Phenomenology and Thought Experiments Thought ...philsci-archive.pitt.edu/12334/1/Wiltsche_PittArchive.pdf1 Phenomenology and Thought Experiments Thought Experiments as Anticipation

1

PhenomenologyandThoughtExperimentsThoughtExperimentsasAnticipationPumps

HaraldA.Wiltsche

DepartmentforPhilosophyUniversityofGraz,[email protected]

Theaimofthispaperistopresentanoutlineofaphenomenologicaltheoryofthoughtexperiments (henceforth:TEs). Indoing so, I amdealingwitha topic that is currentlystarting to receive increased attention from philosophers with phenomenologicalleanings. However, since no serious attempt has been made to tackle the issue in asystematic fashion, I will not merely review existing phenomenological work on TEs(suchasMohanty1991;Kunjundzic1995;Froese&Gallagher2010;Fehige&Wiltsche2013;Wiltsche2013;Hopp2014).Forthemostpart,mypaperisprogrammatic:itsaimistosuggestsomebasicdirectionsinwhichaphenomenologicaltheoryofTEsshouldbedeveloped.The paper is structured as follows. I will begin by saying a word or two on whatphenomenologyisandonwhyIbelievethatitcancontributetotheongoingdebateonTEs.Ishallthenintroducefivephenomenologicalconceptsthatwillprovecrucialforanunderstanding of TE-reasoning: fulfillment, frustration, horizon, anticipation andbackground knowledge. My strategy is to first give a brief sketch of how knowledgeacquisitionworks in the perceptual realm. It is against this backdrop that Iwill thenoutlinehowknowledgeisgeneratedintherealmofTE-reasoning.ItisoneofmymainthesesthatTEsshouldbeunderstoodasanticipationpumps.BeforeIbegin,however,twoqualifyingremarksareinorder.Thefirstconcernsmyuseoftheterm“phenomenology”:Justasthereismorethanoneanalyticmethod,thereisnogeneral agreement within the phenomenological community on what thephenomenological method precisely is. It is thus important to bear in mind that thefollowingportrayalofphenomenologyreflectsmyownbackgroundin(aliberalreadingof)EdmundHusserl’sphilosophy.Secondly,muchofwhatIwillhavetosayaboutTEsconcerns TE-reasoning in science and not TE-reasoning in philosophy. To a certainextent, this limitation again reflectsmyownbackground in thephilosophyof science.However,sinceIbelievethatscientificandphilosophicalTEsdifferonly indegreeandnot in kind, I am confident that the proposed framework can also be applied to non-scientificTEs.

1. GettingPhenomenologyofftheGround

When phenomenology entered the philosophical stage in the early 20th century, Neo-Kantianismwasstillinfullswing.OttoLiebmannhadcoinedtheunifyingmotto“BacktoKant!” under which the Neo-Kantians sought to overcome the dispute betweenmaterialism and idealism that had dominatedmuch of the philosophical discourse inGermanyandAustriaupuntil themid-19th century. Itwasagainst thismotto that thefirst generation of phenomenologists directed their own slogan “Back to the thingsthemselves!”. Yet, itwasn't particularly Kant ofwhomHusserl and his followerswerecritical.Thefirstwaveofphenomenologistsdisapprovedofanykindofphilosophythat

2

looks at its problems through the glasses of pre-established theories, systems orschemes. Instead of forcing problems into a particular (and potentially artificial)theoreticalmould,phenomenologistswere (andstill are)drivenbyadeeprespect forthe phenomena, i.e. the things exactly as they are given in experience. On aphenomenologicalview,manyphilosophicalproblemscouldbesolved–orevenbetter:made to evaporate – if we resisted the temptation to interfere with ready-madetheoreticalschemesandputmoreeffortinafaithfuldescriptionofthephenomena.Paradigmatic for thisapproach isHusserl’ssixthLogical Investigation(Husserl2001b,177-348):Theaimofthisbook-lengthtreatiseistotackleoneofthemostfundamentalissues in philosophy, namely to understand the concepts “truth”, “knowledge” and“knowledgeacquisition”,aswellastherelationsbetweenthem.OnHusserl’sview,thetroublewithmanyexistingapproachesto this issue is that theyare infectedwithpre-establishedtheoreticalschemesalreadyat the levelof the formulationof theproblem.Forinstance,acommonwaytostartistoaskhowaself-enclosedsubjectisabletoreachtheexternalworldinordertoattainjustifiedtruebeliefsaboutrealitybyrelyingonactsof visual, auditory or olfactory perception. However, built into this question are anumberof far-reachingassumptionssuchascommon-senserealismortheviewthatagapbetweentheinteriorityofthementalandtheexteriorityoftheworldinfactexists.Husserl accepts that these and similar other assumptions may be useful or, in somecases,evenindispensableincertainpracticalcontexts.Yet,sincethejobofphilosophyisnot to simplyadopt existingassumptions, but rather to scrutinize them,philosophicalanalysis must proceed from a standpoint that is not already contaminated with pre-establishedtheoreticalschemes.ItisforthisreasonthatHusserlactively“brackets”1allkinds of metaphysical, scientific and commonsensical assumptions and starts with acarefulandunbiaseddescriptionofthetypesofconsciousexperiencesinwhichdifferenttypesofobjectsbecomepresent froma first-personpointofview.Tobesure,nothingpreventsthephenomenologistfromeventuallytakingsidesintraditionalmetaphysicaldisputes about, say, dualism, realism or physicalism. But whatever she will end upsayingonsuchmattersmustresultfromadescriptionofthephenomenaandnotfromthemerestipulationofbasicmaximsorprinciples.As I have pointed out, phenomenologists are driven by a deep respect for thephenomena,i.e.forthethingsastheyappearinconsciousexperience.Thisgeneraltenet–which, inmyview, capturesmuchofwhatmakesaphenomenological account trulyphenomenological2– also yields important consequences for how phenomenologists

1Thenotionof“bracketing”comesfromHusserl’smathematicalbackground,specificallyfromtheconceptofabsolutevalue(thenotation“|x|”wasintroducedbyHusserl’steacherKarlWeierstrass).Theabsolutevalue|x|ofarealnumberxisthenon-negativevalueofxwithoutregardtoitssign.So,forinstance,3and-3havethesameabsolutevalue3.Hence,tobracketanaturalnumbermeanstobeindifferentwithregardtoanumber’spropertyofbeingpositiveornegativeandtofocusonitsmagnitudeinstead.Accordingly,tobracket assumptionsneithermeans to endorse, nor todeny them. Itmerelymeans to focusonwhat isgivenindependentlyfromallassuming.2It should be emphasized that it is mainly for the sake of brevity that I am working with a ratherrestricted conception of phenomenology here. For the purposes of this article, I am conceiving of thephenomenologicalmethod roughly along the lines of Charles Siewert’s “plain phenomenology”. On thisview, one is doing phenomenology if (1) one explains mental distinctions, if (2) one shows how suchexplanations have significant theoretical consequences, if (3) one’s explanations relies on a source ofwarrant special to some first-person applications of the distinctions explained and if (4) one does notassumethatsuchfirst-personwarrantasonereliesonisderivedfromthird-personevidence(cf.Siewert2007, 202).However, I’d also like to stress that this understanding of phenomenologydoesnot do fulljustice to the all the subtleties of Husserl’s account. Readers interested in a more comprehensive

