Brains in a Vat: Different Perspectives Author(s): Yuval Steinitz Reviewed work(s): Source: The Philosophical Quarterly, Vol. 44, No. 175 (Apr., 1994), pp. 213-222 Published by: Wiley-Blackwell for The Philosophical Quarterly Stable URL: http://www.jstor.org/stable/2219742 . Accessed: 19/08/2012 03:29

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

.

Wiley-Blackwell and The Philosophical Quarterly are collaborating with JSTOR to digitize, preserve and extend access to The Philosophical Quarterly.

http://www.jstor.org ON MCDOWELL ON THE CONTENT OF PERCEPTUAL EXPERIENCE 213 in givingan expositionof one of thehorns of thedilemma. One hornof thedilemma is thatif we take consciousnessto be inseparablefrom the norm-governednotion of self-consciousnessand fromreflective and self-criticalattitudes towards oneself, then it does seem hard to accept thatDennett's framework, in whichconscious contents are a selectionof thecontents that go to and frobetween subsystems, will capture all of thephenomena; but by the same tokena frog'svisual experience will not countas contentfuland conscious,as McDowell claimsit does. But,on the otherhorn, if we shrinkthe phenomenonof consciousnessto exclude self-consciousnessand thereby includethe experienceof unself-consciouscreatures such as frogsas conscious,then it seemsto me thedual strategyI adopted - of(l) generalizingDennett's idea ofaccess relations,and (2) defusingMcDowell's epistemologicalalarms which motivatehis criticismof Dennett - providesa quiteadequate defence of Dennett against McDowell.

ColumbiaUniversit

BRAINS IN A VAT: DIFFERENT PERSPECTIVES

BY YUVAL STEINITZ

In his'Brains in a Vat" Putnamargues with great panache against ontological, external- worldscepticism. Ultimately, however, his argumentsare uncompelling.In fact,they seem disingenuousin places. This paper has two parts. In the firstI argue that even if Putnam's linguistic argumentsare enoughto convincea modernlinguistic sceptic (like Putnam himself), theyare certainlyinsufficient to win over an old-school'ontological sceptic' - that is, a scepticwho sees ontologyas a fieldof philosophymore fundamentalthan the theoryof reference(like Descartes at the outsetof his Meditations).As I shall explain, Putnam'sreliance on causal semanticsto achieve his anti-scepticalresult would beg the centralquestion against certain kinds of sceptics.In the secondpart I argue that Putnam's argumentsare not sufficientto win over his own camp of linguistic philosophers,not even when his causal semanticsis takenfor granted. His arguments, ifvalid, might undercut the possibility of what I shall call a comprehensivesceptical scenario,but theyhave littleto say about the possibilityof what I shall call partial scepticalscenarios. Before presentingmy criticism,let me brieflysummarize Putnam's 'linguistic' argumentagainst external-world scepticism. Putnam argues that sceptical propositions like'I am a brainin a vat' (henceforthBIV), thatis, a brainsubjected to a sophisticated computerwhich performs comprehensive simulation of ,are self-refuting.

H. Putnam, Reason, and History(Cambridge UP, 1981), pp. 1-21; see also the related discussion pp. 49-74. All page referencesare to this book. ? The Editorsof ThePhilosophical Quarterly, 1994. 214 YUVAL STEINITZ

His argumentis as follows:in orderfor this proposition to acquire its presumed realisticmeaning, the terms '' and 'vat' mustrefer to realbrainsand vats.However, accordingto causal semantics,a word can referto an object only if that object is somehow causallyconnected with the utteranceor thoughtof the word. Thus in order forthe word 'vat' to referto an actual vat, and not to referto a 'vat in the image' (p. 15), or be a mere clusterof random,meaningless markings, there must be some causal connectionbetween the perceptionof some actual vat and the fact thatthe word 'vat' was written,said or conceived. Returningnow to the BIV proposition:if the speakeris a genuineBIV, thenit is impossiblethat he should ever have observedthe vat he is in, or any vat, brain or computerwhatsoever. Thus the word 'vat' in his usage does not referto real vats, but only to 'vats in the image', and the entireproposition fails to refer.If, on the other hand, the speakeris not a BIV, then his propositionhas reference,but is obviouslyfalse. Hence the proposition'I am a BIV' eitherfails to referor is false. But if it failsto referto the vat containingthe brain, i.e., if it refersto mental images ratherthan to real objects,then it is also false (p. 15). For what it would (actually)tell the assumedBIV is thathe is a BIV in theimage, while what he would (actually)be is a BIV in reality.So it is necessarilyfalse. Taken together,the two partsof thispaper place Putnamin a dilemma.On the one hand,as thesceptical scenario (e.g., the BIV story)is renderedmore comprehensive, it becomes increasinglynecessary, from the sceptic'sperspective, to discard the realistic,causal theoryof referenceupon whichPutnam's anti-sceptical argument is based. On the otherhand, as the scepticalscenario is renderedmore partial, indirect causal connectionswith external reality, which are sufficientfor realistic reference, become increasinglyplausible.

