JohnH.Taylor

IsAttentionNecessaryandSufficientforPhenomenalConsciousness?

HenryTaylor

ThisisthepenultimateversionofapaperthatappearedinTheJournalofConsciousnessStudies2013.20:173-194

Correspondence:jht30@cam.ac.uk

Abstract.

There has recently been a flurry of interest over how attention and phenomenalconsciousnessinteract.FelipeDeBrigardandJessePrinzhavemadetheboldclaimthatattentionisnecessary and sufficient forphenomenal consciousness. If this turnsout tobe true, thenwewillhavetakensignificantstepstowardnaturalisingthemind,whichisaparticularlyexcitingprospect.

Againstthisposition,severalthinkershavepresentedempiricaldatawhichapparentlyshowthatconsciousnessispossibleintheabsenceofattention,andviceversa.InthispaperIarguethattheseresultsdonotharmDeBrigardandPrinz’sposition,butthatthisisunsurprisingbecausetheyuseadefinitionof‘attention’whichmakestheirviewempiricallyself-sealing.Ishallalsoarguethatthe argument in favour of this definition of attention is unsuccessful. I shall close with somecommentsonwhatshouldbedoneforthedebatetoprogress.Ihavethreemainaimsinthispaper:firstly to give an overview of the debate, secondly to thoroughly analyse De Brigard and Prinz’sposition and thirdly (and most importantly) to point out some general and troublesomemethodological issues that beset the debate, which have gone largely unacknowledged in theliterature. Particularly, I will highlight the cross-purposes which stem from participants usingdefinitionsofthekeytermsindifferentways.

1-Conceptualpreliminaries.

In this section, I will get clear on howwe should understand the claim that ‘attention is

necessary and sufficient for consciousness’.1In sections 2 and 3 I shall outline the empirical data

whichhavebeenpresentedagainstthisview,andarguethatthesedatadonotharmDeBrigardand

Prinz’spositionbut that this isunsurprisingbecause theydefineattention ina tendentiousway. I

shalldiscussonlythedatawhicharemostprominentintheliterature.Insection4Iwillarguethat

theirdefinitionofattentionmakestheirviewempiricallyself-sealing.Iwillconsiderpossiblewaysof

refuting theposition, andargue thatnoneof themcanwork. In sections2-4 Iwill not argue that

theirdefinitionofattentionisfalse,Ionlywishtoshowhowthedefinitionservestosweepasideall

1PrinzhasdevelopedtheviewinmoredetailthanDeBrigard,andasaresult,manyoftheclaimsaremadeonlybyPrinz,notboth.Forthisreason,IwillattributesuchviewsonlytoPrinz.ItisunclearwhetherDeBirgardwouldstillholdtheviewthatattentionissufficientforconsciousness.Forexample,inhis(2012)heclaimsthatattentionisinsufficientforconsciousrecollectionofmemories,andhehasinformedmethatheisalsomorescepticaloftheviewthatattentionissufficientforperceptualrepresentationstobecomeconscious.

JohnH.Taylor

2

possible empirical evidence that could be presented against their position. In section 5 I will

considerPrinz’sargument in favourofhisdefinitionofattentionandarguethat it iscircular. Iwill

alsoconsideranargumentPrinzgiveswhich I call the ‘threatofeliminativism’andargue that this

fails.Insection6Isuggestsomestepsforresolvingthesedifficulties.

De Brigard and Prinz say this: ‘[w]e claim that attention is necessary and sufficient for

perceptual representations to become conscious’ (2010, p.51). We can understand this as a

conjunctionofthefollowingtwoclaims,whichIcallNTandST:

i) Thenecessitythesis(NT):

Attention to some item(s) isnecessary fora representationof that item (or those

items)tobephenomenallyconscious.2

ii) Thesufficiencythesis(ST):

Attention to some item(s) is sufficient for a representationof that item (or those

items)tobephenomenallyconscious.

STandNTallowforthepossibilitythattheattentioninquestionmaybeinward-directed(at

ourownmental statesandevents)oroutward-directed (at theworld).Relatedly,STandNT leave

openthepossibilitythattheitemthatthesubjectisattendingtomaybeself-representing.Thatisto

say,itmaybethecasethatsometimes(forexample,whenwepayattentiontoourownperceptual

states)thosestatesrepresentthemselvesinaphenomenallyconsciousmanner.

Wecanbrieflylistthemainpositionsinthedebate:3

1) Attentionisnecessaryandsufficientforconsciousness(DeBrigardandPrinz,2010and

Prinz,2010,2011and2012).

2) Attentionisnecessarybutnotsufficientforconsciousness(Cohenetal.2012).4

2Itakephenomenalconsciousnesstobeatermthatappliestoallandonlythosementaleventsthatthereis‘somethingitislike’toundergo(seeNagel,1974).3Thislistisobviouslynotexhaustive.

JohnH.Taylor

3

3) Attention is sufficient but unnecessary for consciousness (Smithies, 2011 and Mole,

2008).5

4) Attention is not sufficient for consciousness (Kentridge, 2011 Kentridge et al. 1999,

2008aand2008bandNormanetal.inpress).

5) Attention is neither necessary nor sufficient for consciousness (Lamme, 2003, 2010;

KochandTsuchiya,2007).6

Wecannowaskwhyweshouldcarewhetherornotattentionisnecessaryandsufficientfor

consciousness. Themain reason is that if position (1) were true, then we would have gained an

importantinsightintothenatureofphenomenalconsciousness,andhowitinteractswiththerestof

the mind. Such an insight would have important implications for philosophy, psychology and

neuroscience. Position (1) also opens the door to other important theses about the interaction

betweenattentionandconsciousness,forexample,itmaybethatoneofthemcausestheother,or

thatoneconstitutes theother (in the senseofbeinganecessarypartof it). It couldevenbe that

attention and consciousness are identical. 7 Another important reason to care about whether

attentionisnecessaryandsufficientforconsciousnessisthatthisisthecentralclaimofPrinz’swhole

‘AIR’theoryofconsciousness,soitshouldbeexaminedcarefully.

DeBrigardandPrinz’sarguments infavourofposition(1)arebaseduponinferencetothe

best explanation. They argue that in certain cases, attention to a stimulus can bring it to our

consciousness,andthatalackofattentiontoastimuluscanexcludeitfrombeingconscious(seeDe

Brigard and Prinz, 2010, pp.53-54 and Prinz, 2012, pp.81-87). They draw upon cases such as

4Tye(2010)defendsathesissimilartoNT.DeBrigardhastoldmethathisviewisnowprobablycloserto(2)than(1).5ThoughMolegivesadifferentviewinhis(2011).6(5)isthemostnaturalreadingofBlock(2013).7Thoughsuchanidentitytheorycancomeinvariouskinds.Prinzhimselfidentifiesconsciousexperienceswithmentaleventsthatfulfilacertainfunctionalrole(beingattendedto)andthenarguesthatvectorwavefiringwithinthegammarangerealisesthisfunctionalrole(2012,ch.9).Prinz,however,thinksthatitisacontingentfactthatvectorwavefiringrealisesthisrole,sotheidentitybetweenconsciousness,attentionandvectorwavefiringdoesnotholdwithmetaphysicalnecessity.ThegeneralstructureofPrinz’sviewissimilarinmanywaystothatofLewis(1966).

