Thereis，however，anotherwayinwhichthehabitofmindnecessarilyresultingfromthecontinualpracticeofhedonisticcomparisonissometimesthoughttobeunfavourabletotheattainmentofthehedonisticend:fromasupposedincompatibilitybetweenthehabitofreflectivelyobservingandexaminingpleasure，andthecapacityforexperiencingpleasureinnormalfulnessandintensity。Anditcertainlyseemsimportanttoconsiderwhateffectthecontinualattentiontoourpleasures，inordertoobservetheirdifferentdegrees，islikelytohaveonthesefeelingsthemselves。Theinquiryatfirstsightseemstoleadtoirreconcilablecontradictioninourviewofpleasure。Forifpleasureonlyexistsasitisfelt，themoreconsciousweareofit，themorepleasurewehave:anditwouldseemthatthemoreourattentionisdirectedtowardsit，themorefullyweshallbeconsciousofit。OntheotherhandHamilton’sstatementthat``knowledgeandfeeling’’cognitionandpleasureorpainarealways``inacertaininverseproportiontoeachother’’，correspondsprimafacietoourcommonexperience:forthepurelycognitiveelementofconsciousnessseemstobeneitherpleasurablenorpainful，sothatthemoreourconsciousnessisoccupiedwithcognition，thelessroomthereseemstobeforfeeling。
Thisview，however，restsontheassumptionthatthetotalintensityofourconsciousnessisaconstantquantity；sothatwhenoneelementofitpositivelyincreases，therestmustpositively——aswellasrelatively——diminish。Anditdoesnotappeartomethatexperiencegivesusanyvalidgroundformakingthisgeneralassumption:itratherseemsthatatcertaintimesinourlifeintellectandfeelingaresimultaneouslyfeeble；sothatthesamementalexcitementmayintensifybothsimultaneously。
Stillitseemstobeafactthatanyverypowerfulfeeling，reaching，tothefullintensityofwhichourconsciousnessisnormallycapable，iscommonlydiminishedbyacontemporaneousstrokeofcognitiveeffort:henceitisageneraldifficultyinthewayofexactobservationofouremotionsthattheobjectcognisedseemstoshrinkanddwindleinproportionasthecognitiveregardgrowskeenandeager。Howthenarewetoreconcilethiswiththepropositionfirstlaiddown，thatpleasureonlyexistsasweareconsciousofit？Theanswerseemstobethatthemereconsciousnessofapresentfeeling——apartfromanydistinctrepresentativeelements——cannotdiminishthefeelingofwhichitisanindispensableandinseparablecondition:butinintrospectivecognitionwegobeyondthepresentfeeling，comparingandclassifyingitwithrememberedorimaginedfeelings；andtheeffortofrepresentingandcomparingtheseotherfeelingstendstodecreasethemerepresentativeconsciousnessoftheactualpleasure。
Iconclude，then，thatthereisarealdangerofdiminishingpleasurebytheattempttoobserveandestimateit。Butthedangerseemsonlytoariseinthecaseofveryintensepleasures，andonlyiftheattemptismadeatthemomentofactualenjoyment；andsincethemostdelightfulperiodsoflifehavefrequentlyrecurringintervalsofnearlyneutralfeeling，inwhichthepleasuresimmediatelypastmaybecomparedandestimatedwithoutanysuchdetriment，Idonotregardtheobjectionfoundedonthisdangerasparticularlyimportant。
Moreserious，inmyopinion，aretheobjectionsurgedagainstthepossibilityofperforming，withdefiniteandtrustworthyresults，thecomprehensiveandmethodicalcomparisonofpleasuresandpainswhichtheadoptionoftheHedonisticstandardinvolves。Icannotindeeddoubtthatmenhabituallycomparepleasuresandpainsinrespectoftheirintensity:thate。g。whenwepassfromonestateofconsciousnesstoanother，orwheninanywayweareledtorecallastatelongpast，weoftenunhesitatinglydeclarethepresentstatetobemoreorlesspleasantthanthepast:orthatwedeclaresomepleasantexperiencestohavebeen`worth’，andothers`notworth’，thetroubleittooktoobtainthem，orthepainthatfollowedthem。