3

shouldapproachtheissueofscientificTEs.Letmebeginbybrieflysummarizinghowthediscussionisusuallyframed:ItisbeyonddisputethatTEsarepartofscientificpracticeincoredisciplinessuchasphysicsorbiology.MostcommentatorsalsoagreethatatleastsomescientificTEsaresuccessful ingeneratingknowledgeabout theempiricalworld.Buthow is thispossible?Since,bydefinition, theyareexperiments inthought,TEsdonotseemtoestablishdirectcognitivecontactwiththeworld.ThismakesthesuccessofTEsindeedpuzzling:Howcanwelearnsomethingabouttheworldbymerelythinkingabout it? It has become common to refer to this question as the paradox of TEs (c.f.Horowitz&Massey1991,1).Much of the contemporary discussion about scientific TEs is devoted to the task ofsolvingthisparadox.Hence,theexistenceofapparentlysuccessfulTEsisusuallytreatedastheexplanandumtowhichsuitabletheoreticalexplanationsaredirected.Ononesuchtheory,thesuccessof(some)TEsisexplainedbymeansofaPlatonistconstrualofthelawsofnatureandbytheancillaryepistemologicalthesisthatTEsallowustodirectlygrasptheselawswithourmind’seye(e.g.Brown2004,2011).AnothertheoryexplainsthesuccessofTEsbydeclaringthemtobeargumentsindisguise.Onthisview,scientificTEs lead us to knowledge because they are, in reality, chains of inductive and/ordeductiveinferences(e.g.Norton1996,2004).Onathirdpopularaccount,thesuccessofTEs is explained by means of a psychological theory according to which themanipulation ofmentalmodels allows us to trigger knowledge formations that aren’tcognitivelyavailableotherwise(e.g.Miščević1992,Nersessian1993).AgooddealofinkhasbeenspilledovertheseproposalsandIwilladdtothisdiscussionbelow.For themoment,however, Iwant toemphasize thatmuchof theplausibilityofeachof these theoriesdependsonquitesubstantialmetaphysical, epistemologicalandontological assumptions whose scope goes well beyond their actual subject matter.Take, for instance,Platonism:TheplausibilityofaPlatonistexplanationofTEs largelydepends on our willingness to accept a rationalist epistemology, an ontology thatincludes universals and the view that the laws of nature are contingent necessitationrelations between universal properties. If one accepts these assumptions, thenPlatonismisaperfectlygoodexplanationfortheexistenceofTEs insciencehistoryaswell as of their apparent success. If, on the other hand, one thinks thatwe should beparsimoniousinontologicalandmetaphysicalmattersandthatsenseperceptionistheonly source of knowledge about the world, then one will likely opt for one of theempiricistalternatives toPlatonism.Hence,although itwouldbe toostrongofaclaimthattheavailabletheoriesonTEsareevaluatedonlyonthebasisoftheircommitmentstocertainmetaphysicalandepistemologicalstances, it isat leastsafetosaythatthesecommitments play a decisive role in how the discussion normally evolves. And,supposedly, it is also due to these commitments that certain strands of thecontemporarydebateonTEsappeartobeameresidelineoftheage-oldbattlebetweenrationalismandempiricism.Now,mypointisnottodenythattheTE-debatemayhaveimplicationsforthequarrelbetweenrationalistsandempiricists.SinceitisatleastinitiallyplausibletoregardTEsasprimafaciecasesofknowledgeacquisitionminussenseperception,theonusseemstobeonphilosopherswithempiricistleaningstocomeupwithadeflationaryaccountonTE-reasoning.Rather,mypoint is that, ifwewish toaddressTEsphenomenologically,treatment may either refer to Crowell (2006) for a highly instructive introductory essay or to Smith(2007)foramoredetailed,book-lengthstudy.

4

the issue should be approached in a way that is not already contaminatedwith pre-established assumptions concerning metaphysical, ontological and epistemologicalmatters. To construct one’s theory on the basis of either rationalist or empiricistintuitions not only clashes with the phenomenological demand to proceed from anunbiased description of the phenomena; it also harbours the danger of begging thequestion against the respective opponent, of distorting the subject matter before theactualanalysiscanevenbeginandofendingup inastalematebetween irreconcilableepistemologicalandmetaphysicalsystems.Hence,whatsetsphenomenologyapartfromothertheoreticalalternativesis,firstofall,its point of departure: Instead of treating TEs as explananda that call for theoreticalexplanationswhosecredibilitydependsonmoregeneralmetaphysical,ontologicalandepistemological intuitions, phenomenology actively brackets such intuitions andproceeds fromanunbiaseddescriptionof thephenomenonofTE-reasoning instead. Inotherwords:Themainobjective,atleastattheinitialstage,isnottoconstructatheorythat seeks to bring TEs in line with certain pre-established standpoints. The mainobjective of a phenomenological account is rather to give a faithful description of theactualperformanceofTEsfromafirst-personperspectiveandtogoonfromthere.

2. FillingUpthePhenomenologicalToolboxAsIhavepointedout,phenomenologydiffersfromothertheoriesinitsbasicapproach:AphenomenologicalaccountonTEsbracketsallkindsofmetaphysical,epistemologicalandscientificassumptionsandproceedsfromafaithfuldescriptionofthephenomenonofTE-reasoning.However,whatdoesthismeanexactly?WheredowehavetostartifwewishtodescribetheperformanceofTE-reasoningfromafirst-personperceptive?HowdoweidentifycertainmentalepisodesasTEs?Andhowcouldwepossiblytellifthesementalepisodesembodyinstancesofsuccessfulknowledgeacquisition?Inthefaceofthesequestions,mystrategyintheremainingpartsofthispaperwillbeasfollows: Since I am interested in the question ofwhether TEs fall under the rubric ofsuccessfulknowledgeacquisition,Iwillhavetosayafewwordsonaphenomenologicaltheory of knowledge first. It ismainly through the discussion of examples that I willintroduce five concepts that are crucial for a phenomenological understanding ofknowledge acquisition in the perceptual realm. These concepts will form thebackgroundagainstwhichIwillthentakeacloserlookattheimagination,3themediumin which TEs are performed. My aim is to highlight the similarities as well as thedifferencesbetweenperceptionandimaginationinordertoelucidatehowknowledgeisgeneratedthroughtheuseofTEs.

3“Imagination”isanumbrellatermthatisnotoriouslyhardtodefine(cf.e.g.Kind2013).Inwhatfollows,Iwill take it to denote a type ofmental event inwhichwe are intentionally directed towards imaginedobjectsorscenariosandnot,forinstance,towardspropositions.Hence,Iamdistinguishingbetweenactsof imagining (inwhich imaginedobjects or scenarios are given in a quasi-sensorymanner) and acts ofconceiving (in which the objects of our intentional directedness are propositions). It should be noted,however,thatmyterminologydepartsfromHusserl’swhousesthenotion“phantasy”inordertodenotewhatIcall“imagination”.

5

2.1 FulfillmentandFrustrationWhat is knowledge? When are we dealing with cases of successful knowledgeacquisition?Inlinewiththegeneraltenetsofthephenomenologicalresearchprogram,these questions, too, are to be answered by way of a description of those types ofconscious experience in which the relevant distinctions become evident. Take thefollowingtwocasesasastartingpoint:1)IamjudgingthatmybikeisintheofficewhileIamstillinthecafeteria.2)IamjudgingthatmybikeisintheofficewhileIamstandingrightinfrontofit.Phenomenologicallyconstrued,thesetwocaseshavealotincommon.Tobeginwith,bothexperiencesareexperiencesofthesametype,namelyexperiencesofjudging(incontrasttoexperiencesofdoubting,imagining,fearingetc.).Inaddition,bothexperiences also have the same intentional object, namelymy bike in the office. But,clearly, there is a striking difference as well: While the judgement about my bike’swhereaboutsmay just be awild guess as long as I am still in the cafeteria, the directacquaintancewithmybike(i.e.mystandingrightinfrontofit)warrantsthejudgementaboutmybike’slocationbeyondallreasonabledoubt.The acknowledgement of this difference allows us to make a first step towards thenotionoffulfillment,oneofthemainpillarsofphenomenologicalepistemology:If–likein the first case – I am directed towards an object in its absence, thenmy intentiontowardstheobjectisempty.If,ontheotherhand,Iamdirectedtowardsanobjectinitsactualpresence– if theobject isgiven in its “‘bodily’ selfhood” (leibhaftigeSelbstheit)(Husserl1983,pp.9-10;translationmodified)–,thenmyintentiontowardstheobjectisfulfilledbythepresenceofthatveryobject.Fulfillment,i.e.thecongruencebetweentheobjectasitisemptilyintendedandtheobjectasitisintuitivelygiven,istheideallimittowardswhichourjudgingstrives,“themeasureofitssuccessorfailure”(Crowell2006,14).Ifthiskindofcongruencenotonlytakesplace,butisalsoregisteredbyacognizingsubject,thenwearedealingwithwhatWalterHoppaptlycalls“knowledgeatitsbest”(Hopp2011,chapter7).4Letmesummarize:Fulfillmenttakesplace if there isarelationofcongruencebetweenanobject as it is emptily intendedand theobject as it is intuitivelygiven.Conversely,frustrationoccurswhentheintuitiveexperienceofanobjectisnon-congruentwiththeempty intention towards thatsameobject. Inan idealcaseof fulfillment, I realize thattheobjectisexactlylikeIthoughtitwouldbe.Incasesoffrustration,theintendedobjectturns out to be different than initially thought. On a phenomenological view, allepistemicconceptssuchasevidence,justificationorfalsificationultimatelyleadbacktothebasicconceptsoffulfillmentandfrustration.

