The Unity ofConsciousnessTim Bayne13Great Clarendon Street, Oxford ox2 6dpOxford University Press is a department of the University of Oxford.It furthers the Universitys objective of excellence in research, scholarship,and education by publishing worldwide inOxford New YorkAuckland Cape TownDar es SalaamHong Kong KarachiKuala Lumpur Madrid Melbourne Mexico City NairobiNew Delhi Shanghai Taipei TorontoWith ofces inArgentina Austria Brazil Chile Czech Republic France GreeceGuatemala Hungary Italy Japan Poland Portugal SingaporeSouth Korea Switzerland Thailand Turkey Ukraine VietnamOxford is a registered trade mark of Oxford University Pressin the UK and in certain other countriesPublished in the United Statesby Oxford University Press Inc., New York#Tim Bayne 2010The moral rights of the author have been assertedDatabase right Oxford University Press (maker)First published 2010All rights reserved. No part of this publication may be reproduced,stored in a retrieval system, or transmitted, in any form or by any means,without the prior permission in writing of Oxford University Press,or as expressly permitted by law, or under terms agreed with the appropriatereprographics rights organization. Enquiries concerning reproductionoutside the scope of the above should be sent to the Rights Department,Oxford University Press, at the address aboveYou must not circulate this book in any other binding or coverand you must impose the same condition on any acquirerBritish Library Cataloguing in Publication DataData availableLibrary of Congress Cataloging in Publication DataData availableTypeset by SPI Publisher Services, Pondicherry, IndiaPrinted in Great Britainon acid-free paper byMPG Books Group, Bodmin and Kings LynnISBN9780199215386135 7 910 8 6 4 2ContentsPreface xNote to the Reader xiiiPart I. FoundationsChapter 1. The Phenomenal Field 31.1 Consciousness 41.2 Unity relations 91.3 Conceptions of the unity of consciousness 111.4 The unity thesis 141.5 Conclusion 18Chapter 2. Phenomenal Unity: Mereology 202.1 The mereological model 202.2 Experience: the tripartite account 212.3 Objections and replies 292.4 Partial unity 362.5 Conclusion 45Chapter 3. Phenomenal Unity: Closure 473.1 Closure introduced 473.2 Inconsistency 523.3 Visual experience 583.4 Inter-sensory unity 613.5 The distinct kinds objection 643.6 Implications 693.7 Conclusion 71Part II. Is Consciousness Unied?Chapter 4. Motivating the Unity Thesis 754.1 The unity judgement 764.2 Beyond the reach of introspection 804.3 Hurleys objection 834.4 From the unity judgement to the unity thesis: I 884.5 From the unity judgement to the unity thesis: II 904.6 Conclusion 92Chapter 5. How to Evaluate the Unity Thesis 945.1 The positive moment 945.2 The negative moment: representational disunity 1055.3 The negative moment: access disunity 1075.4 Probe-dependence 1115.5 Imperial versus federal models 1155.6 Conclusion 119Chapter 6. Fragments of Consciousness? 1226.1 The microstructure of perception 1226.2 The emergence of thought 1306.3 Minimally responsive patients 1366.4 The ickering ame 1416.5 Conclusion 146Chapter 7. Anosognosia, Schizophrenia, and Multiplicity 1487.1 Anosognosia: blind to ones blindness 1487.2 Schizophrenia: an orchestra without a conductor 1567.3 Multiplicity: the delusion of separateness 1627.4 Conclusion 171Chapter 8. Hypnosis 1738.1 Thehidden observer 1738.2 The zombie model 1758.3 The two-streams model 1808.4 The switch model 1848.5 Conclusion 187Chapter 9. The Split-Brain Syndrome 1899.1 Setting the stage 1909.2 The case for disunity 1949.3 Everyday integration 1999.4 Partial unity 2059.5 The switch model 2099.6 Objections and replies 2149.7 Conclusion 220viii CONTENTSPart III. ImplicationsChapter 10. The Quilt of Consciousness 22510.1 Atomism versus holism 22510.2 Binding 22910.3 Inter-sensory integration 23310.4 Implications of the unity thesis 23610.5 Correlates, causes, and counterfactuals 23810.6 Towards a plausible holism 24410.7 Conclusion 247Chapter 11. The Body 25011.1 Bodily self-consciousness 25011.2 Egocentric space 25211.3 The fragmented body 25711.4 Multiple embodiment 26011.5 Conclusion 267Chapter 12. The Self 26912.1 Selves 26912.2 The biological self 27112.3 The psychological self 27512.4 Selves and streams 28112.5 Virtual phenomenalism 28912.6 Conclusion 293References 295Index 337CONTENTS ixPrefaceOneincursagreatnumberofdebtsinwritingabook.Eachofthefollowingindividuals provided me with crucial assistance at some point, and I am extrem-elygrateful tothemall: NedBlock, StephenBraude, AlexByrne, StephanieCarlson, Philippe Chuard, Andy Clark, Axel Cleeremans, Mike Corballis, JillianCraigie, MartinDavies, JeromeDokic, Christof Koch, UriahKriegel, DanielFarnham, Daniel Friedrich, Rami Gabriel, AlvinGoldman, JakobHohwy,Susan Hurley, Graham Jamieson, Joseph Jedwab, Mark Johnston, Sid Kouider,TimLane,NeilLevy,CalebLiang,FionaMacpherson,ChrisMaloney,FaridMasrour, Peter Menzies, David Milner, Elisabeth Pacherie, JosephPerner,HannaPickard,David Rosenthal,GaiaScerif,TobiasSchlicht,NicholasShea,Christopher Shields, Henry Shevlin, JeroenSmeets, HoustinSmit, CharlesSpence, Maja Spener, John Sutton, Frederique de Vignemont, Patrick Wilken,Ernan Zaidel and Adam Zeman.I am also grateful to a number of institutions for their support. The Centre forConsciousness Studies at theAustralianNational UniversityandtheInstitut JeanNicod, CNRS (Paris) provided nancial assistance for the research that forms thebasis of Chapters 8 and 9 respectively. Support of another kind was provided byC1, Cafe Boreal, Cafe Cubana, The Cup, Ecabur, Hernandez, The Lemon & OnionCafe,TheMissingBean,PablosVice,RagingSage andTobysEstate.Thanksforthe coffee.ThereareanumberofpeopletowhomIoweaspecial debtofgratitude.AndrewBrook, Barry Dainton, Ian Phillips, and Lizzie Schechter eachprovidedmewithaveryusefulsetofcommentsontheentiremanuscript,insomecasesatveryshortnotice.JerryLevyandColwynTrevarthenanswerednumerous questions about their split-brain research and provided the gures forChapter 9. Stuart Crutcheld and Oliver Rashbrook provided invaluableassistance as I was preparingthe volume for publication. Myeditor, PeterMomtchiloff,nudgedmejustwhenIneededit,andElmandiDuToitshep-herdedmethroughtheproductionprocess.Myparents,AndrewandBridgetBayne, taught metoreadandwrite, andmanyyears laterdissuadedmefrom opening a bagel shop. Harvey Brown invited me to lunch at Wolfson andremindedmethatbooksarenevernishedbutonlyabandoned;withoutthatremark, this book would indeed have been abandoned. David Chalmerssupervisedthedissertationoutof whichthisbookdevelopedandhasbeenatremendous source of encouragement and inspiration ever since. His inuenceon my thinking has been immeasurable.But my deepest thanks go to Nishat, who not only took the photograph thatgracesthecoverbutalsoputupwithmewhilstIwasstrugglingtowritethepages that lie behind it. I dedicate this volume to her with all my love.PREFACE xiSomeof thematerial that appears inthis bookhas appearedinpublicationelsewhere:Bayne, T.2000. TheUnityof Consciousness: ClaricationandDefense, AustralasianJournal of Philosophy, 78: 248-54.Bayne, T. 2004. Self-Consciousness and the Unity of Consciousness, The Monist, 87/2:22441.Bayne, T. 2005. DividedBrains &UniedPhenomenology: AReviewEssay onMichael Tyes Consciousness and Persons, Philosophical Psychology, 18/4: 495512.Bayne, T. and Chalmers, D. 2003. What is the Unity of Consciousness?, in A. Cleere-mans (ed.) The Unity of Consciousness. Oxford: Oxford University Press (pp. 2358).Bayne, T. andDainton, B. 2005. Consciousness as aGuidetoPersonal Persistence,Australasian Journal of Philosophy, 83/4: 54971.Bayne, T. 2007. Hypnosis andthe Unity of Consciousness, inG. Jamieson(ed.),Hypnosis and Conscious States. Oxford: Oxford University Press (pp. 93109).Bayne, T. 2007. Conscious States and Conscious Creatures: Explanation in the ScienticStudy of Consciousness, Philosophical Perspectives 21 (Philosophy of Mind): 122.Bayne, T. 2008. The Unity of Consciousness andthe Split-BrainSyndrome, TheJournal of Philosophy, 105/6: 277300.Note to the ReaderNoonehastimetoreadawholebookthesedays,especiallynotoneaslongasthis. Hereis someadvicefor thoselookingtodipintothis book. Readersinterestedintheanalysis of theunityof consciousness wouldbeadvisedtoreadChapters1through 3, inthat order. Readersinterestedinthequestionof whether consciousness remains unied in the context of particularsyndromessay, hypnosisorthesplit-brainsyndromewouldbeadvisedtoskim Chapters 1 and 5 before turning to the chapter (or part thereof ) in whichthesyndromeof interest is examined. Readers interestedintheconnectionbetweentheunityofconsciousnessandtheoriesofconsciousnessshouldreadChapters 1 and 5 before turning to Chapter 10. Finally, readers with an interestintherelationshipbetweentheunityofconsciousnessandtheselfwouldbeadvised to focus on Chapters 1 and12, but might also wish to browse Chapters 4and 11.. . . in so far as we have only one simple thought about a given object at any one time,theremustnecessarilybesomeplacewherethetwoimagescomingthroughthetwoeyes, orthetwoimpressionscomingfromasingleobjectthroughthedoubleorgansofanyothersense,cancometogetherinasingleimageorimpressionbeforereachingthesoul, sothat theydonot present toittwoobjectsinsteadof one. Wecaneasilyunderstand that these images or other impressions are unied in this gland by means ofthe spirits which ll the cavities of the brain.ReneDescartes, The Passions of the SoulP A R TI . F O U N D A T I O N S1The PhenomenalFieldConsciousness, as William James noted, is of a teeming multiplicity of objectsandrelations ( James 1890/1950: 219). James was right: consciousness doespresent us withateemingmultiplicityof objects andrelations. But as Jamesalsorecognized, inpresentingus withateemingmultiplicityof objects andrelationsconsciousnessitselfremainsunied.Thisbookisaboutthatunitythe unity of consciousness.Mytreatmentoftheunityofconsciousnessisataletoldinthreeacts.Therst act unfolds over Chapters 1 to 3 and is concerned with questions of analysis.Some theorists assert that consciousness is uniedindeed, that it is necessarilyunied.Othersholdthatalthoughconsciousnessistypicallyuniedthereareconditions inwhichthis unityis lost. Still others assert that consciousness isgenerally disunied. Withouta clearconception ofwhat itmight mean tosaythatconsciousnessisuniedwecannotbesurewhatexactlyisatstakeinthisdisputeindeed, wecannot besurethat it is substantiveratherthanmerelyverbal. In order to articulate competing conceptions of the unity of conscious-ness we need to rst identify the various unity relations that can be found withinconsciousness. That task forms the focus of this chapter. I single out phenomenalunity as the most fundamental of the various unityrelations that permeateconsciousness. Withthe notionof phenomenal unity inhandI goontoadvance ananalysis of whatit mightmean to saythat consciousness is unied.I call this claim the unity thesis. Roughly speaking, the unity thesis is the claimthatahumanbeingcanhaveonlyasinglestreamofconsciousnessatanyonepoint in time. The unity thesis forms the backbone of this project.The second act unfolds across Chapters 4 to9, and is concerned withthequestionofwhethertheunitythesisistrue.InChapter4Iprovidesomeinitial