JohnH.Taylor

4

inattentionalblindness,visualpop-out,visualneglectandotherphenomenatomakethiscase.From

thistheyconcludethatattentionisnecessaryandsufficientforconsciousness.

Iwillnotanalysethearguments infavouroftheircentralclaim.This isfirstlybecausethey

arenotthefocusofthispaper,andsecondlybecausetheyhavebeenaddressedelsewhere.8Ishould,

however,mentionthatsincetheirargumentismainlybaseduponanempiricalcorrelationbetween

phenomenalconsciousnessandattention,andsince their claim isauniversalone (theyclaimthat

attentionisalwaysnecessaryandsufficientforphenomenalconsciousness)weonlyneedonecase

whereeitherattentionoccurswithoutconsciousness(disprovingST)orwhereconsciousnessoccurs

in the absence of attention (disproving NT) in order to disprove their claim that attention is

necessaryandsufficientforconsciousness.

2-SupposedevidenceagainstST.

In this section, Iwill argue thatDeBrigardandPrinz’sdefinitionof their key termsmakes

theirpositionimpervioustofalsificationbytheempiricalresultssofarputforth.Mymainfocuswill

be on their definition of ‘attention’ but I will also raise some important issues concerning their

definitionof‘workingmemory’insection4.2.Myaiminthissectionisnottoarguethattheirviewis

false but rather to show how their definition of attention serves to insulate their position from

falsification. I will assess whether there is good reason to accept their definition of attention in

section5.

2.1-Theevidence.

In a series of studies,9the subject GY (who suffers from blindsight)10was tested in an

attempttodeterminewhetherhecouldattendtostimuliinhisblindfield.Theexperimenterssetup

a test where they would present target stimuli in the blind area of GY’s visual field. Since the

8E.g.Mole(2011,ch.7).9SeeKentridgeetal.(1999,2008aand2008b).SeealsoNormanetal.(inpress)andKentridge,(2011).10Blindsightisaconditionwheresubjectsdenyawarenessofitemsplacedincertainareasoftheirvisualfield,butaresusceptibletoprimingandothersubliminaleffectswhichshowthatinformationabouttheitemsisprocessedintheirvisualsystem.

JohnH.Taylor

5

stimuluswas inGY’sblindarea,hedeniedseeing it.A tonewas soundedonoccasion, sometimes

coincidingwith thepresentationof the target stimulus inGY’sblindarea, sometimesnot.GYwas

encouragedtorespondasquicklyaspossibleafterthetonewassounded ifhe felt thatthetarget

had been presented in his blind field. Itwas found thatGYwas quicker to respond to the target

stimulus if the location of the stimulus was prior indicated by the use of a cue. The cue was

presentedinthehealthyareaofGY’svisualfield.11TheexperimentersconcludedthatGYwaspaying

attentiontothetargetstimulus,eventhoughhehadnophenomenalrepresentationofit:‘attention

could selectively modulate the processing of a target without that target’s entering awareness’

(2011,p.240).

Following this, studieswereperformed in anattempt todemonstrate attention to certain

items in the absence of phenomenal consciousness of the attended items in non-neurologically

impairedsubjects.Itwasfoundthatthesesubjectsweremorelikelytobeprimedbyanunconscious

stimuluswhenarrows (whichwerepresented ina locationvisible to thesubjects)pointed toward

thatstimulus(seeKentridgeetal.2008bandKentridge,2011,p.240).12

2.2-Whytheresultsdonotdamagetheposition.

WecanseehowDeBrigardandPrinz’sposition isnotdamagedbytheseresultswhenwe

look at how they define attention.Here is Prinz’s definition: ‘attention can be identifiedwith the

processesthatallowinformationtobeencodedinworkingmemory.Whenastimulusisattended,it

becomesavailabletoworkingmemory,andifitisunattended,itisunavailable’(Prinz,2011,p.184.

cf.Prinz,2012,p.93andDeBrigardandPrinz,2010,p.52).

11Insomemanipulations,cueswereusedwhichfellinGY’sblindfield(inthesemanipulations,hewasawareofneitherthecuenorthetarget).Hisreactionspeedalsoincreasedinthesemanipulations(seeKentridgeetal.,1999,pp.1805ff.).Theseperipheralcueswerepairsofbarspresentedaroundthelocationwherethetargetwouldappear.12Thestimuliunderwentmeta-contrastmaskingtoensurethatthesubjectswerenotphenomenallyconsciousofthem.Maskingisaprocesswheretheouteredgesofaninitialstimulus(the‘masked’stimulus)coincidewiththeinneredgesofasubsequentlypresented‘mask’stimulus.Theresultisforsignalsfromthemaskedstimulustocoincidewithsignalsfromthemaskintheearlystagesofprocessinginthevisualcortex,resultinginthemaskedstimulusnotbeingconsciouslyperceived.

JohnH.Taylor

6

So,attentionistobeidentifiedwiththefacultythatmakesinformationavailabletoworking

memory,allandonlyinformationthatisavailabletoworkingmemoryisattendedto.Howdothey

understand‘workingmemory’?

Workingmemory, as Prinz understands it, is ‘a short-term storage capacity, but one that

allows for “executive control”… Once something is encoded in working memory, it becomes

availabletolanguagesystemsforreporting,andwithsystemsthatalloweffortfulserialprocessing’

(2011, p.184). Prinz also expands the idea thus: ‘[t]he attended stimulus becomes available for

processes that are controlled and deliberative. For example, we can report the stimulus that we

consciouslyperceive,wecanreasonabout it,wecankeep it inourmindsforawhile,andwecan

wilfullychoosetoexamineitfurther’(2012,p.92).

Theimportantthingtonoteaboutthesedefinitionsofworkingmemoryistheimportanceof

reportability.That is, Prinznotes thatoneof the functionsofworkingmemory is tomake certain

information reportable. Information which is not available to workingmemory is not reportable;

informationwhichisavailabletoworkingmemoryisreportable.NoticealsothatPrinzmentionsthat

informationavailabletoworkingmemorycan(bydefinition)beusedforcontrolledanddeliberative

action,soifasubjectshowsprimingaffectsafterbeingexposedtoastimulus,i.e.ifshecanreport

thestimuluswhengivenaforcedchoice,butdeniesseeingit,andclaimsonlytobeguessingwhat

wasthere,thenthestimulusdoesnotcountashavingenteredworkingmemory.