2.2 HorizonandAnticipationThe concepts of fulfillment and frustration are crucial for the phenomenologicalunderstandingofknowledgeandknowledgeacquisition.Fulfillment,i.e.thecongruence

4Itmustbenotedthattheconceptoffulfillmentisnotrestrictedtotherealmofperceptualexperiences.Althoughitistruethatveridicalperceptionsaremodelcasesoffulfillment(cf.e.g.Husserl1983,5-6,82-83,154,327),fulfillmentisafunctionalconceptthatgoesalongwithafunctionalobjectconcept.Thisistosay that fulfillment takesplacewheneveranempty intention towardsanobject is incongruencewithafulfilling intention towards the same object. However, as Husserl makes clear, “an object […] may asreadilybewhatisrealaswhatisideal,athingoraneventoraspeciesofamathematicalrelation,acaseofbeingorawhatoughttobe”(Husserl2001a,p.145).

6

betweenanobject-as-intendedandtheobject-as-given,istheideallimittowhichallofour knowledge-claims aspire.On closer inspection, however, this turnsout tobe onlypartofthestory,especially–butnotonly–withrespecttotheacquisitionofempiricalknowledge.Consideragain theexampleof theveridicalperceptionofmybike.At firstglance,thedirectperceptualgivennessofmybikeseemstoclearlyfulfillthejudgmentthatmybikeisintheoffice.Butamoreaccuratedescriptionrevealsthatwhatisreallyexperientiallygiveninthissituationisnotsimplymybike,butonlyonesingleprofileofmybike,itscurrentfrontside.Tobesure,Icouldaltermypositionandmakethecurrentbacksidethenewfrontside,andviceversa.Butthisdoesn’tchangethefactthatmybikeisalwaysgiven inperspectives and that,moregenerally, thingsalwaysandnecessarilyhavemoreparts,functionsandpropertiesthancanbeactualizedinasingleintentionalact.Mybike–asitisintended–istranscendent,notonlyinthesensethatitcanbeseenfromindefinitelymanymoreperspectivesthanIcantakeupatagivenpointintime.Itisalso transcendent in thesense that ithas, for instance,amomentarilyhidden internalstructure, a history, certain practical functions or many properties that aren’t in thecenterofattentionrightnow.So,a closer lookathowthingsappear tous reveals thatour intentions towards thesethingsalways“transcend”or“gobeyond”theactualexperiencesthatgiverisetothem.As theexampleofmybikeshows, there isadescribablediscrepancybetweenwhat ismeant through a particular intentional act (my bike over there) and what isexperientially given (my bike’s facing side with its momentarily visible features).Phenomenologicallyconstrued,thisdiscrepancydoesnotrepresentaproblemthatmustbe somehow remedied, e.g. by proposing a theory that explains how a number ofseemingly disconnected profiles add up to a homogeneous thing to which we thenattributetheseprofiles.Thefactthatourintentionstowardsthingsalwaystranscendthesphere of intuitive givenness is rather to be treated as a phenomenologicallydiscoverable feature of experience itself: Intending is, as Husserl puts it, always andnecessarily an “intending-beyond-itself” (Husserl 1960, 46). In being intentionallydirectedtowardsmaterialthings,we“know”thatthereismoretothemthanisrevealedinonesingleglance.This“knowing”isnomatterofinferentialbelieforjudgementoverand above the experiences in which things are perspectivally given; it is rather anessentialpartofanysuchexperience.The important lesson to draw from these considerations is that “[e]ach individualpercept isamixtureof fulfilledandunfulfilledintentions”(Husserl2001b,221).Or,toput it in an alternative terminology: Intentional experiences are always embedded inimplicithorizonsofintentionsthataremomentarilyunfulfilled,butthatcouldbefulfilledinthecourseoffurtheracts.EventhoughIcannowonlyseemybike’sfacingsidewithits momentarily visible features, my bike appears to me as something that could beexplored more fully. I “know” that I could alter my vantage point and explore itsmomentarilyhiddenbackside. I “know”that Icould lookmorecloselyandexplore itssurfaceinmoredetail.I“know”thatIcouldcuttheframeinhalfandexploreitsinternalstructure.Itistheseandindefinitelymanyotherpotentialitiesthatadduptotheimplicithorizon against the background of which singular intentions towards things alwaysstandout.Phenomenologicaldescriptionsrevealthatintentionalactstowardsthingsalwayspointtoimplicithorizonsofemptyintentions.Hence,onaphenomenologicalview,experienceis never exhausted by what is actual; experience is always already saturated with

7

implicit references to future experiences that are possible insofar as they could beactualized in the course of further acts. However, in order to get a more completeunderstandingofthephenomenologicalnotionofhorizon,twomoreaspectshavetobestressed:The first thing tonote is thatanact’shorizon is,as Ihavealready indicated,implicittotheactitself.Thisistosaythatthehorizonisnotheoreticalconstructthatisretrospectivelyascribed to the initial act. An act’s horizon is rather co-givenwith theinitialacteventhoughweusuallyaren’tawareofthis.Thatmybikehasabacksideisn’tsomethingthatcanonlybeassertedafterIhavechangedmyvantagepoint.Itisalsonotsomethingthatistheproductofsomesortofinferentialprocess.Rather,itissomethingthatbelongstotheverymeaningofbeingintentionallydirectedtowardsmaterialthings.Thesecondimportantaspect isthis:OnthebasisofwhatIhavesaidsofar,onecoulddefinehorizonsassetsofemptyintentionsagainstthebackgroundofwhichparticularfulfilled intentions always and necessarily stand out. But this definition is somewhatmisleading:Althoughitiscorrecttosaythatahorizonconsistsofemptyintentionsandthuscanbedescribedasa“haloofemptiness”,itiscrucialtostressthat“thisemptinessis not a nothingness”, but rather that “the sense of this halo […] is aprefiguring thatprescribesaruleforthetransitiontonewactualizingappearances”(Husserl2001c,42;myemphases).WhatHusserlissayingherecanbeelucidatedwiththehelpofmyearlierexample:IfIperceivemybike,myintentiontowardsthebike’sfacingsideisconjoinedwithahorizonofemptyintentionsandthuswiththeanticipationthatthereismoretothe bike than is revealed in one single act. However, this more is far from beingindeterminate: In perceiving my bike, I implicitly anticipate concrete courses ofexperiences that are compatiblewithwhatwasoriginally intended.Theperceptionofthe greyish blue of my bike’s facing side, for instance, comes with the implicitanticipationthatthemomentarilyhiddenbacksidewillexhibitthesamecoloraswell.Tobesure, it ispossible that thisanticipation is frustratedby futureexperiences.But if Iwere to findout thatmybike’sbackside isn’t greyishblue, but coatedwithdiamondsandrubies,thenIwouldprobablystarttowonderwhetherIamreallydealingwithmybikeatall.So, typically,horizonsarenot indeterminateinthesensethatagivenactpointstotheentirety of acts that are logically compatible with the initial act. Horizons are ratherstructured: they prescribe implicit rules of anticipation that restrict how things couldappearifthecorrespondingexperienceswerestilltoqualifyasexperiencesofthesamething.