motivationfor theunitythesis, arguingthat rst-personacquaintancewithconsciousnessprovidesuswithreasontotakeitseriously.Butthisprimafacie case infavour of the unity thesis must be weighedagainst evidencefor thinking that the unity of consciousness canbreakdown. Inorder toevaluate thatevidence weneedasuitable framework.I develop suchaframe-workinChapter 5; subsequent chapters applythis frameworktoa varietyof ndingsdrawn from both normal and abnormal forms of consciousnessthat might be taken to show that consciousness is not always unied. Chapter 6examinespotential counter-examplestotheunitythesisthataredrawnfromthestudyof perceptual experience(}6.1); behavioural control ininfants andyoungchildren(}6.2); andpatients inminimallyresponsivestates (}6.3and}6.4). Chapter 7 examines the case for thinking that the unity of consciousnessmight belost intheclinical syndromes of anosognosia(}7.1), schizophrenia(}7.2),anddissociativeidentitydisorder(}7.3).AndChapters 8and9addressobjectionstotheunitythesisthatderivefromthestudyof hypnosisandthesplit-brain syndrome respectively. The conclusion of this section of the volumeis that the unity thesis remains unrefuted by these conditions.The nal act is concerned with implications of the unity thesis and tidies upsomelooseends. InChapter10Isuggestthattheunityof consciousnesshasimportantandwidelyoverlookedimplications for theorizingabout con-sciousness. If consciousness is indeed necessarily unied, then models that takespecic, ne-grainedstatesofconsciousnessasthefundamental unitsofcon-sciousness must berejectedinfavour of holisticmodels that beginwiththesubjects overall conscious state. InChapter 11I explore certainpoints ofcontact between the unity of consciousness and the experience of embodiment,and I argue against the claim that the unity of consciousness might be groundedinthephenomenologyof bodilyself-consciousness. ThevolumecloseswithChapter12, inwhichIarguethat prominentapproachestotheself must berejected on the grounds that they fail to account for the essential link betweenthe unity of consciousness and the self. In their place I sketch a view of the selfthat does do justice to this link.So,thatswherewearegoing.First,however,thecentralcharactersinourplay must be introduced.1.1 ConsciousnessThis is a book about the unity of consciousness rather than consciousness per se,but ones viewof the unity of consciousness cannot but be informed byonesviewofconsciousnessitself.InthissectionIpresenttheframeworkthatIadoptinthinkingaboutconsciousness.Thisframeworkisrestrictedtowhatwe might think of as the surface features of consciousness. Importantly, it doesnot make any assumptions about the deep nature of consciousness, such as therelationshipbetweenconsciousnessandthephysical world. Idoassumethatconscious states are grounded in neural states, but I will make few controversialassumptions about the nature of this grounding relation.4 THEPHENOMENAL FI ELDThe notion of consciousness that I am interested in here is that of phenomenalconsciousness. Phenomenal consciousness is thekindof consciousness that acreatureenjoyswhenthereissomethingthatitislikeforthatcreaturetobethecreaturethatitis(Farrell 1950;Nagel 1974).Phenomenalconsciousnessand only phenomenal consciousnessbrings with it an experiential perspectiveorpoint of view. Thereissomethingthat it islikeformetobeme, andI presumethere is something that it is like for you to be you. An accountofconsciousnessisnothingmorenorlessthananaccountofwhatisinvolvedinhaving such a point of view.Creatures,ofcourse, arenever merely conscious;instead,they areconsciousinparticularways.ManyofthesewaysinvolvewhatIwillcallspecicconsciousstates.Specicconsciousstatescanbedistinguishedfromeachotherbyrefer-encetotheirphenomenal character(orcontent). Heres adescription, as Irememberit, of someof thespecicconsciousstatesthatIenjoyedduringashort episode of consciousness some years ago:ImsittingintheCafeCubana(47RueVavin,Paris).Ihaveauditoryexperiencesofvarious kinds: I can hear the bartender making a mojito; I can hear the dog behind mechasing his tail; and theres a rumba song playing somewhere on a stereo. I am enjoyingvisual experiences of various kinds: I can see these words as they appear in mynotebook; I cansee the notebookitself; andI have a blurryvisual impressionofthose parts of the room that lie behind the notebook. Co-mingled with these auditoryand visual experiences are olfactory experiences of various kinds (I can smell somethingroasting in the kitchen); bodily sensations of various kinds (I am aware of my legs undermychair; Icanfeel myngersonthetable); andarangeof cognitiveandaffectiveexperiences.Thebartenderistalkingtoanoldwomanatthebar,andIhaveavaguesenseofunderstandingwhathessaying.Iamsoontoembarkonalengthytrip,andasenseof anticipationcolours mycurrent experiential state. Finally, Iamenjoyingconsciousthoughts.Irealizethatthebarisabouttoclose,andthatIwillbeaskedtoleaveifI stayfor muchlonger.Despiteits detailednature, this vignettebarelytouches onthetremendousvarietyof conscious statetypes that canbefoundwithinthestreamof con-sciousness. There are experiences within the ve familiar perceptual modalities.There are bodily sensations of various kinds, each with its distinctive phenom-enal character. Stayingwithinthesensoryrealm, thereareconscious statesassociatedwithimageryandmemory, not tomentionthe conscious statesassociatedwithvariouskindsof affect. Thereisthephenomenologyof bothdirected and undirected moods and emotions. Moving from the sensory to thecognitive we can identify states of fringe phenomenology, such as tip-of-the-tongue and deja` vu experiences. There are various experiential states associatedTHEPHENOMENAL FI ELD 5withagency, suchas thefeelingthat oneis incontrol of ones movements.Andthereareconsciousthoughtsof familiarkinds, suchasdesire, intention,andjudgement.Thereareclearlyavastnumberofwaysinwhichonecanbeconscious.Althoughfewwoulddisagreewiththeclaimthat consciousness manifestsitself in many forms, many would reject the claim that all forms of consciousnessare forms of phenomenal consciousness. According to a number of theorists, weneed to distinguish those conscious states that are phenomenally conscious fromthosethatareconsciousinsomeothersenseoftheterm.Theoristswhoholdthis view typically regard bodily sensations, perceptual experiences, and affect-iveexperiences(plusorminusabit)asmodesof phenomenal consciousness,but theydenythat thoughtsareeverphenomenallyconsciousassuch. Thesetheorists usually allowthat conscious thoughts are accompanied by variousphenomenally conscious statesinner speech and visualization of variouskinds, for examplebut theydenythat thoughts themselves are modes ofphenomenal consciousnessthat theyhaveadistinctiveor proprietaryphe-nomenal character in the way that bodily sensations or perceptual states do. Thisconservative view of the reach of phenomenal consciousness can be contrastedwith a liberal perspective. Liberals hold that conscious thoughts possess a whatitslikenessinpreciselythesamesenseinwhichperceptualstatesandbodilysensations do. There is, the liberal insists, something that it is like to get a joke,tobe puzzledabout a problem, andtosee that anargument is fallacious.Accordingtothisperspective,itissimplyamistaketorestrictthedomainofthe phenomenal to the sensory. Sensory states may be among the most obviousand arresting examples of phenomenal consciousness, but the sensory does notexhaust the domain of the phenomenal.1Thecontrastbetweenthesetwoviewshasavitalbearingonapproachestotheunityofconsciousness.Fromtheliberalperspectiveitisnaturaltoexpectthat therewill beasingleaccount of theunityof consciousness that mightaccommodateconsciousstatesofallstripes,althoughliberalsmightallowthatthis account will besupplemented bysubsidiaryanalyses that applyonlytorestricted classes of phenomenal states (say, perceptually conscious states).Conservatives, bycontrast, havenoreasontoexpect anysingleapproachtodothe heavylifting inaccountingfor the unityof consciousness, for theconservativedenies that conscious states formauniedcategory. Instead, theconservative is likely to expect at least two accounts of the unity of consciousness:1For conservatism see e.g. Carruthers (2005); Lormand (1996); Nelkin (1989); Prinz (2011); Robin-son (2005); Tye & Wright (2011); for liberalism see Flanagan (1992); Horgan & Tienson (2002); Kriegel(2007); Pitt (2004); Siewert (1998); Strawson (1994).6 THEPHENOMENAL FI ELDonethatappliesto phenomenallyconsciousstatesandanotherthatappliestonon-phenomenally conscious states. Indeed, conservatives are likely to also need a thirdaccountof theunityof consciousnessinordertoaccommodatetheunitythatholdsbetweenphenomenallyconsciousstatesandnon-phenomenallyconsciousstates.Thedebatebetweenconservatismandliberalismisacomplicatedone,andI will not attempt to settle it here. (Thats a job for another booksee Bayne &Montague 2011.)Instead,Iwillsimplyassumethatliberalismistrue.Inotherwords,Iwilltakephenomenalconsciousnesstobepleonastic:allconscious-ness is phenomenal consciousness. Iusetheterms conscious, experiential,andphenomenal as synonyms, andmychoiceof onetermoveranotherismade on purely stylistic grounds. Conservatives are advised to take note of thisfact, for some of the arguments in the following chaptersparticularly those inChapter 2assume liberalism.Anal dimensionof consciousness involves what arevariouslyknownasbackground states or levels of consciousness. The notion can perhaps be bestintroducedbymeansofexamples.Considerthecontrastbetweenbeingcon-sciousinthecontext of thenormal wakingstateontheonehandandbeingconsciousinthecontextofextremedrowsiness, delirium, hypnosis, orREMdreaming on the other. Over and above the ne-grained differences in consciouscontents that might accompanythis contrast, it will alsobecharacterizedbydifferences in the overall tone of ones conscious state. Ones general consciousstate will be modulated in one way in the context of normal wakefulness and avery different way in the context of (say) extreme drowsiness.Althoughwidelyrecognized,thenotionofabackgroundconsciousstatehasnot received the critical attention that it deserves, and it is something of an openquestionhowbesttoconceptualizeit.Ithinkthatsuchstatesarebestthoughtofasregionsinacomplexstate-space,theparametersofwhichdetermineboththeselectionandfunctional roles of thesubjects specicconscious states (orcontents). Consider rst the role that background states play in the selection ofspecic conscious states. Although some types of conscious states can occur withinthe context of various background statesfor example, one can have a pinprickexperience whilst normally awake, dreaming, or hypnotizedothers are restrictedtoparticularbackgroundstates. Thekindsof consciousreectionandthoughtthat arepossiblewithinthestateof normal, clear wakefulness arenot tobefoundwithinthestates of cloudedconsciousness that occur indeliriumorREM sleep. Moreover, where a specic conscious state (or content) can occurin the context of different background states it might play quite different functionalroles. The ways in which a specic conscious state can be employed in the controlof thought andactionwill dependonthebackgroundconscious stateof theTHEPHENOMENAL FI ELD 7creature in question. Subjects who are in a state of normal wakefulness might beabletoemploythecontentsof theirspecicconscious statesintheserviceofmultipleformsofcognitiveandbehavioural control,whereassubjectswhoare(say)deliriousmightbeabletoemploythecontentsoftheirspecicconsciousstates in only a restricted range of cognitive and behavioural tasks.Within the neuroscientic and clinical literatures the contrast between differentbackgroundstates of consciousness is oftenreferredtoas acontrast betweendifferent levels of consciousness. In my viewthe termlevels should be avoided,foritsuggeststhatdifferencesbetweenbackgroundstatesof consciousnesscanbe mapped onto differences in degrees of consciousness. This assumption is prob-lematic. Someone who has made the transition from the state of delirium to thenormal waking state might be conscious of more than they were before, but theyare not thereby more conscious than they were before. Inthis respect being consciousresembles being alive. In the same way that it is not possible for one creature to bemorealivethananother, sotooitisnotpossibleforonecreaturetobemoreconsciousthananother.