Withthisdefinitionofattention,wecannowseehowthestudiesinquestiondonotharm

ST.Theimportantdetailliesinthewaythattheexperimenterscanestablishthatthesubjectdidnot

haveaphenomenalrepresentationofthestimulustheywere(supposedly)attendingto.Inorderto

establish this, the experimenters had to ask the subjects whether they saw the stimulus. The

subjectsinquestion,ofcourse,deniedthattheysawthestimulus.Thisistakenasevidencethatthey

werenotphenomenallyconsciousofit(Kentridge,2011,p.230).Oncethishasbeenestablished,the

JohnH.Taylor

7

experimentersattempttoestablishthatthesubjectspaidattentiontothestimulithattheydenied

seeing.Iftheycanestablishboth,thenwehaveevidenceagainstST.

Theproblem is that if the subjects do indeeddeny seeing the stimulus (which is how the

experimentersestablishedthatthestimuluswasunconscious),then itwill followthatthestimulus

wasnotavailable toworkingmemory,and thus (byDeBrigardandPrinz’sdefinition of attention)

that it was not attended. So immediately the data will not count as an instance of attention to

something which was not phenomenally conscious. So the data in question will not count as a

counterexampletoST.

Indeed,Prinzseemstomakejustsuchanargument(thoughhedoesnotstateitinsuchstark

terms) when he says the following: ‘[i]fmy earlier analysis of attention is right, attention entails

availability to working memory. Availability is clearly absent in blindsight, so GY cannot be

instantiatingalloftheprocessesnecessaryforattention’ (2012,p.115cf.2011,p.194).Prinzrelies

uponhisearlierdefinitionofattention inordertoherereject theKentridgestudiesasshowingan

instanceofattentionwithoutphenomenalconsciousness.

What we see, then, is that De Brigard and Prinz’s position is completely immune to

falsificationbytheresults thatKentridgeetal.put forth.Ordinarily,ofcourse, thiswouldbegood

newsfortheirposition.However, there isclearlysomethingsuspecthere.Theproblemis thatthe

work is being done by their definition of attention. De Brigard and Prinz’s position derives a

substantialamountofitsforcefromthewaythattheydefinetheirterms.

2.3-Orienting.

IhavearguedthatDeBrigard’sandPrinz’sdefinitionofattentionallowsthemtodismissthe

evidencepresentedagainstST.However,thereisanotherroutethatDeBrigardandPrinzcouldtake,

which is to attempt to explain the data without committing themselves to the claim that GY is

attendingtosomethingofwhichheisnotconsciouswithoutrelyingontheirtendentiousdefinition

JohnH.Taylor

8

ofattention.Indeed,sucharesponsehasbeengiven(seePrinz,2011,pp.193-196and2012,pp.113-

118).PrinzsuggeststhatGY’sperformancecouldbeduetoGYorientingtothetarget,ratherthan

actuallyattendingtoit.13

Thereare twothings that Iwill saywith respect to thispossible response.The first is that

there is now reasonably good empirical evidence that the orienting response is untenable.

Unfortunately, I cannot give full details here, but the main issue is that orienting responses do

nothingtofavourtheprocessingofatargetwhichoccurswithinthesameobjectasthecue,butthis

preferential processing is just what we find in some of the experiments in question. There is a

relatedpointtobemadeherethatitisquestionablewhethertheitemsinquestioncouldtriggeran

orientingresponse,duetotheirextremelylowsalience.14

Secondly,whether theorienting response isviableornot, itwillnotaffect themainpoint

thatIwishtomake,whichisthatitdoesnotmatterexactlywhethersuchresponsesareplausible,

because De Brigard and Prinz’s position can already dismiss the empirical evidence presented

againsttheirview,simplyinvirtueoftheirdefinitionofattention.ThemainproblemIwishtoshow

isthatnomatterhowtheempiricalevidenceturnsout,DeBrigardandPrinz’sdefinitionofattention

will always be sufficiently tendentious to brush aside the GY results as not really instances of

attentionwithoutconsciousness.Thedefinitionofattentionsuppliedwillalwaysservetoprotectthe

positionfromtheresultsinquestion.

InthissectionIhavearguedthatDeBrigardandPrinz’sdefinitionofattentionmakestheir

position impervious to falsification by the results in question. In section 4 I will argue that their

definitionof attentionmakes theirposition immune from falsification fromany possibleempirical

results,butfirstwemustconsidertheevidenceputforthagainstNT.13‘Orienting’isthetermPrinzgivestothecollectionofprocesseswhichcontrolwhatinformationentersthevisualsystem(2012,pp.113-4).Notethattheresultscannotbeexplainedintermsofovertmovementsoftheeyetowardthetarget,becauseeyemovementwasfixedintheexperiment(thiswasverifiedwithaneye-tracker).Rather,theclaimisthattheneuronsthatrespondtotheareaofthevisualfieldwherethetargetwaspresentedbecomemoresensitive,andthatthisincreaseinsensitivitycanexplainsubjects’performance.14ForamoredetailedexplanationofthesedataseeNormanetal.(inpress)andMcCarleyetal.(2002).

JohnH.Taylor

9

3-SupposedevidenceagainstNT.

3.1-Theevidence.

IhavearguedthattheempiricaldataagainstSTareunabletoharmDeBrigardandPrinz’s

position,duetotheirdefinitionofattention.Ishallnowshowhowasimilarsituationariseswhenwe

areconsideringthedatathathasbeenputforwardagainstNT.

The most prominent argument against the claim that attention is necessary for

consciousness is based on empirical results by Li et al. (2002). In these experiments, subjects are

askedtoperformataskwhichrequiresalotofattention(theyhadtoworkoutwhetheracollection

ofletterscontainedan‘L’ornot)andatthesametimeastheywereperformingthistask,animage

ofsomethingwasflashedupinanareaoutsidethefocusoftheirattention.Subjectswereaskedto

releaseabuttonwhentheydetectedatargetstimulusintheperipherallocation(i.e.outsideofthe

focusof attention). Itwas found that subjects could report the gist of the image flashedupeven

though theywereconcentrating theirattentionon the letter identifying task.15Theauthorsof the

article themselves take this to be an instance of phenomenally conscious experience in the near

absenceofattention,thoughothers16haveheldittobeaninstanceofphenomenalconsciousnessin

the complete absenceofattention. It is these resultswhichhavemostprominentlybeen takenas

evidenceagainstNT.

Thecore featureabout theseexperiments thatwemust focusupon, is that inaddition to

demonstrating that attention to the peripheral stimulus was absent (or nearly absent), the

experimenters must also demonstrate that the subjects had a phenomenal experience of the

peripheralstimulus. Inordertodothis,ofcourse,theyhadtoaskthesubjectwhethertheycould

15Thiswasa‘freereport’task,makingitveryunlikelythattheresultscanbeattributedtounconsciousprimingofthesubjectsbytheperipheraltargets(cf.AzzopardiandCowey,1997).16NedBlock,inconversation.

JohnH.Taylor

10

see the stimulus. The reportability of the stimulus is taken as evidence that the subjects had a

phenomenalrepresentationofthestimulus.

3.2-Whytheevidencedoesnotdamagetheposition.

JustaswasthecasewithST,theseresultsdonotharmthepositionofPrinzandDeBrigard.

