Perception and Cognition Are Largely Independent, but Still ...

By 'cognitive penetration' most authors intend a general category of cognitive influences on how perceptual information is processed by sensory ... ThisarticleispartoftheResearchTopic Pre-cueingEffectsonPerceptionandCognitivePenetrability Viewall 10 Articles Articles AthanassiosRaftopoulos UniversityofCyprus,Cyprus MariaOlkkonen DurhamUniversity,UnitedKingdom RobertL.West CarletonUniversity,Canada Theeditorandreviewers'affiliationsarethelatestprovidedontheirLoopresearchprofilesandmaynotreflecttheirsituationatthetimeofreview. Abstract Introduction:EvolutionaryArgumentsforaPerceptionandCognitionInterface DefininganInterfaceforCognitivePenetrationthatdoesnotJeopardizeEarlyPerception ConceptAcquisition CadasaFrameworkofDistinctionsforEmotion,Perception,andJudgment CognitivelyDrivenAttention:Feature-Based,Syntactic,andSemantic Conclusion AuthorContributions Funding ConflictofInterestStatement Acknowledgments References SuggestaResearchTopic> DownloadArticle DownloadPDF ReadCube EPUB XML(NLM) Supplementary Material Exportcitation EndNote ReferenceManager SimpleTEXTfile BibTex totalviews ViewArticleImpact SuggestaResearchTopic> SHAREON OpenSupplementalData HYPOTHESISANDTHEORYarticle Front.Psychol.,24January2017 |https://doi.org/10.3389/fpsyg.2017.00040 PerceptionandCognitionAreLargelyIndependent,butStillAffectEachOtherinSystematicWays:ArgumentsfromEvolutionandtheConsciousness-AttentionDissociation CarlosMontemayor1*andHarryH.Haladjian2 1DepartmentofPhilosophy,SanFranciscoStateUniversity,SanFrancisco,CA,USA 2LaboratoirePsychologiedelaPerception,CNRS,UniversitéParisDescartes,Paris,France Themainthesisofthispaperisthattwoprevailingtheoriesaboutcognitivepenetrationaretooextreme,namely,theviewthatcognitivepenetrationispervasiveandtheviewthatthereisasharpandfundamentaldistinctionbetweencognitionandperception,whichprecludesanytypeofcognitivepenetration.Theseoppositeviewshaveclearmeritsandempiricalsupport.Toeliminatethispuzzlingsituation,wepresentanalternativetheoreticalapproachthatincorporatesthemeritsoftheseviewsintoabroaderandmorenuancedexplanatoryframework.Akeyargumentwepresentinfavorofthisframeworkconcernstheevolutionofintentionalityandperceptualcapacities.Animplicationofthisargumentisthatcasesofcognitivepenetrationmusthaveevolvedmorerecentlyandthatthisiscompatiblewiththecognitiveimpenetrabilityofearlyperceptualstagesofprocessinginformation.Atheoreticalapproachthatexplainswhythisshouldbethecaseistheconsciousnessandattentiondissociationframework.Thepaperdiscusseswhyconcepts,particularlyissuesconcerningconceptacquisition,playanimportantroleintheinteractionbetweenperceptionandcognition. Introduction:EvolutionaryArgumentsforaPerceptionandCognitionInterface Thispapercriticallyassessestheviewthattherearesystematicandrobustinfluencesfromcognitiononperceptionattheearlystagesofprocessing,whichcouldbeconsideredcasesofcognitivepenetration.Whileweagreewiththecriticismsthatthereareempirical“pitfalls”intheexperimentsallegedlyreportingcognitivepenetration(seeFirestoneandScholl,2016),therealsoaredifficultiesregardingtheviewthatthereisasharpdistinctionbetweenperception(theprocessingofsensoryinformationthatoccursatseverallevels)andcognition(thejudgingofrepresentationalcontentsrelatedtoreasoning).Besidesbeingproblematictheoretically,theassumptionthatasharpdistinctionbetweenallcognitionandallperceptionmustbeanessentialaspectofthemindmayevenbeempiricallyfalse.Thecriticismsaroundthenotionofpenetrabilityneedtobemorebalancedsothatitaccountsforanarchitectureconsistingofsomecognitivelyimpenetrablemodules(characteristicofearlyperception)alongwithothersthataresusceptibletotop-downinfluences(characteristicoflateperception).Suchvariedeffectsmustbeavailableinperceptiontounderstandabilitiessuchaspredictivecodingandconceptualattention. Wefocusonconceptacquisitiontoexplaintheinterfacebetweencognitivelypenetrableperceptionandcognitivelyimpenetrableperception,andparticularlyonthefactthatconceptacquisitionisalsoaperceptual,ratherthanastrictlycognitiveprocessinvolvingonlyreasoningorjudgment.Evenifthebrain’sarchitectureisorganizedinamodularandencapsulatedway,therecanstillbeaconceptualinterfacebetweenperceptionandcognition.Itisatthisconceptualinterface,whichisalsoresponsibleforexplicitordiscursivejudgmentandinference,wheremostinteractionsbetweenperceptionandcognitionwilloccurthatcancontaininstancesofcognitivepenetration.Wewillexploretheissueofconceptacquisitionatdifferentstagesofprocessingandexplainhowitrelatestotop-downpre-cueing.Thisrelationwillreinforceourpointthatabalancedcombinationofanypossiblecognitivepenetrabilityandearlyimpenetrabilityiscritical.Infact,weaimtoshowthatconceptualinterfacesbetweencognitionandperceptionarecrucialforunderstandinghowourspeciesdevelopedsophisticatedformsofattention. Oneapproachtoachievethisbalancedinterfaceperspectiveisbasedontheconsciousnessandattentiondissociation(CAD)framework(MontemayorandHaladjian,2015).Thisframeworkcharacterizestherelationshipbetweenconsciousnessandattention,andclaimsthatattentionissignificantlydissociatedfromconsciousness,withdifferentlevelsofinteractionsbetweenattentionandconsciousawareness.Thisdistinctionisimportantbecausethereisabundantevidenceofcognitiveeffectsonattentionwithoutconsciousawareness—anunconsciousformofcognitiveguidance—aswellascasesinwhichmotivationalstatesguideimplicitly,sometimesagainsttheconsciousjudgmentsofsubjects,asincasesofimplicitbias(seeMontemayorandHaladjian,2015,foradiscussionoftheevidenceinvision).Thesecasesofguidanceandselectioninperceptionmaybeconceivedasattentionroutines,andmanywillbeindependent,andevendisagree,withconsciousperception.Crucially,forthetopicofcognitivepenetration,CADallowsforthesystematicguidanceoflateperceptionbycognitivelydrivenattention,whilealsoallowingforthecognitiveimpenetrabilityofearlyperception. Thesedifferenttypesofguidanceandinfluenceonlateperception(voluntaryandinvoluntary,consciousandunconscious)helpclarifysomeproblemsconcerningextantdiscussionsoncognitivepenetration.Someallegedcasesofcognitivepenetrationmayreadilybeexcluded,forinstancecasesofexplicitvoluntaryjudgmentonperceptualcontentsthatarenotevenindirectlyinfluencedbybeliefsordiscursiveinference.Somemotivationalandemotionalformsofguidancearemoreproblematic,astheytypicallyoccurindependentlyofexplicitpropositionalattitudes,althoughtheycaneasilybeunderstoodaspartoftheattentionalguidanceonlateperception.Butitseemsthatifallimplicitformsofmotivationalandcognitiveguidanceareexcluded,asdefendedbythesharpdelineationview,thenitistooeasytoconcludethatperceptionisneverpenetrablebycognition.Wewillresistthisconclusionbyarguingthatitisanimplausibleviewofthecomplexityofperception—andofitsevolution.Wewillalsoargueagainsttheoppositeviewthatcognitivepenetrationiswidespread,assomeproponentsofcognitivepenetrabilitypropose.Someformsofperception,specificallyearlyperceptualprocessing,mustbeimpenetrable.Theproblemisoneofbalance:theremustbesystematicformsofinfluenceonperceptionwithoutmajordisturbancestotheevolutionarilydevelopedandrequiredperceptualinvariancesforsuccessfulnavigationandmotorcontrol.Thedissociationbetweenconsciousnessandattentionprovidesthismorenuancedtheoreticalapproach,anditadvancesthedebatebeyondthestrictdichotomybetweencognitionandperception. Inaddition,theCADframeworkisparticularlywellsuitedtoaddresscognitivepenetrationbecauseitissupportedbyavastamountoffindings,specificallyinvisionscience(MontemayorandHaladjian,2015).The‘earlyversuslateperception’distinctionwasintroducedinordertointerpretfindingsinvisionscience.Earlyvisualperceptionincludessensoryprocessesthatarespecializedforhandlingspecifictypesofinformationusedinconstructingrepresentationsindependentlyofbeliefs(Pylyshyn,1999).Lateperceptioninvolvesselectiveprocessingbytop-downattentionandothercognitiveprocesses(Raftopoulos,2015b).Justlikethisdistinctiongeneralizestootherperceptualmodalitiesandtothemoregeneraldistinctionbetweencognitionandperception,CADalsogeneralizestoallkindsofdissociationsbetweensubjectivelyconsciousexperienceandattentionroutinesthatdonotnecessitateconsciousawareness,includingemotionsandmemory.ThecentraltenetoftheCADframeworkisthattheremustbesomedissociationbetweenattentionandphenomenalconsciousness(subjectiveexperience)withsomeextanttheoriesindicatingahighdegreeofdissociation.Thus,CADisaframeworktobetterunderstand,model,andintegratefindingsandtheoriesonconsciousnessandattentionbasedonhowtheyaredissociatedfromeachother.InthispaperwepresenttheimplicationsofCADforthetopicofcognitivepenetration. Thecrucialargumentwemakeinsupportoftheseclaimsconcernsevolution.AlsobasedontheCADframework,theargumentisasfollows. (a)Perceptualsystemsevolvedfrombasictocomplexformsofprocessing,andsomearelesscognitivelypenetrablethanothers(e.g.,earlyperceptualstagesarecognitivelyimpenetrable). (b)Perceptualstatesproducedbysuchsystemscanbedefinedintermsofintentionality(thewayinwhichmentalrepresentationsareaboutthingsandfeaturesintheworld):perceptionisalwaysaboutfeaturesoftheenvironmentthatcaninprinciplebeinterpretedpropositionally(althoughtheyneednotbepropositionaltobeintentional). (c)Ifperceptualsystemsevolved,thenintentionalityalsoevolved. Therefore,someformsofintentionalityaremorecognitivelypenetrablethanothers,andaninterfaceforpenetrabilityisneededforconceptacquisitionandglobalaccess(includingaccesstopropositionalcontent). Thisargumentshowswhyevolutionmatterstothedebateoncognitivepenetration,andwhypenetrabilityismorecomplicatedthanpreviouslythought.CADcanhelpexplaintherelationshipbetweencognitionandperception,andindicatewherecasesofpenetrabilitymayoccur.Forinstance,onepossibilityisthattheremayneedtobetwointerfacesbetweencognitionandperception,oneconcerningphenomenallyconsciousexperiencesandanotherconcerningnon-phenomenallyconsciousperceptualcontents.Suchinterfaceswillbecriticalforallkindsofconceptualandpre-conceptuallearningthatguideattentionroutines. Adiscussionaboutwhatismeantby‘cognitivepenetration’isrequiredtofullyunderstandtheimplicationsofthisargument.By‘cognitivepenetration’mostauthorsintendageneralcategoryofcognitiveinfluencesonhowperceptualinformationisprocessedbysensorymechanisms,whichincludescasesinwhichthebeliefsanddesiresofperceiverssomehowdeterminewhattheyperceive.Thisofcoursecanbeinterpretedinmanyways.Thedemarcationbetweencognitivelypenetrableandimpenetrableperceptionwasoriginallyproposedtounderstandcognitivearchitecture,butitnowencompassescasesinwhichtop-downattentioninfluencesbottom-upearlyattentionroutines,independentlyofspecificcommitmentsregardingarchitecture(VetterandNewen,2014).Asmentioned,viewsatoneendofthepossibledegreesofpenetrabilitydenythatcognitivepenetrationcapturesatrulyuniquetypeofinfluenceofcognitiononperceptualprocessing(e.g.,FirestoneandScholl,2016).Suchviewswouldneverconsidersystematicinfluencesofcognitiononperceptionaslegitimatecasesofcognitivepenetration.Ontheotherhand,viewsthatstatethatthereisnoboundarybetweencognitionandperceptiondenythatcognitioncouldbedissociatedfromperception(e.g.,Clark,2013). Thus,acriticalissueishowtoclearlyspecifylegitimatecasesofcognitivepenetration—casesinwhichtheinfluenceofcognitiononperceptionisnottrivialoreasilyexplainedbyappealtoinference(FirestoneandScholl,2016),orsomeothercognitiveprocesssuchasjudgmentorinterpretation.Thisbecomesespeciallyimportantwhenauthorsarguingforthecaseofpenetrabilitydothisbygivingexamplesofchangestohigherlevelsinperception,thosethatarebeyondtheinitialstagesofsensoryprocessing.Forexample,somefindingsindicatethatthroughoutthestagesofperceptualprocessingtherearebothforwardandbackwardneuralprojectionsthatcontributetoperception(e.g.,VetterandNewen,2014).Yet,thesedonotnecessarilyindicatethatearlyperceptionispenetrablebycognition.Wearguethatthemoreinterestingcasesofcognitivepenetrationwouldnotbeatthehigherlevelofperceptualjudgmentortheinterpretationoftheoutputfromsensoryprocessing.Norwouldtheybecaseswherevoluntaryattentionsimplychangestheperceptualstimulusorinput(e.g.,lookingtotheleftbasedonmydesiretochangemygazeshouldnotcountasacaseofcognitivepenetration).Radicalcasesofpenetrabilitywouldinfluenceperceptualprocessingdirectlyatearlystages,andnotsimplyatahigherattentive(orcognitive)level. Morespecifically,themostproblematicformofcognitivepenetrationwouldhavetooccuratthelevelofprocessingcalled‘earlyvision’orearlyperceptionmoregenerally(seePylyshyn,1999).Instancesofradicalcognitivepenetrabilityshouldshowthatperception,particularlyearlyperception,cannot“resist”theinfluenceofcontentcomingfrominferences,beliefs,ordesires.Thiscouldhappenquiteselectively:notallbeliefsanddesirescandirectlyaffectperception,butonlysomespecificonesinspecificsituations.Whatiscrucialisthatifradicalcognitivepenetrationexists,thenthereisthepossibilityofcausalinfluencesfromcognitionthatdirectlymodifyperception,evenwhenallelseisbeingequalatthesensoryinputlevel,includinghowattentionisbeingallocated.Thiscausalinfluencemustexplaindirectlyhowearlyperceptionisprocessed—otherwise,purelyconceptualinfluencescouldexplaincognitivepenetration(seeRaftopoulos,2014,pp.605–606fordiscussion).Weshallargueagainstthisradicalformofcognitivepenetration. Cognitivepenetrationisacrucialtopicinphilosophyofperceptionbecauseofhowitrelatestocontroversialissuesinepistemologyorthetheoryofknowledge.Forinstance,thereistheviewthatthecontentsofperceptionarepropositional(i.e.,theyhavetruthconditions,justlikethepropositionsexpressedbysentences),andthatperceptionisakintobelief—akindofpropositionalattitude(Byrne,2005).Thereisalsotheviewthatperceptionneednothavepropositionalcontent(Crane,2009).Thisissueisclearlyrelatedtothetopicofnon-conceptualcontentinperception.Inthesedebates,itisgenerallytakenforgrantedthatthefocusofanalysisisperceptualconsciousexperience.ButCADshowsthisisanassumptionthatshouldnotbetakenforgrantedbecausewhatistrueaboutphenomenallyconsciousperceptionneednotbetrueaboutperceptioningeneral—therearetypesofnon-phenomenallyconsciousperceptionasinblindsight(e.g.,seeKentridge,2011).Moreimportant,CADexplainswhytheseapparentlyoppositeviewscouldbetrueaboutdifferenttypesofperception—onecognitivelypenetrableatthepropositional,laterperceptuallevelandtheothercognitivelyimpenetrableatthenon-conceptual,earlyperceptuallevel.Aswearguebelow,thisisactuallyaconsequenceoftheargumentfromevolution. ToillustratetheimportanceofCADtounderstanddifferenttypesofcognitiveimpenetrability,considerthemostbasickindofconsciousexperience,forinstanceofcolor.OnepossibilityCADallowsforisthatearlycolorperceptionisexperiencedintheexactsamewayasinotherorganismsthatlackthetop-downroutinesdependentoncognitivecapacities.Thispossibilityplaysamajorroleinmotivatingthenotionofphenomenalconsciousness,particularlyfor“firstorder”theorists,whodenythatexperiencesmustbepartofathoughtorrepresentationforthemtobeconscious.Thisapproachsuggeststhatmanyspecies,certainlymammals,musthavephenomenalexperiencesthatareanalogoustohumanphenomenalconsciousness.Forsuchoverlapinexperiencesofcolor,itseemsnecessarytoadopttheviewthatearlyvisioncolorisimpenetrable(fordissentseeMacpherson,2012).Sowhataboutcolorperceptionthatisprocessedattheinterfacewithworkingmemory,conceptualcategorization,andmotivationalguidance(e.g.,perceptuallyjudgingthetypicalcolorofanobjectorevaluatingthebeautyofacombinationofcolors)?Atthislevel,itisclearthatcolorperceptionwouldbesusceptibletodifferentkindsoftop-downeffects,andthesecouldcountascognitivepenetrationatlaterstagesofprocessing.Inhumans,thesetwotypesofperceptualprocessingcomeapart,andonlyCADmakessenseofthispossibility:consciousearly(bottom-up)visionwithouttop-downattentionmodulationandconceptualizedcolordetection,susceptibletocognitiveandmotivationalmodulation.Anintriguingpossibility,entailedbytheargumentfromevolution,isthatsomeanimalsexperiencecolorinamodular,andmoreencapsulatedway,becausetheylacktheconceptualinterfacesrequiredforlateperceptualmodulationandjudgment. Theconsciousnessandattentiondissociationthushelpsusunderstandthecognitiveimpenetrabilityofearlyperceptualprocesses,withoutmaintainingthatthereisnoroomforcognitivepenetrabilityatmoreintegratedlevelsofperceptionandcognition,inawaythatgeneratesaninteractionbetweentheselevels.Italsofacilitatesthetheoreticalcharacterizationofcognitiveinfluencesonunconsciousperceptionthatplaynoroleinconsciousexperience,andviceversa.Combinedwiththeargumentfromevolution,CADjustifiestheimpenetrabilityofearlyperceptionbasedontheimportanceofperceptualinvariancestonavigatetheenvironment,forexample,whichmusthaveevolvedearlyon,independentlyofcognitiveandmotivationalinfluences.Itispreciselybecausedifferentkindsofintentionalityevolvedatdifferenttimesthattheremustbeinterfacesbetweenperceptionandcognition,someofwhichneednotbefullyfledgedconceptualinference.Thisiswhyprocessesinvolvedinconceptacquisitionarerelevantforstrikingabalancebetweenthe‘pervasivecognitivepenetration’and‘nocognitivepenetration’views. Likeanytheoreticalcategory,thatof‘earlyvision’(whichcanbeextendedtoearlyperception)hasfuzzyboundaries.Thereisagreement,however,thatearlyvisionmustincludemodularlyspecific(cognitivelyimpenetrable)featuredetection,suchascolor,motion,ororientation,typicallybeforetheinvolvementofworkingmemory.Itmayalsoinvolveobjecthood,withoutthecognitiveimprintofconceptualcategories.Onemaysaythatattheveryfirststagesofperception,thereissentienceofphenomenallyexperiencedfeatures,structuredspatiallyandtemporally,whichcanbecross-modallyintegratedbyfeaturemaps.Thisprocessingmustpreserveexternalinvariancesconcerninglightreflectance,shape,distance,andduration(amongmanyotherinvariancesthatallowforreliablyaccuratenavigationandcoordinatedmotorcontrol).Inthissense,perceptualinvariancesarepreservedbycognitiveimpenetrabilityfrommotivationalandconceptualattentionmodulation(atleastinhumans).Thelaterinvolvementofworkingmemoryallowsforsuchcognitiveandemotionalmodulation,andwhatwasconsciouslyexperiencedwithouttheimprintofcategorizationisnowexperiencedunderaconceptualormotivationalinfluenceorcategory.Thiscognitivetransitionhasimplicationsforhowtounderstandperceptioninotherspeciesandalsowithrespecttotheevolutionofourownperceptualsystem.ThisisoneofthereasonswhyCADandtheargumentfromevolutionmustinformourunderstandingofcognitivepenetration. Basedontheseconsiderations,itseemsthattherearetwokindsofcognitiveimpenetrability:phenomenallyconscious(basicfeatureperception)andnon-phenomenallyconscious(featuredetectionoutsideofawareness).Likewise,theremightbetwokindsofcognitivepenetrability,onephenomenal(motivationalinfluencesonperception)andtheothernon-phenomenal(conceptualinfluencesinblindsight-likedetection).Onceconceptualcapacitiesareinthepicture,however,onecanalwaysinterpretperceptualcontentsbyprovidingapropositionalexplanationorinterpretation.Considerthecontrastbetweenexplaininganddirectlycausingthecontentsofperception.Intypicalcasesofautomaticoreffortlessinference,youcaninferthatsomeoneislatebylookingattheirfacialexpressionorhowtheyarelookingattheirwatch,butthisdoesnotmeanthatyouareseeing“lateness.”Emotionperceptionismorecomplicated,butitmightbesusceptibletosimilarinterpretativetreatments(fordissent,seeSiegel,2006;NewenandVetter,2017).Wecaninfersomeone’sjoythroughtheirfacialexpressions,butwedonotnecessarilyseetheactualfeelingofjoy.Inthissense,inferencecaninfluencewhatsomeoneperceiveswithoutchangingradicallyhowthevisualsystemperceivesenvironmentalfeatures,whichwouldremainimpenetrable.Whatcausesthecontentsofperceptionatearlystagesremainsuntouchedbytop-downmodulation. Suchinferentialinfluencescouldbeimplicitandnotdependonanykindofvoluntaryguidance.Thenotionof‘inference’isflexibleenoughthatitcouldoccuratallstagesofperceptualpredictiveprocessinginperception(seeClark,2013),wheresuchprocessingcanbeinfluencedbythestatisticalpropertiesofexperiencesorcontexts(e.g.,seeYuilleandKersten,2006).Thismoreflexiblenotionseemstoproblematizethedistinctionbetweenimpenetrableandpenetrableperception,butoncetheCADframeworkisinplace,onecanargue,basedpartlyontheargumentfromevolution,thatearlyperceptualstatisticalprocessingneednotbeconsideredsusceptibleofanytop-downinfluence.Suchprobabilisticinformationaboutperceptualpropertiesiscompatiblewithencapsulation(Raftopoulos,2015a). Acriticalpointthatdeservesemphasisisthatcognitivepenetrationshouldnotjeopardizethestableinvariancesofperception.Thisconstraintisparticularlyimportantforresultsthataimtoshowputativeformsofpenetrabilityconcerningbasicinformationfornavigation,suchasinformationconcerningdistanceanddepth.Ifpenetrationoccursinthesecases,itmustbeshownthattheyarenotpervasivetothedegreethatsomeonewhoissimplywalkingoutofaroomwouldbedisorientedbythechangesinsize,distance,anddepththatarebasedonherbeliefsanddesires.Ifcognitivepenetrationentailedthiskindofdisruptionofbasicperceptualinvariances,thensuchcasesofpenetrabilitywouldbejustasdisorienting,ifnotmoredisorienting,thanhallucinations.Typically,hallucinationsareexplainedintermsofchangesinphysiology(e.g.,adeliberateneurophysiologicalchangecausedbyingestingcertaindrugs),ratherthansimplechangesinbeliefanddesire.Thus,animportantconstraintisthatcognitivepenetrationshouldnotbeconceivedinwaysthatwouldentailradicalalterationstoperception,analogoustothosecausedbyphysiologyfromexternalsources.Perception(e.g.,earlyvision)mustpreserveinvariancesreliably.Fortrulyradicalcognitivepenetrationtooccur,theremustbeevidencethattop-downconceptualinformationinfluencestheearlystagesofvisualperceptionbeyondsimplyfacilitatingtheprocessingofvisualinformation(e.g.,attentionaleffects)(Raftopoulos,2015b). Asmentioned,anotherimportantconsiderationisthenotionofintentionality(i.e.,thewayinwhichmentalrepresentationsareaboutthingsandfeaturesintheworld)andhowevolutioncanexplainit.Intentionalitymaybeverybasic,processedinamodularfashion,andresponsivetoimmediateinformationfromtheenvironment,oritcanbemoreabstract,categorical,andinfluencedbyjudgmentsandinferences.Variousformsofintentionalitywillcorrespondtotheevolutionaryrecordofsuchcapacities,aswellashowwidespreadtheyareacrossspecies(theearlier,themorewidespread).Intentionalitywillrequireaconceptualinterfaceatsomelevel,atleastinhumans,especiallywhenfacedwithnovelstimuliorsituationsthatdemandcategorization.Itisthisareaofconceptualdevelopmentthatrequiresscrutinyintermsofpotentialinterfacesforcognitivepenetrabilityoflateperceptualstagesofprocessing. Conceptacquisitionofperceptualcategories,wepropose,isthebestexampleofwhyaninterfacebetweenperceptionandcognitionisneeded.Interestingcasesofcognitivepenetrationcouldbedefinedintermsofsuchinterfacesconcerningconceptacquisition,andthisisthestrategywefollowhere.Animportantquestioniswhethertherearepre-cuingeffectsonconceptacquisition.Sincepre-cueingdetermineshowattentionisallocatedandcanchangethebackgroundneuralactivityinawaythathelpsdeterminewhatisperceived,itmayalsodetermineorbiashowaconceptisobtainedorcategorizedthroughperception.Therelationbetweencategoricalreasoningandcategoricalcapacitiesbasedonwhatethologistscall‘fixedactionpatterns’isonethatdeservesattentioninthisregard.Athoroughevaluationoftheevolutionofintentionalityacrossdifferentspeciesshouldincludeanexaminationofpre-cuingeffectsontheseproto-conceptualintentionalrepresentations. DefininganInterfaceforCognitivePenetrationthatdoesnotJeopardizeEarlyPerception