2.3 BackgroundKnowledgeThepointofthepreviousconsiderationsisthatourexperienceofthingsis“thoroughlyinterwovenwithanticipations”(Husserl2001c,47).Thingsarealwaysandnecessarilygiven inhorizonsand thesehorizonsprescribe rules throughwhichweanticipate thecourseof futureexperiences.Undernormalcircumstances, theserulesremain implicitand largely unnoticed. It is mainly in cases of frustrated anticipations that theirexistenceevenbecomesmanifest.5

5Here is an example that illustrates the issue: Imagine you are asked to bring a shopping bag into thekitchen.Yougotothecar,youseethebaginthetrunk,yougetreadytoliftitand–Oops!–yourarmsgoupway tooeasyandyoualmost toppleover.Whathappened?Well, thebag issignificantly lighter thanyouanticipatedbecauseitisfilledwithpapertowels.Whatisremarkableaboutthiscaseis,firstofall,thefunny feeling that you hadwhen you lifted the bag. Since you have liftedmany light items before, this

8

Giventheviewoutlinedsofar,anobviousquestionarises:Wheredotheaforementionedanticipationscome from?Howare theygenerated?Andwhat is theirstatus?AsDavidWoodruffSmithandRonaldMcIntyrehavepointedout(Smith&McIntyre1982,chapter5; Smith 2004, chapter 5), both horizons and the rules of anticipation that are giventhroughthemaredependentonhighlyheterogeneousstocksofbackgroundknowledgethat,intheirtotality,makeupourfundamentalbackgroundimageoftheworld.Inlargepart,thisbackgroundknowledgeisempiricalandthushighlycontingentinnature.Justthinkoftheearlierexampleofmyanticipationthatmybike’sbacksidewillexhibitthesame color as its facing side:Clearly,myanticipation in this casedependson the factthat I’vehad countlesspreviousencounterswithmybike.Hence, amoredeterminatehorizonwithmoredeterminateanticipationsisco-givenwithmyexperiencesofmybikethanwouldbeco-givenwithyourexperiencesofmybike.Otherpartsofourbackgroundknowledgearenotempirical,but theoretical innature:Psychologists conducted experiments in which students were asked to observe thefallingoftwoobjects(ametalsphereandaplasticsphereofthesamediameter)andtorecord their observations (Gunstone &White 1981). Students who initially held thatheavierobjectsfallfasterweremuchmorelikelytoreportobservationsthatsupportedtheirtheory.Experimentssuchasthesesuggestthatouranticipationsoffurthercoursesof experience are sometimes strongly influenced by theoretical components of ourbackgroundknowledge.Finally, as phenomenologists typically stress, there is a third type of backgroundknowledge that belongs to neither of the two aforementioned categories: Take, forinstance,theanticipationthatmybikewillexhibit,notaparticularcoloronitsbackside,butrathersomekindofrearsideatall.Sincewefailtoevenimagineamaterialthingthatdoesnotexhibitabacksideofsomesort,phenomenologistsclaimthatthecorrespondinganticipationisnotgovernedbyacontingentpieceofbackgroundknowledge;itisrathergovernedbya“necessaryla[w]whichdetermine[s]whatmustnecessarilybelongtoanobjectinorderthatitcanbeanobjectofthiskind”(Husserl1973,352).Thatmaterialthings, unlike, say, geometrical objects, are necessarily given in perspectives isdeterminedby“auniversalessencewhich[…]prescribesanintellectuallyseengenericalruleforeveryparticularobjectbecomingintendedtoinmultiplicitiesofconcretementalprocesses”(Husserl1983,341).

3. PerceptionandImagination:TheReturnoftheParadoxMyapproachintheprevioussectionswastouseexamplesfromtheperceptualsphereinorder to introduce five phenomenological key concepts. Yet, on closer inspection, theapplicability of these concepts is not restricted to the perceptual realm. Consider, forexample, “horizon” and “anticipation”: As I have pointed out, material things arenecessarilygiveninhorizonsthroughwhichweanticipatefurthercoursesofexperience.Horizontalgivenness,however,isbynomeansexclusivetoperception:“Imaginings[…]present their objectswithin exactly the samehorizons” (Husserl 1973, 169). Imagine,

feeling is certainly not a consequence of the fact that you lifted something light. What makes theexperiencestandoutisratherhowitcomparedtoyour(frustrated)anticipations.Whatismore,caseslikethese also highlight the fact thatmany of our anticipations correspond to pieces of non-propositional,practicalandsensorimotorknowledgeabouthowthingsaredoneandhowweuseourbodiestodothem(cf.,e.g.,Husserl1997;Noë2004).

9

for instance,BartSimpsonwriting “Noonecareswhatmydefinitionof ‘is’ is”onMrs.Krabappel’s blackboard. If you do so, then the object of your attention is a yellowcartoon character that is scribbling on a blackboard. Yet, co-given with this initialexperienceisasetofemptyintentionsthroughwhichyouanticipatefurtherfeaturesofthe imagined scenario.Although theyaren’t in the centreof yourattention right fromthe start, you automatically anticipate further features such as the spikiness of Bart’shair,thewallbehindtheblackboardorthecolorofBart’spants.Ifthefocusofattentionshiftsandobjectsfromthemarginmovetothecentre,certainotherobjectsfadeintothebackground,without,however,disappearingcompletely.Likeinthepreviousexamplesfromtheperceptualsphere,you“know”thatthereismoretotheimaginedobjectsthancanbegraspedinonesingleact.Horizontalgivennessisnottheonlycommonalitybetweenperceptionandimagination.Rememberthecrucialdistinctionbetweenemptyandfulfilledintentions:IfIammerelythinkingofmybike,thenmyintentiontowardsmybikeisempty.If,ontheotherhand,Iperceivemybike,myintentiontowardstheintendedthingisintuitivelyfulfilledbythething’s presence. But is perception the only means by which the empty intentiontowardsmybikecanbefulfilled?Notatall.IfIimaginemybike,thenthis,too,fulfillsthemerethoughtofmybike–whatwasemptybefore(thebikeastheobjectofmythought)is now fulfilled by the quasi-experience of my imagined bike. Hence, perception andimagination reveal a close parallelism not only with respect to the structuralcharacteristicofhorizontalgivenness:Since “[t]heobjective intentiondirected towardthe imagined object has its filling in the experienced phantasms, just as the objectiveintention in perception has its filling in sensations” (Husserl 2005, 93; translationmodified), the parallelism between perception and imagination involves the crucialconceptoffulfillmenttoo.Thereare,aswehaveseen,importantessentialfeaturesthatarecommontoperceptionandimagination.But,ofcourse,therearecrucialdifferencesaswell.Themostobviousconcerns the fact that,unlike imagination,perception isan“originallypresentivemode[ofconsciousness]”(Husserl1982,327).Whatthismeanscanagainbeillustratedwithrecoursetotheearlierexample:IfIperceivemybikestandinginmyoffice,thenthisactisself-givingwithregardtoactualthings(mybike)andstatesofaffairs(thatmybikeisstandinginmyoffice).Thus,itisnotonlythecasethattheperceptualgivennessofmybikeintuitivelyfulfillsthethoughtthatmybikeisinmyoffice.Evenmoreimportantly,the perceptual act “gives its object itself in the flesh” (Husserl 2001c, 140) and thusexhibits a particular quality that is discussed in the contemporary analytic literatureunder labels such as “presentational feel” (Foster 2000, 112), “scene-immediacy”(Sturgeon2000,24)or“presentationalphenomenology”(Chudnoff2013,chapter1.2).Onaphenomenologicalview,thisqualityisamainreasonwhywetakeperceptualactstoconfer justificationonempiricalbeliefs. I seemybike leaningagainst thebookshelfanditinstantlyseemstomethatwhatIseeisatruthmakerfortheproposition“Mybikeisintheoffice”.Thefactthatthevisualexperienceofmybikeinstantiatesthequalityofhavinga“presentationalfeel”explainswhythisisso.Butnowcomparethiswiththecaseofimagination:IfIimaginemybikestandingintheoffice, then this, as I have pointed out, also counts as a case of fulfillment:Whatwasempty before is now fulfilled by the quasi-sensory experience of my imagined bike.However,clearly, thequasi-experienceofanimaginedbike isno justifier forthebeliefthatmybikeisinmyoffice.Phenomenologicallyconstrued,thisisbecauseimaginative