(Thisissoeveniftherearecasesinwhichthereisnofact of the matter as to whether or not a creature is conscious.) That being said, talkof levels does point to the fact that background states can be ordered in some kindofroughhierarchyaccordingtothekindsofspecicconsciousstatesthattheyadmit and the functional roles that they support. Ordinary waking consciousnessinvolves a higher level of consciousness than that which is seen in (say) delirium,for not only does ordinary wakefulness admit forms of consciousness that do notoccur withinthecontext of delirium(suchas self-conscious thought), it alsoinvolves forms of cognitiveandbehavioural control that arenot tobefoundwithindelirium. However,inrecognizingthatthesebackgroundstatescanbeordered in this way we should avoid the temptation to think that those in the stateof ordinary wakefulness are more conscious than those who are delirious.Letusrecap.Ihavedistinguishedthreebroadaspects ofphenomenal con-sciousness. We might call the rst of these three aspects creature consciousness.As I use it, this termdoes not pick out a particular kind of consciousness that mightbecontrastedwithphenomenal consciousness, but merelyrefers tothemostgeneral(or determinable) property associated withconsciousness: the propertyof beingphenomenallyconscious. Creatureconsciousness takes twokinds ofdeterminates: backgroundconscious states andspecicconscious states. Back-ground states of consciousness can be contrasted with each other in terms of thekinds of conscious states that they admit and the functional roles that those statesplay within the subjects cognitive economy. Specic states of consciousness canbe contrasted with each other in termsoftheirphenomenalcharacter(orcon-tent)thatis,whatitisliketobeinthem.Aswewill see,inaccountingforthe unity of consciousness we will need to take all three aspects of consciousnessinto consideration.8 THEPHENOMENAL FI ELD1.2 Unity relationsI turnnowfromconsciousness per se tothe various unities that structureconsciousness.Thissectionprovidesanintroductiontothecentralunityrela-tions with which I will be concerned in this book. I will return to each of thesenotions at various points in later chapters.Subject unityThe self or subject of experience provides one form of unity to be found withinconsciousness. My conscious states possess a certain kind of unity insofar as theyareall mine; likewise, your conscious states possess that samekindof unityinsofar as they are all yours. We can describe conscious states that are had by orbelong to the same subject of experience as subject unied. Within subject unityweneedtodistinguishtheunityprovidedbythesubjectofexperienceacrosstime (diachronic unity) from that provided by the subject at a time (synchronicunity). Inkeepingwithmyfocus ontheunityof consciousness at a time,referencestosubjectunityshouldbetakenaspickingoutsynchronicsubjectunity unless otherwise specied.The notion of subject unity is a formal one, and just what the notion amountstomateriallywill dependinnosmall part onhowthesubject of experience(orself )isunderstood. Weneedtoaskwhat rolethenotionof asubject ofexperiencemightplaywithinourconceptualscheme,andwhatkindofthingmight actually play that role. There is famously little agreement with respect tothesequestions. Itwouldbeanexaggerationtoclaimthatthereareasmanyconceptions of the subject of experience as there are theorists who have writtenonthetopic, but it wouldnot bemuchof anexaggeration. Sometheoristsregardselvesasimmaterialsubstancesorsouls.Othertheoristsregardselvesasbrains or some part thereof. Still others regard selves as psychological networksthatarerealizedinorconstitutedbybrainsbutareinprincipledistinctfromthem. Yet anotherapproachconceivesof theself as akindof virtual entity,albeit one that may play an indispensable role in theoretical and practical reason.A nal approach identies selvesat least the kind of selves that we arewithorganisms.InthenalchapterIwillarguethatweoughttothinkoftheselfinvirtualterms. However, for the majority of this project I will employ an organismic (orbiological) conception of the self, according to which the self is nothing otherthan an organismin our case, the human animal. Adopting this conception ofthe self provides a particularly useful framework in which to explore questionsrelating to the unity of consciousness. Unlike other conceptions of the self, theTHEPHENOMENAL FI ELD 9biological account enables us to determine the boundaries between selves withrelative ease.Representational unityA second kind of unity to be found within consciousness concerns not the subjectof conscious states but theirobjects orintentional contents. Let us saythatconsciousstatesarerepresentationallyuniedtothedegreethattheircontentsareintegratedwith eachother.Representationalunity comes in avarietyof forms.A particularly important form of representational unity concerns the integrationof the contents of consciousness around perceptual objectswhat we might callobject unity. Perceptual features are not normally represented by isolated statesof consciousness but areboundtogether intheformof integratedperceptualobjects. This process is known as feature-binding. Feature-binding occurs not onlywithinmodalities butalso betweenthem,forweenjoymultimodalrepresenta-tions of perceptual objects. Representational unityextends beyondperceptualobjects to include the perceptual elds in which those objects are located. As Kantemphasized, oneexperiencestheobjectsof perceptionas bearingdeterminatespatial relations, bothtoeachotherandtooneself. So, representational unityinvolves multiple layers of structure: features are bound together into objects, andobjects are bound together into scenes.Representionalunityisntrestrictedtothecontentsofperceptualstatesbutcan also be found within conscious thought. A set of thoughts that is consistentwith each other is more representationally unied than a set of thoughts that isnot; anda set of thoughts that is bothconsistent andstructuredaroundacommonthemeisevenmorerepresentationallyuniedthanasetofthoughtsthat is merely consistent. Self-consciousness and various forms of metacognitivemonitoring bring with them additional opportunities for representational unityanddisunity. Wemight saythat acreatureenjoys representationallyuniedstates to the extent that it has an accurate awareness of its own rst-order statesof consciousness, and that it enjoys representationally disunied experiences tothe extent that it misrepresents its own rst-order states of consciousness.Phenomenal unitySubject unity and representational unity capture important aspects of the unityof consciousness, but they dont get to the heart of the matter. Consider againwhat its like to hear a rumba playing on the stereo whilst seeing a bartender mixamojito. Thesetwoexperiencesmightbesubjectuniedinsofarastheyarebothyours.Theymightalsoberepresentationallyunied,foronemightheartherumbaas comingfrombehindthebartender. But overandabovetheseunities is a deeper and more primitive unity: the fact that these two experiencespossessaconjoint experiential character. Thereissomethingitisliketohearthe10 THEPHENOMENAL FI ELDrumba, thereis somethingit is liketoseethebartender work, andthereissomethingit isliketoheartherumbawhileseeingthebartenderwork. Anydescription of ones overall state of consciousness that omitted the fact that theseexperiences arehadtogether as components, parts, or elements of a singleconscious state would be incomplete. Let us call this kind of unitysometimesdubbed co-consciousnessphenomenal unity.Phenomenal unity is often in the background in discussions of the stream oreld of consciousness. Thestreammetaphorisperhapsmostnaturallyasso-ciatedwiththeowof consciousnessits unitythroughtimewhereas theeldmetaphor moreaccuratelycaptures thestructureof consciousness at atime. We can say that what it is for a pair of experiences to occur within a singlephenomenal eld just is for them to enjoy a conjoint phenomenalityfor theretobesomethingit is likeforthesubject inquestionnot onlytohavebothexperiencesbut tohavethemtogether. Bycontrast, simultaneousexperiencesthat occur within distinct phenomenal elds do not share a conjoint phenome-nal character. This claim is stipulative. It is not to be taken as a substantive thesisabouttherelationshipbetweenphenomenal unityandthephenomenal eld,but as a way of eshing out the notion of phenomenal unity.Although there is no denying that phenomenal unity is a puzzling feature ofconsciousness,Itaketheexistenceofsomesuchrelationtobebeyonddoubt.The multiplicity of objects and relations that we experience at any one point intime are not experienced in isolation from each other; instead, our experiencesof themoccur as components, aspects, or elements of moreinclusivestatesofconsciousness.Itisthisfacthoweverexactlyitistobeunderstoodthatthenotionof phenomenal unityattempts tocapture. I will saymuchmoreabout phenomenal unity in the following chapters, but I trust that the foregoingsufces to provide readers with an initial grip on the notion.1.3 Conceptionsof the unity of consciousnessAlthough it is common to speak of the unity of consciousness, this locution canbemisleading. As wehavenoted, multipleunityrelations canbediscernedwithinconsciousness, anditispossibletodrawoneachof theserelationstoconstruct aconceptionof what it wouldbeforconsciousness tobeunied.Indeed, various conceptions of the unity of consciousness have been constructedfromtheserelations. Wewill get totheseconceptions of theunityof con-sciousness shortly, but let us rst reect onwhat we might want fromaconception of the unity of consciousness.THEPHENOMENAL FI ELD 11Whatwewantoratleast,whatIwantisathesisthatmightcapturethethoughtthatthesimultaneousexperiencesofasinglesubjectmustbearsomekindof unityrelationtoeachother. Wemight call suchathesis theunitythesis. Ideally, we want our unity thesis to have three properties. First, it mustbe substantive. If true it must not be a trivial truth. Secondly, it must beplausible.Thereisnopointinexploringtheviabilityofaunitythesisthathaslittlechanceof beingtrue. Thirdly, it