Thecrucialpointisthatthesubjectswereabletoreportthepresenceoftheimagethatwasflashed

upoutsidethe focusofattention.So itmust followthat the imagewasattendedto (becauseonly

representations available to working memory are reportable in this sense, and all and only

representationsavailabletoworkingmemoryareattended,accordingtoDeBrigardandPrinz).So,

theLietal.resultsdonotcountasaninstanceofconsciousnesswithoutattention,invirtueoftheir

definitionofattention.

Aswas the casewith ST, itmayprovepossible to accommodate thesedataby turning to

otherinterpretationsthatdonotrelyexplicitlyonDeBrigardandPrinz’sdefinitionofattention.One

option(e.g.Cohenetal.2012andCohenandDennett,2011)istoclaimthattheperipheralstimuli

mayhavebeensubject toakindof ‘distributed’ rather than focalattention,and thus theLietal.

resultsdonotcountasaninstanceofconsciousnessintheabsenceofattention.However,asbefore,

even if thiswere plausible, thiswould not affectmymain point,which is that itdoes notmatter

whether such responses are viable, because De Brigard and Prinz’s position is invulnerable to

falsificationfromtheseresults,simplyinvirtueoftheirdefinitionofattention.

4-CouldDeBrigardandPrinz’sclaimsbedisproven?

HereIshallarguethatDeBrigardandPrinz’sclaimsareimperviousnotonlytofalsification

by theavailable evidence but that it is hard to see howany evidence could possibly falsify their

position. I am not arguing that their definition is false (though I shall later argue against the

argumentgiveninfavouroftheirdefinition)Iamarguingthattheirdefinitionmakestheirposition

JohnH.Taylor

11

empiricallyself-sealing.InthissectionIshallalsoexaminearelatedissue,whichisasetofdifficulties

arisingfromPrinz’sdefinitionof‘workingmemory’.

4.1-Isthepositionempiricallyself-sealing?

Theproblemselucidatedinsections2and3aresymptomaticofadeeperproblem,whichis

thatDeBrigardandPrinz’spositionappears tobeentirely impervioustoempirical falsification.To

seethis,considerthefollowingargument:

1) Inexperimentalsettingsinvolvinghumansubjects17reportabilityisalwaysusedinorder

to establish phenomenal consciousness. That is to say, if subjects can report the

presence of a stimulus then they are taken to be conscious of it, and if they cannot

reportit,theyaretakennottobeconsciousofit.

2) InordertodisproveNTorST,wewouldrequireacasewhereattentionandphenomenal

consciousnessdissociate.

3) De Brigard and Prinz have linked their account of attention with reportability (by

definingitintermsofworkingmemory).

4) (Therefore)anythingwhichwecanestablishisconsciouswillalsocountasattendedto,

and anything that we can establish as unconscious will count as unattended to, as

definedbyDeBrigardandPrinz.

5) (Therefore)Theirpositioncannotbeempiricallydisproved.

Inordertosupportthisargument,considerwhatacounterexampletotheirpositionmight

looklike.WhatwouldberequiredtodisproveSTisasubjectwhosaidthattheycouldseesomething,

freelyreportit,anduseinformationaboutittoguidetheiractioninawilfulanddeliberatemanner,

and interact with it just as a normal human can, but who still denies having any phenomenal

consciousnessofthatthing.Thiswouldbeanexampleofsomethingwhichwasavailabletoworking

17Iincludethisqualifierinordertodistinguishbetweenconsciousnessstudiesonhumansfromstudiesonmonkeys,whichdonotutiliseverbalreports(seeLogothetisandSchall,1989).

JohnH.Taylor

12

memory(andthusattendedto)butwhichwasnotphenomenallyconscious,andwouldthusserveas

acounterexampletoST.

Unfortunately,thisclearlysetsthebartoohigh.Whatwewouldeffectivelybeaskingforis

somethingapproachingaphilosophicalzombie.Notevenanepiphenomenalistwouldbelikelyclaim

thatsuchazombiewasphysicallypossible.18Indeed,notonlywoulditclearlybeaskingtoomuchto

requestsuchextravagantempiricalevidenceagainstST,butevenifwedidhavesuchasubject,then

wewouldbefarmorelikelynottobelieveherclaimthatshewasnotphenomenallyconsciousofthe

iteminquestion.Rather,wewouldprobablyconcludethatshewaslying,ordeluded.19

Similar things go for any possible counterexamples to NT. In order to demonstrate a

counterexample to NT, what we would require for NT to be falsified is a case of a conscious

experiencewhichthesubjectinquestionwastotallyunawareof,andactuallydeniedhaving.Again,

such a case is extremely fanciful, and toomuch to hope for. It is hard to see howwe could ever

establishthatthereexistedsuchaphenomenalexperience,giventhatsubjectswouldactuallydeny

havingit.20

4.2-Unconsciousworkingmemory?

It may be argued that there already exists an empirical example of information that is

availabletoworkingmemory,butwhichisunconscious.21 Ifthisweretrue,thenitwouldserveas

anexampleofsomethingwhichisattendedto(bythedefinitionofattentionatissue)butwhichwas

unconscious, thus disproving ST. The relevant example comes from Soto et al. (2011) who

established that subjects can perform above chance at a task comparing the orientation of a

subliminallypresentedGaborpatch (ofwhich thesubjectswerenotconscious) toa supraliminally

18Isay‘physicallypossible’tosetthemodalstrengthoftheclaimasidefrom‘metaphysicalpossibility’,whichiswhatisofimportinthezombiedebates.19Cf.Dennett(1995).20ThoughseeLamme(2010).21Thankstoananonymousrefereeforemphasisingthis.

JohnH.Taylor

13

presentedGaborpatch (ofwhich thesubjectswereconscious).22Theexperimentersconclude that

the subliminally presented patch was encoded in working memory (and thus must have been

availabletoworkingmemory)butwasunconscious.Ifthisweretrue,itwouldclearlybeintolerable

toDeBrigardandPrinz’sposition.AretheSotoetal.resultsacounterexampletotheirclaim?

Theanswerisno,andthereasononceagainlieswithdefinitions.Inordertoseethis,wewill

needtostepaway fromthemain issueof thispaper,which is theuseof theterm ‘attention’and

considertheuseofsomeoftheothertermsinthedebate,specificallytheuseoftheterm‘working

memory’.AsIwillnowargue,therearesimilarissuesheretothosethatIhaveexplainedabove.