10

quasi-experiencesdonotinstantiatethepropertyofhavinga“presentationalfeel”withrespecttoactualthingsandstatesofaffairs.While“[p]erceptionmakesapresentrealityappeartousaspresentandasareality[…],[imagination],ontheotherhand,lackstheconsciousnessofrealityinrelationtowhatis[imagined]”(Husserl2005,4).Itisforthisreasonthattheimaginationdoesnotcountasadirectsourceofjustificationofempiricalbeliefs.Perceptionisanactthatgivesitsobjectasactual.Anactofimaginationgivesitsobjectasnon-actual.This,inanutshell,isthereasonwhytheimaginationdoesnotcountasadirect source of empirical justification. But there is a second, evenmore fundamentaldifferencethatconcernstheconditionsunderwhichanticipationsandexpectations6arefrustrated in each respective realm. Here is an example that illustrates the point:Suppose that I perceive the greyish blue of my bike’s facing side. And supposefurthermore that – for whatever reasons – the horizon in which this perception isembeddedcomeswith theanticipation thatmybike’sbackside is coated inpuregold.Whatwillhappen?Sinceabikethatisbothmineandridiculouslyexpensiveisn’tpartofthe inventory of the actual world, this anticipation will be frustrated by furtherperceptual experiences. And, obviously, there is nothing I can do about this: By andlarge,ourperceptualexperiencesarenotunderourvoluntarycontrol;theiroccurrenceaswellastheircontentissystematicallyfixedbyourexternalenvironment.Butthingsareentirelydifferentintheimagination.SupposethatIimaginemybike.Andsuppose furthermore that thehorizon inwhichmy imaginedbike isgivencomeswiththeanticipationthatitsbacksideiscoatedinpuregold.Obviously,nothingpreventsmefrom imagining my bike in a way so that the anticipation of a gold-coated backsidewouldbe intuitively fulfilled.Generally speaking,weknow thatwe cando all sorts ofcrazythingsinthesphereofimagination:Icouldimaginemybikeasalmosttransparentortalllikeamountain.Icouldimaginethatitisacceleratedto90%ofthespeedoflight.OrIcouldimaginethatmybiketransmutesintoahoneybeeandfliesaway.Itisinmyfreedom to project these and indefinitely many other scenarios with the help of myimaginativecapacities.Hence,whilethepersistencyofanticipationsandexpectationsisstrictlyregulatedbytheexternalworldintheperceptualsphere,theimaginationseemstobedistinguishedby“itsoptionalcharacter [and] therefore,speaking ideally, [by] itsunconditioned arbitrariness” (Husserl 2005, 642). “[I]t remains”, as Husserl puts itelsewhere, “within theprovinceofour freedom toallow the indeterminatenessof […]horizonstobequasi-fulfilledinanarbitrarywaybyimagining.”(Husserl1973,171)Given this rough-and-ready comparison, one could come to the following conclusion:Imaginationislikeperceptioninmanyways.Butbeneaththesurfaceofsimilarities,twocrucial differences remain: First, perception gives its objects as actual. The objects ofimagination,ontheotherhand,exhibitthe“characteristicofinactuality”(Husserl2005,320). Secondly, perceptual anticipations and expectations are fixed by the external

6A word on terminology: Anticipations are non-independent parts of perceptual experiences. If, forinstance, Ienter thekitchenand flip the lightswitch, theexperienceof the flippingof theswitchcomeswiththetacitanticipationthatthelightwillgoon.Thisanticipationisneitheranactinitsownrightnorisit consciouslyexperienced in thenormalcourseofevents.However,anticipationscanbecomethebasisforseparateactsofexpectation. If, for instance, Ihave justrepairedthe lightswitchandcheck it fortheveryfirsttime,theflippingoftheswitchmaycomewiththeexpectationthatthelightwillgoon.Inthiscase,theexpectationisanactinitsownrightandhasthestatusofanexplicitprognosisoffutureevents.

11

world. Imaginative anticipations and expectations, on the other hand, seem to beessentiallyunrestricted.If we take this result as conclusive, then this also has serious consequences for thephenomenological understanding of TEs. In particular, the paradox of TEs seems toreturnwithinthe frameworkofphenomenology in itsoriginal force:ScientificTEsaresupposed to justify beliefs about the empirical world. But since TEs rely on ourimaginative capacities, and since the imagination is no direct source of empiricaljustification,weare inneedofaplausiblestoryabouthowTEscouldpossiblyachievethisaim.Butwhateverstorywetell,wewillhavetoconfrontthefollowingworry:Theprincipal reason forregardingphysicalexperimentsasepistemicallysignificant is thatourexpectationstowardsthemcanbeandquiteregularlyarefrustratedbyperceptualexperienceswhose occurrence and content is not controlled by us. To put it bluntly:Whenweperformphysicalexperiments,realitysometimes“kicksback”atus.Andif itdoes, we learn something utterly important, namely that the world is different fromwhat we had expected. But what could possibly “kick back” at us in TEs? Since TEsexploit our imaginative capacities, and since, apparently, these capacities arecharacterizedbytheunconditionalfreedomtoimaginewhateverwelike,theconditionsoffulfillmentandfrustrationinTEsappeartobecompletelyunderourcontrol.Ifthatisthe case, however, then it is hard to see how TEs could possibly qualify as a seriousmethod of knowledge acquisition. Viewed in this light, TEs are similar to children’sgames.Ofcourse,theyarefuntoplay.Butsincetheirrulescanchangeerraticallyatanytimeduringtheprocess,theyarealsosomehowpointlessfromanadult’sperspective.

4. ThreeTypesofConstraintsHereistheupshotoftheprevioussection:InordertosolvetheparadoxofTEs,weneedtotellastoryaboutwhytheimaginationcanberegardedasasourceofstableevidence.But there is an obstacle to any such story:One of the principal reasons for regardingperceptionas epistemicallyvaluable is thatwhat isperceived isnot entirelyup tous.The occurrence aswell as the content of our perceptual experiences is systematicallyfixedbyourexternalsurroundings.Imaginings,ontheotherhand,seemtolackexternalconstraints: Apparently,we are free to imaginewhateverwe like,wheneverwe like.7Hence, the conditions of fulfillment and frustration appear to be entirely under ourcontrol in the sphereof the imagination.But if this is true, then it is unclearhow theimaginationcouldpossiblygeneratestableevidence.Onthebasisofwhathasjustbeensaid,askepticalattitudetowardstheepistemicvalueofimagination(and,consequently,ofTEs)isfuelledbytheviewofimaginationasbeingessentiallyunrestricted.But isthisviewcorrect?Inwhat followsIwill trytoconvinceyouthatitisn’t.Myaiminthissectionistogothroughacoupleofexamplesinordertomake clear that, how and to what extent the course as well as the content of ourimaginingsaresubjecttothreedifferenttypesofexternalconstraints.Gettingclearon

7Sartreseemsto followthis lineofreasoningwhenhecomesto theconclusion“that it is impossible tofindinthe[mental]imageanythingmorethanwhatoneputsintoit;inotherwords,the[mental]imageteaches nothing” (Sartre 2004, 103). But similar views can also be found outside of phenomenology:Consider,forinstance,LudwigWittgenstein’sremarkthat“[i]tisjustbecauseimagingissubjecttothewillthatitdoesnotinstructusabouttheexternalworld”(Wittgenstein1980,§80).OrtakeAlanWhite’smorerecentassessmentthat“onecan’tbesurprisedbythefeaturesofwhatoneimagines,sinceoneputthemthere”(White1990,92).