But，grantingthis，itmaystillbemaintained1thatthiscomparisonasordinarilymadeisbothoccasionalandveryrough，andthatitcanneverbeextendedassystematicHedonismrequires，norapplied，withanyaccuracy，toallpossiblestateshoweverdifferinginquality；and2thatascommonlypractiseditisliabletoillusion，ofwhichwecannevermeasurethepreciseamount，whilewearecontinuallyforcedtorecogniseitsexistence。ThisillusionwasevenurgedbyPlatoasagroundfordistrusting，theapparentaffirmationofconsciousnessinrespectofpresentpleasure。Platothoughtthattheapparentintensityofthecoarserbodilypleasureswasillusory；becausethesestatesofconsciousness，beingprecededbypain，werereallyonlystatesofrelieffrompain，andsoproperlyneutral，neitherpleasantnorpainful——examplesofwhatIhavecalledthehedonisticzero——onlyappearingpleasantfromcontrastwiththeprecedingpain。
Tothis，however，ithasbeenanswered，thatinestimatingpleasurethereisnoconceivableappealfromtheimmediatedecisionofconsciousness:thatherethePhenomenalistheReal——thereisnootherrealthatwecandistinguishfromit。Andthisseemstometrue，insofarasweareconcernedonlywiththepresentstate。Butthen——apartfromthedifficultyjustnoticedofobservingapleasurewhileitisfeltwithouttherebydiminishingit——itisobviousthatinanyestimateofitsintensitywearenecessarilycomparingitwithsomeotherstate。Andthislattermustgenerallybearepresentation，notanactualfeeling:forthoughwecansometimesexperiencetwoorperhapsmorepleasuresatonce，wearerarelyinsuchcasesabletocomparethemsatisfactorily:foreitherthecausesofthetwomutuallyinterfere，sothatneitherreachesitsnormaldegreeofintensity；or，moreoften，thetwoblendintoonestateofpleasantconsciousnesstheelementsofwhichwecannotestimateseparately。Butifitisthereforeinevitablethatonetermatleastinourcomparisonshouldbeanimaginedpleasure，weseethatthereisapossibilityoferrorinanysuchcomparison；fortheimaginedfeelingmaynotadequatelyrepresentthepleasantnessofthecorrespondingactualfeeling。Andintheegoisticcomparison，thevalidityofwhichwearenowdiscussing，theobjectsprimarilytobecomparedareallrepresentedelementsofconsciousness:forwearedesiringtochoosebetweentwoormorepossiblecoursesofconduct，andthereforetoforecastfuturefeelings。
Letusthenexaminemorecloselythemannerinwhichthiscomparisonisordinarilyperformed，thatwemayseewhatpositivegroundswehaveformistrustingit。
Inestimatingforpracticalpurposesthevalueofdifferentpleasuresopentous，wecommonlytrustmosttoourprospectiveimagination:weprojectourselvesintothefuture，andimaginewhatsuchandsuchapleasurewillamounttounderhypotheticalconditions。Thisimagination，sofarasitinvolvesconsciousinference，seemstobechieflydeterminedbyourownexperienceofpastpleasures，whichareusuallyrecalledgenerically，orinlargeaggregates，thoughsometimesparticularinstancesofimportantsinglepleasuresoccurtousasdefinitelyremembered:butpartly，too，weareinfluencedbytheexperienceofotherssympatheticallyappropriated:andhereagainwesometimesdefinitelyrefertoparticularexperienceswhichhavebeencommunicatedtousbyindividuals，andsometimestothetraditionalgeneralisationswhicharethoughttorepresentthecommonexperienceofmankind。
Nowitdoesnotseemthatsuchaprocessasthisislikelytobefreefromerror:and，indeed，noonepretendsthatitis。
Infactthereisscarcelyanypointuponwhichmoralisershavedweltwithmoreemphasisthanthis，thatman’sforecastofpleasureiscontinuallyerroneous。Eachofusfrequentlyrecogniseshisownmistakes:andeachstillmoreoftenattributestootherserrorsunseenbythemselves，arisingeitherfrommisinterpretationoftheirownexperience，orfromignoranceorneglectofthatofothers。