I. THE ONTOLOGIST'S PERSPECTIVE

The BIV hypothesisis an updated versionof the scepticalarguments presented in Descartes' firstMeditation, that the entireworld might be only a self-madefigment of his imagination,or rathera fictiongenerated by an evil genius.The firstline of defencefor the scepticshould be to stressthe disjunctivenature of thisclaim: I cannotknow whether (a) I am a bodilyperson who perceivesindependent real- ity,or (b) everythingis onlya dreamin my . Adaptingthis formulation to the BIV hypothesis,the scepticaldisjunction can be formulatedthus:

I cannotknow whether (a) I am a bodilyperson who perceivesindependent real- ity,or (b) I am a BIV. An unprejudicedconsideration, the scepticwill argue,will find the two disjunctsto be equallyplausible. There can be no reasonfor preferring one overthe other,since any possibleexperience would be compatiblewith both. Does thisemphasis on the disjunctivenature of scepticismpose a new challenge fromPutnam's perspective?At firstglance it seems not. Putnam could elegantly ? The Editorsof ThePhilosophical Quarterly, 1994. BRAINS IN A VAT 215 tacklethe above disjunctionmuch as he respondedto the BIV propositionalone: he mightargue that,for a BIV, the disjunction'I am a bodilyperson or a BIV' fails to acquireits assumed reference and - because,for a BIV, theterms 'body', 'brain', 'computer'and 'vat' failto refer.It followsthat if the scepticaldisjunction is meaningfulin the presupposedcommon sense, i.e., if it acquires its assumed referenceto a real body, brain,computer and vat, ratherthan referringto purely mentalentities, then disjunct (a) mustbe true. But here,I contend,we can see how froman old-styleontologist's perspective, e.g., fromthe perspectiveof a Cartesianfor whom ontologyis the most fundamental philosophicalfield, Putnam's argument indeed begs thequestion. Being a philosopher of languageat heart,Putnam decides the questionof the theoryof referencebefore decidingmajor ontologicalissues, such as the existenceof an external,perceivable world.As Casati and Dokic2briefly remarked, Putnam presupposes a 'meta-language' whichis inappropriatefora real BIV, and onlythen addresses the scepticaldisjunction. It turnsout thathis apparentrefutation of the BIV propositionsrests upon the prior assumptionthat the speakeris, afterall, a bodilyperson who has a directperceptual contactwith reality, or thathe at leasttacitly believes himself to be such a person,or, alternatively,that he at least tacitlybelieves most members of his societyto be such persons.3For it is onlyby relyingon such realisticpresuppositions (or on some internal realisticpresuppositions, as argued by Dell' Utri4),that the assumedcausal theoryof referencebecomes acceptable. And suchrealistic premises do indeedbeg thequestion. Of coursePutnam might respond that, from his linguistic perspective, this 'begging of the question'is no liabilityat all, but rathera signof thevalidity of his argument. For,as Geach remarks,5itis themark of a validdeductive argument that its conclusion is impliedby its premises. But if Putnam's argument is intendedto convincea sceptical ontologist,he should recognizethat the above 'question-begging'will, for such a sceptic,entirely undercut the forceof his arguments.For the ontologistmaintains, as I shall soon explain in greaterdetail, that choosing a theoryof referenceis conditionedby one's chosen ontology,and not viceversa. Consider,for example, the above scepticaldisjunction. The Cartesianapproach sees thisdoubt about the natureof man, whichmay be compressedinto the well- knownconundrum 'What am I?', as a key issue of .I shall now argue thatthis fundamental scepticism, which Putnam claims to have renderedimpossible, 2 R. Casati andJ. Dokic, 'Brains in a Vat, Language and Meta-Language', Analysis51 (1991), pp. 91-3. 3 On the social background of Putnam's linguisticargument see Massimo Dell' Utri, 'Choosing Conceptions of Realism: the Case of the Brains in a Vat', Mind 99 (1990), pp. 79-90 (cf. p. 83). 4 Dell' Utri p. 90. His reconstructionsuggests that Putnam's argumentagainst scepti- cism presupposesinternal realism. I have opted, however,to agree withBrueckner, 'Brains in a Vat', Journalof Philosophy 83 (1986), pp. 148-67, note 2, that this argumentshould be sharplydistinguished from Putnam's internalrealism as expressed in the rest of the book (all the more so since Brueckner'sinterpretation was confirmedby Putnam himself). I was thus surprisedto discover in Realismwith a HumanFace (Harvard UP, 1992), ch. 7, that Putnam reiterateshis BIV argumentas an outcome of internal realism. I contend, however,that my argumentagainst Putnam works against both interpretations. 5 P. Geach, Reasonand Argument (Oxford UP, 1976), p. 18.