12

the nature of these types will finally bring us closer to a phenomenologicalunderstandingofhowknowledgeisgeneratedthroughTEs.Suppose,thatIaskyoutoimagineBartSimpsonwriting“Iwillnotuseabbrev.”onMrs.Krabappel’s blackboard. If you do so, then the object of your attention is a yellowcartoon character that is scribbling on a blackboard. Yet, co-given with this initialexperience is the anticipation of further features of the imagined scenario: Althoughthese features aren’t in the centre of attention right from the start, you automaticallyanticipatefurtheraspectssuchasthespikinessofBart’shair,thepurplewallbehindtheblackboard,Bart’sdepravedbehaviourorhisredpants.Butwait!Whilethesequenceofimaginingsevolves inmoreandmoredetail,yousuddenlyrealize thatsomething isn’tright.Eventhoughyoucan’tquitefigureoutwhatit is,thewholescenariodoesn’tfeellikeitshould.Itisonlythroughcarefulreflectionthatyoufinallynoticewhatbotheredyou:Bart’spantsareblue,notred!What happened in the situation just described? To begin with, by following myinstructiontoimagineBartSimpson,youimmersedyourselfinthequasi-worldof“TheSimpsons” and thus accepted certain limitations to your imagining.The concept “BartSimpson”containswhathasbecomeknowntoyouandyourepistemiccommunityaboutthekindofobjectinquestion.Hence,byapplyingtheconcept“BartSimpson”inordertodetermine what you were about to imagine, you activated certain parts of yourbackgroundknowledgethat,inturn,motivatedahorizonthroughwhichfurtherquasi-experienceswereanticipated.Itisofcoursetruethattheimagination,unlikeperception,gives you the freedom to produce quasi-experiences that are at odds with what iscontained in the initial concept:Youcouldeitherconsciouslydecide to imaginequasi-experiencesthatarenotcompatiblewiththeconcept“BartSimpson”.Oryoucouldjustinadvertently fail tostaywithintheboundariesofwhattheconceptprescribes.But, inany case, if you imagine awhite cartoon character that has a yellowbill andwears asailor shirtwitha redbow tie, thenyouobviously failed to immerseyourself into thequasi-worldof“TheSimpsons”.Therearefurtherlessonstodrawfromthisexample:First, itunderscoresthefactthatnormally,whenwe employour imaginative capacities,we arenot imagining series ofdisconnected figments. Rather, most imaginative processes consist of sequences ofrelatedimaginingsthat,intheirtemporalsuccession,formacoherentwhole.And,giventheexampleabove,itiseasytoseewherethecoherenceiscomingfrom:itisensuredbythebackgroundknowledgethatenterstheimaginationthroughtheconceptswithwhichwe determinewhatwe intend to imagine and fromwhich our imaginative processestake their basic direction. This is not only true when we are asked to imagine BartSimpson.ItisalsotruewhenweareusingourimaginativecapacitiestodecidewhetherthesofaatIkeawillfitthroughthehallwayathome.Inbothofthesecases,theconceptsthroughwhichwedeterminewhatweintendtoimaginerefertochunksofbackgroundknowledgethat,inturn,putconstraintsonhowourimaginingcanevolve.Tobesure,wecould always break the coherence by imagining quasi-experiences that stand in norelation to what has been imagined before. But, given our practical ends at Ikea,imagining a sofa that suddenly transforms into a swarm of butterflies just isn’t therationalthingtodo.What the example of Bart Simpson also shows is that, secondly, frustration is by nomeansimpossibleinthesphereoftheimagination.Otherthanintheperceptualrealm,

13

however,anticipationsandexpectationsarenotfrustratedbyquasi-experiencesalone.Imaginative anticipations are frustrated by way of a comparison between quasi-experiences and the background knowledge that enters the imagination through theconceptswithwhichwe determine the course of our imagining. This is exactly whathappened in the example above: Initially, the content as well as the course of ourimaginative endeavours were determined by the concept “Bart Simpson”. And it wasthrough reflection on the background knowledge to which the concept refers thatcertainquasi-experiencescouldbesingledoutasbeingatoddswithwhattheconceptprescribes.Hence, thereturnsouttobeacloserelationshipbetweentheconditionsoffrustrationontheonehandandtheconceptsthatdeterminethecourseofourimaginingon theother.Given thisrelationship,wecan formulatean importantconclusion: Ifwewanttolearnabouttheconditionsoffulfillmentandfrustrationinagivenimaginativeprocess, we have to look closely at the concepts through which the horizontalanticipationsaredetermined.8What I have tried to do so far in this section is take some first steps towardsunderminingtheviewaccordingtowhichtheimaginationisepistemicallyineptduetothelackofexternalconstraintstotheconditionsoffulfillmentandfrustration.Followingmy analysis, the imagination indeed gives us the freedom to produce random quasi-experiencesthataren’trelatedtoearlierimaginings.However,itisalsopossibletouseourimaginativecapacitiesdifferently:Wecanchoosetoimmerseourselvesinaquasi-world by staying within the boundaries that are prescribed by the concepts throughwhichwedeterminewhatweactuallywishtoimagine.Andifwedoso,theconditionsoffulfillment and frustration are externally fixed; they are fixed by the backgroundknowledge that enters the imagination through the concepts that give our imaginingstheirbasicdirection.Assumeforthemomentthattheforegoinganalysisiscorrect.Evenifitis,however,onecouldstillwonderwhetheritreallycaptureswhatistrulycharacteristicofscientificTEs.One could argue as follows: It may be true that the conditions of fulfillment andfrustration are relatively well defined as long as we use the imagination in a fairlyconservativefashionbystickingtoconceptssuchas“BartSimpson”or“bike”.Conceptslike these prescribe clear rules of anticipation because they refer to ratheruncontroversial chunks of empirical background knowledge that have been acquiredthrough common experiential sources. In the case of “Bart Simpson”, for instance,weknowwhatto imaginebecausewehavespentmanyhourswatchingtherespectiveTVshow.But isn’t it characteristicofTEs toemployour imaginativecapacities inamuchmoreexceptionalmanner?Of course, at firstglance, scientificTEsalso seemtoutilizefairly conventional concepts such as “bucket”, “car”, “train”, “cat” or “tower”. But thepoint ofmost scientificTEs is to imagine theseobjectsunder conditions towhich themore common parts of our background knowledge do not apply. Our commonbackground knowledge may prescribe clear rules of anticipation if, for instance, weimagineacarthatisrushingtowardsagaragewith100km/h.Butwhatdeterminesthe8This,ofcourse,isnotonlytrueoftheimagination.Supposethatyouseeanobjectthatlookslikeabarn.Ifyouapplytheconcept“barn”totheseenobject,thenyouwillanticipatetheobjecttohaveabacksidethatresemblesitsfacingside.Accordingly,theexperienceofabacksidethatdoesn’tresemblethefacingside(forinstance,abacksidethatischaracteristicoffakebarns)willfrustrateyourinitialintention.If,ontheotherhand,youapplytheconcept“fakebarn”totheseenobject,abacksidethatdoesn’tresemblethefacingsideisexactlywhatyouanticipate.Hence,acloserelationbetweenconceptsandtheconditionsoffulfillment and frustration is not only to be found in the sphere of the imagination, but also in theperceptualrealm(cf.alsoHopp2011,chapter2.1).

14

conditions of fulfillment and frustration if we are imagining a car rushing toward agaragewith90%ofthespeedoflight?Evenacursory lookat thepracticeofTE-reasoning insciencereveals thatTEs indeedemployourimaginativecapacitiesinaquiteexceptionalway.InorderformostscientificTEstobeperformed,wehavetoimmerseourselvesinquasi-worldsthatdifferfromouractualworldinmoreorlessdrasticways.Asthefollowingexamplesshow,“immersion”consistsintheactivebracketingand/ormodificationofcertainpartsofourbackgroundknowledge:Immersionintothequasi-worldofGalileo’sshipconsists inbracketingthebackgroundknowledgeaccordingtowhichshipsatseaarealwayssubjectedtorockingmotions. Immersion into the quasi-world of Stevin’s chain consists in bracketing thebackground knowledge according to which objects moving down inclined planes arealways subjected to kinetic friction and air drag. Immersion into the quasi-world ofNewton’sbucketconsists inbracketingthebackgroundknowledgeaccordingtowhichthematerialuniverseconsistsofmoreobjectsthanjustawaterbucketandarope.As we shall see in the next section, the ability to be selective with regard to thebackground knowledge that determines the content as well as the course of ourimaginings is indeed crucial for TEs in science. In the present context, however, thisability is still a reason for concern. Our problem was this: In standard cases ofimaginativeactivity,theconditionsoffulfillmentandfrustrationaredeterminedbytherelevantpartsofourbackgroundknowledge thathavebeenacquired in the courseofpreviousencounterswiththeworld.Butwhatdeterminestheconditionsof fulfillmentand frustration in cases where the imagination takes us well beyond our ordinaryexperientialgrasp?Theobviousanswer is thatmuchofwhatweanticipateandexpect inscientificTEs isdetermined by the theories that work in the background of these TEs. Theaforementionedcar/garage-TEfromSpecialRelativityisacaseinpoint:Ifweareaskedtoimagineacarthat isrushingtowardsagaragewith90%ofthespeedof light, then,obviously, our anticipations and expectations aren’t guided by previous perceptualencounterswiththiskindofscenario.Whatdeterminesthecourseofourimaginingsisrather a theory according to which the measured length of an object decreasesnoticeablyatvelocitiesclosetoc.9Thepointisevenmoreobviousconsideringexamplessuch asHeisenberg’s gamma-raymicroscope or EPR: In cases like these, not only thecourse, but even the content of our imaginings is almost entirely determined bytheoretical components of our background knowledge. To put it bluntly: We justwouldn’tknowwhattoimagineifwedidn’tknowathingaboutquantummechanics.Many constraints that impinge on our imaginative activities are due to empiricalcomponents of our background knowledge.Others are due to theoretical parts of ourbackground knowledge. This, in a nutshell, is the upshot of this section. However,phenomenologicallyconstrued, there isyeta third typeofconstraints thatareneitherempirical nor theoretical in nature. Here is an example that might help to make my

9Oncloserinspection,thecar/garage-TEisparticularlyinterestingbecauseitisnotatallclearwhichpartofSpecialRelativityshouldactuallydeterminewhatwearesupposed to imagine in thisTE:On theonehand,SpecialRelativitytellsusthatobjectsmovingwithvelocitiesclosetocareLorentz-contracted.Ontheotherhand,weknowsincethe1930iesthatLorentz-contractedobjectswouldnotappearcontracted,but rotated (Lampa 1924; Terrell 1959; Penrose 1959). This raises the question as to whether it isrelevantforthesuccessofTEstoimagineobjectsrealistically(cf.Brown2013).