Prinz’sresponsetotheSotoetal.dataistoclaimthatthesubliminalGaborpatchwasnot

really encoded in working memory (2012, p.96). When we examine the definitions of working

memory at issue, we can see why this is. Recall that when Prinz defines working memory, he

mentionsreportabilityandexplicitlysaysthatifsomethingisunreportable,itmustbeunavailableto

workingmemory.IntheSotoetal.task,theexperimentershadtoaskwhetherthesubjectssawthe

subliminally presented Gabor patch, and rate their awareness of it on a scale (2011, R912). If

subjects reported that they had ‘no awareness’ of the subliminally presented Gabor patch, the

experimentersconcludedthatitwasunconscious.Now,theproblemwillbeobvious.Ifthesubjects

denyseeing theGaborpatch, then itwill follow fromPrinz’sdefinitionofworkingmemory (which

includes reportability) that itwasnotavailable toworkingmemory. So the subliminallypresented

Gabor patcheswill not count as an instance of unconscious representations that are available to

workingmemory,andthusnotanexampleofsomethingunconsciousbutattended,andthusnota

counterexample to ST. Notice that here,much of thework is being done by Prinz’s definition of

workingmemory,whichinturnaffectshisdefinitionofattention.

4.3-Theneuralcorrelatesofworkingmemory?

22AGaborpatchisarippledtexture,tiltedtoaspecificorientation.

JohnH.Taylor

14

Felipe De Brigard has suggested23that there is a hypothetical experiment which could

disproveDeBrigardandPrinz’sclaim.Inthishypotheticalexperiment,onegroupofsubjectswould

attendtoastimulus,and(presumably)reportawarenessofit.Thenasecondgroupwouldperform

the same task whilst undergoing some kind of brain manipulation which interrupts the neural

circuitswhichPrinzclaimsunderpinattention. If thesecondgroupstill report seeing thestimulus,

thiswillbeevidenceagainsttheposition.

However, thiswillnotdisproveNT,because thesame issueasweencounteredabovewill

reemerge,which is that if the subjects in the secondmanipulation report seeing the stimuli, then

those stimuli will automatically count as available to working memory regardless of the neural

details and thus they will count as attended to, and so the experiment will not count as a

counterexampletotheposition.Prinz’sdefinitionofattentionisfunctional,soifsubjectsfulfilthat

functionalrole(bybeingabletoreportthestimuli)thentheywillcountasattendingtothestimuli,

nomatterwhatishappeningintheirbrains.

Itmay be replied that such an experiment could disprove someaspects of Prinz’s theory,

however. Specifically, Prinz claims that what he calls or ‘gamma-locked oscillations’ or ‘gamma

synchrony’aretheneuralentitiesthatfulfiltheroleassociatedwithattention.Itmaybesaidthatif

it transpires that the subjects in the second manipulation still could report the presence of the

stimuli, despite the brain manipulation interrupting the relevant brain properties, then this

disprovesPrinz’sclaimabouttheneuralcorrelatesofattentionbeinggammasynchrony.24

However,herewemustbecareful.SuchevidencewoulddisproveapartofPrinz’s theory,

butitwouldnotdisprovetheclaimsthatwearereallyinterestedin,whichareNTandST.Toseethis,

wecandeconstructPrinz’stheoryintoseveraldifferentclaims:

i) Attentionisnecessaryandsufficientforconsciousness.

23Personalcommunication.24ThankstoJamesStazickerforsuggestingthistome.

JohnH.Taylor

15

ii) Attentionisdefinedastheprocessthatunderpinsavailabilitytoworkingmemory.

iii) Workingmemoryshouldbedefinedacertainway.

iv) Gammasynchronyrealisestheroleassociatedwithattention.

Now,theexperimentoutlinedabovemaybeabletodisprove(iv),butitcouldnotaffectthe

claims that this paper is primarily concerned with, which are (i-iii), because they can all be true

independentlyof(iv).Soifitdidindeedtranspirethatavailabilitytoworkingmemorycandissociate

fromgammasynchrony,thenallthatwouldfollowwouldbethatavailabilitytoworkingmemoryis

not in fact realised by gamma synchrony, but this would only show that we need to find some

different brain properties that fulfil the role of availability to working memory. The claim that

attention (defined functionally) should be identifiedwith availability toworkingmemory and the

relatedclaimthatNTandSTaretruewouldnotthemselvesbedamagedbythisevidence,anditis

after all these claims that we are interested in. Relatedly, all of the functional definitions of

‘attention’ and ‘working memory’ which make De Brigard and Prinz’s version of NT and ST self-

sealingwouldstillstand.

IshouldemphasisethatIamnotarguingthatallaspectsofPrinz’srichandvariedtheoryof

consciousnessareempiricallyself-sealing, IonlyclaimthathisversionofNTandSTareempirically

self-sealing,andthatthisisprimarilyduetohowhedefines‘attention’,otheraspectsofhistheory

(suchaswhichpropertiesofthebrainhethinksrealisetheroleofavailabilitytoworkingmemory)

couldprovefalse,butthiswouldnotendangerthemainthesesthatwehavebeendiscussing.

In summary, the basic problem is that Prinz does not allow attention, reportability and

availabilitytoworkingmemorytodissociate,andalsotakesreportabilityasevidenceofphenomenal

consciousness.For this reason, it isalmostanalytic toclaim that there isevidenceofphenomenal

consciousness when there is evidence of attention, and it is unsurprising that all proposed

counterevidencetotheclaimdoesnothitthemark.

JohnH.Taylor

16

5-IstheregoodreasontoacceptDeBrigardandPrinz’sdefinitionofattention?

Perhapswewouldbewillingtoaccepttheseproblemsiftherewereagoodargumentwhich

showedthatDeBrigardandPrinz’sdefinitionofattentionwasindependentlyplausible.DeBrigard

andPrinzdoputforthsuchanargument,25butPrinz(2012)hasworkeditoutinthemostdetail,soI

shall concentrate upon his formulation of it. In section 5.1 I shall argue that Prinz’s argument is

circular.Insection5.2IshallconsideranotherargumentthatPrinzgives,andargueagainstthat.

5.1-Prinz’sargument.

Prinzproposes thatwe list ‘paradigm’ instancesofattentionandthenattempt todiscover

whetherthereisacommonbrainmechanismthatunderliesthemall.Ifwefindsuchamechanism,

wecanidentifyitwithattention.Prinzthengoesontolistsomeinstancesofcaseswhereattention

seems tomake information available toworkingmemory (2012, pp.90-95). These include studies

whichlinkattentionwithshorttermmemoryretention(RockandGutman(1981))aswellasstudies

which show thatwhenworkingmemory is full, itbecomesharder toattend (FougnieandMarois,

2007).

Prinz thenmakes the following claim: ‘[s]uch interactionsbetweenattentionandworking

memorysuggestanintimaterelationship.Thesimplestexplanationforthisrelationshipisanidentity

claim: attention can be identified with the processes that allow information to be encoded in

workingmemory’(2012,p.93).Prinzthengoesontoclaimthat‘[t]heideaofavailabilityunderliesall

ofthephenomenathatwecallattention’(2012,p.95)andthat‘thefolk-psychologicalinsightimplicit

in the range of phenomena that we call attention can map onto the empirical construct of

availabilitytoworkingmemory’(2012,p.95).26

25SeeDeBrigardandPrinz(2010,pp.51-53).26Prinzseemstorelyonthefolkpsychologicalconceptofattention,andthenattemptstofindthephysicalentityinthebrainwhichfulfilstherolesassociatedwiththatconcept.Hismethodthereforehasastrongresemblancetothe‘Ramsification’methodmadefamousbyLewis(1966and1970).