Amoreessentialstartingpointistodefinewhatismeantbyperceptionandcognition.Perceptionistheprocessingofexternalinformationbythesensorysystems,suchasvisualorauditoryinformation.Ithasvariousstages,andcanbebroadlycategorizedbetweenearlyperception,whichiscomprisedofencapsulatedsensoryprocessingmodules(e.g.,seePylyshyn,1999;Raftopoulos,2015b),andlateperception,whichincludesmulti-modalintegration,eventperception,andobjectrecognition(e.g.,seeCavanagh,2011).Perceptualinformationprocessingoftenleadstothesubjectiveexperienceofthatinformation,forexample,ofseeinganobjectorhearingasound.Yetsensoryprocessingdoesnotneedtoenterconsciousawarenesstobeperceptuallyregistered—alotofitcanhappeninthebackground.Importantly,perceptionisconsideredtobeessentiallya“belief-independent”process(particularlytheearlykind).Akeyquestion,notoriouslydifficultinepistemology,ishowcansuchbelief-independentprocessesjustifybeliefs?Again,thisissueconcernstheinterfacebetweenperceptionandcognition. WhenIseeanapple,forexample,myvisualsystemisprocessinginformationaboutthefeaturesofthisobject,buthowexactlyissuchprocessingrelatedtothejustificationofmybeliefthatIamseeinganapple?IfallIperceiveisshapeandcolorthenthejustificationofmybeliefismostlyindependentofperceptionanditmustbesomekindofinference.ButthereisnoprobleminsayingthatIseeanapple(orthatIseeanobjectasanapple),andthatwhatIseejustifiesmybeliefsbecauseofthetop-downmodulationofconcepts.Thisiscompatiblewiththeencapsulationofcolorandshapeperception,andCADisparticularlyhelpfulinexplaininghowthisispossible.Thishelpssolvetheproblemofhowepistemicallyunjustifiedearlyprocessinggivesrisetoperceptuallyjustifiedbeliefsbythetop-downinfluencesofconceptsonlateperception. Cognitioninvolvesmoredeliberatemodulationbytop-downprocesses,likeusingfocusedattentiontosearchforaspecificobject,andincludesaction-planning,self-reflection,andabilitiesrelatedtolanguage.Alloftheseprocessesarecloselylinkedtoconsciousnessandpropositionalcontent(specificallytheso-called‘accessconsciousness’—Block,1995).Theseprocessesaregenerallyepistemic,buttheycanalsoincludemorecomplexformsofcognitionandconsciousexperience,likeaestheticandmoraljudgments.Theimplicationofradicalcognitivepenetrationisthatsuchgoal-orientedhigher-levelprocessescandirectlyaffectthewayinwhichinformationisinitiallyprocessedbysensorysystemssuchthatitaffectsfeaturedetection(e.g.,thecoloroftheobjecttobefound).Weshallarguethattheycanonlyalterthemindirectly,bythemodulationoflateperception. Thequestionatissueisjusthowmuchcancognitionaffectlow-levelperceptualprocesses?Willthisbeaformofpre-cueingthatsimplydirectsimpenetrablemodulesandroutines,ordoesitactuallyaffecttheprocessingofperceptualinformationwithinthemodule(beyondattentionaleffects)?Isanyaspectoflow-levelperceptiontrulycognitivelypenetrable?Giventheconstraintsmentionedabove,aswellastheargumentfromevolution,theansweristhatcognitivepenetrationcannotbepervasive,andifithappens,ithastohappenattherightlevel(e.g.,lateperception,aftertheinterventionofatleastworkingmemory)sothatperceptualinvariancesarenotaffectedandbasicabilitiesnecessaryforsurvival,suchasnavigation,arepossible.Toreiterate,earlyperceptionisnotlikelytobesusceptibleofanykindofcognitivepenetration.Onepossibility,compatiblewithCAD,isthataccessconsciouspenetrationofperceptionmayoccurwithoutphenomenallyconsciouspenetrationonearlyperceptualexperiencesandviceversa.Withrespecttophenomenalconsciousness,asimilardistinctionisunproblematic:earlyphenomenalconsciousvisionmaybenon-conceptualandthenphenomenalconceptsaredeployedtocategorizeexperiences(seeLoar,1997). Asmentioned,someauthorsarguethatcognitivepenetrationnevergenuinelyoccurs.Instead,whatfallsunderthecategoryof“penetration”isjudgmentorcognition,anditneveraffectsperceptionassuch(seeFirestoneandScholl,2016).Otherauthorsdefendtheviewthatcognitivepenetrationaffectsperceptioninallsortsofways,suchthatbeliefsystematicallyaltersperception(e.g.,Siegel,2006,2010;Stokes,2012).Thisisarguedtooccurevenattheearlieststagesofprocessing.Giventheamountoftop-downinfluenceonperceptualprocessingonaneurallevel(e.g.,seeGilbertandLi,2013;VetterandNewen,2014),thisviewisnotimplausible.Accordingtothispervasive-penetrabilityview,ourbeliefs,desires,andgoalsaffectperceptioninmultipleways.Whatweperceive,therefore,issusceptibletoavastarrayofcognitiveinfluences. Thepervasive-penetrabilityviewpresentsadifficultchallenge.Ifcognitivepenetrationalwaysandsystematicallyoccurs,perceptionwouldinformusalmostalwaysaboutwhatwealreadybelieveorfeel,insteadofinformingusaboutfeaturesoftheworld(particularlywhenweencounternovelobjectsorevents).Thisisaproblemthatisespeciallyworrisomeforepistemology(Stokes,2012).Clearly,therewouldneedtobevarietiesofperceptualpenetrationwithvaryingdegreesofpenetrability.Ifexperiencesareanalogoustobeliefsinthesensethattheyrequirecriticaljudgmentandjustification,thenonemustreflecton,aswellassystematicallyanalyze,whatoneperceives.Thisreflectiveanalysiswouldconstituteaneffortfulandhighlytop-downformofattention(perhapseveneffortfulvoluntaryattentiontoexplicitlyjudgedperceptualcontents).Problematically,suchabelief-basedattentiveprocesswouldneedtodominateallotherformsofperceptualattentionforpervasivepenetrationtooccur. Theconsciousnessandattentiondissociationandtheargumentfromevolutionofferawayoutofthischallenge.Itcouldbethatcognitivepenetrationonlyaffectsaccessconsciousness(i.e.,accesstoinformationavailableforthought,memory,andaction,butwithoutsubjectiveexperience)athigherlevelsofcognitiveintegration.Alleffectsofcognitiononperceptualexperiencescanbeexplainedbyappealtoconcepts,beliefs,orinferences,andperceptualcontentsremainimpenetrableattheearlystages.Itcouldbe,therefore,thattop-downattentionroutinesoperateindependentlyfromphenomenallyconsciousperception.Motivationaleffectsmaybeexplainedathigherlevelsofintegration,whichneednotmodifythecontentsofearlyphenomenallyconsciousperception.Theformsofperceptualexperiencethatevolvedearly,suchasexperiencesofcolor,wouldbeimpenetrable.Thistheoreticalpossibilitywouldsolvetheepistemicproblempresentedabove.CADcouldalsoexplainwhythepervasivepenetrationofconceptualizedcontentsinaccessconsciousnessneednotentailthepervasivepenetrationofphenomenallyconsciousperception(subjectivelyexperiencedperception). Thereare,however,goodreasonstobelievethattheviewattheotherextremethatrejectsanyformofpenetrabilityisalsotooradical.Forexample,socialinteractionsrequireperceptualprocessingandanunderstandingofthesituationalcontext(includingotheragents)inordertosucceed.Categorizingnewobjects,events,orsituationsalsorequiresalevelofcognitiveinfluencethatmaydependonpreviousexperienceorknowledge.TheviewofperceptionasBayesianinference,forexample,presentsmodelsofhowperceptioncanbeconstrainedbypriorexperience,biasingdetectionofmorelikelyfeaturesandlimitingthepossibleinterpretationsofthisinformation(e.g.,seeKerstenetal.,2004;YuilleandKersten,2006).Althoughwewouldarguethatthissortofbiasingisnotaformofcognitivepenetrationofearlyperceptualprocessing,itcaninfluencehowthisprocessingoccursandparticularlyinfluencehowthecontentsofperceptionareinterpreted.Suchreasonsexemplifywhytheremustbeaninterfaceforcognitivepenetration.Thesewouldbeepistemicallyfundamentalcasesofcognitivepenetrationatlaterstagesofperception,wherethecognitiveintegrationofemotion,cognition,andperceptionisatwork.HerewetrytostrikeabalancebetweentheseoppositeviewsbyappealingtotheCADframeworkandtheargumentfromevolution(seeHaladjianandMontemayor,2015).Amorenuancedviewisrequirednotonlytosolvetheepistemicproblemmentionedabove,butalsotoachieveacomprehensivetheoryofperceptionthataccountsfortheepistemicandmotivationalsignificanceofperception,andtheBayesianapproachisparticularlyhelpfulhere. Howexactlyshouldtheevolutionofintentionalitybeunderstood,particularlywithrespecttoCADandcognitiveimpenetrability?Onepossibilityisthathumansandotherspeciessharemanyformsofearlyperception,withnon-conceptualintentionalcontent,whichcouldbeunderstoodintermsofPeacocke’s(1992)accountof“scenariocontent.”AsCrane(2009)clarifies,suchscenariocontentmustbeinterpretedintermsofbeinginastatewithnon-conceptualcontent—arepresentationalstatesuchthatbeinginitdoesnotrequirethepossessionofconcepts—eventhoughsuchcontentscouldbeproperlycharacterizedbyconceptsbyacreaturewithconceptualcapacities,suchashumans.Wecannotbecertainabouthowanimalsexperiencesuchcontents,butitishighlylikelythattheymusthavesimilarexperiences.Animalsnavigate,identifyobjects,reacttocolor,andhavesimilarsensorialsystems.Atsomepointinourevolution,ourbrainscreatedroutinestocognitivelyguideattention,buttheseroutinescannotdirectlychangeearlyperceptionduetotherequirementoffeatureconstancyforsurvival,whichincludesfeaturessuchascolorandtime(LisiandGorea,2016).Then,evenlaterinourevolution,welearnedtoexplicitlyinterpretourperceptualexperiencesandtolinguisticallyarticulatesuchinterpretationsintermsofdiscursiveinference(acapacitythatseemstobeexclusivelyhuman).Thus,genuinecasesofcognitivepenetrationshouldnotappealtoexplicitinference,aswhenone“sees”thatsomeoneislate.Butperceptionathigherlevelsofcognitiveintegration(e.g.,aboveearlyvision)maypresentinterestingcasesofcognitivepenetrationbyconceptualization.Thiswouldleaveearlyprocessingencapsulatedandimpenetrable,anditwouldalsoopenthedoortointerfacesbetweenpreconceptualperceptionandcognitivelyguided,conceptualperception. Formsofcognitiveintegrationalsoevolved,andtheymatterforthewayinwhichperceptualcontentsareprocessed.Forexample,thecross-modalintegrationofinformation(e.g.,auditoryandvisual)canindicateinfluencesfromonemodalityonanotherwhenattentionisdirectedinacertainway(e.g.,PalmerandRamsey,2012).Suchcross-modalintegrationisoften,thoughnotalways,relatedtoconsciousexperience,withsometheoriesofconsciousnessrelyingontheintegrationofinformationfrommultiplesourcestoproducetheunifiedexperienceofconsciousness(e.g.,Tononi,2012).Thismulti-levelapproachcouldhelpmodelpossibleformsofcognitiveintegrationintermsofdifferentinterfacesthatevolvedatdifferenttimes.Earlyperceptionremainsimpenetrabletoguaranteestability,butinthecourseofevolution,contentsareaccessedandintegrated,withoutaffectingearlyperception.Thenmemoryandmotivationalsystemsarealsointegratedintomorecomplexcognitivestates,guidedbycognitivelydrivenattention. Earlyperceptualprocessesmust,aboveall,providereliableinformationabouttheenvironmentindependentlyofmotivationorcognitivemodulation.Theyincludefeature-basedandobject-basedattention(Treisman,1988),andmotiontrackingmechanisms(Pylyshyn,1989;Cavanaghetal.,2001).Top-downpre-cuingandcognitiveguidanceoperateathigherlevels,afterearlyselectionmechanismsofattentionhaveoccurred(YehandChen,1999;Theeuwes,2010).Thus,earlyvisionprovidesabasicrealmofperceptualexperiencesthatinformnavigation,immediateengagementwiththeenvironment,andevenformsofplanningthatcanbefoundinotherspecies,suchasbirds(ClaytonandDickinson,1998).Asmentioned,thisformofintentionalitymaybeunderstoodintermsofthenotionof‘scenariocontent’—anintentionalstatethatneednotbeconstitutedeitherbyconceptsorpropositionalcontentsforittoberepresentational.Navigationinmanyspeciesseemstodemandthiskindofintentionalityanditmusthaveevolvedearly(fordiscussiononhowthistopicrelatestothedistinctionbetweenanaloganddigitalformatsofmentalrepresentation,seeMontemayor,2013,chapter3).Itisverylikelythatincreatureswithphenomenalconsciousness,scenariocontentisdeeplylinkedtobasicexperiencesthatinformthemabouttheenvironmentmuchinthesamewayastheyinformus.Althoughmanyskillfulreactionstotheenvironmentoccuroutsidephenomenalconsciousness,consciousexperienceisourmostimmediateguideforaction.Accesstocontent,ontheotherhand,requireshigherlevelsofintegrationandtheinterventionofpropositionalattitudes,suchasbeliefs. Thereisarelatedissueconcerninghowattentionworksoutsideconsciousawarenessinspeciesthatmaynothavephenomenalconsciousness.Non-humanspecieswithcomplexattentivesystems,suchasdragonflies(WiedermanandO’Carroll,2013),arealsonotlikelytoaccessnavigationalinformationpropositionally(intermsofaccessconsciousnessandconceptualjudgment).HereCADpresentsaninterestingpossibility.Perhapsthoseattentioncapacitiesfornavigationandobjecttrackinginspecieslikeinsectsareextensionallyequivalenttothoseoforganismsthatrelyonphenomenalconsciousness(theyoverlapintermsoftheirreferenceandhowtheorganismreactstostimuli).Butforsuchextensionaloverlaptobepossible,theseearlyperceptualprocessesmustbeimpenetrable,orattheveryleast,theimpenetrabilityofsuchperceptualprocessesisthebestexplanationwehavefortheiroverlapacrossspecies.Obviously,understandingexactlyhowmuchperceptualguidancehappensoutsideconsciousawarenessisanempiricalissue.Theclaimwedefendhereisthatthedistinctionbetweencognitivelyimpenetrableperceptionandcognitivepenetrationisfundamentaltoaccountforthecomplexityofperceptionanditsevolution.Thechallengeistounderstandtherelationbetweencognitivelyimpenetrableperceptionandcognitivelypenetrableperception.Tothisend,wenowproceedtodiscussconceptacquisition—oneoftheclearestinstancesinwhichaninterfacebetweencognitionandperceptionmustoccur. ConceptAcquisition Thesharpdistinctionbetweencognitionandperception,whichsomecriticsofcognitivepenetrationtheorizeasacentralfeatureofthemind(seeFirestoneandScholl,2016),confrontsaparticularlypressingproblemattheheartofthecognitivesciences:conceptacquisition.Infact,theclaimthatsuchastrictdemarcationisanessentialaspectofthenatureofthemindmayevenbeempiricallyfalse(Kosslyn,1980,1994).Forourpurposes,wewillfocusonlyonhowthesharpdemarcationbetweencognitionandperceptiongeneratesproblemsfortheissueofconceptacquisition.Weaimtoshowthatalthoughthepervasivecognitivepenetrationviewcannotbetrue,asarguedabove,theoppositeviewthatclaimsthatnocognitivepenetrationeveroccursisalsowrong.Animportantclarificationisthatcognitivepenetrationcanoccurinlateperception(afterearlyperceptualprocessing),andthatpreconceptualprocessesplayamajorroleinprovidinganinterfacebetweencognitionandperceptionatthatlevel.Thuswedefendtheviewthatearlyperceptioncannotbedirectlyaffectedbycognition,butthatthereisaninterfacethatmakeslate(penetrable)perceptionpossibleand,infact,systematic.Themaindifficultyistoexplaintheacquisitionofperceptuallybasedconceptsthatarecriticalforbasicrecognitiontasks. Justasweneedtobeclearaboutthesenseinwhichcognitiondeterminesperception,wealsoneedtobeclearaboutwhatismeantby‘conceptualcognition.’First,considerthedistinctionsbetweenmemory,recognition,andseeing.Rememberingisclearlydifferentfromseeingandmemory-basedattentionaleffects.Althoughmemorymaybecrucialtoguideperceptionandcategorizenovelobjects(e.g.,Vlach,2016),itdoesnotdeterminewhatwesee.Butwhyshouldconceptuallybasedrecognitionbeonparwithmemoryasanon-perceptualprocess?Takeforinstancetheevolutionarilycrucialskillofrecognizingkinandenemies.Thisfundamentalcapacityseemstobepartoftheperceptualsystem,anditseemstobetheresultofitsevolution(Millikan,2005).Additionally,recognizingsomethingdoesnotalwaysrequireafullperceptionofit,sinceinferentialprocessingcanusekeyfeaturestoinformtherepresentationbasedonmemory,whichindicatesthatrecognitionalabilitiesinanimalsmustbeacombinationofperceptualandpreconceptualcapacities.Becauseofhowbasictheseskillsareforsurvival,twoformsofrecognitioncouldbepostulated:onedependentonmemoryandtheotherfundamentallyperceptual(e.g.,theautomaticreactiontosensoryinputs).Thispossibilitywouldnotbecompatiblewiththesharpdemarcationmodel(e.g.,favoredbyFirestoneandScholl,2016),sincerecognitionalcapacitiesseemtodetermineperceptualprocessinginsuchcases. Inwhatsensecanpreconceptualstatesthatarenotcognitivelypenetrableleadtoattentionmodulationthatiscognitivelydriven?Asmentioned,onepossibilityisthatconsciousandunconsciousnon-conceptualstatesoverlapsystematicallywithcontentsthatcanbedescribedcategoricallybyanorganismwithconceptualcapacities.Giventheaccuracyandreliabilityofthemechanismsthatproducesuchpreconceptualstates,onecouldthinkofthesestatesasarepresentationalframeworkthatstructuresaninterfaceformoreabstractrepresentations.Languageseemstobepresentonlyinhumans,andasamatterofmethodology,itisbestnottoattributeconceptualcapacitiestootherspecies(Bermúdez,2003,callsthisaminimalistapproachtonon-linguisticthought).Takingaminimalistapproachisfundamentaltoexplainmanynavigationalcapacitiesthatarebestunderstoodeitherasmeasurement-basedrepresentationsorscenariocontents.Itwouldbeinappropriatetocharacterizetheserepresentationsintermsoflanguage,concepts,orlinguistic-propositionalattitudepsychology.Actually,someauthorsthinkthateveninthecaseofpropositionalattitudeattributiontherearereasonstobeskepticalaboutadoptingalinguistic-propositionalmodelinsteadofamoreminimalistone(Matthews,2007). Theproposalmentionedpreviously,thataccessconsciousnessmayberesponsibleforcognitivepenetrationwithoutcausallyanddirectlychangingthecontentsofearlyperception(includingphenomenallyconsciousperception),cannowbespelledoutinmoredetail.Earlyprocessingiscognitivelyimpenetrable,intentional,andrepresentational,anditcaneitherbephenomenallyconscious(producingexperiencesofasensorialkind)oroccurunconsciously—inaccordancewithCAD.Theseearlyperceptualstateshaveacontentthatcanbecharacterizedasnon-conceptualornon-propositional(fordiscussionofhowtocharacterizetherepresentationalnatureofthesestatesseeMontemayor,2013).Thenworkingmemoryand,inthecaseofhumansatleast,conceptualrepresentations,caninfluence,guide,andindirectlydeterminethecontentsofperceptionatlaterstages.Workingmemoryprocessescanalsohelpmaintainrepresentationsoftask-relevantfeaturesbyactivatingearlyfeatureselectionregionsofthevisualcortex(Serencesetal.,2009),whichsuggestsatop-downinfluenceonearlyvisionactivations.Infact,variousstudiestestingthememoryforsensorysignalssuggestthatthecircuitryunderlyingtheworkingmemoryinvolvedinthesetasksincludescorticalareasthatdotheprocessingofthesesignals(forareview,seePasternakandGreenlee,2005).Nevertheless,suchmodulationsofearlyvisionareconsistentwiththeCADapproach.AlsoconsistentwithCADistheindirectguidanceoflateperception,whichmaydependnotonlyonaccess-consciousstateswithpropositionalcontent,butalsoonothermotivationalandphenomenologicallypowerfulstates,suchasemotions.Thisisallconsistentwithearlyperceptionbeingcognitivelyimpenetrable.Buttheinterfacebetweenearlyandlateperceptionshowsthattheinteractionbetweenperceptionandcognitionisvitalforconceptacquisition.Thisclarificationisimportant,becauseonewayofinterpretingFirestoneandScholl’s(2016)proposalisthatsuchaninterfaceisneverpossibleandthatthereisnokindofcognitivepenetration,evenatlaterperceptualstages. Asacognitivephenomenon,conceptacquisitionseemstocriticallydependonperceptualprocessesonsomelevel.Fodor(1983,1998),whoisaprominentproponentofthemodularandencapsulatedarchitectureviewthatisputativelyincompatiblewithpenetrability,explainsconceptacquisitionasfollows:“WehavethekindsofmindsthatoftenacquiretheconceptXfromexperienceswhoseintentionalobjectsarepropertiesbelongingtotheX-stereotype”(Fodor,1998,pp.137–138;hisemphasis).Thesepropertiesarenotbasedonstoredmemories,otherwisehowcouldoneevenacquireaconcept?WhatFodorcallsa‘stereotype’isnotajudgment,butastatisticalnotionthatcapturesperceptualregularities(Fodor,1998,p.138).Fodorinsiststhatperceptualexperiencesarenecessaryforconceptacquisition.Ifonlyjudgmentswerenecessaryforthis,howcouldoneacquireaperceptualconceptinthefirstplace?Soconceptualrecognitionseemstobeanessentiallyperceptualprocess.Evenifoneholdsthatconceptsareinnate,perceptualprocessesarestillnecessarytoacquiresuchconcepts(obviously,forthosewhodenyinnatism,perceptualprocessessufficetoexplainconceptacquisition).Conceptacquisitionisneitherexplicitjudgmentnormerelyunconsciousinference,andfavoringamodularandencapsulatedarchitecture(e.g.,Pylyshyn,1999,2003)canstillbecompatiblewithhavingaconceptualinterfacebetweencognitionandperception. Below,wedrawadistinctionbetweenlinguisticlabelsandconceptualcategories,whichfurtherclarifiestheprocessesunderlyingconceptacquisition.First,wewanttoexpandonhowthedistinctionbetweenearlyandlateperceptionrelatestotraditionalissuesinepistemology.Whenyouseearedcup,seeingitasacupthathasthepropertyofbeingredobviouslymeansthatyoupossesstheconcepts‘red’and‘cup.’Butyourperceptualsystemcanbeinaphenomenalstatewiththeredcupaspartofitscontent,independentlyoftheseconcepts(asitoccurswithinfants,andpresumablyinotherspecies).Inotherwords,yourperceptualsystemcanhaveavisualexperienceoftheredcupwithoutseeingitasanobjectthatfallsunderthecategory‘redcup.’Forthisreason,itseemsthattheoriesincognitivesciencemustallowforthedistinctionbetweennon-epistemicandepistemicseeing(e.g.,seeingabundleoffeaturesversusseeingsomethingasaninstanceofaconceptualcategory). Casesofexpertisegenerateaninterfacenotonlywithconcepts,butalsowithlargerrepertoiresofjudgmentsandbeliefs.Lookingoutyourwindow,youseeabirdlandonanearbytreelimbandyounoticeitsgrayandblackcolors.Yourexpertfriend,anornithologist,seesnotonlythebirdanditscolors,butalsoseesitasahoodedcrow.Thiscontrastcanbeinterpretedinseveralways:youseeanobjectanditscolors,andafterattendingtoitcarefullyyouseethatitisacrow;oryouseeabirdandwhileyouseeitasacrow,yourexpertfriendseesitasahoodedcrow.Inthelattercaseitseemsclearthatyouandyourexpertfriendseethesamebird(butseeSiegel,2010,fortheclaimthatthesemightbedifferentperceptualexperienceswithdifferentcontents).Intheformercaseyouseethebirdandapplytheconcept‘crow.’Otherspeciesmayseethebirdandbeinaperceptualstatethatdisposestheanimaltobehaveasifitwerereferringtocrowsinparticular,butwithoutneedingtobeinaconceptualorpropositionalstate.Youandyourfriend,however,areaccessinginformationdifferentlyeventhoughthecontentofyourearlyperceptualexperiencesverylikelyoverlap.Thisiswhyaccessconsciousnessisassociatedwithmorecomplexformsofcognitiveintegrationthatoccuratlaterstagesofperceptualprocessing.Youpossestheconceptcrowandbird,butonlyyourfriendcandrawtheinferencethatthisisaspecifickindofcrow. Expertise(and/orpriorexperience)canchangehowweseesomethingconceptually,attheaccessconsciousnesslevel,butnotperceptually,attheearlyphenomenallyconsciouslevel.Itcanaffectperceptionatlaterperceptualstages,aswhenperceptualcontentsareintegratedwithmotivationalstates.Beinganexpertmighthelpyounoticethenuanceddetailsofabirdthatenableyoutoidentifyitasacertainspecies,comparedtothenaïveobserverthatjustseesitassomekindofbird(i.e.,attentiontothedetailmightdiffer,thoughthesameperceptualcontentsareavailabletobothobservers).Expertisecouldprovideaformofpre-cueingeffect.Forexample,bytuningthenervoussystemtointegratedcontents,musiciansareabletorespondtomultisensorystimulimoreefficiently(LandryandChampoux,2017).Theseeffectsmodulateorguideattention,ratherthandeterminewhatoneperceivesbyaffectinghowinformationisprocessed.Evenincasesofsensoryphenomena,suchasadaptationornegativeafter-images,changesinperceptionareduetotheunusualandconsistentactivationofvisualneurons(e.g.,byforcingaconstantfixation,astimulusintheperipherycandisappearduetoneuralfatigue),andwouldnotbeconsideredcasesofcognitivepenetration.Infact,thesechangesinadaptationoccurbecausegazeisdirectedinsuchawayastoinducethesephenomena,whichareexamplesofhowthemodulesofperceptioncanbedirectedinwaystoexploittheirinherentcharacteristics,andnotanexampleofcognitiondirectlychangingtheprocessingwithinthemodulesofearlyperception(seeCliffordetal.,2007).Theseadaptationeffectsoccuratseverallevelsofperceptionthatincludelateones,asinthecaseoffaceperception(WebsterandMacLeod,2011).Itisthemodulationbasedonconceptsandpropositionalcontentthatisdistinctlycharacteristicofaccessconsciousness,whichaccordingtoCAD,neednotcharacterizephenomenalconsciousness,includingsubjectivelyexperiencedadaptationeffects,therebyallowingforthecognitiveimpenetrabilityofearlyperceptualstates. Conceptacquisitionbeginswithperceptualprocessesthatprovidecontentsthatneednotbeconceptualizedtobeinformative.Thenlaterperceptualstagesinterfacewithconceptualinformationandthenstorecategoricalinformationintomemory.Suchinterfacesarecriticalforconceptualcognition.Ofcourse,onecancombineexistingconceptstoformnewonesindependentofdirectperception(e.g.,a“Pegacorn”canbeeasilyimaginedifoneisfamiliarwithPegasusandunicorns).Conceptacquisitionisaproductofvariousprocesses,someofwhicharepurelyperceptualandotherspurelycognitive,andmanythatareacombinationofthetwo.Partlybecauseofthis,webelievethatneithertheabsoluteimpenetrabilitynorthepervasivepenetrabilityviewsareentirelycorrect. TheCADframeworkalsoallowsforgradeddistinctionsthatexplainwhy,ontheonehand,earlyperceptualprocessesaresostableregardlessofbackgroundbeliefsandemotions,andontheotherhand,whyhighlyintegratedinformationissusceptibletodistortionsbasedonbeliefsandemotionsatlaterperceptualstages.Thisisaconsequenceoftheargumentfromevolution.Sinceintentionalityevolved,aninterfacebetweencognitivelypenetrableperceptionandcognitivelyimpenetrableperceptionmusthaveevolved.Theperceptionofmagnitudesoffersaparticularlyinterestingcase.Perceptuallyrepresentedmagnitudesformotorcontrolandnavigation(e.g.,duration,distance,orrate)differfromconsciousattentiontothedurationofsensationsandemotions,includingexperiencedeffort.Theformerareveryreliableacrossspecieswhilethelatteraresusceptibletowell-confirmeddistortioneffects(Kahneman,2000).Partlybecauseofthedifferenceinintegrationandsusceptibilitytodistortioneffects,therearetwomodelsinassessmentsofexperiencebasedontheirdurationorintensity:thememory-basedandmoment-experiencebasedmodels(Kahneman,2000,p.692).Thiscontrastbetweentheearlyperceptionofmagnitudesandmorerecentinterfacesbetweenperceptualmagnitudesandconceptualizedexperienceshasclearimplicationsforagencyandplanning,anditsuggeststhatdifferentspeciesmustrepresentthemselvesintimedifferently(Montemayor,2010). Perhapsamongtheevolutionarilyoldestformsofearlyperceptionistheperceptionofmagnitudesfornavigation.Perceptualcapacitiesfornavigationareamongthemostreliableskillsthathavebeenverifiedacrossspecies,includinginsects(Gallistel,1990).Theseperceptualcapacitiesrelyonrepresentationsthatarenon-conceptual,andcanbeexplainedintermsofscenariocontent(seealsoMontemayor,2013,fordiscussionofwhythesearerepresentational).Conceptualizedemotions(andtheirdurationandintensity),however,aremuchmoredifficulttoverifyinotherspeciesandcannotbeassumedtobepresentinmanyofthem(e.g.,ininsectsthatcanreliablynavigateandattendtomagnitudes).Presumably,specieswiththeoryofmindcapacitieshaveamorecomplexinterfaceforperception,emotion,andcognition,asthedistinctionbetweenempathicandnociceptivepainshows.Thepossibilityforcognitivepenetrationcorrelateswithevolutionaryhistory,astheargumentfromevolutionentails,andalsowiththecognitiveintegrationrequiredforaccessingpropositionalcontents.Thecontrastbetweentheperceiveddurationofemotionsandthemorebasicperceptionofmagnitudes(e.g.,time,distance,andrate)caneasilybeaccommodatedbytheCADframework:thereisaninterfacefortheintegrationbetweenemotionsandjudgmentsconcerningintensityandvalueatmuchlaterstagesofperception,butearlyperceptualprocessingofmagnitudeperceptioniscognitivelyimpenetrable.Thisguaranteesreliability,asmentionedbefore.Inhumans,thereisalsoaconceptualinterfacefortheintegrationofperceptualmagnitudesandnon-perceptualconcepts,suchasmathematicalconceptsconcerningspace,time,andrate.Thisinterfaceisassociatedwithaccessconsciousnesswhiletheinterfacewithemotionsisacombinationoflateconceptualperceptionandthephenomenologyofemotions.Whilesomestudiesshowthatmagnitudejudgmentscanbecalibratedsystematically(e.g.,IzardandDehaene,2008),thesewouldbecasesofmodulatingtheinterpretationoftheoutputratherthanacognitivepenetrationofthemagnitudeestimationmechanismitself. AninterestingconsequenceoftheargumentfromevolutioninthecontextofCADisthatcompetingviewsaboutconceptsmaybecorrectlydescribingdifferentlevelsofperceptualprocessing.Conceptualstructureofthekindhumanshaveismoreabstractthananysetoffeaturesorsimpleperceptualattentionroutines—ithasalogicalstructurethatallowsfornegation,validinference,andcompositionality.Suchconceptscannotbereducedtothesumsoftheexpectedprobabilitiesoffeaturesgivenaperceptualscene,buttheearliest,cognitivelyimpenetrablestagesmaybereducibletosuchfeatureorprototype-basedanalysis.Thisleadstotwofurtherimplicationsoftheargumentfromevolutionconcerningconceptsinparticular.First,thehigherthedegreeofcognitiveintegrationandpenetration,themorelogicalstructureisneededforcognitiveinfluence.Second,thehigherthedegreeofinferentialintegration,themoreabstractandamodaltheconceptsare.Thishigher-levelofcognitiveintegrationistheonetypicallyassociatedwithexplicitjudgment(i.e.,explicitjudgmenthaslogicalstructure).Thisopensthepossibilityfordifferenttypesoffeatured-basedprototypesoperatingatearlystages,andmorecharacteristicallyabstractconceptualrepresentationsplayingdifferentrolesatdifferentinterfaces,allowingfordifferentformsofintegrationandde-modularizationatlaterperceptualstages.Theseinterfaceswouldbeconsistentwithempiricalfindings,suchasthecross-speciesfindingsontheperceptionofmagnitudesandthefindingsonthedistortionofdurationjudgmentsregardingemotionsinhumans.Finally,onefindsasimilardistinctionbetweenprototype-basedcategorizationandmoreabstractconceptsinhumandevelopment(e.g.,Keil,1989).Developmentalstudiesindicatethatinfantscanobtainperceptualconceptsbeforecomplexformsofabstractconcepts(Spelke,1988;SpelkeandKinzler,2007;Carey,2009).Itiswiththismoreadvancedtypeofconceptualinterfacewherewecouldfindcognitivepenetrability,atlaterstagesofperceptualprocessingthatareintegratedwithcognitivelydrivenattentionmodulation.Theseinterfacesare,intheveryleast,evidencefortheinterrelationbetweenperceptionandcognitionatlaterstages.Thus,postulatingdifferenttypesofinterfaces,basedontheCADframeworkandtheargumentsfromevolution,mayhelpexplaincasesofcognitivepenetrationatlaterstageswhilepreservingthecognitiveimpenetrabilityofearlyperception,strikingabalancebetweentheprevailingoppositeviews. CadasaFrameworkofDistinctionsforEmotion,Perception,andJudgment Emotionscomplicatethepictureconsiderably.Theyareanimportantaspectofsocialcognitionandinteractions,particularlyintermsofdevelopingempathyandhelpingtounderstandothers.Forsuchreasons,emotionalprocessingmustbeanintegralcomponentofhumanperceptionandcognition.Newen(2016),forinstance,arguesthatemotionscanbeperceivedsimilartothewayperceptualfeaturesareperceived.Studiessuggestthatemotionscanberecognizedinthesamewayaspatternrecognitioninothersensemodalities,drivenbyevolutionarynecessityandrequiringaninteractionofbottom-upandtop-downprocesses(seeNewen,2016).Similarly,sociallyrelevantinformationseemstobeprocessedautomatically,thuscallingintoquestionwhetherperceptionshouldincludeattentiontosocialcues(Neufeldetal.,2016).Ifitistruethatemotionsandsociallyrelevantinformationareprocessedlikeperceptualfeatures,thisviewwouldstronglyfavoraveryrobustkindofcognitivepenetrationbecausewenotonlyseethebasicperceptualconstanciesthatgroundobject-andfeature-basedattention,butalsoemotionalandsociallyrelevantcontent.Inotherwords,ifthisviewiscorrect,thenemotionalandsocialbeliefswoulddetermineasubstantialportionofperception.Itisimportanttonoticethatevenifthiswerethecase,itwouldstillbecompatiblewithearlyperceptionbeingcognitivelyimpenetrable. Themainproblem,however,isthatthisexampleofpenetrabilitycouldsimplymeanguidance.Thereisgoodreasontobelievethattheneuralsystemsthatsupportemotionoverlapwithcognition(Pessoa,2008),andemotionalstatesmaybeconsideredaformofpre-cueing.Forexample,anemotionalstate,likefear,canbiashowonedirectsattention(e.g.,tomorethreateningaspectsofenvironment)andthusimproveinteractingwiththeenvironment(LeDoux,2012).Thisabilityalsoincludesnon-consciousperceptionofemotionalstimuli(seeTamiettoanddeGelder,2010).Ifthesepre-cuingeffectsareveryrobustandsystematic,thereisaveryclearsenseinwhichtheydeterminewhatoneperceives,thusfavoringsomelevelofpenetrabilityatlaterstagesofprocessing. Justhowpowerful,exactly,cancognitivepenetrationbeinthecaseofemotionswithoutbeingcognitivelypernicious(e.g.,byalteringtoomuchthecontentsofperceptionandrenderingcrucialperceptualinvariancesunstableandunreliable)?CADalsohelpselucidatethisissue.Emotionshaveanenormousimpactonconsciousawareness,butthisimpactneednotbeeitherfullyperceptualorinferential.Webelievethisisasignificantsourceofconfusion.Emotionshaveasignificantimpactonanindividual’soverallphenomenology,buthavingtoomuchimpactonawarenesscandistractfromormayevensuppresswhatoneperceives.Insuchcases,thephenomenonisoneofinterferenceorhindranceofperceptionratherthanadeterminationofperception(e.g.,aswithpost-traumaticstressdisorder).Inothercasesitmayenrichperception—notbydeterminingit,butbyaddingvivacitytotheoverallphenomenologicalexperience.Aestheticexperiencesandthevivacityofcertainautobiographicalmemoriesaregoodexamplesofthisphenomenon(MontemayorandHaladjian,2015,pp.150–165).Allthesecasesarebestunderstoodaslateperceptualcognitivepenetration(perhapsmotivationalpenetrationisabetterterm),ratherthancognitivepenetrationofearlyperception(forinstance,earlyvision). ColorperceptionfurtherelucidatestheimportanceofCADtorigorouslydefinecasesofcognitivepenetrationatlaterstagesofperceptualprocessingfromcognitivelyimpenetrableearlyvision.Colorperceptioninvolvestwodistinctneuralcircuits,oneforcolordetectionandanotheronerelatedtocircadianregulationandemotion(Pauersetal.,2012).Doweperceiveemotionswhenweperceivecolor?Thisdoesnotseemplausible.Rather,wedetectandexperiencecolorinearlyvision,andwealsoexperienceacomplexstateofperceptualandemotionalcontentsatlaterstagesofprocessing.Eveninthecaseofanindividual’smemoryofanobject’sexpectedcolor,whichcaninfluencetheperceivedcolorappearanceofanobject(seeHansenetal.,2006),suchfindingsdonotconclusivelyindicatecognitivepenetrationofearlyvisualperception,butratherthestagethatincludestheinterpretationofthesignalsfromearlyvision.Tocomplicatethingsfurther,someaspectsoffeature-detectionmayoccuroutsideconsciousness—theyaremostlyindependentintheirneuralcorrelates(KochandTsuchiya,2012).Primingofcolorcanoccurathigherlevelsofprocessingevenwithoutconsciousperceptionofthecolor,asinstudiesthatusebackwardmaskingtotestprimingofresponsestocolorsthatarenotconsciouslyseen(Normanetal.,2014).ToaccommodatethisfactweneedagradedframeworklikeCADratherthanasharpdistinctionbetweencognitionandperceptionorapervasiveformofcognitivepenetration.Colordetectionandcolor-basedemotionsdointeractsystematicallyatthelaterstagesofperceptualprocessingthatarealsophenomenallyconscious,butthisdoesnotentailthatemotionpenetratescolordetectionorearlyvisualcolorexperiences. Regardingthecasesofautomaticsocialdetection(e.g.,Neufeldetal.,2016),thesecouldbesimilartodetectionpatternsassociatedwithsocialplanningroutinesthatoperateindependentlyofexperiencedemotionsandfeelings.Thus,basedonCAD,itisnotsoeasytosaythatemotionisdetectedaspartofperception,becausesuchroutinescouldbemodeledeitherasunconsciousprocessingorasspecificattentionroutinestriggeredbyspecificperceptualconceptualizedcontents,ratherthanbeingconstitutiveofearlyperception,sincethisdetectionisnotalteredbyoverallphenomenology(apointentirelyanalogoustothedistinctionbetweenmagnitudesandemotionintensitymentionedabove).Generally,itmaybethatsuchpatterndetectionofsocialcueswouldnotentailsystematicpenetrabilitybecausetheymayactuallyoccuratlatelevelsofcognitiveprocessingornotbeperceptualatall(e.g.,theycouldbeinferentialorstrictlymnemonic). TheCADframeworkcanexplainmanychangesinperceptionatlaterstageswhilejustifyingtheimpenetrabilityofearlyperception.MerelyappealingtophenomenologyandhowswitchingfromoneattentiontasktoanothervarywhatoneexperiencesdoesnotsufficetoprovepenetrabilitypreciselybecauseofthedistinctionsbasedonthelevelsofCAD.Moreover,eventhephenomenologyofperceptionfavorsstabilityandcontinuityinexperience,ratherthanvariabilitycausedbyconstantcognitivepenetration.Forexample,asonemovesaroundaroom,theexperiencedcolorandshapeconstanciesofthewallsandfurnitureremainthesamedespitethemanyinferentialtriggers,actualandpotential,thatonehasatanysinglemoment.Strikingly,thisalsoseemstobethecaseindreamswherethereisagenerallycoherentexperience,nomatterhowabsurditmaybe.Therefore,appealstoconsciousexperiencemaynotprovidedecisiveevidenceforcognitivepenetrationbecauseoverallphenomenologydependsoncognitiveintegrationatlatestagesofperceptualprocessinginawaythatiscompatiblewithearlyperceptualimpenetrability.Whatoneneedstoshowinordertoverifyradicalandpervasivecognitivepenetrationisthatcognitiondeterminesperceptionatanessentiallevel,attheearlieststages,causingchangesinperceptioninadirectway.CADshowsthattheevidencecanbeunderstoodinawaythatavoidsthisinterpretationbecauseCADdemonstratesthatcognitionandperceptioncanbeindependentandyetinteractinsystematicways.Inparticular,conceptacquisitionofbasicperceptualcategoriesisagoodplacetoidentifyclearcasesofcognitivepenetrationbeyondtheinitialstagesofearlyperception. CognitivelyDrivenAttention:Feature-Based,Syntactic,andSemantic Itisimportanttorestatewhyresistingpervasivecognitivepenetrationisnotonlyplausiblebecauseoftheargumentfromevolution,butalsoasageneraltheoreticalcommitment.Onereasonistheproblemoftheimpossibilityofcommongroundamongperceivers.Ifthereisnocommonground,howcanoneexplainreliablecoordinationamongmultiplesubjectsformotorcontrolandattentionaltasks(e.g.,thatareexecutedwhenplayingteamsports)?Onesolution,offeredbyCAD,isthatwhiletherearesignificantlevelsofcognitivepenetrationathighlyintegratedlevelsofcognitionandperception,thereisnocognitivepenetrationatearlyconsciousandunconsciousperception.Butitisalsoimportanttoexplainhowexactlytop-levelprocessesinfluenceperceptualexperience.ThisiswhatCADallowsfor:cognitiveimpenetrabilityofearlyprocessingwithrichinfluencefromcognitionathigherlevelsofcognitiveintegration(e.g.,attentiontotheintensityofemotions,ortheimportanceofanautobiographicalmemory),whichcorrespondtomoreevolutionaryrecenttypesofattention(foracriticismagainsttheviewthattop-downpathwaysentailcognitivepenetration,seeRaftopoulos,2001a,b). Thereareseveralpossibleareasofhigher-levelcognitionthatcouldbesusceptibletocognitivepenetration.AccordingtoCAD,phenomenalconsciousnessvariessystematicallywithemotionalandbackgroundknowledgecontents—itisempathicallystructured(MontemayorandHaladjian,2015).Howsusceptiblethemoresemanticaspectsofthemindaretoinferenceandemotionalinfluencemaydependontheconceptsaspecieshasandthedegreeofinformationintegration—hencetheimportanceofconceptacquisition.Torepeat,earlyperceptioniscognitivelyimpenetrable,whichallowsforreliableandpredictablemotorcontrolandcoordinationwithexternalobjects.Thesecontentsareprocessedindependentlyoftheempathicandintegrativeinfluencesofcognitionandemotion.Thisstructuralrequirementisrelatedtoadaptivenecessity,andlikelyappearedinotherspeciesthatareevolutionarilyclosetohumans(Zentall,2005).Furthermore,human-likeconsciousawarenessseemsdependentonaglobalfunctionalconnectivityamongbrainmodules(Wu,2014;Godwinetal.,2015),andthismayindicateaformofpenetrabilityatlater,moreintegratedstagesofperception. Attheearlystages,perceptualfeaturesareprocessedindependently,withminimaltop-downmodulation,inordertoreliablyandaccuratelystructuretheperceptualscenes(e.g.,auditoryorvisualscenes).Thisrepresentationalscaffoldingsupportslatercognitiveguidanceandcanbecharacterizedasscenariocontentorpreconceptualsensorialrepresentation.Then,atlaterstagesthatlikelydependontheinterventionofworkingmemory,featurebased-attentioncanbeguidedandorientedbycognitivelydrivenformsofattentionthathighlightsomeperceptualfeaturesandsuppressorinhibitothersbasedoncognitiveandmotivationalinformation.Someofthesecognitivelydriventypesofattentionlikelyevolvedatdifferenttimes.Someofthemmodulatedetection;othersexclusivelyconcernconceptualinformationandcanonlybefoundwithoutcontroversyinhumans.Therangeofinfluenceofcognitiononperceptionisquitevastanditincreaseswiththedegreeofcognitiveintegration,characteristicoflateperceptualprocessing.AccordingtoCAD,thereisatypeofattentionatlatestagesofintegrationthatisfullyindependentofspecificperceptualexperiencesandthatisexclusivelydriventowardaccesstopropositionalcontents.Wehavearguedthatthiskindofcognitivelydrivenattentionplaysanimportantroleinspecifyingthecontentsoflateperception,butthatitcannotdirectlychangeearlyperception,includingtheperceptualexperiencesassociatedwithearlystagesofperception. Theconsciousnessandattentiondissociationalsohelpsaddressthepreviouslymentioneddifficultythatperceptionmayoccuroutsideconsciousness,independentlyofwhetherornotthecontentsofperceptionaresusceptibletocognitivepenetration.ConsidertheresultbyVishtonetal.(2007)concerningtheEbbinghausillusion,inwhichtheinstructiontograspthestimulusreducestheillusion.Thiseffectofareductionintheillusionwasfoundinpreviousstudieswhereactingonastimulusproducinganillusionindicatedamoreaccurateinternalrepresentationthanwhatwasconsciouslyperceived.Thatis,whilethephenomenologyofperceptionistrickedbyanillusion(e.g.,theMüller-Lyerillusion),perceptionforactionisnot(StöttingerandPerner,2006).Also,thiskindofperformancecanbeaffectedbyemotionalstates(vanUlzenetal.,2008),whichindicatesthatemotioncaninfluenceconsciousexperience,butonlyatahigherlevelofintegrationandprocessing,astheargumentfromevolutionentails.Similarly,desiringsomethingmightaffecthowitisconsciouslyperceived;forexample,anappealinglocationmightseemcloserthananunappealingonethatisatthesamephysicaldistancefromtheobserver(AlterandBalcetis,2011).Suchstudiesareexamplesofhowconsciousperceptionsmaybeaffectedbycertainmentalstates,andthattherecanbeadissociationbetweentheinformationthatentersawarenessandtheunconsciousinformationusedforotherperceptualprocesses.Itisunlikely,however,thattheseeffectscouldinfluencetheperceptual-navigationalsystem(e.g.,thesystemweusetowalkacrossaroom),ortheexperiencesproducedbyearlyperception. Anotherexampleofhowfeaturedetectioninconsciousperceptioncandifferfromthatusedtoexecutemotoractionsisseeninanexperimentinvestigatingthedouble-driftillusion.Thisillusionoccurswhenanobjectmovesintheperipheryofthevisualfieldalongaspecifictrajectory,butbecausetheobjecthasatexturethatmovesorthogonaltothistrajectory,theoverallperceivedmovementoftheobjectdoesnotcorrespondtotheveridicalpath.Inotherwords,anillusorypathisperceivedbecauseofthecombinationofmotioninformationfromtheinternalmotionoftheobjectaswellasitsactualtrajectory.InarecentstudybyLisiandCavanagh(2015),participantswereaskedtomakeaneyemovementtooneofthesemovingobjects(thatdisappearedassoonastheeyemovementbegan),andtheyfoundthattheeyeslandedclosertotheveridicalpathasopposedtotheperceivedillusorypath.Thissuggeststhattheinformationsenttothemotorsystemisnotsusceptibletotheillusion,sincethemotorsystemcanexecutecorrecteyemovements,eventhoughtheillusionisconsciouslyperceived.Animplicationoftheseresultsisthatunconsciousperceptioncanbehighlyaccurateaswellasintegratedwithcognitive-drivengoals. TheCADframeworkalsoallowsforamoreusefuldistinctionthatcanpotentiallyclarifyambiguities.ConsiderKravitzandBehrmann’s(2011)findingconcerningfacilitationbyaconcept:fasterresponsetimestodetect‘h’basedonpriorexposureto‘H.’Thiskindofcognitivelydrivenattentiontosyntacticfeaturesshouldnotbeconsideredcognitivepenetration.Forsimilarreasons,semanticprimingshouldalsobeconsideredanattentionaleffectthatiscognitivelydrivenandthatoccursatlaterstagesofprocessing.Intheevolutionofthevisualandotherperceptualsystems,itislikelythatfeature-basedattentionandbasicformsofobject-basedattentionevolvedfirst,andonlylatercanonefindcomplexformsofsemanticallydrivenattentiontofeaturesrelevanttoexpertiseandpropositionalcontents(seeHaladjianandMontemayor,2015).Attentionbasedonpropositionalcontentis,therefore,akindofcognitiveguidancethatmustoccuratlaterstagesofperceptualprocessingandwhichmusthaveevolvedmorerecently.Thiskindofcognitiveguidanceatlaterstagescaninfluenceinference,memory,objectrecognition,andconceptcategorization.Inthecaseofhumancognitionandperception,thiskindofcognitivelydrivenattentiontosemanticcontentsisthemostimportantcomponentthatfacilitatesapowerfulinterfacebetweencognitionandperception,anditprovidesthebasicscaffoldingforconceptacquisitionofallkinds.Asmentionedbefore,conceptacquisitionallowsformanykindsofcognitivepenetrationatlaterstagesofprocessing,anditisfundamentaltounderstandhumanperception. Thereisyetanother,andperhapsevenmorerecent,kindofcognitivelydrivenattentionthatmodulateslateperceptualcontents:attentiontosyntacticallystructuredperceptualpatterns.Thecomplexhierarchicalstructureofhumanlanguagemustbesomehowperceived.Thequestionisexactlyhow.IfBerwickandChomsky(2016)areright,thecapacitytodetectsyntacticpatternsevolvedquiterecentlyinourspecies.Infact,ifitistruethatthecapacitytoarticulateandcombinestringsofsymbolshierarchicallyisasrecentas200,000to150,000yearsago(BerwickandChomsky,2016,p.54,indicatethatitisonly60,000yearsagothatitcertainlyemerged),thenitmustbeoneofthemostrecenteventsinourcognitiveevolution.Whilesyntaxprocessinghasaverysignificantimpactonhumancognition,itneednotoperatebyconstantlyinfluencingwhatweperceive(unlikeconceptuallybasedlateperception,whichisessentialforepistemicseeingandepistemicperceptionmoregenerally).Rather,itmayoperateinthewaymotorcontroloperates:inahighlyautomaticandreliablefashionthatcannotbemadeexplicitthroughdiscursivejudgmentandwhichprocessesinformationbeyondconsciousaccess.Ifso,eveninspiteofitsveryrecentevolution,syntaxprocessingmaynotprovidearobustinterfaceforcognitionandperception,andinterestingcasesofcognitivepenetrationatlateperceptualstagesmaybelimitedtosemanticprocessing.Thisisanissuethatneedstobestudiedinmoredetail. Thisbringsustothelastpointwewanttomake.Thefactthatperceptionisstableandinvariantattheearlystagessupportsnotonlyourcognitivesystemsbutalsomotorcontrolandaction.Earlyperceptionplaysthecriticalroleofmakingthispossible,bynotallowingdirectcasualinfluencesfromcognitionoremotionontheprocessingofthemostbasicstagesofperceptualscenestructuring.Basicperceptualexperiencesarealsostableinthiswayand,moreover,theyareexperiencedinawaythatdoesnotnecessitateconceptualorpropositionalguidance.Consciousaccesstocontents,ontheotherhand,likelyrequiresahighlevelofintegrationofinformationwithinthebrain,whichistheargumentmadebyglobalworkspacetheoriesofconsciousness(e.g.,DehaeneandNaccache,2001;Baars,2005),withincreasedfunctionalconnectivityamongdifferentneuralmodules(ratherthanwithinmodules)beingassociatedwithsuchconsciousawareness(Godwinetal.,2015).Cognitivepenetrationislikelytobefoundattheselaterstagesofperceptualprocessing,andcrucially,attheinterfacebetweenearlycontentsanddifferentformsofconceptformation.Afurtherquestionistheextenttowhichempathicandmotivationaleffectsguidelateperception.Withtherichconceptualframeworkofhumancognition,theinterfacebetweenemotion,cognition,andperceptionallowsformanykindsofcognitivepenetrationattheselaterstagesofprocessing.Semanticandsyntacticguidancethroughcognitivelydrivenattentionisacriticalpartofthisprocess. Acquiringconceptsdoesnotdirectlyaffecthowfeaturesaredetectedattheearliestlevel,buttheydodeterminewhatweepistemicallyperceive(e.g.,asamemberofacategory).Havingaspecificconceptisnotasurgentasrespondingtoafeatureessentialforsurvival,butbasiccategorization,evenifitisofapreconceptualkind,canhelpinurgentsituations,asthealarmcallsofsomeanimalsshow.Itcanalsoleadtocomplexformsofplanning,mentaltravel,andeventheoryofmind,asthequasi-conceptualcapacitiesofbirdsdemonstrate.Fullyfledgedconcepts,asfoundinhumans,leadtoacognitiveframeworkthatallowsnotonlyforepistemicseeing,butalsoforinferentialjudgment(includingdiscursiveinference),andepistemicjustification.BasedonCADandtheargumentfromevolution,itisusefultothinkofthesecapacitiesasfallingunderdifferentlevelsofcognitiveintegrationathigher-levelsofperceptualinformationprocessing.Toreiterate,thisisallcompatiblewiththecognitiveimpenetrabilityofearlyperception. Thus,CADhelpsclarifyhowthefactthatperceptionisdeeplyrelatedtocognitionandemotioniscompatiblewiththecognitiveimpenetrabilityofearlyperceptualprocessing.Conceptualinterfacesareatthecenteroftherelationbetweencognitionandemotion.Theseconceptualinterfacesmanifestinformsofperceptualpre-cuing,biases,modulation,andguidancethroughthemechanismofcognitivelydrivenattention.Theseinterfacesalsoprovidetheframeworkforthetypeofconsciousnessassociatedwithaccesstopropositionalcontents,whichaccordingtoCAD,isdissociatedfromtheexperiencesproducedbyearlyphenomenallyconsciousperception.Earlyperceptionguaranteesstabilityandreliability,aswellasaperceptualcommongroundwithotherorganisms.Lateperceptualprocessingprovidesarichframeworkofpossibilitiesthatenrichperceptioninmanyways.Finally,semanticandsyntacticinfluencesinlateperceptionincreasethesepossibilitiesinwaysthatcannotbefoundinanyotherspecies,andmakeshumanperceptiontherichmanifoldofcontentsthatmakepossibletheverycomplexbehaviorthatcharacterizeshumanity. Conclusion Howtheworldappearstouscandependlargelyonourexpectations,beliefs,anddesires.Thedebateoncognitivepenetrationhasexploredthisissueinthelastfewdecadesfromdifferentperspectives,particularlythoseconcerningcognitivearchitectureandsemanticcontent.Theconclusionmanyauthorsreachisthatcognitivepenetrationiseitherlargelypervasiveorinexistent.Wearguethatamorenuancedperspectiveisrequired.TheCADframeworkallowsforsuchaperspective,informedbyfindingsfromtheresearchonconsciousnessandattention,andtheirevolution.Morespecifically,CADhelpsexplainwhyalthoughtheremaybemanycasesofcognitivepenetrationinlateperception,earlyperceptionmustbecognitivelyimpenetrable. WiththeCADframework,amorebalancedapproachtocognitivepenetrationisfeasible.Aninterestingquestionis:couldasimilarbalancebeachievedwithoutit?Wecannotexplorethisissueindetailhere,butwebelievethatattheveryleast,CADisthebestwaytoachievethisbalance.Itmaybetheonlywaytoachievesuchabalanceinarigorousway,butwewillnotargueforthisstrongerclaimhere.However,weleavethisconsiderationinfavorofCAD:theevidence,includingevolution,doesnotsupportasstronglyaninterfacewithoutCAD.Forinstance,suchaninterfacecouldconcernonlyunconsciousprocessing(e.g.,constitutedbyHelmholtzianinferentialabilities).Alternatively,thisinterfacecouldinvolveexclusivelyconsciousinformation,requiringsubjectivelyexperiencedintegrationforanyperceptualprocess.Theevidenceindicatesthatneitheroftheseoptionsislikelytrue.Thus,theinterfacebetweencognitionandperceptionseemstobefundamentallystructuredintermsofCAD. GiventheimplicationsofCADandtheargumentfromevolution,wearguedthatconceptacquisitionisaparticularlyimportanttopicwithrespecttocognitivepenetration,withramificationsfortheintegrationofemotions,inferentialreasoning,andrecognitionalprocesses.Perceptionandcognitionmaybelargelyindependent,andtheyarefullyindependentatearlystages,buttherearesystematicwaysinwhichtheyinteract.Themorecognitiveintegrationthereis,themorecognitivepenetrationonefinds.Perhaps,assuggestedabove,theremayevenbemorethanoneinterfaceforcognitivepenetrationbecausetherearemanykindsofcognitivemodulationinlateperception.Yetdespitethissystematicinteractionbetweencognitionandperceptionatsuchlatestages,cognitivepenetrationisnotpervasive. Besidesprovidingpositivesuggestionsforaddressingtheproblemofpenetrabilityinamorethoroughtheoreticalway,thispaperalsoraiseschallengingquestions.Whatkindofconceptualorepistemiccapacitiesunderliedifferentformsofpenetrability?Whichcapacitiesnecessitatecognitivepenetration?Howcanoneverifysuchcapacitiesacrossdifferentspecies?Howisitpossibletointegratethefindingsonconsciousnessandattention,aswellastheirdissociation,inawaythataddressestheproblemofcognitivepenetration?Thefindingsonanimalcognitionandfutureresearchonhowourowncapacitiescomparetothoseofotherspecies,particularlyinthedevelopmentofsemanticandconceptualguidance,isfertilegroundforexploration.Theargumentfromevolution,especiallyasitconcernsthedevelopmentofdifferentformsofintentionality,shouldhelpguidefutureinvestigationsinthisarea. AuthorContributions Allauthorslisted,havemadesubstantial,directandintellectualcontributiontothework,andapproveditforpublication. Funding HHreceivedpostdoctoralresearchfundingfromtheEuropeanResearchCouncilundertheEuropeanUnion’sSeventhFrameworkProgramme(FP7/2007-2013)/ERCgrantagreementNo.AG324070awardedtoPatrickCavanagh. ConflictofInterestStatement Theauthorsdeclarethattheresearchwasconductedintheabsenceofanycommercialorfinancialrelationshipsthatcouldbeconstruedasapotentialconflictofinterest. Acknowledgments WeareespeciallyindebtedtoAnasthasiosRaftopoulosforextensive,detailed,andenormouslyhelpfulfeedback.Thispaperimprovedsubstantiallybecauseofhim.WewouldalsoliketothankAlbertNewenforvaluablediscussion,andtworeviewersfortheirinsightfulsuggestionsandrecommendations. References Alter,A.L.,andBalcetis,E.(2011).Fondnessmakesthedistancegrowshorter:desiredlocationsseemcloserbecausetheyseemmorevivid.J.Exp.Soc.Psychol.47,16–21.doi:10.1016/j.jesp.2010.07.018 CrossRefFullText|GoogleScholar Baars,B.J.(2005).Globalworkspacetheoryofconsciousness:towardacognitiveneuroscienceofhumanexperience.Prog.BrainRes.150,45–53.doi:10.1016/S0079-6123(05)50004-9 PubMedAbstract|CrossRefFullText|GoogleScholar Bermúdez,J.L.(2003).ThinkingwithoutWords.NewYork,NY:OxfordUniversityPress. GoogleScholar Berwick,R.C.,andChomsky,N.(2016).WhyOnlyUs:LanguageandEvolution.Cambridge,MA:MITPress. GoogleScholar Block,N.(1995).Onaconfusionaboutafunctionofconsciousness.Behav.BrainSci.18,227–247.doi:10.1017/S0140525X00038188 CrossRefFullText|GoogleScholar Byrne,A.(2005).“PerceptionandConceptualContent,”inContemporaryDebatesinEpistemology,edsE.SosaandM.Steup(Oxford:Blackwell),231–250. GoogleScholar Carey,S.(2009).TheOriginofConcepts.NewYork,NY:OxfordUniversityPress. GoogleScholar Cavanagh,P.(2011).Visualcognition.VisionRes.51,1538–1551.doi:10.1016/j.visres.2011.01.015 PubMedAbstract|CrossRefFullText|GoogleScholar Cavanagh,P.,Labianca,A.T.,andThornton,I.M.(2001).Attention-basedvisualroutines:sprites.Cognition80,47–60.doi:10.1016/S0010-0277(00)00153-0 PubMedAbstract|CrossRefFullText|GoogleScholar Clark,A.(2013).Whatevernext?Predictivebrains,situatedagents,andthefutureofcognitivescience.Behav.BrainSci.36,181–204.doi:10.1017/S0140525X12000477 PubMedAbstract|CrossRefFullText|GoogleScholar Clayton,N.S.,andDickinson,A.(1998).Episodic-likememoryduringcacherecoverybyscrubjays.Nature395,272–274.doi:10.1038/26216 PubMedAbstract|CrossRefFullText|GoogleScholar Clifford,C.W.G.,Webster,M.A.,Stanley,G.B.,Stocker,A.A.,Kohn,A.,Sharpee,T.O.,etal.(2007).Visualadaptation:neural,psychologicalandcomputationalaspects.VisionRes.47,3125–3131.doi:10.1016/j.visres.2007.08.023 PubMedAbstract|CrossRefFullText|GoogleScholar Crane,T.(2009).Isperceptionapropositionalattitude?Philos.Q.59,452–469.doi:10.1111/j.1467-9213.2008.608.x CrossRefFullText|GoogleScholar Dehaene,S.,andNaccache,L.(2001).Towardsacognitiveneuroscienceofconsciousness:basicevidenceandaworkspaceframework.Cognition79,1–37.doi:10.1016/S0010-0277(00)00123-2 PubMedAbstract|CrossRefFullText|GoogleScholar Firestone,C.,andScholl,B.J.(2016).Cognitiondoesnotaffectperception:evaluatingtheevidencefor“top-down”effects.Behav.BrainSci.39,1–77.doi:10.1017/S0140525X15000965 PubMedAbstract|CrossRefFullText|GoogleScholar Fodor,J.A.(1983).TheModularityofMind:AnEssayonFacultyPsychology.Cambridge,MA:MITPress. GoogleScholar Fodor,J.A.(1998).Concepts:WhereCognitiveScienceWentWrong.Oxford:OxfordUniversityPress. GoogleScholar Gallistel,C.R.(1990).Representationsinanimalcognition:anintroduction.Cognition37,1–22.doi:10.1016/0010-0277(90)90016-D PubMedAbstract|CrossRefFullText|GoogleScholar Gilbert,C.D.,andLi,W.(2013).Top-downinfluencesonvisualprocessing.Nat.Rev.Neurosci.14,350–363.doi:10.1038/nrn3476 PubMedAbstract|CrossRefFullText|GoogleScholar Godwin,D.,Barry,R.L.,andMarois,R.(2015).Breakdownofthebrain’sfunctionalnetworkmodularitywithawareness.Proc.Natl.Acad.Sci.U.S.A.112,3799–3804.doi:10.1073/pnas.1414466112 PubMedAbstract|CrossRefFullText|GoogleScholar Haladjian,H.H.,andMontemayor,C.(2015).Ontheevolutionofconsciousattention.Psychon.Bull.Rev.22,595–613.doi:10.3758/s13423-014-0718-y PubMedAbstract|CrossRefFullText|GoogleScholar Hansen,T.,Olkkonen,M.,Walter,S.,andGegenfurtner,K.R.(2006).Memorymodulatescolorappearance.Nat.Neurosci.9,1367–1368.doi:10.1038/nn1794 PubMedAbstract|CrossRefFullText|GoogleScholar Izard,V.,andDehaene,S.