15

point:SupposethatIaskyouto imagineabrickthat isthrownagainstawindowpane.And suppose furthermore that I ask you to bracket everything you know about thebehaviourofbricksandwindowpanes.Hence,whatIaskyoutoimagineisaquasi-worldin which bricks and windowpanes look like actual bricks and windowpanes, but inwhich their behaviour is absolutely unpredictable.Now the question is this: Are youranticipations and expectations concerning the behaviour of the brick and thewindowpaneindeterminate?Aretheconditionsoffulfillmentandfrustrationentirelyuptoyou?At first sight, this questionmight seem odd. If you are serious about bracketing yourbackground knowledge about bricks and windowpanes, then, apparently, there isnothing left that could possibly determine your anticipations and expectations in thescenariodescribedabove.Youcouldimagineaquasi-worldinwhichwindowpanesareshattereduponimpact.Butyoucouldalsoimagineaquasi-worldinwhichitisthebricksthat crumble todustwhenever theyhitawindowpane.Oryoucould imagineaquasi-worldinwhichwindowpanestransmuteintospraysofflowerswhenevertheyarehitbybricks. In short: After bracketing your background knowledge, no anticipation is toooutlandishtobeinprincipleunfulfillablebymatchingquasi-experiences.Orso,atleast,itseems.InfactIthinkthisviewisdescriptivelywrong.Therearerestrictionstowhatwecanorcannotimagine,evenafterwehavebracketedourentirebackgroundknowledgeabouttheobjectsinquestion.Consider,forinstance,theanticipationofawindowpanethatiscompletely shattered and completely unshattered, both at the same time. As you caneasily verify by trying to actually imagine this scenario, it is impossible to intuitivelyfulfill this anticipation. Or consider the anticipation of a brick that is red and blue allover. Here too, we fail to imagine a fulfilling quasi-experience. Or consider theanticipation of a brick that is seen from all sides at once or whose redness can besmelled. Again, it is impossible to imagine a quasi-world inwhich these anticipationswouldbefulfilledbymatchingquasi-experiences.Inallofthesecasestheproblemisnotthataparticularquasi-experiencecontradictsaruleofanticipationthatsuggestsitselfinthelightofpreviousencounterswiththeactualworldorinthelightofatheory.Rather,theproblemisthatcertainquasi-experiencesareimpossibletoimagine,nomatterhowhardwetry.Inalloftheaforementionedcaseswehavetorealizethatthe“freedom[ofvoluntarilyproducingquasi-experiences]islimitedinsofarasessentiallawsofpossiblequasi-fulfillment are inherent here within the boundaries of the unity of an identicalpossible objectivity understood as intentional and still indeterminate” (Husserl 2005,671). Hence, the content and course of our imaginings is not only determined byempirical and theoretical components of our background knowledge.Phenomenologicallyconstrued,our imaginingsarealsoconstrainedby“necessarylawswhichdeterminewhatmustnecessarilybelong toanobject inorder that it canbeanobjectofthiskind”(Husserl1973,352).Ina recentpaper,WalterHopphasdrawnparticularattention to the role thiskindofessentialknowledgeplays inTE-reasoning.OneofHopp’smainthesesis“thatthoughtexperiments,at theirbest,are in fact foundedonactsof fulfillment inwhichwe intuituniversals and the relations among them, and that the actual instantiation of thoseuniversalsandrelationsisimmaterial”(Hopp2014,81).Inadvancingthisclaim,HoppemphasizesthesimilaritiesbetweenBrown’sPlatonismandaphenomenologicaltheoryofTEs.AccordingtoHopp’sanalysis,bothaccountsagreeintheiracknowledgmentofa

16

knowledge-yieldingfacultythatisindependentofthefivesensesandthatisoperativeinatleastsomeTEs.10What Ihave tried todo in this section is toundermine theviewof the imaginationasessentially unrestricted. If my analysis is correct, then our imaginative capacities aresubject to three types of constraints: First, there are constraints that are due to theempiricalpartsofourbackgroundknowledge.Secondly,thereareconstraintsthatresultfrom the theories thatwehave incorporated intoourbackgroundknowledge.Thirdlyand finally, there are constraints that aredue to “essential lawswhich govern acts asintentional experiences, in all their modes of sense-giving objectivation, and theirfulfillingconstitutionof‘truebeing’”(Husserl2001b,319).

5. PuttingthePiecesTogetherWiththeabovefindingsinplace,wecanfinallyturntotheheartofthispaper.Drawingon the results from the previous sections, I will now take a first stab at aphenomenologicaldescriptionofTE-reasoninginscience.Onmyview,TE-reasoningisaprocess thatoccurs in threestages: thepreparatorystage, theperformancestage,andtheconclusionstage.Iwillcommentoneachofthesestagesinturn.5.1ThePreparatoryStageBeforewecanevenbegintoperformaTE,weneedtoknowacoupleofthings.Firstofall,andmostobviously,wemustknowtowhatendtheTEisperformed.ThisistosaythatweneedasufficientlycleargraspofthetargetthesisthattheTEismeanttorefute,to corroborate or to clarify. Secondly, and no less importantly, we need informationaboutthekindofscenariowearesupposedtoimagine.Thisinformationisencapsulatedin the TE-narrative in which the details of the TE-setup is specified. Of course, TE-narrativesdiffersignificantlywithrespecttodetail,complexityandstyle.Someofthemare text-only. Others alsomake use of visualizations. Some of them are colourful andfiligreed. Others are strictly technical and reduced to the bare essentials. In any case,however, TE-narratives pursue a twofold task:On the one hand, TE-narratives tell ussomethingabouttheobjectsofwhichtheTE-scenarioiscomposed.Ontheotherhand,TE-narrativesspecifythecharacteristicsofthequasi-worldsinwhichTE-scenarioistobe imagined. Consider, for instance, the following excerpt from aTE-narrative from astandardphysicstextbook:

“Amanhasanl=5mlonggarageandbuysandl0=7mlongcar(properlengthof the car). He reasons that, if he drives sufficiently fast, the carwill fit in thegarageduetolengthcontraction(ignorethefactthatheisgoingtoruinhisnewcarbysmashingitagainstthegaragewall).”(Faraoni2013,25)

10It must be noted, however, that Hopp is well aware of the differences between a phenomenologicaltheoryofTEsandBrown’s(cf.Hopp2014,89-90).WhilephenomenologistswilltypicallynotbeatoddswiththerationalistpartofBrown’sstory–particularlywithhisclaimthatsenseexperienceisnotthesolesourceofknowledgeabouttheworld–,theywillmostcertainlytakeissuewithhisconstrualofthelawsofnatureandwithhisPlatonisttwo-worldontology.Hence,buildingonDavies’usefuldistinctionbetweenextreme and moderate deflationism on the one hand and (two kinds of) moderate and extremeinflationismontheother(Davies2007,37-42),phenomenologypositionsitselfsomewherebetween(thesecondkindof)moderateandextremeinflationism.