JohnH.Taylor

17

The core premise of Prinz’s argument in favour of his definition is that attention and

availability to working memory always coincide in folk psychological discourse.27However, Prinz

does not consider all of the borderline caseswhere attention appears to occur in the absence of

availabilitytoworkingmemory.Forexample,inRonaldRensink’staxonomyofattention(2013),one

form of (visual) attention that is listed is what Rensink calls ‘sampling’ which is ‘the pickup of

information by the eye’. Another is ‘filtering’ which determines which information that the eye

receivesthentravels furtheralongthevisualsystemforprocessing.Bothoftheseprocessesoccur

before information becomes available toworkingmemory (Rensink himself claims that filtering is

operativeincasesofsubliminalperception(2013,§1.4ii)).

Now,herewehaveacasewhereaproposedinstanceofattentionoccursintheabsenceof

availability toworkingmemory,howcanPrinz respond to suchaproposedexample?Presumably,

Prinzwillhavetoinsistthatthesearenot‘really’instancesofattention,butwhatreason(otherthan

merelysavinghistheory)couldwehavetomakesuchapostulation?

There is the threat of circularity in the offing. In order to argue that we should define

attention in terms of availability to working memory, Prinz needs to argue that attention and

availability to working memory always coincide, but in order to do this, Prinz must claim that

processes suchas ‘filtering’ and ‘sampling’ arenot really instancesofattention (because they can

occurintheabsenceofavailabilitytoworkingmemory).Butitishardtoseewhatgoodreasonwe

might have for thinking that they are not ‘really’ instances of attention, unless we were already

convincedthatattentioncannotoccurintheabsenceofavailabilitytoworkingmemory.Butthatis

precisely the conclusion that Prinz’s argumentwas supposed to show, soPrinz’s argument comes

closetobeingcircular.

27DeBirgard(2010)arguesthatthefolkpsychologicalconceptofattentionisinfactimperfectlydelineated.ThismaybeanotherpossibleavenueofattackagainstPrinz.

JohnH.Taylor

18

OnepossibleavenuethatPrinzmaypursue inordertoarguethat ‘sampling’and‘filtering’

arenotreallyinstancesofattentionistoclaimthatwedonotthinkofsuchthingsas‘attention’in

ourfolkpsychologicaldiscourse.However,thisseemsunclear.Mostfolkpsychologicalspeakersare

unawareofthesubtledistinctionsbetweenprocessessuchas‘sampling’and‘filtering’,andsoit is

very unlikely that normal naïve subjectswill have clear cut intuitions about themonewayor the

other.Prinzwouldbeonextremelyshakygroundattemptingtoarguethatnormalnaïvefolkhada

unifiedviewonsuchmatters.

IshouldsaythatIamnotclaimingthatsuchborderlinecasescertainlydocountasinstances

ofattention,allIamsayingisthatPrinz’sargumentwillnotgothroughunlessheassumesthatthey

donot,but thereseems little reasontoassumethat theydonotunlesswearealreadyconvinced

thatPrinz’sdefinitionofattentioniscorrect,whichiswhatPrinz’sargumentwasintendedtoshowin

thefirstplace.

ButwhatoftheempiricalresultsthatPrinzcites?Dotheynotshowthatthereisanintimate

relationshipbetweenworkingmemoryandattention?Inresponsetothis,Ithinkwecanacceptthat

attention andworkingmemory often interact closely (no onewould deny this), butwe need not

committotheconclusionthatattentionmustbeidentifiedwithavailabilitytoworkingmemory.For

thesereason,IconcludethatPrinz’sargumentinfavourofhisdefinitionofattentionisinconclusive.

Afurtherissueinthevicinityisthatifwedoacceptthatattentionandavailabilitytoworking

memory shouldbe identified, thenallof the literature surrounding thequestionofhowattention

andworkingmemory interact (e.g. Fougnie, 2008)28wouldbecome trivial. Ifweare to accept the

claimthattherearesubstantivequestionsabouthowworkingmemoryandattentioninteract,and

that there are substantial empirical discoveries that can be made and have been made on this

question, then Prinz’s definition of attention will not do, because Prinz’s definition of attention

deliversasimpledefinitionalidentitybetweenthetwoprocesses.

28Watzl(2011)makesarelatedpoint.

JohnH.Taylor

19

5.2-Thethreatofeliminativism.

ThereisafinalargumentthatPrinzgives,whichmaybethoughttounderminesomeofmy

claims.Prinzsaysthis:

‘[t]heremaybeacommondenominator[whichappliestoallandonlyinstancesofattention]

thatcanbeempiricallydiscovered.Ifsuchacommonmechanismwerefound,wemightsay

that“attention” refers to thatmechanism. If thesephenomenasharenothing incommon,

thenwemightsaythat“attention”shouldbedroppedasatermfromscientificpsychology.

Wemightbecomeeliminativists’(2012,p.91).

Prinz ishereexpressingaviewthat isheldbyother thinkersonattention.Smithies (2011,

p.251)similarlyclaimsthatifwedonotfinda‘uniquelocusofattentionalselection’thenthismay

‘yielda formofeliminativism’.Equally,Allport (1993,p.203) claims that,because ‘there isnoone

uniformcomputationalfunction,ormentaloperationingeneral’thatwecanidentifywithattention,

then‘therecanbenosuchthingasattention’.

Prinzmayusethispointtoclaimthatifweweighuptheavailableoptions, it isbetterthat

we accept his own (albeit tendentious) definition of attention, rather than become eliminativists

aboutit.PerhapsPrinz’sviewisthelesseroftwoevils.

Iamunconvincedby thisappeal to thethreatofeliminativism. In responseto it, Iwill say

threethings.Firstly,IshouldpointoutthatIamnotclaimingthatthereisnocommonmechanismin

the brain that we should identify attention with, just that Prinz’s argument is insufficient to

demonstratethatattentionshouldbeidentifiedwithavailabilitytoworkingmemory.

Secondly, it is not even clear that Prinz shouldbe concerned if attention turns out to be

underwrittenbymanydifferentsystemsinthebrain.29Toseethis,notethatPrinzdefinesattention

functionally, sonot findinga common ‘mechanism’ thatunderwritesall formsofattention should

29ThankstoFelipeDeBrigardforpressingmeonthis.

JohnH.Taylor

20

notworryus.SolongasallofthesemechanismsfulfilthefunctionalrolesthatPrinzassociateswith

attention,thentheywillallcountas‘attention’,usingthefunctionalanalysisinquestion.

Thirdly, and most importantly, I think the claim that if we do not find one common

mechanism in the brain that we can identify attention with then we will be threatened with

eliminativism is very implausible. In order to see how we can avoid eliminativism, consider an

analogouscase,whichisthatofmemory.30

In psychological study of memory the concept of ‘memory’ will likely be divided into

differentsubsystems,suchasepisodicmemory,declarativememory,non-declarativememory,long

termmemory,workingmemory,iconicmemoryandsoon.31Weknowthatinsomecases,different

kinds ofmemory operate relatively independently of each other, using differentmechanisms and

operatingindifferentways.Forexample,oneformoficonicmemoryoperatesintheretina,dueto

the fact that retinalcellscontinue to firebrieflyafter theeyelidhasclosed (LongandSakitt,1980

andBlock,2011,p.571).Conversely,certainkindsoflongtermmemoryoperateinthehippocampus,

usinglongtermpotentiationofsynapsesinordertostoreinformationforrecall.

‘Memory’isthusapsychologicalcapacity,underwrittenbymanyheterogenoussubsystems,

manyofwhichdonotshare‘commonmechanisms’,butitwouldbeverystrangetoclaimthatthis

impliesthatmemorydoesnotexist,anditisnotasthough‘memory’hasbeeneliminatedfromour

psychologicaldiscourse. If itdoestranspirethatthere isnoone ‘commonmechanism’thatcanbe

identifiedwithattentionthenattentionwill stand innoworsepositionontologically thanmemory

doesnow.Itsimplydoesnotfollowfromthefactthatthereisnocommonmechanismthatwecan

identify attention with that attention does not exist. It may well be that our folk psychological

conceptofattentiondoesnotmatchupperfectlywithonemechanisminthebrain,butthisshould

30Chunetal.(2011)andDeBrigard(2012)alsolikenattentiontomemory.31Seee.g.Baddeleyetal(2009).SeealsoSligteetal.(2008and2009)forarecentbifurcationintheconceptofshorttermvisualmemory.

JohnH.Taylor

21

notsurpriseus,thatiswhatfolkpsychologicalpredicatesarelike,andwecertainlydonotneedto

leaptoeliminativismifthatisthecase.

Another point thatDe Brigard and Prinzmaymake is thatwhat I have been discussing is

merelyaverbalissue,overhowtodefine‘attention’.Itmaybeclaimedthatthethinkersinthisfield

aresimplyworkingwithdifferentideasofwhat‘attention’is,butthatthisismerelyalinguisticissue,

notasubstantiveone.

In response to this, I claim that if thedebate inquestion is tohave any real substance, if

there really is to be a definitive answer to the question of whether attention is necessary and

sufficientforconsciousness,thenobviouslywearenotfreetodefine‘attention’inanywaythatwe

choose.Ifwesimplyclaimthatthetheoristsinquestionarejustworkingwithdifferentconceptsof

attention, and that this is simply a linguistic issue, then we are dangerously close to saying that

‘really’ there is no answer to the question of whether attention is necessary and sufficient for

consciousness, because different theorists will deliver different answers depending upon their

definitionof‘attention’.Tomakethisclaimistoconcedethatthequestionofwhetherattentionis

necessaryandsufficientforconsciousness is itselfmerelyverbal,which isessentiallytogiveupon

thewholedebate.

6-Whattodo.

I havebeenelucidatinghowapparently innocuousdefinitionsof terms suchas ‘attention’

(and also ‘working memory’) can have substantial (though usually unacknowledged) force in the

debatesoverSTandNT.InthissectionIwillbrieflyoutlinetwopossibleroutesthatwemaytaketo

makeprogressontheseissues.

ThefirstwouldbetoshiftthefocusofthedebateawayfromthequestionofwhetherSTor

NT is true,andfocus insteadonthenatureofattention itself.32The ideawouldbethatweshould

32SeeWatzl(2011)forasurveyofthisissue.

JohnH.Taylor

22

reflect upon our concept of attention and attempt to come up with a unified account of what

attentionis.Onlythencanweassesswhetherattentionisnecessaryandsufficientforconsciousness.

Thisoptionmaybeviable,but it reliesuponaheavyassumption,which is thatwewillbe

abletoformulatesuchaunifyingaccount.Thereareseveralreasonsthatwemaybescepticalofthis.

Firstly,wemaynotputmuchfaithinconceptualanalysisatall(e.g.Quine,1951).Secondlywemay

think that theword ‘attention’most likely covers a heterogenous range of different phenomena,

whichmakestheprojectofobtainingaunifyingaccountofallofthemappearquixotic(cf.Duncan,

2006).33

Anotherapproach(suggestedbyDeBrigardhimself(2010,p.200))ismoreamenabletothe

approach of empirical psychology, which is that we develop operationalist definitions of the

phenomenainquestion,designedtomakeSTandNTempiricallytestable.Suchdefinitionsmaywell

betailoredtospecificexperimentalparadigms.Obviously, itwillbeofcentral importancethatone

uniform definition of the phenomena be used by different interlocutors in the debate. The

advantages of this approach inmaking the questions empirically tractable are obvious, though it

seems likely that such a definitionwill depart from a normal folk psychological understanding of

attention, in which case we will have to ask serious questions about whether philosophers and

empirical psychologists are really talking about the same thing. In any case, one thing that seems

likelyisthatDeBrigardandPrinz’sdefinitionwillnotservetheseempirically-focussedpurposes,as

theirdefinitionappearsself-sealing.

7-Summary.

IhavearguedthatDeBrigardandPrinz’sclaimthatattentionisnecessaryandsufficientfor

consciousness is virtually impervious to any empirical falsification, but that this is due to their

33Thisismyownview.

JohnH.Taylor

23

definition of attention. I have considered the reasons given for accepting their definition, and

rejectedthem.

My main concern in this paper was to highlight the difficulties which often go

unacknowledgedinthesedebates,whicharecausedbythedefinitionsofthetermsinvolved.Ihope

tohaveshownat thevery least thatweneed tobemorecarefulwhenapproaching these issues,

becauseoftheproblemsembodiedinourunderstandingoftermssuchas‘attention’andalsoterms

suchas‘workingmemory’.Onethingthatisclearisthatprogressinthisdebateseemsunlikelyifwe

donotpaymoreattentiontothedefinitionsofkeytermsinvolved.34

References.

Allport, A. (1993) Attention and Control. Have we been asking the wrong questions? Acritical review of twenty-five years. In Meyer, D. E. and Kornblum, S. (eds.) Attention andPerformanceXIV.(Cambridge,MA:MITPress.

Azzapardi,P.andCowey,A.1997.Isblindsightlikenormal,near-thresholdvision?PNAS94,pp.14190-14194.

Baddeley,A., Eysenck,M. andAnderson,M. (2009)Memory.NewYork: PsychologyPress.Reprinted2010.

Block,N.(2011)Perceptualconsciousnessoverflowscognitiveaccess,TrendsInCognitive

Sciences.15(12),pp.567-575.

Block,N.(2013)TheGrainofVisionandtheGrainofAttention.Thought:AJournalofPhilosophy.doi:10.1002/tht3.28.

Chun,M.,Golomb,J.andTurk-Browne,N.B.2011.Ataxonomyofinternalandexternalattention.AnnualReviewofPsychology.62,73-101.