must beinterestingit musthavethepotential tosignicantlyinformour understandingof consciousness. Let usexamine some candidate unity theses with these desiderata in mind.Some theorists conceive of the unity of consciousness in terms of therepresentational integrationof its contents. Baars adopts asomewhat dea-tionary version of this approach, claiming that the unity of consciousness is thethesis that the ow of conscious experience . . . is limited to a single internallyconsistent content at anygivenmoment (Baars 1993: 285). Drawingonamore expansive conception of the representational structure of consciousness,Shoemaker suggests that theunityof consciousness is inpart a matter ofonesvariousbeliefsforming,collectively,auniedconceptionoftheworld(Shoemaker 1996: 184). He goes on to say that perfect unity of consciousnesswouldconsist of a uniedrepresentationof theworldaccompaniedbyauniedrepresentationofthatrepresentation(1996: 186).What should we make of these conceptions of the unity of consciousness? Itis prettyclear that Baars conceptionof theunityof consciousness does notprovideuswithaplausibleunitythesis.Althoughperceptualexperiencedoesexhibitsomekindofdrivetowardsconsistency,thereisgoodreasontothinkthat asingle subject of experience can enjoyconscious states that areinconsis-tentwitheachother(see }3.2). Forexample, thevisual experiencesthatoneenjoysonlookingatthetwolinesofaMuller-Lyerillusionmightbeatoddswith ones judgements about their relative lengths. And even if it were true thatthecontents of acreatures simultaneous conscious states must beconsistentwitheachother,thisconceptionoftheunityofconsciousnesswouldlackthescope that we might want from a robust conception of the unity of conscious-ness, for at best consistency is only a necessary condition on unity.WhataboutShoemakers suggestionthattheunityofconsciousnesscanbethoughtofintermsof representational integration?Thiscertainlycapturesaviable notionof the unityof consciousness. The conscious states of somesubjects of experience will exhibit more representational integration thanothers, andit wouldbenatural todescribeasubject whoseconscious statesexhibitedahighdegreeof representational unityas havingamoreuniedconsciousness thanasubject whoseconscious states exhibitedalowdegreeofrepresentationalunity.ButalthoughShoemakersconceptionoftheunity12 THEPHENOMENAL FI ELDof consciousness might besuitablefor somepurposes, it does not suit ourpurposes forit does not captureaconceptionof theunityof consciousnessaccordingtowhichit is plausibletosupposethat thesimultaneous experi-encesof asinglesubjectarenecessarilyunied. Thisisbecauseit ispossiblefor asubject tohaveconscious states that exhibit verylittleinthewayofrepresentational integration (see }6.3). Suitably qualied, representationalunitymightprovideuswithsomekindofregulativeideal onconsciousness,but it does not provideus withaconstraint onunitythat everysubject ofexperiencemust meet.Another version of the representational approach to the unity of consciousnessappealstoself-consciousness.Rosenthalholdsthattheso-calledunityofcon-sciousness consists in the compelling sense we have that all our conscious mentalstates belongtoasingleconscious subject (2003:325). This claimsuggests acertainconceptionof what itisforconsciousnesstobeunied, accordingtowhich it is a necessary truth that any subject of experience will be aware of theirconscious states as their own conscious states.As his use of the termso-calledsuggests,Rosenthalhimselfdoesnotthinkthatconsciousnessisuniedinthissense,buthisapproachmightnonethelessprovideuswithaviableanalysisofwhat it would be for consciousness to be unied.This conception of the unity of consciousness certainly captures an importantaspect of theunityof consciousness. It gives us asubstantiveconceptionoftheunityof consciousness, for it is notrivial truththat subjects will alwaysexperience each of their conscious states as their own states. It also provides uswithaninterestingconceptionoftheunityofconsciousness,forifconscious-ness were unied in this sense then we would have an important constraint ontheories of consciousness. But does it also provide us with a conception of theunityof consciousness accordingtowhichit might beplausibletosupposethat consciousness is necessarily unied?Opinions will differonthis point. Thosetheorists whoholdthat all con-sciousstatesinvolvesomekindofself-consciousnessmaybeinclinedtothinkthat it captures a sense in which consciousness might be necessarily unied. Butthose theorists who doubt whether consciousness necessarily involves any formof self-consciousnessevenonethatmightbeimplicit orpre-reectiveare very unlikely to think that this captures a sense of the unity of consciousnessthatmightholdnecessarily. ThissuggeststhatRosenthalsconceptionof theunity of consciousness does not capture the basal phenomenon that we are after.Wewantaconceptionof theunityof consciousnessthatmightbeendorsedevenbythosewhodenythatsubjectsof experiencesareinvariablyawareoftheir conscious states as their conscious states.THEPHENOMENAL FI ELD 13It might be suggested that a plausible modication of Rosenthals conceptionof the unity of consciousness can be found by drawing on Kants famous claimthat it must be possible for the I think to accompany all my representations(1781/1787/1999: B131-2).JustwhatKantreallymeantbytheunityofcon-sciousness is notoriously obscure, but we might take him to have held that themembersofasetofconsciousstatesareuniedwitheachotherexactlywhenthe subject of those states is able to self-ascribe them.2Might this conception ofthe unity of consciousness be the one that we are after?TherstquestiontoaskisjustwhatthisKantianconceptionof theunityinvolves.Doesitrequirethatthesubjectinquestionbeabletoself-ascribeitsconscious states right here and now, or does it require only that the subject be abletoself-ascribeits conscious states under certainidealizedconditions?Does itrequirethatthoseconsciousstatesmustbeconjointlyself-ascribable,ordoesitrequireonlythattheybeindividuallyself-ascribable?Whatevertheanswerstothese questions, it seems unlikely that this conception of the unity of conscious-ness will do a better job of meeting our desiderata than Rosenthals conceptiondid. AlthoughtheKantianconceptionof theunityof consciousness is bothsubstantive and interesting, it too fails to capture a thesis that might be generallyregarded as plausible. Those who deny that conscious states are always experi-enced as the subjects own states will often also deny that subjects are invariablyin a position to self-ascribe their conscious states. Awake adults will typically beable to self-ascribe their conscious states but they might lose this capacity in thecontext of certain pathologies of consciousness. Moreover, young children andnon-linguistic animals might not have ever had this capacity to begin with. Thepotentialforself-ascriptionundoubtedlycontributestotherichsenseofcon-scious unitythat weenjoy, but it does not provideus withthebasis for aconception of the unity of consciousness that meets the constraints that I haveset out.1.4 The unity thesisTheconceptionoftheunityofconsciousnessthatIamafterhasatitsheartthenotion of phenomenal unity. Consider again the vignette of my experience in theParisian cafe (see p.5). Although it is something of an open question just how richthe streamof consciousness can be at any one point in time (see }4.1), let us suppose2For discussionof Kants conceptionof theunityof (self)-consciousness seeBlackburn(1997),Brook (1994), Keller (1998), Powell (1990), and van Cleve (1999).14 THEPHENOMENAL FI ELDthat there was a period during which I simultaneously enjoyed a visual experienceof thebartender infront of me, anauditoryexperienceof arumba, avaguesenseof embodiment, anda backgroundexperienceof dull anxiety. Eachofthese experiential elements were, I suggest, phenomenallyuniedwitheachotherwithinmyoverall phenomenal eld. Thisunitywasnot somethingthatIreectedoninenjoyingthis streamof experienceit wasnot itself anobjectof experiencebut it was nonetheless a feature of my experience. Irrespective oftheirmodality, eachof theexperiencesthatIwasenjoyingataparticulartimewere phenomenally unied with each other.Let us say that a subject has a unied consciousness if, and only if, every one oftheir conscious states at the time in question is phenomenally unied with everyotherconsciousstate. Wecanthinkof suchsubjectsasfullyunied. Whereasubject is fully unied, we can say that they enjoy a single total conscious state. Atotal conscious state is a state that is subsumed by nothing but itself, where oneconsciousstatesubsumesanotheriftheformer includesthelatterasapartorcomponent. This total state will capture what it is like to be the subject at thetimeinquestion. Inspecifyingthesubjectstotal consciousstateonetherebyprovides a full specication of the subjects specic conscious states. By contrast,if we are dealing witha creature whose consciousness is disunied, thenthere willbe no single conscious state that subsumes each of their specic conscious states.How common is it for human beings to have a unied consciousness in thissenseoftheterm?Quitecommon,itseemstome.Takeanysetofconsciousstates that youarecurrentlyenjoyingvisual experiences, auditoryexperi-ences,emotionalexperiences,bodilysensations,consciousthoughts,orwhat-ever. Irrespective of the degree to which these states might be representationallyunied with each other, they willI wagerbe mutuallyphenomenally uni-ed with each other. You might not have been aware of this unity until I drewyour attentiontoit, but havingdrawnyour attentiontoit I trust that yourecognize it in your own experience.Generalizingsomewhat, we mighthazard theguessthat unity in thissenseisa deep feature of normal waking experience. Indeed, we might even go further,and suggest that this kind of unity is not just a feature of normal waking experiencebutalsocharacterizesother kindsofbackground states ofconsciousness,such asthosethat arepresent inREMdreaming, hypnosis, andvarious pathologiesof consciousness. Onemight evenhazardthethought that this kindof unityis anessential featureof consciousnessonethat it cannot lose(at least whenit comestocreatures like us).AsSearle puts it, allof theconscious experiencesat anygivenpoint inanagents lifecomeas part of oneuniedconsciouseld (Searle2000: 562). Indeed, Kanthimself gesturesatthisideainreferringTHEPHENOMENAL FI ELD 15totheonesingleexperienceinwhichall perceptionsarerepresented (1781/1787/1999: A110).All of this leads directly to a conception of the unity of consciousness that canbe captured by appeal to the following thesis (Bayne & Chalmers 2003).Unity Thesis: Necessarily, for any conscious subject of experience (S) andanytime (t), the simultaneous conscious states that Shas at t will besubsumed by a single conscious statethe subjects total conscious state.Here, I think, we have found anacceptable characterizationof the unityofconsciousness.First,thisthesisissubstantive:iftrue,itisnottriviallytrue.Secondly, this thesis is interesting, for it offers us anexactingconstraint ontheories of consciousness. Thirdly, it is plausible: although it faces plenty of hardcases(asweshallsee),therearenoobviouslydecisivecounter-examplestoit.All up, the unity