? The Editorsof ThePhilosophical Quarterly, 1994. 216 YUVAL STEINITZ

assumesthat any decisionsconcerning the righttheory of referenceare absolutely dependentupon decisions about central ontologicalissues. Therefore,from the ontologist'spoint of view,the ontologicalissues involved have clearprecedence. Putnammight still resist by claimingthat since ontological sceptics state their views and argue forthem, the questionof whetherthey assert anything - whethertheir wordshave real meaningand reference- mustin some sense be priorto whether what theysay is true. This, I believe, is partiallytrue about meaning,but it must be wrong about reference.As Quine emphasized,a reasonabletheory of referencemust allow us to say what thereis not. And since therecan be no referenceto what thereis not, anyonewho deniesthe existenceof the externalworld (or who deniesthe existence of ghosts,unicorns, etc.) mustnot be disturbedby the obvious factthat his words 'externalworld', 'ghosts' and 'unicorns'do notrefer. This seemsin linewith Putnam's own argument,for though he accuses theBIV-sceptic of being incapable of referring to the real vat which containshim accordingto his story,he does not accuse him of being incapable of referringto anythingwhose existencehe denies,e.g., his hands, his friends,etc. (and we shall see in the imagineddialogue below where it would lead him if he did). Coming back to the ontologist'spoint of view, I thinkthe best way to illustrate the significancethe above differenceof approach mightacquire is by considering Berkeley'sempirical idealism. Obviously,the theoryof referencederived from an ontologicalstance such as Berkeley'sis incompatiblewith Putnam's realistic causal theoryof reference. Indeed,both Berkeley and Putnamwere looking for a way to escape thescepticism Descartes unfoldsin the firstchapters of the Meditations.But Berkeleypreaches a differentsolution. He maintainsthat the bestmethod of overcomingscepticism is to neutralizeit by givingup realism and then re-adjustingthe theoryof reference accordingly.6He would thusclaim thatwords such as 'brain', 'vat', etc.,should not be conceivedas denotingobjective objects which are independentof any perceiver,nor as denotingsubjective objects, i.e., objectswhich owe theirexistence, viz., their internal reality,to the linguisticpractice of any particularsociety, but ratheras referringto the commonperceptions of God and man, perceptionsprojected by God onto His creatures'. If Berkeley'srebuttal of scepticism is internallycoherent, then Putnam's argument, claimingany surrenderof realismto be self-refuting,is not responsiveto it. Putnam, quite surprisingly,does not endeavour to demonstratethat idealisticor solipsistic approachescannot be accompaniedby compatibletheories of reference which make theirexternal-world scepticism self-coherent.7 These considerationssuggest that there is anotherand indeed a major difficulty withPutnam's (and Dell' Utri's)responses to traditionalscepticism: the BIVsto?y does

6 Principlesof Human , secs 87-90. 7 A transcendentalargument for that purpose is set forthby Kant, who assertsthat our concept of self-identityacquires its meaningonly against the assumed backgroundof exter- nal reality:Critique of Pure Reason B 276. This view actuallyimplies that the Cogitois already sufficientto confuteexternal-world scepticism.

