobeingequivalent，asheadds，to’orderinplace’and’orderintime’；orfourthly，wemaypredicate’resemblance’；orfifthly，andthisisonlytobestatedprovisionally，wemaypredicate’causation。’19*
Sofar，Mill’sviewcorrespondstothepsychologyoftheAnalysis，whichgivesasimilaraccountofthevarioustermsemployed。J。S。Millhasnowthestandinggroundfromwhichhecanexplainthewholedevelopmentofknowledge。Atthispoint，however，hehastodivergefromhisfather’sextremenominalism。
Predication，accordingtotheelder，isaprocessofnaming。A
predicateisanameofthesamethingofwhichthesubjectisaname；andtopredicateissimplytoassertthisidentityofnames。Thisdoctrine，asMillthinks，isequallyimpliedinthedictumdeomnietnullowhichistakenastheexplanationofthesyllogism。Wehavearbitrarilyputanumberofthingsinaclass，andto’reason’issimplytorepeatofeachwhatwehavesaidofall。Thisistoputthecartbeforethehorse，ortoassumethattheclassificationprecedesthereasonforclassification，thoughprobablythetheory，thusnakedlystated，wouldnotbegrantedbyanyone。20*What，then，isthetruetheory？Thatisexplainedbythedistinctionbetween’connotation’and’denotation，’whichMillacceptedthoughinvertingtheuseofthewordsfromhisfather。Ageneralnamesuchas’man’denotesJohn，Thomas，andotherindividuals。Itconnotescertain’attributes，’suchasrationalityandacertainshape。When，therefore，IsaythatJohnisaman，Isaythathehastheattributes’connoted’；andwhenI
saythatallmenaremortal，Iassertthatalongwiththeotherattributesofmangoestheattributeofmortality。21*
Predication，then，ingeneral，involvestheattributeof’relation。’Wemayassertthesimpleexistenceofa’quality，’
or，whichisthesamething，ofa’sensation’；buttosaythatJohnisaman。orthatmenaremortal，ortomakeanyofthegeneralpropositionswhichconstituteknowledge，istoassertsomeofthose’relations’whichareperceivedwhenweconsidertwoormorethingstogether。
’Things，’then，sofarasknowableareclustersinHartley’slanguageof’attributes’；andtheattributesmaybeequallyregardedas’feelings。’Topredicateistoreferathingtooneoftheclusters，andthereforetoassertitspossessionoftheattributesconnoted。Iwillonlynoteinpassingthatbydecliningtogointothemetaphysicalquestionastothedifferencebetween’attributes’and’sensations，’orthoughtsandthings，Millleavesanobscurityatthefoundationofhisphilosophy。Butleavingthisforthepresent，itisenoughtosaythatwehaveourfivepossibletypesofpredication。22*Allpropositionsmaybereducedtooneoftheforms。Thingsexistorcoexistorfolloworresembleorarecauseoreffect。23*Thenextproblem，therefore，is，Howarethesepropositionstobeproved？or，bywhattestsisourbelieftobejustified。
WhatmaybethenatureofbeliefitselfisaquestionwhichMillleavestotheanalyticalpsychologist，24*who，asheadmits，willprobablyfinditpuzzling，ifnothopeless。Butasweallagreethatsomehoworotherweattainknowledge，wemayinquirewhatisimpliedintheprocess。Now，somepartofourknowledgeobviouslydependsupon’experience。’Weknowofanyparticularfactfromthetestimonyofoursenses。WeknowthatLondonBridgeexistsbecausewehaveseenandtouchedit；anditwouldbeobviouslyhopelesstotrytodeducetsexistencefromtheprincipleoftheexcludedmiddle。LondonBridgewouldthenbesomethingindependentoftimeandplace。Butdowenotwantsomethingmorethanbareexperiencewhenwelaydownageneralruleisalawofnature？Thenwenotonlysay’is，’but’mustbe’；andthis，accordingtotheIntuitionist，markstheintroductionofsomethingmorethananappealto’experience。’
Therearetruths，hesays，whichrepresent’lawsofthought’；
whichareself-evident，orperceivedby’intuition’；orthecontraryofwhichis’inconceivable。’Withoutsomesuchlaws，wecouldnotbindtogethertheshiftingdataofexperience，oradvancefrom’is’to’mustbe，’orevento’willbe。’Weloseallcertainty，andfallintothescepticismofHume，whichmakesbeliefamere’custom，’regardsallthingsasdistinctatomsconjoinedbutnotconnected，andholdsthat’anythingmaybethecauseofanything。’Mill’saimistoexplodetheintuitionswithoutfallingintothescepticism。Necessarytruths，heholds，aremerefigments。Allknowledgewhateverisoftheempiricaltype。’Thishasbeen’justifies’thiswillbe。’Empiricaltruthsclearlyexist，andareheldundoubtingly，althoughtheyhavenofoundationexceptexperience。Nobodyeverdoubtedthatallmendie；yetno’proof’ofthefactcouldbeeversuggested，beforephysiologywascreated，exceptthebarefactthatallmenhavedied。Ifphysiologyhasmadethenecessitymoreevident，ithasnotappreciablystrengthenedtheconviction。Weallbelieveevennowthatthunderwillfollowlightning，thoughnobodyhasbeenabletoshowwhyitshouldfollow。Theultimateproofincountlesscases，ifnotinall，issimplythatsomeconnectionhasbeenobserved，and，inmanysuchcases，thebeliefreachesapitchwhichexcludesallperceptibledoubt。Asafact，then，beliefofthestrongestkindcanbegeneratedfromsimpleexperience。Theburthenofproofisuponthosewhoassumedifferentoriginsfordifferentclassesoftruth。25*
II。SYLLOGISMANDDEFINITION
Thismainthesisleadstotwolinesofargument。Firstofall，Millseekstoshowthatthemethodsofproofexpoundedbyhisadversariesdonotreallytakeusbeyondexperience；and，secondly，heseekstoshowthatexperiencegivesusasufficientbasisofknowledge。Letusfirstnotice，then，howthegroundisclearedbyexaminingpreviousaccountsoftheprocessofinference。Theoldtheoryofreasoningdependsuponthesyllogism。Thatgivesthetypeofthewholeprocessbywhichknowledgeisextended。Allmenaremortal；Socratesisaman，thereforeSocratesismortal。StewartandBrownhadbothattackedthesyllogismonthefamiliargroundthatitistautologous。Themajorhasalreadyassertedtheminor。Tosaythatonemanismortalwhenyouhavealreadysaidthatallmenaremortal，ismerelytorepeatyourself。Therecanbenorealinference，andnoadvancetonewknowledge。Solongasthesyllogismistobeexplainedontheoldterms，Millthinksthiscriticismfatal；butheholds，too，thatbyadifferentinterpretationwemayassignarealandvitallyimportantmeaningtothisvenerableformofargument。Inseveralplaces26*hegivesaviewwhichseemstobemuchtothepurpose。Thesyllogism，itwouldseem，correspondsreally，nottoamodeofreasoning，buttoasystemofarguing。
Whenadisputantbasessomestatementuponaninference，wemaychallengeeitherthetruthoftheruleorthestatementoffact。