thesis provides us with just what we have been looking for.This analysis of the unity of consciousness provides us with a deeper analysisof theunityof consciousnessthanthat whichisprovidedbytherepresenta-tionally based approaches that we examined in the previous section. As we willsee in the following chapters, there are many ways in which the representationalunityofconsciousnesscanbreakdown.Patientswhosufferfromintegrativeagnosia no longer experience objects as unied and coherent wholes (see also}3.3).Patientswhosufferfromanosognosiahavedifcultytrackingthecon-tents of their own conscious states (}7.1). And patients who suffer from certaindissociative disorders seem to have lost the sense that their conscious states aretheirown(}7.3). But despitethebreakdownsinrepresentational unitythatcharacterizetheseconditions, thereis everyreasontosuspect that patientsmightnonethelessretainamoreprimitiveformofconsciousunity.Itisthiskindof unitytheunityof thephenomenal eldthat theunitythesis isconcernedwith.Although we have located a viable conception of the unity of consciousness,we cannot bring this chapter to a close just yet, for the unity thesis raises threeissues that require comment. The rst of these issues concerns the notion of thesubject of experience. It is clear that theplausibilityof theunitythesis willdependinnosmall part onjust howweconceiveof subjectsof experience.As I indicated earlier, I will work with an organismic conception of the subject,according to which subjects of experience are animals of a certain kindin ourcase, human beings. We could of course employ some other conception of thesubject of experience (see Chapter 12), but any other view would enormouslycomplicatethe task ofassessing the unitythesis. So, in the interests of keepingthediscussionmanageableIwillassumethatwecancountsubjectsofexperi-ence by counting human beings.16 THEPHENOMENAL FI ELDThe second of the three issues that needs to be addressed concerns the kind ofnecessity that governs the unity thesis. In what sense is the unity of conscious-nessanecessaryfeatureofhumanexperience?Iwillarguethatweneverhavedisuniedexperiences.Notonlydoweretainauniedconsciousnesswithinnormaleverydaycontexts,weretainthiskindofunityeveninthecontextofthe most severe impairments of consciousness. The mechanisms underpinningconsciousness functioninsuchawaythat theconscious states theygeneratealways occur as the unied components of a single phenomenal eld. However,Idonotclaimthatitisaconceptual ormetaphysical truththatourconsciousstates are always unied; indeed, I do not even claim that the unity of conscious-ness is grounded in the laws of nature. Perhaps there are surgical innovations orevolutionarydevelopments that couldbringabout adivisioninthestreamofconsciousness; perhaps there are other species in which the unity of consciousnesscan be lost. My only claim is that we have no good reason to think that any suchdivision has actually occurred in the members of our own species.The third of the three outstanding issues concerns the temporal structure ofconsciousness. The unity thesis asserts that the simultaneous conscious states of asingle subject will be phenomenally unied with each other. The idea, roughly,is that any instantaneous snapshot of a subjects experience will reveal it to befully unied (in the sense identied). Some will object to the idea that we mightbe able to take snapshots of the stream of consciousness, forthey might sayto do this would be to impose a static structure on something that is fundamen-tallydynamicandtemporallyextended. Ithinkthat this objectionmisses itstarget.Evenifconsciousnessisessentiallytemporallyextended,wecannone-theless takeasliceof thestreamof consciousness andinvestigateits internalstructure. (What we cannot assume if consciousness is essentially dynamic is thatthecontent andstructureof theslicethat weselect is independent of thecontentandstructureofthestreamfromwhichitistaken.)Takingtheunitythesis seriously doesnt presuppose a na vely static metaphysics of experience.But what sort of time-slice shouldwe take? The temporal structure ofconsciousnesshastwoaspects.Ontheonehand,consciouseventsthemselves(ortheirneuralrealizers)havelocationsinobjectivetime.Wecanaskofanyparticular conscious event when it happened. This is the temporal structure ofthe vehicles of consciousness. On the other hand, conscious events also representevents as occurringat particular times. This is thetemporal structureof thecontents of consciousness. AsDennett andKinsbourne(1992)pointedout, inprinciplethetemporal relationsbetweenthevehiclesofconsciouseventscandissociate from the temporal relations between their contents. The possibility ofthis kind of dissociation is most evident in the context of conscious thought, foritisobviousthatthetemporalcontentofathoughtcancomeapartfromtheTHEPHENOMENAL FI ELD 17temporal location of the thought itself. But it is no less important to distinguishcontentfromvehiclewhenitcomestoperceptual experience.Considertwoperceptualevents,e1ande2.Inprinciple,e1mightoccurbeforee2eveniftheintentionalobject ofe1 is represented as happening after the intentionalobjectof e2 is represented as happening. Just as there is no a priori requirement that thebrain use space to represent space, so too there is no a priori requirement that ituse time to represent time. So, we need to ask whether the unity thesis is bestunderstoodintermsofthetemporalstructureofthevehiclesofconsciousnessor in terms of the temporal structure of its content.It is clear that the motivation behind the unity thesis is not captured by an appealto the contents of conscious thought. Suppose that at 9 a.m. this morning I formedtheintentiontobooksomeairlineticketsatnoontoday,andthatwhennoonarrived I remembered this intention and as a result booked some tickets. Now, theintention to book the tickets and the actual booking of the tickets all involved areference to the same time (namely, noon), but it is clear that the conscious statesinvolvedintheseprocesses werenot phenomenallyuniedwitheachother.Furthermore, I might now remember that I booked the tickets at noon, and thismemory would not be phenomenally unied with either the conscious intentionto book themor the actual act of booking them. These states might (or might not)have occurred within the same temporally extended stream of consciousness, butthere was no single phenomenal state that subsumed them all. This suggests thatwhen it comes to conscious thought, the frame of reference that concerns us mustbethat of thestates themselves ratherthantheircontents. Andindeedthat isprecisely howthe unity thesis is formulated: the temporal framework in question isthat of clock-time, not that of the contents of experience.Althoughthedistinctionbetweenthetemporal structureof consciousnessitself and that of its content is of deep theoretical interest, it will be of little directconcerntoushere. Theremaybenoapriori requirementthatthebrainuseneural time torepresent time, but inpractice the experience of temporalrelationsistightlyconstrainedbythetemporal relationsbetweenexperiencesthemselves. In light of this, in the pages that follow I will frequently move backand forth between claims about simultaneous experiences on the one hand andclaims about the experience of simultaneous events on the other.1.5 ConclusionThere are many ways in which consciousness might be said to be uniedor,as thecasemaybe, disunied. Insomesenses of unity theclaimthat con-sciousness is necessarily unied is clearly implausible; in other senses of unity it18 THEPHENOMENAL FI ELDis well-nightrivial. Themainbusiness of this chapterhas beentoidentifyaconceptionof theunityof consciousness accordingtowhichtheclaimthatconsciousness is unied is substantive, plausible, and of some interest.I began the search for such a conception by rst considering the central unityrelations that structure consciousness: subject unity, representational unity, andphenomenal unity. Subject unity does not itself provide us with a conception ofthe unity of consciousness that we are after, for it is trivial that each of a subjectssimultaneousconsciousstateswillbesubjectuniedwitheachother.Variousformsofrepresentational unityappearedtobemorepromising,andcertainlymanytheoristshavetakensomeformoranotherof representational unitytocapturewhatitisforconsciousnesstobeunied. However,Isuggestedthatrepresentational analysesoftheunityofconsciousnessfail toprovideuswithwhat we are after. Instead, we were able to nd a viable conception of the unityof consciousness by putting together the notions of subject unity and phenom-enal unity:whatitisforasubjectsconsciousnesstobeuniedisforeachoftheir simultaneous conscious states to be phenomenally unied with each other.Puttingthesamepointindifferentterminology,itisforthesubjecttohaveasingleconsciousstateatotal consciousstatethatsubsumeseachandeveryoneoftheconsciousstatesthattheyenjoyatthetimeinquestion.Wemightidentifythistotal consciousstatewithaphenomenal eld.Andwhatitisforconsciousnessitself tobeuniedis,Isuggested,forittobethecasethatnohumansubjectcanhaveadisuniedconsciousness.Thisclaimisencapsulatedin the unity thesis.The unity thesis is the gravitational centre aroundwhichthe followingchaptersorbit,althoughsomechaptersaremoretightlyboundtothatcentrethan others. Chapters 2 and 3 focus on the notion of phenomenal unity, and aresomewhat independent of the unity thesis. Chapters 4 and 5 examine the unitythesis in the abstract: the former provides a rst-person based case for taking theunitythesisseriously, whilethelatterprovidesathird-personframeworkforevaluatingpotential counter-examples toit. Chapters 6through9applythisframework to a variety of phenomena drawn from the study of both normal andabnormal forms of experience. The remaining three chapters explore some of theimplications of the unity thesis: for theories of consciousness (Chapter 10), treat-ments of embodiment (Chapter 11), and accounts of the self (Chapter 12).THEPHENOMENAL FI ELD 192PhenomenalUnity: MereologyInthe previous chapter I introducedthe termphenomenal unity for therelationof conjoint phenomenalitythe unity relationthat permeates ex-perience and provides us with a plausible conception of what it is for conscious-ness to be unied. Discussions of the unity of consciousness often treatphenomenalunity(orco-consciousness,asitissometimes called)as aprimi-tive. That is perfectlynefor somepurposes, but athoroughtreatment oftheunityof consciousness needs togiveanaccount of phenomenal unity.Mytreatment of phenomenal unitycomes intwoinstalments. This chapterdevelopsanddefendsamereologicalaccountofphenomenalunity,accordingtowhichconsciousstatesarephenomenallyuniedinvirtueof thefactthatthey occur as the parts of a single conscious state. The following chapterexaminesthe question of whether phenomenal unitycan (also) be understoodin terms of relations between the representational contents of consciousness.2.1 The mereologicalmodelIn seeking to account for phenomenal unity it is natural to invoke the notion ofsubsumption (Bayne & Chalmers 2003). We might say that two conscious statesare phenomenally unied when, and only when, they are co-subsumed. Whatit is to experience a headache and the sound of a trumpet togetherwhat it