? The Editorsof ThePhilosophical Quarterly, 1994. BRAINS IN A VAT 217

notconstitute an adequate version of old, external-world scepticism. Putnam claims that a real BIV is incapable of expressingscepticism because his words fail to refer.Or, as suggestedby Dell' Utri'sreconstruction, an actual BIV would not have the (internal) realisticlanguage requiredfor expressing his fearsand doubts.But what forcedoes thisclaim have against,say, an idealisticsolipsist, who claims the entireworld to exist in his mind, withoutassuming a singleexternal entity? Such a solipsistcan thinkof himself,i.e., his soul, as the only thingthat exists,without contradicting himselfby appealingto realisticlanguage. Thus Putnam'sargument against external-world scepticism (as well as Dell' Utri's versionof it) is a bit tricky:the representativesceptic is held to doubt everything outsidehis mind,except his brain,vat and computer,which do lie outsidehis mind. He is thenfound incapable of formingdirect or indirectperceptual connections with thoseisolated islands of externalreality left behind. And thisrenders his scepticism self-refuting. This shows that Putnam's refutationof external-worldscepticism is heavily dependentupon, and hence limitedto, a speciallytooled scepticalhypothesis, in whichthere is an externaldeceiving agent. Where this deceiving agent is furtherassumed to be of a materialisticnature, e.g., thevat-computer system, Putnam can arguethat it has to be perceptuallyconnected to the sceptic,in orderto be referredto. Now, in orderto see moreclearly why Putnam's argument is entirelyineffective against other, more conclusiveexpressions of external-worldor materialscepticism, consider the resultswere he to tryto apply it to Berkeley'sidealism: Berkeley:All thereis are souls and ideas. Putnam:This is self-refuting.Indeed, any external-worldscepticism is self-refuting, as I have demonstratedin the case of BIV, forif it is true,then the sceptic's words and ideas fail to referto those material,external objects which constitutehis deludingenvironment. So ifwe are brainsin a vat,we cannot say or thinkthat we are. Berkeley:But mystory is different!In mystory there are no material,external objects, and hence no causal connectionwith such objects is necessaryfor their description. [Putnamwould do well to end the discussionat thisjuncture. For were he to press his point,he would be forcedto argue thatthe existenceof an external,objective referentis a necessarycondition not just forreference but formeaningfulness too. He would have to do thisin orderto extendhis linguisticargument to Berkeley's idealism,that is, in orderto doubtthe meaningfulnessof Berkeley'sentirely negative existentialpropositions about an objective,external reality. But such a view has unwelcomeontological consequences.] Putnam:This is wrongtoo. True, yourstory does not containany external,material objects. But have you not just spoken about such objects,while denying theirvery existence? Thus, accordingto mycausal theoryof meaning, there mustbe somecausal, perceptualconnection between your words or ideas of materialobjects and at least one exemplarof such an object,in orderfor ? The Editorsof ThePhilosophical Qyarterly, 1994. 218 YUVAL STEINITZ

