ThetheoryismostexplicitlyworkedoutintheGrammarofAssent。Newmanexertsallhisskillinexpoundingaverysounddoctrine。Asamatteroffact，weforminnumerablejudgmentsbywhathecallsthe’illativesense’；thatistosay，notbyformalargument，butbyacomplexsystemof’implicit’reasonings。
’Logic，’ashesays，’doesnotreallyprove。Itenablesustojoinissuewithothers……itverifiesnegatively’；andfor’genuineproofinconcretematterwerequireanorganismmoredelicate，versatile，andelastic，thanverbalargumentation。’244*Logicisachainwhich’hangslooseatbothends，’245*forthefirstprinciplesmustbeassumed，andtheabstractconceptneverfitstheactualcomplexityofconcretefact。Bythe’illativesense，’again，wereachinnumerabletruths。WeholdthatEnglandisanisland，orthatthemanwhomweseeisourbrother，withafaithindistinguishablefromabsoluteconviction。Wegofurther；webelievethatafriendishonest，or，say，thatCaesarcrossedtheRubicon，withoutadmittingtheslightestscrupleofdoubt。Allknowledgewhateveroffactplainlyimpliessomethingdifferentfromformallogic；
and，sofar，theonlyquestionseemstobewhysopalpableatruthneedssoelaborateandgracefulanexposition。TheanswerisindicatedbythepolemicagainstLocke。Lockehadproposed，asatestofaloveoftruth，therefusaltoholdanyproposition，withagreaterassurancethantheproofsitisbuiltonwillwarrant。’246*ThestatementseemstobenotonlyunassailablebutinconformitywithNewman’sdoctrine。ShouldwebelieveEnglandtobeanisland？WhenJuliusCaesarlanded，itwasnotproved；andhewouldhavebeenwrongtobecertain。Whendiditbecomerighttobecertain？Surelyatwhatevermomentitwasadequatelyproved。Itisneversoprovedthattodenyitwouldbeself-contradictory，butbythistimeitisasmuchprovedasanyfactcanbeproved。Lockewouldsimplyjustifyhimselfbysayingthatinthiscaseour’assurance’doesnotexceedthe’proofsonwhichitisbuilt。’Theapproximationtodemonstrationisindefinitelyclose，thoughneverabsolute，andthedifferencebecomestoosmalltobeperceptible。Adifficultyemergesonlyifweatonceadmittherightnessofbeliefanddenythesufficiencyoftheevidence。
Newman，havingshownthatwebelieveinconcretetruthsnotprovedbyabstractlogic，arguesthatwealsoassumemanytruthsnotprovedevenbysufficientempiricalevidence。WehavewhatLockecalleda’surplusageofassurance。’Thefact，again，isundeniable。Webelieveimplicitlyincountlessthingsuponinsufficientevidence。This，asLockewouldadd，isonemainexplanationoftheprevalenceoferror，andalsoaproofthaterrormaybeinnocent。Itisadutytobecandid；itcannotbeadutytoberight。Wemustlistentoreason；buttheeffectofreasoningmustdependupontheconstitutionofourminds，andthevariousbeliefswithwhichtheyarealreadystored。NowtoNewmanthisdoctrinealwaysseemstobesceptical。Itamountstothe’liberalising’viewthatallcreedsareequallygoodifonlytheybeequallysincere。Hencehelaysstressuponthedoctrinethat’assent’isavolitionalaswellasan’intellectualact。’Itisourdutytoobeythereason；andwhenthe’illativesense’
declaresthetruthofaproposition，weareboundtoan’activerecognition’ofthetruth。247*Locke，onthecontrary，holdsthatifwelistentoreason，theassentfollowsautomaticallybyanon-voluntaryact。