isfor these two experiences to possess a conjoint phenomenal characteris forthere to be a single experience that in some way includes both the experience ofthe headache and that of the trumpet. Whereas treating phenomenal unity as aprimitiveprovidesuswithabottom-upapproachtotheunityofconscious-ness, one that starts with the multiplicity in consciousness, taking subsumptionas our primitive is to adopt a top-down approach to the unity of consciousnessand begin with the unity that subsumes this multiplicity.Howshouldwe thinkof subsumption? It is tempting tothinkof it inmereological termsthatis, intermsof partsandwholes. Whatitisforoneexperience to subsume another is for the former to contain the latter as a part.My total experiential state is a whole that includes within itself various experi-ential parts, suchas myoverall perceptual experience, myoverall auditoryexperience, and my experience of the diesel trucks outside my window.Ones overall phenomenal eldis anexperiencethat contains withinitselfother experiences, nestledlikeRussiandolls withineachother. Indeed, wemight venturethethought that total phenomenal states arehomoeomerous:all the parts of which they are composed share their experiential nature.Mereological languageisoftenusedinconnectionwiththeunityof con-sciousness. To take just a couple of examples, Lockwood says that experiencesareco-conscious whentheyareparts of acomplexexperience (Lockwood1989:88); Siewertsuggeststhatphenomenal unitynormallyrelatesthecon-stituentexperiencesofasinglevisualeldtooneanother, aswell asthosemaking up a single temporal stream of thought and imagery (Siewert 2001:548); while Shoemaker holds that conscious states are co-conscious when theyare parts of a unied state of consciousness (Shoemaker 2003: 59). But althoughthe mereological approachhas some claimtobe regardedas the standardconceptionof phenomenal unityit has rarelybeendevelopedinanydetail.The aim of this chapter is to do precisely that.2.2 Experience: the tripartite accountThemereological viewtreats phenomenal unityas arelationbetweentokenexperiencesthat is, betweenparticularmental states orevents.1Sometokenexperiencessuch as my headache experience and my auditory experience of thetrumpetarepartsof asinglecompositeexperienceandhencephenomenallyunied with each other; other token experiencessuch as my experience of thetrumpet andyour experienceof itarenot phenomenallyuniedwitheachother, for there is no experience that contains both of these experiences as parts.Thisaccountisatoddswithcertainconceptionsofexperiences,suchasthatwhich Tye develops under the one-experience label. According to Tye, the onlyexperiences that human beings have are entire streams of consciousness, where astream of consciousness is a period of consciousness between one state of uncon-sciousness and the next (2003: 97). The mereological model would be untenableifTyesconceptionofexperiencewereright;infact,Tyesaccountentailsthatno two experiences can be phenomenally unied with each other. Tye provides1I do not invest anything in the distinction between states and events, and use one term rather thanthe other only on grounds of stylistic convenience.PHENOMENAL UNI TY: MEREOLOGY 21both positive and negative arguments for his view. His negative arguments areintendedtounderminetheappealofsubsumptiveapproachestophenomenalunity. I examine those arguments in }2.3. Here, I focus on the positive motivationthat he provides for identifying experiences with entire streams of consciousness.2Tyes conception of experience is revisionary. It is not an analysis of our pre-theoretical notionof experience, for ordinarythought has nodifculty intakingastreamof consciousness tocontainmultipleexperiences, bothat atime and through time. This fact doesnt mean that it should be rejectedafterall, revisionary analyses canbe well motivatedbut it does meanthat hisproposal starts off on the back foot: we should identify experiences with streamsof consciousness only if there are good reasons to do so.Central toTyes casefor theone-experienceviewis a supposedparallelbetweenexperiences ontheonehandandclouds andstatues ontheother(2003:30, 99). Clouds containundetachedcollections of watermolecules asproper parts, but such collections are not (typically) themselves clouds. Similar-ly,statuesofclaymaycontainasproperpartsundetachedchunksofclay,butsuchchunksdont(typically)constitutestatuesintheirownright.Tyegrantsthatexperiencescanhaveexperiential stages,butheclaimsthatexperientialstages are no more bona de experiences than undetached cloud parts are bonade clouds or undetached parts of statues are bona de statues.Idont ndthis proposal persuasive. First, apickypoint: althoughclouds(statues) dont typicallycontainclouds (statues) as proper parts, thereis noprincipledreasontodenythattheycould.Onecanbuildastatuethatcontainsother statues as parts. But Tyeis certainlyright topoint out that wedonttypically regard undetached regions of water molecules contained within cloudsas clouds intheir ownright, nor doweregardundetachedregions of claycontained within statues as statues in their own right. So why should we regardarbitrarycomponentsofastream ofconsciousnessas experiencesintheirownright?Well,Imnotsurethatweshouldregardeveryundetachedcomponentofastream of consciousness as an experience in its own right. Consider the streamof consciousness that beganwhenyouwokeupthis morningandthat willconcludewhenyou fallintoadreamlesssleeptonight.The component of thisstreamthatoccupiedtherstwakinghourofyourdayisnotitselfanexperi-enceat least not in the sense that I am considering here. I dont deny that thereisauseof thetermexperience that canbeusedtopickout arbitrarilylargestretches of consciousness. (Tell me about your time in the Gobi desert. It was2For further discussion of Tyes position see Bayne (2005) and Dainton (2004).22 PHENOMENAL UNI TY: MEREOLOGYgreatareallyfantasticexperience.)Inthisverybroadsenseofexperiencethe chunk of my stream of consciousness that occupied the rst waking hour oftoday might indeed qualify as an experience in its own right. But this is not thenotionof experiencethat Iamafterhere. Experiences, asIaminterestedinthem here, are states that can be enjoyed all at once.Nonetheless,althoughIwouldresisttheclaimthateverypartofastreamofexperience is an experience in its own right, it does seem to me that many partsof a stream of experience qualify as experiences in their own right. My currentconscious state is an experience in its own right, as are the various ne-grainedconscious statesthe pain in my left leg; my olfactory experience of the coffee;myauditoryexperienceof thedogbarkinginthealleythat arecontainedwithin it.We should not be misled by Tyes discussion of clouds and statues, for thereareimportant differences betweenclouds andstatues ontheonehandandexperiences on the other. For one thing, clouds and statues are physical objects,whereas experiences are events. Even if it is generally true that the proper partsof an object of kind K do not constitute objects of kind K in their own right (andthereareexceptionstothisgeneralizationconsidercomputers), eventsoftencontainasproperpartseventsofthesamekind.Argumentscontainasproperpartsotherarguments,battlescontainasproperpartsotherbattles,trafcjamscontain as proper parts other trafc jams, and stories can contain as proper partsother stories. (Consider Hamlet.)Furthermore, thereisusuallylittleneedtodistinguishthepartsof acloudfromeachotherorthepartsofastatuefromeachother, butthereisindeedsomepointindistinguishingthepartsofastreamofconsciousnessfromeachother. For one thing, we need some way of referring to the distinct instances ofa particular experiential type that occur within a single stream of consciousness.Consider again the stream of consciousness that began when you woke up thismorning and will conclude when you fall into a dreamless sleep tonight. Now,supposethatthisstreamcontainsoneinstanceofheadachephenomenologyatits beginning and another towards its end. It is natural to say that this stream ofconsciousness contains twoheadache experiences. Indeed, it becomes verydifcult todescribetheinternal structureof astreamof consciousness if weinsistonidentifyingexperienceswithentirestreamsofconsciousness.Tye,ofcourse,couldsaythatsuchastreamofconsciousnesscontainstwostagesthatincludeheadachephenomenology, but at this point his experiences stageswould appear to be experiences in all but name. I suppose that one could insiston distinguishing experiences from experiences stages, but I dont see anythingto be gained from doing so. All in all, I see no reason to follow Tye in reservingPHENOMENAL UNI TY: MEREOLOGY 23the termexperience for entire streams of consciousness. In fact, on theconception of experiences that I will embrace, the typical stream of conscious-ness does not constitute an experience in its own right.Tyes one-experience conception of consciousness can be set to one side, butitdoesraisetheimportantquestionofhowweshouldthinkoftokenexperi-ences. Iamnotconvincedthatthereisanysinglewayinwhichexperiencesshouldbeindividuated.Countingexperiencesisarguablymorelikecountingthe number of objects in a room or the number of events that took place duringa meeting than it is like counting the number of beans in a dish: one has someidea of how to go about ones business, but the idea that there is only one wayin which to proceed is somewhat farcical. The notion of a token experience iselastic,anddifferentapproachestotheindividuationofexperiencesmightbeappropriate in different contexts.That said, I dothinkthat thereis oneconceptionof experiences that isparticularly well-suited for addressing the questions raisedby the unity ofconsciousness: thetripartite conception. Accordingtothetripartiteconceptionof experience, experiences are to be individuated in terms of subjects ofexperience, times, and phenomenal properties.3In other words, token experi-ences must differ from each other in terms of whose experiences they are, whenthey occur, or the kinds of phenomenal properties that they involve. AlthoughI have introduced the tripartite analysis by reference to phenomenal properties,we canalsothinkof it interms of phenomenal events, for events canbeunderstood in terms of the instantiation of properties (Kim 1976).There is a natural t between the tripartite conception of experience and themereological conceptionofphenomenal unity,forwithinthestreamofcon-sciousness wecanidentifybothmoreandless complexexperiences. Anex-perience producedby tasting a strawberry will involve a range of distinctphenomenal properties, suchas tanginess, sweetness, anddare one sayitstrawberryness. Wecanthinkof eachof thesephenomenal propertiesasinvolvingdistinctexperiencesthatarepartsofthemorecomplexexperienceoftastingthestrawberry. That experience, inturn, will beapart of evenmorecomplex experiences.Some of these more complex experiences will be modali-ty-specic, otherssuch as the experience that corresponds to ones overall per-ceptual phenomenologymight include content drawn frommultiple modalities.Indeed, ones overall phenomenal eldwhat it is like to be you right nowis, in3ThenotionoftemporallocationthatIamworkingwithhereallowsbothtemporal pointsandbrief periods of time. Isuspect that therearenoinstantaneous