yourwords to acquire theirprevalent meanings, though you believe them to have no referenceat all. So eitherthere are materialobjects or thereare not. If thereare, your external-worldscepticism is false.If thereare not, thenyou could not say and mean thatthere are not. Berkeley:I now see the true nature of your argument.Your demand for a causal connectionwith the object we speakabout, or at leastfor a causal connection withother objects of the same kind,is none otherthan a disguisedversion of the ancientParmenidean doctrine that 'It cannotbe said thatanything is not'.8Indeed, you ostensiblyalter it by claimingthat 'It cannot be said that anythingis causally unconnected'.But since therecan be no causal connectionwith what thereis not, your demand boils down to the same ontologicalconsequences, that is, theinflated ontology of objects(or at least kindsof objects)that issues from the old riddleof non-being.9 In the last few pages we have seen why Putnam's anti-scepticalargument is ineffectiveagainst widely comprehensive sceptical scenarios which permit no external reality(), or at least no materialreality (idealism). Let us summarizethe discussionso farby stressingtwo major points: 1. Putnamargues that sceptical assertions such as 'Realityis an all-encompassing dream',or 'I am a BIV', are self-refuting,and triumphantlyclaims that he has revealed a substantialinconsistency in ontologicalscepticism. But I havesuggested that Putnam's satisfactionwould be warrantedonly if his realisticcausal theoryof referencewere truecome what may. I have also demonstratedthat Putnam might find it quite difficult to forcea stubbornontological sceptic to accept his causal theoryof referencein the firstplace. For the ontologicaldoubts raised by such a scepticare not restrictedto ontology.Rather, they spread to the prevalentcausal semanticsas well. Note, then,that the contradiction that Putnam claimed to have revealed,between thecausal theoryof referenceon the one hand, and some major scepticalarguments on the other,would not immediatelywin over a trueontological sceptic. In fact,this contradictionmight only serve to emphasize how extensivethe ramificationsof traditionalontological scepticism really are. 2. Insteadof being true come what may, and insteadofbeing an elementaryhypothesis or an axiom about the theoryof reference,Putnam's causal theoryof referenceis clearlydependent upon an implicitontological stance: either upon externalrealism, as suggestedby Brueckner,or upon internalrealism, as suggestedby Dell' Utri. Well,why should not Putnamadopt one of these'conceptions of realism' (as Dell' Utriputs it) as his premise?Take, forinstance, internal realism: is it not possible,at least in principle,that thistheory is indeed necessaryfor thinking and referring,as Putnamand Dell' Utri suggest?'0

8 See M. Furth,'Elements of Eleatic Ontology',Journal of the History of Philosophy 6 (1968), p. 116. 9 These unwelcomeconsequences of Putnam'sargument were first pointed out in Philosophia 16 (1986) by P. Tichy, 'Putnam on Brainsin a Vat', pp. 144-5, and by F.B. Farrell,'Putnam and the Vat-People', p. 154. '0 Putnam ch. 3; Dell' Utri sec. V.

? The Editorsof ThePhilosophical Quarterly, 1994. BRAINS IN A VAT 219

This is not to thepoint here. Assuming that external realism and internalrealism are false,then Putnam's causal semanticsgoes too. In that case, Putnam's anti- scepticalargument, and Dell' Utri's reconstructionof it, go as well. Assuming,on the otherhand, thatone of theseontological stances happens to be true,then their argumentsagainst scepticism are valid but useless.For ifexternal or internalrealism is true,then all formsof external-worldscepticism, e.g., the BIV story,Solipsism, Idealism,the Evil Genius story,etc., are instantlyfalsfied by this very assumption. The questionhas been explicitlybegged at the outset.There is no need for Putnam's BIV referencestory.

II. A BIV PERSPECTIVE

In the firstpart of thispaper I showedthat Putnam's argument is ineffectiveagainst old-styleontological scepticism. This is because scepticismof thissort might become so comprehensivethat it must influence the theory of reference preferred by the sceptic. In the followingpages I shall evaluatethe forceof Putnam'sarguments against a more moderatesceptic, who considersa singleBIV placed in a real, rich material surrounding,and who happensto takefor granted Putnam's causal theory of reference. Must he accept,under those assumptions, Putnam's solution of BIV-scepticism? Putnampresents the BIV hypothesisas a storyabout a mad scientistwho performs the experiment.The followingstory provides a reasonablesequel, in the sense that it does not contradictany major physicalprinciples or the laws of logic. Its purpose is to showthat the hermetic separation between a BIV, on theone hand,and objective reality,on the other,might not be as air-tightas Putnammaintains. It would follow thateven if a scepticalBIV wereto acceptPutnam's realistic causal theoryof reference (as no CartesianBIV would),still there might be circumstancesin whichhis assertion 'I am a BIV' would acquire itsassumed realistic reference, and hence be meaningful and succeed in referringin the ordinarysense. Imagine a brain in a vat named Alex. Apart fromthe unfortunatefact that his nerve-endingsare connectedto a computer,Alex is a normalperson in all other respects.Like mostpeople, Alex does not belongto the categoryof thosewho waste theirtime on philosophicalmeditations. Doubts such as 'Am I a BIV?' have never enteredhis mind. One day it occurs to the mad scientistto documenthis sensationalexperiment. He buysa video camera,films the BIV systemand the computercontrolling it, and thatvery evening screens the film for his friends.Suddenly he realizesthat his circle of friendsincludes, in some weirdsense, Alex as well. It would be unfair,he thinks, to preventAlex fromseeing the film: 'After all, haven'tI hurthim enough by putting him into the vat?' But how could Alex watchthe film? And how could he be preventedfrom realizing thathe himselfis thefilm's hero, which might jeopardize some psychologicalaspects of the experiment? In orderto accomplishthis, the mad scientistdecides thatit would be best ifthe computermade Alex meethim somewhere, after which he would invitehim to watch the filmat his home. Since the filmshows nothingabout the way the brain was