Cohen,M.andDennett,D.(2011)Consciousnesscannotbeseparatedfromfunction,TrendsInCognitiveSciences,15(8),pp.358-364.

Cohen,M.A.,Cavanagh,P.,Chun,M.andNakayama,K.(2012)Theattentionalrequirementsofconsciousness,TrendsInCognitiveSciences.16(8),pp.411-417.

34ThankstoFelipeDeBrigard,E.J.Loweandananonymousrefereeforcommentsonpreviousdraftsofthepaper.ThanksalsotoNedBlockandJamesStazickerforinterestingdiscussion.SpecialthankstoBobKentridgeformanypatientdiscussionswhilethepaperwasbeingwritten.

JohnH.Taylor

24

DeBrigard,F.(2010)Consciousness,attentionandcommonsense.JournalofConsciousnessStudies.17(9-10),pp.189-201.

DeBrigard,F.(2012)Theroleofattentioninconsciousrecollection.FrontiersinPsychology.3pp.1-10.

DeBrigard,F.andPrinz,J.(2010)Attentionandconsciousness.Wileyinterdiscerplinaryreviews:Cognitivescience.1(1),pp.51-59.

Dennett,D.C.(1995)ThePathnotTaken.Reprintedin(1997)Block,N.,Gulzeldere,G.andFlanagan,O.(eds.)TheNatureofConsciousness:PhilosophicalDebates.USA:MITPress.

Duncan,John.(2006)Brainmechanismsofattention.TheQuarterlyJournalofExperimentalPsychology.59,pp.2-27.

Fougnie,D.(2008)Therelationshipbetweenattentionandworkingmemory.In.Johansen,N.B.(ed.)NewResearchonShortTermMemory.(NewYork:NovaSciencePublishers).

Fougnie,D.andMarois,R. (2007)Executive load inworkingmemory induces inattentionalblindness.PsychonomicBulletinandReview.14(1)pp.142-147.

Kentridge,R.(2011)Attentionwithoutawareness:abriefreview.InMole,C.,Wu,W.andSmithies,D.(eds.)Attention:PhilosophicalandPsychologicalEssays.NewYork:OxfordUniversityPress.

Kentridge,R.,Heywood,C.A.andWeiskrantz,L.(1999)Attentionwithoutawarenessinblindsight.ProceedingsoftheRoyalSociety(London)SeriesB.266pp.1805-1811.

Kentridge,R.,de-Wit,L.H.andHeywood,C.A.(2008a)Whatisattendedinspatialattention?JournalofConsciousnessStudies.15(4)pp.105-111.

Kentridge,R.,Nijober,T.C.W.andHeywood,C.A.(2008b)Attendedbutunseen:Visualattentionisnotsufficientforvisualawareness.Neuropsychologia.46(3)pp.831-69.

Koch,C.andTsuchiya,N.(2007)ConsciousnessandAttention:Twodistinctbrainprocesses.TrendsinCognitiveSciences.11(1),pp.16-22.

Lamme, V. (2003) Why visual attention and awareness are different, Trends in CognitiveSciences.7(1),pp.12-18.

Lamme,V.(2010)Howneurosciencewillchangeourviewonconsciousness,CognitiveNeuroscience.1(3),pp.204-240.

Lewis,D.(1966)Anargumentfortheidentitytheory,TheJournalofPhilosophy,63(1),pp.17-25.

Lewis,D.(1970)Howtodefinetheoreticalterms,TheJournalofPhilosophy,67(13),pp.427-446.

JohnH.Taylor

25

Li,F.F.,VanRullen,R.,Koch,C.andPerona,P.(2002)RapidNaturalSceneCategorizationintheNearAbsenceofAttention,ProceedingsoftheNationalAcademyofSciencesoftheUnitedStatesofAmerica.99(14),pp.9596-9601.

Logothetis,N.andSchall,J.(1989)Neuralcorrelatesofsubjectivevisualperception,Science,245(4919),pp.761-763.

Long,G.M.andSakitt,B.(1980)Theretinalbasisoficonicmemory:EriksenandCollinsRevisited,AmericanJournalofPsychology.93(2),pp.195-206.

McCarley,J.,Kramer,A.andPeterson,M.(2002).Overtandcovertobject-basedattention.PsychonomicBulletinandReview.9(4),pp.751-758.

Mole,C.(2008)Attentionandconsciousness.JournalofConsciousnessStudies.15(4),pp.86-104.

Mole,C.(2011)AttentionisCognitiveUnison:anessayinphilosophicalpsychology.NewYork:OxfordUniversityPress.

Nagel,T.(1974)WhatisitliketobeaBat?Reprintedin(2004)Heil,J.(ed.)PhilosophyofMind:

AGuideandAnthology.NewYork:OxfordUniversityPress.

Norman,L.J.,Heywood,C.A.&Kentridge,R.W.(inpress)Object-basedattentionwithoutawareness.toappearinPsychologicalScience.

Prinz,J.(2010)Whenisperceptionconscious?InNanay,B.(ed.)PerceivingtheWorld.NewYork:OxfordUniversityPress.

Prinz,J.(2011)Isattentionnecessaryandsufficientforconsciousness?InMole,C.,Wu,W.andSmithies,D.(eds.)Attention:PhilosophicalandPsychologicalEssays.NewYork:OxfordUniversityPress.

Prinz,J.(2012)TheConsciousBrain:HowAttentionEngendersExperience.NewYork:OxfordUniversityPress.

Quine,W.V.O.1951.Twodogmasofempiricism.InFromaLogicalPointofView:9Logico-PhilosophicalEssays.USA:Harvard.Reprinted1981.

Rensink,R.2013.Perceptionandattention.InReisberg,D.(ed.)TheOxfordHandbookofCognitivePsychology.

Rock,IandGutman,D.(1981)Theeffectofinattentiononformperception.JournalofExperimentalPsychology:HumanPerceptionandPerformance.7(2),pp.275-285.

SligteI.G.,ScholteH.S.,LammeV.A.F.(2008)AreThereMultipleVisualShort-TermMemoryStores?PLoSONE3(2):e1699.doi:10.1371/journal.pone.0001699[09/01/2013].

Sligte, I.G.,Scholte,S.andLamme,V.A.F. (2009)V4ActivityPredicts thestrengthofVisual

Shorttermmemoryrepresentations.TheJournalofNeuroscience.29(23),pp.7432-7438.

JohnH.Taylor

26

Smithies,D.(2011)Attentionisrationalaccess-consciousness.InMole,C.,Wu,W.andSmithies,D.(eds.)Attention:PhilosophicalandPsychologicalEssays.NewYork:OxfordUniversityPress.

Soto,D.,MäntylänT.andSilvanto,J.(2011)Workingmemorywithoutconsciousness.CurrentBiology.21(22),pp.R912-R193.

Tye,M.(2010)Attention,seeingandchangeblindness.PhilosophicalIssues.20(1),pp.410-437.

Watzl,S.(2011)Thenatureofattention,PhilosophyCompass.6(10),pp.842-853.