experiences, but Iseenoreasontoassumethat here. Bya brief periodI meansomethingthat is nolonger thana specious present(howeverlongthatis).24 PHENOMENAL UNI TY: MEREOLOGYmyview,averycomplexphenomenal eventthatcontainswithinittherestofones experiential states.According to this conception of experiences, not everything that has experi-encesaspartsisitself anexperience. Thetypical streamof consciousnessaperiod of consciousness between one state of unconsciousness and the nextisnotbestthoughtofasaphenomenalevent,forthereisnosinglephenomenalproperty that correspondsto a typicalstream of consciousness. There is some-thingit is liketoenjoyatypical streamof consciousness, but this what itslikeness is spread outdistributed across a number of distinct conscious states.It lacks the kind of unity that the phenomenal eld possesses. (The only streamof consciousness that might qualifyas aphenomenal event wouldbeaveryshort streamastreamwhosedurationwasnolongerthanasinglespeciouspresent.)Itmightbeobjectedthat thetripartiteaccountof experienceshascounter-intuitive consequencesthat it individuates experiences in ways that are signi-cantlyat odds withour ordinaryways of countingexperiences. Of course,onemight wonderjusthowmuchweightourpre-theoretical practicesoughttobeaffordedhere,butwaivingthatconcernletusexaminejustwhatkindoffriction there might be between the tripartite approach and common sense.Consider rst colour experience. Suppose that you are experiencing a certainshade of blue. According to the tripartite account, there will be an experiencecorrespondingtotheevent of your instantiatingthis phenomenal property.But, intuitively, one could have multiple experiences of blue at a single point intime, for one can see multiple objects and regions of space as being blue. Here,the tripartite account appears toentail that one wouldhave only a singleexperiencewhereas thereis somepre-theoretical forcetothethought thatone actually has multiple experiences of blue.Therearevariouswaysinwhichwemightrespondtothisobjection, butperhaps the best response is torestrict the tripartite account tomaximallyspecic or ne-grained phenomenal properties, where a phenomenal propertyis maximally specic if it has nodeterminates (Bayne &Chalmers 2003).Thephenomenal propertyof blueis adeterminablethat has as determinatesthephenomenal propertyofblueoccurringinacertainlocationof space, butarguably this phenomenal property has no determinates. And there is an intuitivesenseinwhichonecouldnt havemultipleinstantiations of that phenomenalproperty. Oncerestrictedtone-grained, maximallydeterminatephenomenalproperties the tripartite account faces no objection from this quarter.Another objection to the tripartite account derives fromconsiderationsconcerningthecommonsensiblesthat is, properties that canbedetectedvia more than one modality. Consider a fundamental common sensible:PHENOMENAL UNI TY: MEREOLOGY 25motion. Suppose that you are watching an ant crawl across your skin. You areawareofthe antsmovement intwoways:via visionandviatouch. Yettheobjection runsalthough there are intuitively two experiences of motion here,thetripartiteaccountentailsthatthereisonlyasingleexperience,forthereisonly one phenomenal property in question: a representation of motion. It looksasthoughwewill need toappealtosome furtherelement overandabovethethreecomponentsofthetripartitemodelinordertoaccountforthecontrastbetween visual experiences of motion and tactile experiences of motion.Onelineof responsetothis objectionwouldbetoarguethat thevisualexperience of motioninvolves one phenomenal propertywhile the tactileexperienceofmotioninvolvesanother.Generalizingthisresponse,wemightsay that when different senses represent the same property they will often do soviadistinctwaysthatphenomenalpropertiesaremorene-grainedthantheworldly properties that they represent. This Fregean conception of phenome-nal properties will allow that experiences can be grouped into modality-specicclusters without putting any pressure on the tripartite account.4Althoughtheapproachjustoutlinedwillhaveitsadvocates,Ipreferaratherdifferent response to it. I think the objection is best met by biting the bullet: in thesituation outlined one would have only a single experience of the ants motion. Iminclined to think of phenomenal properties in Russellian termsthat is, in termsof external world properties such as motion itself.5On this view, what it is liketo see the ant as moving at such-and-sucha speed is identical towhat its like tofeelit as moving at such-and-such a speed, for the two cases involve the representationof oneandthesamepropertynamely, movementatsuch-and-suchaspeed.On this view, the experience of the ants motion is amodal when considered inand of itself. Situations in which one is bothvisually and tactually aware of the antsmotion are not best described as situations in which one has two experiences of theantsmotion,butassituationsinwhichonehasasingleexperienceoftheantsmotion that is supported by two streams of perceptual processing (each of which isredundant given the presence of the other).Ofcourse,therearedifferencesbetweenseeinganantmoveandfeelingitmove. For onething, sight andtouchwill rarelyhavethesamedegreeofresolution. Moreover, evenif thesetwosensesweretorepresentexactlythesamespeedof theant, theywoulddosointhecontext of other modality-specic representations. For example, the visual representationof the antsmotion will occur in the context of a visual representation of its colour, whereas4See e.g. Chalmers (2004), Thompson (2009), Kulvicki (2007).5See e.g. Dretske (1995); Harman (1990); Tye (2003).26 PHENOMENAL UNI TY: MEREOLOGYthetactilerepresentationof theants motionwill occur inthecontext of arepresentationofonesownbody.InthiswaytheRusselliancanaccountforour ability to say whether an experience of an ants motion is visual, tactile, orboth visual and tactile. The upshot of the foregoing is that the tension betweenthetripartiteanalysisandourordinarywaysof countingexperiencesismoreapparent than real.Before concluding this section, let me contrast the tripartite approachwith another approach to the individuation of experiences, an approachthat is perhaps its natural rival. The approachinquestionappeals tothephysical-functional basis of experience inorder tosegment the streamofconsciousness intoparts. Toinvokeatermthat has somecurrencyinthecurrentliterature, wemighttrytoindividuateexperiencesintermsoftheirvehicles.Although there may be perspectives which demand a vehicular approach tothe individuation of experience, I think we have reason to prefer the tripartiteapproachwhenitcomestoquestionsconcernedwiththeunityofconscious-ness. Experiential states arestates of organismstheyarenot states of hemi-spheres or parts of brains. There is, of course, a sub-personal account to be givenof why a particular organism might enjoy experiential states at any one point intime,andofwhyitmightenjoytheparticularexperiential statesthatitdoes,but weought tobewaryof thesuggestionthat thesub-personal features ofconsciousness might enter intotheir identityconditions. Theunityof con-sciousness is an experiential aspect of consciousness, and our approach to it musttake this fact seriously.Asecondconcernwithvehicularapproachestoconsciousnessissomewhatpragmaticinnature. Althoughthereis muchtalkof thevehicular natureofconsciousness, there are rather fewworked-out accounts of just what thevehicles of experience are supposed to be or how one might go about pinningthemdown.Theproblemhereisnot( just)thatweknowsolittleabouttheneuro-functional basis of consciousnessalthough that certainly doesnthelpbutthattheverynotionofavehicleofconsciousnessremainsobscure.Onsome accounts the vehicles of consciousness are identical toconsciousstates, andphenomenal propertiesaresimplyentitiesthatthosestatescarry.On other accounts the vehicles of consciousness are not identical to experiencesbuttheydoconstitutethemorformtheirsuperveniencebase. Onstill otheraccounts,vehiclesareneuro-functional statesthatunderwriteacertainkindofexplanation of consciousness. Even within the explanatory conception of vehiclestherearevarious positions that might beadopted, dependingonthekindofexplanation that one is looking for (Hurley 2010). Given this confusing welter ofconceptions, it is verydifcult tosayjust what a vehicular approachtothePHENOMENAL UNI TY: MEREOLOGY 27individuation of experiences might look like. Would it individuate experiencesinne-grainedterms, say, accordingtoperceptual qualities suchas colour,shape, motion, timbre, andsoon, or wouldit individuate experiences incoarse-grainedterms, accordingto(say)perceptual objects, particularmodal-ities, or indeed entire phenomenal elds? Not only do we lack good answers tothese questions, it is rather unclear how we ought to go about determining theanswers to them.Certain readers might be inclined to respond that cognitive neuroscience hasalreadyidentiedthevehiclesof consciousness. Dontweknowsomemaysaythat V4 functions as the vehicle for the visual experience of shape(Pasupathy&Connor 1999, 2001); V4/V8functions as thevehiclefor theexperienceof colour(Bartels &Zeki 2000; Hadjikhani et al. 1998); andV4(MT/MST) functions as thevehiclefor theexperienceof motion(Tootell,Reppas, Dale, et al. 1995; Tootell, Reppas, Kwong, et al., 1995)? But whetherindeed this is something that is known depends very much on just what vehiclesaresupposedtobe. Theseregions mayindeedplayaparticularlyimportantroleinthegenerationof states withtheappropriatecontentsfor example,activation in V4 might explain why the creature in question is enjoyingexperiences of shape rather than some other type of experiencebut we shouldnotidentifysuchexperienceswithactivityinV4(seealso }10.5). Indeed, itisimplausible to suppose that V4 activity is even sufcient for visual experiences(as)ofshape.AsliceofV4thathasbeenplacedinatesttubewontgeneratevisual experiences nomatter howmuchcurrent is runthroughit. Isuspectthatevenwithacompletedcognitiveneuroscienceofconsciousness,wewillstill be left with competing ways of carving up the stream of consciousness intodistinct experiential vehicles.Let us recap. The business of counting experiences is a messy one, and thereis more than one respectable way of going about it. That being said, I suggestedthat experiences should be thought of in tripartite terms: an experience is to beunderstood in terms of the instantiation of a phenomenal property by a subjectatatime.Wecanthinkoftheseinstantiationsasphenomenalevents.Andinlight of this, phenomenal unitycanbeunderstoodinterms of mereologicalrelations between phenomenal events. At any one point in time ones stream ofconsciousness takes the form of a single highly complex phenomenal event thatsubsumes a number of less complex phenomenal events. It is the fact that theseless complex events are proper parts of more complex events that accounts fortheirunity.Themereologicalrelationsbetweenphenomenaleventsmightbereected in mereological relations between their vehicles but they need not be.28 PHENOMENAL UNI TY: MEREOLOGY2.3 Objections and repliesLet us consider now some objections to the mereological account. I begin withTyes objections totheview, beforeturningtocriticisms fromBrookandRaymont on the one