? The Editorsof ThePhilosophical Quarterly, 1994. 220 YUVAL STEINITZ removedfrom its body and put into the vat, thereis no reason forAlex to realize thathe is lookingat himself. And so it was. The computerperformed a perfectsimulation of a partyat which the mad scientistinvited Alex to his home. There he showedhim the filmabout the BIV. (Obviously our amazed hero watched the film through the electrodes permanentlyattached to his visual nerves,just as he saw all the otherthings that constitutedhis world.) Nevertheless,a few days later it occurredto Alex forthe firsttime in his lifeto raise a scepticalquestion: 'Might I be a BIV myself?'. Afterturning this doubt over in his mind day and night,he finallywent to the library(obviously a simulatedlibrary) to read an essayentitled 'Brains in a Vat'. On page 50 he came upon the followingconclusion: 'So, ifwe are Brainsin a Vat, we cannotthink that we are'. 'That poor brain I saw in the film',thought Alex. 'Even ifit had suspectedits truepredicament, that very thought on itspart would be self- refuting.'And so he sighedin relief. The moral: Alex consideredthe possibilityof being a BIV, and conceived the concepts'brain' and 'vat' in hismind, as a resultof some causal, perceptual connection withthe real vat containinghim. Hence his wordsand doubtsseem to acquire their assumed referenceand meaning. Does the factthat his observationtook place by means of a filmprojected from a computermemory, and transmittedto hisbrain via electrodes,prevent the fulfilment of the requiredcausal connection? Putnam,quite surprisingly, claims that it does. Here he sharplydiffers from Kripke, anotherproponent of the causal theoryof reference,who arguesthat even the most indirectand heavilymediated causal connectionis sufficientfor the establishment of fullreferential relations between the conceptsor names and the objects,through what he calls 'a causal (historical)chain'." Accordingto Kripke,a blind man can referto light,even thoughhe has neverseen it, or anythingsimilar to it. Thus, if Alex opted for the Kripkean interpretationof the causal theoryof reference,he would be 'allowed' to referto vats and computers,even thoughhe grasped their conceptsthrough discourse with a computer,on the condition,of course,that those conceptswere programmedinto the computerby someone whose conceptsof vat and computerdo refer. Putnam(pp. 11-12), in contrast,dismisses the referentialvalue of such veryweak causal connections:12

Althoughthe machine does not perceiveapples, fields,or steeples,its creator- designersdid. There is somecausal connectionbetween the machineand the real- world apples, etc., via the perceptualexperience and knowledgeof the creator- designers.But such a weak connectioncan hardlysuffice for reference.

" S. Kripke, Namingand Necessity(Oxford: Basil Blackwell, 1980), p. 139. 12 Putnam's position here concerningthe insufficiencyof such weak connections con- trasts not only with Kripke's, but also with his own earlier writingon this topic: cf. 'Explanation and Reference', reprintedin his Mind, Languageand Reality(Cambridge UP, 1975), p. 202.