(2008).Calibratingthementalnumberline.Cognition106,1221–1247.doi:10.1016/j.cognition.2007.06.004 PubMedAbstract|CrossRefFullText|GoogleScholar Kahneman,D.(2000).“Experiencedutilityandobjectivehappiness:amoment-basedapproach,”inChoices,Values,andFrames,edsD.KahnemanandA.Tversky(NewYork,NY:CambridgeUniversityPress),673–692. GoogleScholar Keil,F.C.(1989).Concepts,Kinds,andCognitiveDevelopment.Cambridge,MA:MITPress. GoogleScholar Kentridge,R.W.(2011).“Attentionwithoutawareness:abriefreview,”inAttention:PhilosophicalandPsychologicalEssays,edsC.Mole,D.Smithies,andW.Wu(Oxford:OxfordUniversityPress),228–246. GoogleScholar Kersten,D.,Mamassian,P.,andYuille,A.(2004).ObjectperceptionasBayesianinference.Annu.Rev.Psychol.55,271–304.doi:10.1146/annurev.psych.55.090902.142005 CrossRefFullText|GoogleScholar Koch,C.,andTsuchiya,N.(2012).Attentionandconsciousness:relatedyetdifferent.TrendsCogn.Sci.16,103–105.doi:10.1016/j.tics.2011.11.012 PubMedAbstract|CrossRefFullText|GoogleScholar Kosslyn,S.M.(1980).ImageandMind.Cambridge,MA:HarvardUniversityPress. GoogleScholar Kosslyn,S.M.(1994).ImageandBrain:TheResolutionoftheImageryDebate.Cambridge,MA:MITPress. GoogleScholar Kravitz,D.J.,andBehrmann,M.(2011).Space-,object-,andfeature-basedattentioninteracttoorganizevisualscenes.Atten.Percept.Psychophys.73,2434–2447.doi:10.3758/s13414-011-0201-z PubMedAbstract|CrossRefFullText|GoogleScholar Landry,S.P.,andChampoux,F.(2017).Musiciansreactfasterandarebettermultisensoryintegrators.BrainCogn.111,156–162.doi:10.1016/j.bandc.2016.12.001 PubMedAbstract|CrossRefFullText|GoogleScholar LeDoux,J.E.(2012).Rethinkingtheemotionalbrain.Neuron73,653–676.doi:10.1016/j.neuron.2012.02.004 PubMedAbstract|CrossRefFullText|GoogleScholar Lisi,M.,andCavanagh,P.(2015).Dissociationbetweentheperceptualandsaccadiclocalizationofmovingobjects.Curr.Biol.25,2535–2540.doi:10.1016/j.cub.2015.08.021 PubMedAbstract|CrossRefFullText|GoogleScholar Lisi,M.,andGorea,A.(2016).Timeconstancyinhumanperception.J.Vis.16:3.doi:10.1167/16.14.3 PubMedAbstract|CrossRefFullText|GoogleScholar Loar,B.(1997).“Phenomenalstates(secondversion),”inTheNatureofConsciousness:PhilosophicalDebates,edsN.Block,O.J.Flanagan,andG.Güzeldere(Cambridge,MA:MITPress),597–616. GoogleScholar Macpherson,F.(2012).Cognitivepenetrationofcolourexperience:rethinkingtheissueinlightofanindirectmechanism.Philos.Phenomenol.Res.84,24–62.doi:10.1111/j.1933-1592.2010.00481.x CrossRefFullText|GoogleScholar Matthews,R.J.(2007).TheMeasureofMind:PropositionalAttitudesandTheirAttribution.NewYork,NY:OxfordUniversityPress. GoogleScholar Millikan,R.G.(2005).Language:ABiologicalModel.NewYork,NY:OxfordUniversityPress. GoogleScholar Montemayor,C.(2010).“Time:biological,intentional,andcultural,”inTime:LimitsandConstraints,edsJ.A.Parker,P.Harris,andC.Steineck(Leiden:Brill),39–63. GoogleScholar Montemayor,C.(2013).MindingTime:APhilosophicalandTheoreticalApproachtothePsychologyofTime.Boston:Brill. GoogleScholar Montemayor,C.,andHaladjian,H.H.(2015).Consciousness,Attention,andConsciousAttention.Cambridge,MA:MITPress. GoogleScholar Neufeld,E.,Brown,E.C.,Lee-Grimm,S.-I.,Newen,A.,andBrüne,M.(2016).Intentionalactionprocessingresultsfromautomaticbottom-upattention:anEEG-investigationintotheSocialRelevanceHypothesisusinghypnosis.Conscious.Cogn.42,101–112.doi:10.1016/j.concog.2016.03.002 PubMedAbstract|CrossRefFullText|GoogleScholar Newen,A.(2016).Defendingtheliberal-contentviewofperceptualexperience:directsocialperceptionofemotionsandpersonimpressions.Synthese1–25.doi:10.1007/s11229-016-1030-3 CrossRefFullText|GoogleScholar Newen,A.,andVetter,P.(2017).Whycognitivepenetrationofourperceptualexperienceisstillthemostplausibleaccount.Conscious.Cogn.47,26–37.doi:10.1016/j.concog.2016.09.005 PubMedAbstract|CrossRefFullText|GoogleScholar Norman,L.J.,Akins,K.,Heywood,C.A.,andKentridge,R.W.(2014).Colorconstancyforanunseensurface.Curr.Biol.24,2822–2826.doi:10.1016/j.cub.2014.10.009 PubMedAbstract|CrossRefFullText|GoogleScholar Palmer,T.D.,andRamsey,A.K.(2012).Thefunctionofconsciousnessinmultisensoryintegration.Cognition125,353–364.doi:10.1016/j.cognition.2012.08.003 PubMedAbstract|CrossRefFullText|GoogleScholar Pasternak,T.,andGreenlee,M.W.(2005).Workingmemoryinprimatesensorysystems.Nat.Rev.Neurosci.6,97–107.doi:10.1038/nrn1603 PubMedAbstract|CrossRefFullText|GoogleScholar Pauers,M.J.,Kuchenbecker,J.A.,Neitz,M.,andNeitz,J.(2012).Changesinthecolouroflightcuecircadianactivity.Anim.Behav.83,1143–1151.doi:10.1016/j.anbehav.2012.01.035 PubMedAbstract|CrossRefFullText|GoogleScholar Peacocke,C.(1992).“Scenarios,conceptsandperception,”inTheContentsofExperience,ed.T.Crane(Cambridge:CambridgeUniversityPress),105–135. GoogleScholar Pessoa,L.(2008).Ontherelationshipbetweenemotionandcognition.Nat.Rev.Neurosci.9,148–158.doi:10.1038/nrn2317 PubMedAbstract|CrossRefFullText|GoogleScholar Pylyshyn,Z.W.(1989).Theroleoflocationindexesinspatialperception:asketchoftheFINSTspatial-indexmodel.Cognition32,65–97.doi:10.1016/0010-0277(89)90014-0 PubMedAbstract|CrossRefFullText|GoogleScholar Pylyshyn,Z.W.(1999).Isvisioncontinuouswithcognition?Thecaseforcognitiveimpenetrabilityofvisualperception.Behav.BrainSci.22,341–365;discussion366–423.doi:10.1017/S0140525X99002022 PubMedAbstract|CrossRefFullText|GoogleScholar Pylyshyn,Z.W.(2003).SeeingandVisualizing:It’sNotWhatYouThink.Cambridge,MA:MITPress. GoogleScholar Raftopoulos,A.(2001a).Isperceptioninformationallyencapsulated?:theissueofthetheory-ladennessofperception.Cogn.Sci.25,423–451.doi:10.1016/S0364-0213(01)00042-8 CrossRefFullText|GoogleScholar Raftopoulos,A.(2001b).Reentrantneuralpathwaysandthetheory-ladennessofperception.Philos.Sci.68,S187–S199.doi:10.1086/392908 CrossRefFullText|GoogleScholar Raftopoulos,A.(2014).Thecognitiveimpenetrabilityofthecontentofearlyvisonisanecessaryandsufficientconditionforpurelynonconceptualcontent.Philos.Psychol.27,601–620.doi:10.1080/09515089.2012.729486 CrossRefFullText|GoogleScholar Raftopoulos,A.(2015a).Thecognitiveimpenetrabilityofperceptionandtheory-ladenness.J.Gen.Philos.Sci.46,87–103.doi:10.1007/s10838-015-9288-6 CrossRefFullText|GoogleScholar Raftopoulos,A.(2015b).“Reframingtheproblemofcognitivepenetrability,”inPhilosophyandCognitiveScienceII:Western&EasternStudies,edsL.Magnani,P.Li,andW.Park(London:SpringerInternationalPublishing),3–20. GoogleScholar Serences,J.T.,Ester,E.F.,Vogel,E.K.,andAwh,E.(2009).Stimulus-specificdelayactivityinhumanprimaryvisualcortex.Psychol.Sci.20,207–214.doi:10.1111/j.1467-9280.2009.02276.x PubMedAbstract|CrossRefFullText|GoogleScholar Siegel,S.(2006).“Whichpropertiesarerepresentedinperception?,”inPerceptualExperience,edsT.GendlerandJ.Hawthorne(NewYork,NY:OxfordUniversityPress),481–503. GoogleScholar Siegel,S.(2010).TheContentsofVisualExperience.NewYork,NY:OxfordUniversityPress. GoogleScholar Spelke,E.(1988).“Whereperceivingendsandthinkingbegins:theapprehensionofobjectsininfancy,”inPerceptualDevelopmentinInfancy,ed.A.Yonas(Hillsdale,NJ:LawrenceErlbaumAssociates),197–234. GoogleScholar Spelke,E.S.,andKinzler,K.D.(2007).Coreknowledge.Dev.Sci.10,89–96.doi:10.1111/j.1467-7687.2007.00569.x PubMedAbstract|CrossRefFullText|GoogleScholar Stokes,D.(2012).Perceivinganddesiring:anewlookatthecognitivepenetrabilityofexperience.Philos.Stud.158,477–492.doi:10.1007/s11098-010-9688-8 CrossRefFullText|GoogleScholar Stöttinger,E.,andPerner,J.(2006).Dissociatingsizerepresentationforactionandforconsciousjudgment:graspingvisualillusionswithoutapparentobstacles.Conscious.Cogn.15,269–284.doi:10.1016/j.concog.2005.07.004 PubMedAbstract|CrossRefFullText|GoogleScholar Tamietto,M.,anddeGelder,B.(2010).Neuralbasesofthenon-consciousperceptionofemotionalsignals.Nat.Rev.Neurosci.11,697–709.doi:10.1038/nrn2889 PubMedAbstract|CrossRefFullText|GoogleScholar Theeuwes,J.(2010).Top-downandbottom-upcontrolofvisualselection.ActaPsychol.135,77–99.doi:10.1016/j.actpsy.2010.02.006 PubMedAbstract|CrossRefFullText|GoogleScholar Tononi,G.(2012).Integratedinformationtheoryofconsciousness:anupdatedaccount.Arch.Ital.Biol.150,56–90.doi:10.4449/aib.v149i5.1388 PubMedAbstract|CrossRefFullText|GoogleScholar Treisman,A.(1988).Featuresandobjects:thefourteenthBartlettmemoriallecture.Q.J.Exp.Psychol.40,201–237.doi:10.1080/02724988843000104 PubMedAbstract|CrossRefFullText|GoogleScholar vanUlzen,N.R.,Semin,G.R.,Oudejans,R.R.D.,andBeek,P.J.(2008).AffectivestimuluspropertiesinfluencesizeperceptionandtheEbbinghausillusion.Psychol.Res.72,304–310.doi:10.1007/s00426-007-0114-6 PubMedAbstract|CrossRefFullText|GoogleScholar Vetter,P.,andNewen,A.(2014).Varietiesofcognitivepenetrationinvisualperception.Conscious.Cogn.27,62–75.doi:10.1016/j.concog.2014.04.007 PubMedAbstract|CrossRefFullText|GoogleScholar Vishton,P.M.,Stephens,N.J.,Nelson,L.A.,Morra,S.E.,Brunick,K.L.,andStevens,J.A.(2007).Planningtoreachforanobjectchangeshowthereacherperceivesit.Psychol.Sci.18,713–719.doi:10.1111/j.1467-9280.2007.01965.x PubMedAbstract|CrossRefFullText|GoogleScholar Vlach,H.A.(2016).Howwecategorizeobjectsisrelatedtohowwerememberthem:theshapebiasasamemorybias.J.Exp.ChildPsychol.152,12–30.doi:10.1016/j.jecp.2016.06.013 PubMedAbstract|CrossRefFullText|GoogleScholar Webster,M.A.,andMacLeod,D.I.A.(2011).Visualadaptationandfaceperception.Philos.Trans.R.Soc.Lond.BBiol.Sci.366,1702–1725.doi:10.1098/rstb.2010.0360 PubMedAbstract|CrossRefFullText|GoogleScholar Wiederman,S.D.,andO’Carroll,D.C.(2013).Selectiveattentioninaninsectvisualneuron.Curr.Biol.23,156–161.doi:10.1016/j.cub.2012.11.048 PubMedAbstract|CrossRefFullText|GoogleScholar Wu,W.(2014).Againstdivision:consciousness,informationandthevisualstreams.MindLang.29,383–406.doi:10.1111/mila.12056 CrossRefFullText|GoogleScholar Yeh,S.-L.,andChen,I.P.(1999).Isearlyvisualprocessingattentionimpenetrable?Behav.BrainSci.22:400.doi:10.1017/S0140525X99602023 CrossRefFullText|GoogleScholar Yuille,A.,andKersten,D.(2006).VisionasBayesianinference:Analysisbysynthesis?TrendsCogn.Sci.10,301–308.doi:10.1016/j.tics.2006.05.002 PubMedAbstract|CrossRefFullText|GoogleScholar Zentall,T.R.(2005).Selectiveanddividedattentioninanimals.Behav.Process.69,1–15.doi:10.1016/j.beproc.2005.01.004 PubMedAbstract|CrossRefFullText|GoogleScholar Keywords:cognitivepenetrability,consciousness,visualattention,evolution,dissociation,language,conceptacquisition Citation:MontemayorCandHaladjianHH(2017)PerceptionandCognitionAreLargelyIndependent,butStillAffectEachOtherinSystematicWays:ArgumentsfromEvolutionandtheConsciousness-AttentionDissociation.Front.Psychol.8:40.doi:10.3389/fpsyg.2017.00040 Received:09September2016;Accepted:06January2017;Published:24January2017. Editedby: AthanassiosRaftopoulos,UniversityofCyprus,Cyprus Reviewedby: MariaOlkkonen,DurhamUniversity,UK RobertLawrenceWest,CarletonUniversity,Canada Copyright©2017MontemayorandHaladjian.Thisisanopen-accessarticledistributedunderthetermsoftheCreativeCommonsAttributionLicense(CCBY).Theuse,distributionorreproductioninotherforumsispermitted,providedtheoriginalauthor(s)orlicensorarecreditedandthattheoriginalpublicationinthisjournaliscited,inaccordancewithacceptedacademicpractice.Nouse,distributionorreproductionispermittedwhichdoesnotcomplywiththeseterms. *Correspondence:CarlosMontemayor,[email protected] COMMENTARY ORIGINALARTICLE Peoplealsolookedat SuggestaResearchTopic>