hand and from Searle on the other.The self-undermining objectionA rst objection to the mereologicalaccount holdsthat it is in some way self-undermining:Consider a maximal phenomenal state(em); that is, a phenomenal statethat is notsubsumedbyanyotherphenomenal state. emincludesastwoof itselementsavisualexperience(e1)andanauditoryexperience(e2). Notonlyarethesetwoexperiencesunied with each other, but each is unied with em. Now, if the unity of each of thesetwo experiences requires that there be a unifying experience that subsumes them, thentheunityofe1(e2)andemseemstodemandthatthesubjecthaveafurtherexperience,bigger thanem, that subsumes bothemande1(e2). But westipulatedthat emwas amaximal phenomenal state, astatenotsubsumedbyanyotherphenomenal state. Soifphenomenal unity is a relationbetweenexperiences, thenthe notionof a maximalphenomenalstate is incoherent. Butthe notionof amaximalphenomenalstateclearlyiscoherent, so phenomenal unity cannot be a relation between experiences. (Tye 2003: 22)6Thenotionofamaximal(ortotal)phenomenalstateis indeedcoherent,butwhyexactlymusttheproponentofthemereological accountdenythis?Thecrucialmoveintheargumentistheclaimthat,onthesubsumptiveview,theunityofe1andemrequiresthattherebeanexperiencethatisbigger(thatis,contains more phenomenal content) than emitself. But we are under noobligationtoacceptthisclaim,forparthoodisareexiverelationanyeventis an (improper) part of itself. The state that unies em and e1 need be none otherthan em itself.Asan analogy,think ofwhatit is for twoactionsto beunied. We mightregard writing a book as an act that subsumes the act of writing the books rstsentence.Whatitisforthesetwoactionstobeuniedisforthelattertobeincluded as a component or proper part of the former. Nonetheless, we can stillthinkofwritingabookasamaximal actionthatisnotoratleastneednotbesubsumed by any more complex action. Similarly, we can think of a totalphenomenal state as a state that subsumes each of the phenomenal states that thesubject hasat the time in question without being subsumed by any stateotherthan itself.6I have substituted Tyes nomenclature for my own.PHENOMENAL UNI TY: MEREOLOGY 29The phenomenal bloat objectionAsecondobjectionholds that the mereological account is facedwiththeproblemof what we might call phenomenal bloat. Various versions of thisobjection have been circulating inthe literature (see Hurley 1998; Bayne2001), but I will focus on Tyes version of it.(1) Suppose that phenomenal unity is a relation between experiences e1e5 (assumptionfor reductio ad absurdum).(2) This unity relation (R1) between experiences must itself be experienced, for if therewere no experience of the unifying relation, then there would be nothing it is likefore1e5tobeunied.(Alternatively,therewouldbenophenomenal differencebetweenasituationinwhiche1e5werephenomenallyuniedandasituationinwhich they are not unied.)(3) If R is itself experienced, it must have its own phenomenology.(4) If Rhasitsownphenomenology, itsphenomenologymust beuniedwiththatof e1e5.(5) Inorder toaccount for thefact that R1is uniedwithe1e5weneedtopositanother unity relation(R2).(6) But of course R2 must itself be experienced, for if there were no experience of theunifyingrelation, thentherewouldbenothingit is likefore1e5tobeuniedwithR1.(7) But now we have embarked on a vicious innite regress.(8) So(1) must befalse: phenomenal unityis not a relationbetweenexperiences.(Tye2003:22)The rst thing to say about this argument is that if it is any good it seems to betoo good, for variants of it threaten to undermine any account of phenomenalunity.Ifphenomenal unityisnot arelationbetween experiences,it lookslikeitmustbearelationbetweenthecontentsofexperience,asTyehasclaimed(see }3.1). Andwhat couldmakeit thecasethat simultaneouslyexperiencedperceptual qualitiesthe loudness of a sound, the smoothness of a surface, andthesweetnessofatasteenterintothesamephenomenalcontent?Arguably,these qualities could enter into a single content only if there were an experien-tial difference between experiencing the loudness of a sound, the smoothness ofa surface, and the sweetness of a taste together as opposed to experiencing theseproperties separately. Buttheobjectioncontinuesthis surelyentails thatthe subject must be conscious that they are conscious of the sound, the surface,andthetaste, andifthatsrightthenwefacethetaskof explaininghowthatexperiential contentisuniedwiththerestof thesubjectsphenomenology.We appear to have embarked on an apparently vicious regress, not unlike the oneoutlined above. It is no accident that Hurleys (1998) version of the phenomenalbloatobjectionwhichshedubbedthejustmorecontentobjectiontargetedcontent-based accounts of the unity of consciousness.30 PHENOMENAL UNI TY: MEREOLOGYLeavingasidethequestionof whethercontent-basedconceptions of phe-nomenal unity might face their own problem of phenomenal bloat, how is themereological account to respond to it?One line of response to the argument is to reject premiss (5). We might supposethat although phenomenal unity is an experience it is a peculiar sort of experience:it is a self-binding experience. Unique amongst experiences, perhaps phenomenalunitybindsexperiencestogethertoformuniedphenomenal wholeswithoutitself needing to be so bound. However, I amnot much attracted to this objection,for why should some experiences be self-binding and others not?Another line of response would be toreject (3). One might hold thatalthoughrelations of phenomenal unityareexperienced, theexperienceofthem doesnotpossessadistinctivephenomenal characterofitsown. Christo-pherHillmayhavehadsomethinglikethisviewinmindwhenhesuggestedthat there is akind ofpure or ghostlyelementin consciousness, anelementthathasnodistinguishingcharacteristicsotherthanitsabilitytounitesensa-tions (Hill 1991: 239). However, Hillcame to reject this view, and I think hewas right to do so. For one thing, phenomenal unity lacks the kind of positivecharacter that wouldallowit tobesingledout viaintrospection. Weseemunable to attend to phenomenal unity as such. One can attend to phenomenallyuniedexperiences,andonecanattendtotheconjointexperiencethatsub-sumes them, but one cannot attend to relations of phenomenal unity orsubsumptionthemselves. Here is another way toappreciate the differencebetweenphenomenal unityandexperiences.Unlikeexperiences,relationsofphenomenal unity cannot themselves be phenomenally unied. But, since anyexperienceisthesortofthingthatcouldbephenomenallyunied,itfollowsthat relations of phenomenal unityarenot experiences. (Judgements aboutrelations of phenomenal unitymight have their ownphenomenology, butthats a different matter.)Inmyview, thearguments awlies withpremiss (2)theclaimthat inorder to make a phenomenal difference phenomenal unity must itself beexperienced. Tyeprovidesthefollowingargumentfor(2): If therewerenoexperienceoftheunifyingrelation,thentherewouldbenothingitisliketohavethesensespecicexperiencesunied (2003:22). Ithinkthisinferenceshould be resisted. Why could there not be something it is like to have a set ofuniedexperiences, without that what its like subsumingor involvinganexperience of the unity relation that binds the experiences in questiontogether? Phenomenal unity is a phenomenal relationinthe sense that itmakes a phenomenal difference, but not in the sense that it has its ownphenomenal characterthatmakesanadditional contributiontowhatitisliketobethesubject inquestion. Wecanthinkof this interms of thedifferentPHENOMENAL UNI TY: MEREOLOGY 31ways of undergoing experiences e1e5. In principle, one can have these experi-ences separately or one can have themtogether, as parts of a subsumingexperience. Unity then is not an object of experience but a manner of experien-cing.7Thelessontobelearnt fromthephenomenal bloat objectionis thatphenomenal relationscanmakeaphenomenal differencewithoutthemselvesbeing either experiences or experienced.The transparency objectionRather thanattackthe mereological account head-on, one might attemptto undermine its epistemic basis. In fact, Tye has done precisely that byappealingtotheso-calledtransparency(ordiaphanousness)of consciousness.Introspection, saysTye, givesoneaccessonlytothecontentsof experience.And this fact, says Tye, undercuts any account of phenomenal unity that treats itas a relation between token experiences:Visual experiences are transparent to their subjects. We are not introspectively aware ofour visual experiences any more than we are perceptually aware of transparent sheets ofglass.Ifwetrytofocusonourexperiences,weseerightthroughthemtotheworldoutside . . . If we are not aware of our experiences via introspection, we are not aware ofthemas unied. Theunityrelationis not giventous introspectivelyas a relationconnectingexperiences. Why, then, supposethatthereissucharelationatall?(Tye2003: 24 f.)8The so-calledtransparency of consciousness is a rather complexmatter, andappealstoitmustbehandledwithcare.9Considerrsttheboldclaimthatwearenotawareofanyofourconsciousstatesviaintrospectionbutonlyoftheircontents. If the notion of a conscious state is understood to include thoughts thenthisclaimissurelyimplausible,forthereisanintrospectibledifferencebetweenconsciously judging that it is sunny and consciously hoping that it is sunny. And ifintrospection can distinguish thoughts that differ only with respect to their modeorattitude,thenitmusthaveaccesstomorethanthecontentsofthoughts.Ofcourse, Tyes claim here is only that visual experiences are transparent to theirsubjects, but once we have allowed that introspection can distinguish betweenthoughtswiththesamecontentwehavesomereasontowonderwhetheritmightnt also allow us to distinguish between perceptual experiences with the7I am indebted to Ian Phillips here.8Tyes one-experience viewof course entails that there are no purely visual experiences, sopresumably he means visual experiential stages or some-such.9The contemporary discussion of transparency begins with Harman (1990). For critical discussion seeKind (2003), Stoljar (2004) and (2007), and Thompson (2008).32 PHENOMENAL UNI TY: MEREOLOGYsame contentindeed, whether we cant evenintrospectively distinguishbetweenvisual experiences withexactlythe same content. Contrast visualperception(as)ofanapplewithvisualimagery(as)ofanapple.Onemightarguethatitispossibletodistinguishthesetwostatesfromeachotheronthebasisofintrospectioneventhoughtheyhaveexactlythesamecontents. Advocatesoftransparencydohaveresponsestothisobjectionforexample,theycanarguethat there must be subtle content-based differences between the perceptual stateandtheimagerystateif indeedtheyare introspectivelydistinguishablebutthe plausibility of such responses is something of an open question.Despitethesereservations, Idothinkthat thereis somethingtoclaims oftransparency. We certainly do not have introspective access to the sub-personalbasis of conscious statestheir natureas neuro-functional states. However,whetherthisimpliesthatwearenotintrospectivelyawareofourexperiencesdepends on just how we conceive of experiences. Suppose, as I have suggested,thatweidentifyexperienceswiththeinstantiationofphenomenal properties.On this conception of experiences, we are aware of our experiences as such inbeingawareof whichphenomenal prop