? The Editorsof ThePhilosophical Quarterly, 1994. BRAINS IN A VAT 221

And he justifiesthis with the followingargument: The machineis utterlyinsensitive to thecontinued existence of apples, fields, steeples, etc. Even ifall thesethings ceased to exist,the machinewould stilldiscourse just as happilyin the same way. That is whythe machinecannot be regardedas refer- ringat all. But thisview is indeed veryodd. It impliesthat when we read in a book about penguinsin Antarctica,we should claim thatthe word penguincannot referto real penguins,as thewritten word and the book are no less insensitivethan the computer to the continuedexistence of penguins.Is the assumptionthat this book might continueto be read 'just as happilyin the same way', even if penguins'ceased to exist',enough to disqualifythe book's referentialvalue? And ifso, thenwhat would be the referenceof the entry 'penguin' in the insensitive1930 edition of the EncyclopaediaBritannica? Putnammight argue thatthe book could in principlebe re-editedand updated, and thatthe 1930 editioncould be replacedby the 2030 edition.But then,in exactly the same way, the mad scientistconducting the BIV experimentcould update the computermemory concerning the extinctions,should theyunfortunately occur, of penguins,apples, or steeples.He could also programthe computerto update itself periodically,perhaps even usingthe Britannica year-books. Incidentally, these volumes are printednowadays from information that is processedand storedin a computer memory,and transmittedto the printervia electronicmediations. Thus we must opt for Kripke's version,and stipulatethat the mere fact that sensationsare processedand transmittedby a computer,or by any artificialor non- artificialmeans whatsoever, cannot eliminate the required causal connectionbetween a word and its supposed referendum,on the conditionthat such a connectiondoes existin the firstplace. In 'a causal (historical)chain' the firstlink is what matters: throughit the whole chain should be perceptuallylinked to reality.All otherlinks need not be perceptual,not even the lastlink which conveys the termlight to a blind man, or theterm vat to a BIV. Therefore,if Alex suspects,as a resultof his computer- mediatedwatching of a film(or even as a resultof a computer-mediatedreading of a descriptionof vats),that he mightbe a BIV, his doubtsare ultimatelymeaningful and capable of referring. What then should our scepticbelieve, if he suspectshis truepredicament, while assumingthe existenceof a rich independentworld of objects outside of his own world? Surely, being sceptical, he cannot reject the possibility(a possibility acknowledgedby Putnamas well) thatthe computer'smemories and programsare at least partlya functionof externalreality. Therefore it would be onlyreasonable forhim to hypothesize(even thoughhe can neververify it) that the words 'brain' and 'vat' in his language mayacquire a real referenceafter all. I thinkit has become evidentby now that the only way to preservePutnam's argumentagainst BIV scepticismis to adopt a more rigidversion of the BIV story, one thateliminates any causal linkagewith external reality whatever. Not surprisingly, indeed, Putnam hintsat this second line of defencewhen he moves to considera strangepossible world in which 'everysentient being is and alwayswas a brain in

? The Editorsof ThePhilosophical Quarterly, 1994. 222 YUVAL STEINITZ

a vat' (p. 14). Such a comprehensivesceptical scenario, he seems to believe,must be eitherself-refuting or false. But thisis cold comfort.First, because it returnsus to thefirst horn of the dilemma, i.e., to our previous conclusion that such a comprehensiveontological scepticism necessitateschanging the theoryof referenceaccordingly. Second, because even if Alex refusedto accept this,and steadfastlyheld to Putnam'srealistic causal theory ofreference, he stillcould not shakeoff scepticism. Sceptics are also concernedabout partialsceptical scenarios. So theactual disjunctionAlex is facingis nota meredouble but a tripledisjunction: (a) I am a personwith body, perceiving objective reality; (b) I am a BIV controlledby a computer,but part of the'computer'ssimulation of realityoriginates in the observationsof its programmers(or, alternatively,in the observationsof attached robotic sensorymechanisms), who do perceive objectivereality; (c) I am a BIV, and thereis nothingelse outsidemy system.(Or, alternatively, all scientistsand programmersare themselvesBIVs.) Of course,Putnam may happilyelaborate upon the possiblerealization of the last comprehensivescenario (c), whichpermits him (on the conditionof neglectingour previousarguments of sectionI) to relyupon his hypothesisthat scepticism, in such a case, is self-refuting.... Yet thisshould not comforthim, nor should it comfortany BIV-sceptic.For as long as disjunct(b) cannot be excluded,scepticism seems to remainas plausibleas ever. 13

Tel-AvivUniversit and Universitof Haifa

'3 I would like to thank Naomi Goldblum, Noah J. Efron and Zev Bechler for helpful comments.

? The Editorsof ThePhilosophical Quarterly, 1994.