Howthenaretheseerrorstobeeliminated？Theobviousansweristhatwemustsubstitutefortheinstinctive，largelyimplicit，inferencejustdescribedamorescientificprocessofreasoning:
bydeducingtheprobabledegreeofourfuturepleasureorpaininanygivencircumstancesfrominductivegeneralisationsbasedonasufficientnumberofcarefulobservationsofourownandothers’experience。Wehavethentoask，first，howfarcaneachofusestimateaccuratelyhisownpastexperienceofpleasuresandpains？secondly，bowfarcanthisknowledgeofthepastenablehimtoforecast，withanycertainty，thegreatesthappinesswithinhisreachinthefuture？thirdly，howfarcanheappropriate，forthepurposesofsuchforecasts，thepastexperienceofothers？
Asregardsthefirstofthesequestions，itmustberememberedthatitisnotsufficienttoknowgenerallythatwederivepleasuresandpainsfromsuchandsuchsources；werequiretoknowapproximatelythepositiveornegativedegreeofeachfeeling；unlesswecanformsomequantitativeestimateofthem，itisfutiletotrytoattainourgreatestpossiblehappiness——atleastbyanempiricalmethod。Wehavethereforetocomparequantitativelyeachpleasureasitoccurs，orasrecalledinimagination，withotherimaginedpleasures:andthequestionis，howfarsuchcomparisonscanberegardedastrustworthy。
Nowformyownpart，whenIreflectonmypleasuresandpains，andendeavourtocomparetheminrespectofintensity，itisonlytoaverylimitedextentthatIcanobtainclearanddefiniteresultsfromsuchcomparisons，eventakingeachseparatelyinitssimplestform:——whetherthecomparisonismadeatthemomentofexperiencingoneofthepleasures，orbetweentwostatesofconsciousnessrecalledinimagination。ThisistrueevenwhenIcomparefeelingsofthesamekind:andthevaguenessanduncertaintyincreases，inproportionasthefeelingsdifferinkind。Letusbeginwithsensualgratifications，whicharethoughttobeespeciallydefiniteandpalpable。SupposeIamenjoyingagooddinner:ifIaskmyselfwhetheronekindofdishorwinegivesmemorepleasurethananother，sometimesIcandecide，butveryoftennot。SoifIreflectupontwomodesofbodilyexercisethatImayhavetaken:ifonehasbeeninamarkeddegreeagreeableortedious，Itakenoteofitnaturally；butitisnotnaturaltometogofurtherthanthisinjudgingoftheirpleasurablenessorpainfulness，andtheattempttodosodoesnotseemtoleadtoanyclearaffirmation。
Andsimilarlyofintellectualexercisesandstatesofconsciousnesspredominantlyemotional:evenwhenthecausesandqualityofthefeelingscomparedaresimilar，itisonlywhenthedifferencesinpleasantnessaregreat，thathedonisticcomparisonseemstoyieldanydefiniteresult。ButwhenItrytoarrangeinascalepleasuresdifferinginkind；tocomparee。g。
labourwithrest，excitementwithtranquillity，intellectualexercisewithemotionaleffusion，thepleasureofscientificapprehensionwiththatofbeneficentaction，thedelightofsocialexpansionwiththedelightofaestheticreception；myjudgmentwaversandfluctuatesfarmore，andinthemajorityofcasesIcannotgiveanyconfidentdecision。AndifthisisthecasewithwhatBenthamcalls`pure’——i。e。painless——pleasures，itisstillmoretrueofthoseevencommonerstatesofconsciousness，whereacertainamountofpainordiscomfortismixedwithpleasure，althoughthelatterpreponderates。Ifitishardtosaywhichoftwodifferentstatesofcontentmentwasthegreaterpleasure，itseemsstillhardertocompareastateofplacidsatisfactionwithoneofeagerbuthopefulsuspense，orwithtriumphantconquestofpainfulobstacles。Andperhapsitisstillmoredifficulttocomparepurepleasureswithpurepains，andtosayhowmuchoftheonekindoffeelingweconsidertobeexactlybalancedbyagivenamountoftheotherwhentheydonotoccursimultaneously:whileanestimateofsimultaneousfeelingsis，aswehaveseen，generallyunsatisfactoryfromthemutualinterferenceoftheirrespectivecauses。