17

Inorder tounderstand thisTE-narrative,wemusthavedifferent typesofbackgroundknowledge at our disposal: While concepts such as “car” or “garage” refer toconventional chunks of empirical background knowledge, concepts such as “properlength”or“lengthcontraction”refertoinformationaboutthekindofphysicsthatformsthetheoreticalframeworkoftheTE.ItisonlyifthesepiecesofinformationarepartofourgeneralbackgroundknowledgethatweareabletosetuptheTEaccordingly.If,forinstance,knowledgeaboutLorentzcontractionis lacking,thenweinevitablyfail tosetuptheTEaccordingtotheintentionsofitspresenter.ButTE-narrativesnotonlygiveusinformationaboutthekindsofobjectsandprocessesthatweneedtoimagineinordertocarryouttheexperiment.TE-narrativesalsocontaininformationabout thequasi-world inwhich theTEought tobeperformed.While thisinformation isonly implicit inmanycases, theabove-quotedTE-narrativehintsat thespecifics of the required quasi-world by remarking that no actual car would remainintactunderthe imaginedconditions.Ofcourse,weknowthatnoactualbrakesystemcoulddeceleratethecarfromvelocitiesclosetoc.Andweknowthatthecarwouldmeltinlessthanablinkofaneye.Buttheseandsimilarothercomponentsofourbackgroundknowledgemustbebracketedifwewanttoimmerseourselvesintothequasi-worldthatsuitstheTEweareabouttoperform.LikemanyotherTEsinphysics,thecar/garage-TEjustwouldn’tworkinaquasi-worldthatresemblesouractualworldtooclosely.TheTEonly works in an idealized quasi-world in which many factors that determine thebehaviouroftheactualworldareneglected.5.2ThePerformanceStageThe purpose of a TE-narrative is to prime its audience for the performance of theensuing TE. The main tool for doing so is the concepts of which the narrative iscomposed.TheseconceptsrefertothekindofbackgroundknowledgethatisnecessaryinorderfortheTEtobesetup.It isonlyiftherightkindofbackgroundknowledgeisalreadyinplacethatweareinapositiontoimaginetheTE-setupaswellasthequasi-worldinwhichtheTEoughttobeembedded.Hence,IfullyagreewithDavidGooding’sobservation that “[a] TE becomes possible [only] when a world is sufficiently well-represented that experimental procedures and their likely consequences can bedescribedwithinit”(Gooding1992,281).WhetherornotaTEmanagestoleadustoitsdesiredoutcome largelydependsonour familiaritywith thequasi-world inwhich theTEoughttobeembeddedaccordingtoitspresenter.11

11ThispointalsohasimplicationsforthehistoriographyofTEs.AshistoricallyinclinedphilosopherssuchasJamesMcAllister(1996)orPaoloPalmieri(2005)haveemphasized,itiscommoninthephilosophicalliterature to make use of the history of TE-reasoning in a rather idiosyncratic, sometimes even“cartoonish” (Palmieri 2005, 223) manner. Not only that case studies are presented without payingsufficientattention to theircontext;manyphilosophersalsoapproachthehistoryofTE-reasoning inanovertly“presentist”or“Whiggish”way.InthecaseofGalileo’sfamoustower-TE,forinstance,ithasbeenobjectedthatitisonlyifwetoday“lookatthisthoughtexperimentfromanhistoricallydistantperspectiveandwith the knowledge ofmodern physics concerning falling bodies in a vacuum [that] the inferencefromthecontradictiontothe‘right’conclusion[…]seemstobeimmediateanduntutoredbyanyempiricalorlogicalreasoning”(Brendel2004,95;cf.,forasimilarcomplaint,Norton1996,344-345).Whatseemsrighttomeaboutthisobjectionisthealmosttrivialtruththatwecannotpresupposeourcontemporarybackground knowledge if we wish to determine the epistemic weight that a TE carried in its originalhistorical context. However, I also believe that the framework of phenomenology is particularly well-suitedtodrawameaningfuldistinctionbetweentheepistemicweightaTEcarriesforus,relativetoourmodern background image of the world, and the epistemic weight a TE originally carried for a given

18

Assume that we have successfully completed the preparatory stage in our currentexample,thecar/garage-TE.Thisistosaythatwehaveasufficientlycleargraspofthetarget thesis,wepossess the rightkindofbackgroundknowledge andweunderstandthespecificsofthequasi-worldinwhichtheTE-scenariooughttobeimagined.WearenowinapositiontoimaginetherequiredTE-setup.Ifwedoso,thentheprimaryobjectof our intentional directedness is a scenario that consists of an imagined car that isabouttorushtowardsanimaginedgarageataspeedclosetoc.However,onthebasisofour earlier reflections on the horizonal structure of intentionality, it is clear that theimagined car and the imagined garage is by no means all that is given to us in thissituation.Co-givenwiththequasi-experienceoftheprimaryobjectisanarrayofemptyintentions that point towards possible future states of the imagined TE-scenario. InimaginingtheTE-setup,asitislaidoutintheTE-narrative,weautomaticallyco-intendahorizon that prescribes anticipatory rules concerning the course of further quasi-experiences. These rules of anticipation aremotivated by the background knowledgethat isnecessary for imagining theTE-setup in the firstplace.And it is through theseanticipatoryrulesthatourbackgroundknowledgerestrictshowtheimaginedscenariocouldevolveifthecorrespondingquasi-experienceswerestilltoqualifyasexperiencesoftheinitialscenario.Thepoint I am trying tomake is this: Ifwe imagineaparticularTE-setup,wearenotmerelyimaginingaparticulararrangementofimaginedobjectsthatareembeddedinaparticularquasi-world.Sincetheobjectsofourintentionaldirectednessarenecessarilygiveninhorizons,wealsoco-intendrulesofanticipationthatrestricthowtheTE-setupcouldevolveiftheensuingphaseswerestilltoqualifyasphasesoftheinitialsetup.Thisway of putting things alsomakes clear what it actually means to carry out a TE: ToperformaTEistoimmerseoneselfintoaparticularquasi-world,toimagineaTE-setupwithinthisquasi-worldandthentolettheTE-setupevolveaccordingtotheanticipatoryrulesthatareco-givenwitheachandeveryquasi-experienceoftheunfoldingTE.Or,toput it in slightlydifferent terms:ToperformaTE is to “live through” theanticipatoryhorizonsagainstthebackgroundofwhichfulfilledintentionstowardsparticularphasesof the imaginative process always stand out.12With each new phase a new horizoncomes to fruition, “a new system of determinable indeterminacy, a new system ofprogressive tendencieswithcorrespondingpossibilitiesofentering intodeterminatelyorderedsystemsofpossible[quasi-]appearances,ofpossiblewaysthattheaspectscanrun their course, together with horizons that are inseparably affiliated with theseaspects”(Husserl2001c,43).Inthiswholeprocessof“livingthrough”theanticipatoryhorizons inwhicheachphaseof the imaginativeprocess isgiven,processesofexplicitinferentialreasoningaremostlyabsent.Thisexplainsoneofthemostintriguingfeaturesof scientific TEs, namely the ease with which they are performed and the apparenteffortlessnesswithwhichwereachthedesiredoutcome.According to view developed so far, the way the initial TE-setup pans out ispredelineatedbytherulesofanticipationthatcomewitheverynewquasi-experienceofthe unfolding TE. This, however, raises an obvious question: At any point during theperformance of a TE, there are far more empty intuitions towards future quasi-

historicalcommunity.Cf.,foracoupleoffirststepstowardsemployingthephenomenologicaltoolofthe“epoché”indealingwithsciencehistory,Arabatzis(2012)andPalmieri(thisvolume).12ThisechoesGooding’sremarkthat“[t]oexplaintheforceofanexperimentithelpstounderstanditasaprocesstobeworkedthrough,ratherthanasalogicalstructure”(Gooding1992,283).

19

experiences than are relevant for the actual outcome of the TE. If we imagine a carrushing towards a garage, for instance, we implicitly anticipate the car to have abackside. Or we anticipate the garage to have a specific color. But, obviously,anticipationssuchastheseareentirelyirrelevantforwhattheTEissupposedtoshow.Hence, the question arises as to how relevant anticipations are distinguished fromirrelevantones.I opened this sectionwith the remark that the aim of TE-narratives is to prime theiraudience for theensuingTE. In the lightof thequestionthathas justbeenraised, thisremarkmust be further clarified:On the one hand, TE-narratives tell uswhatwe aresupposed to imagine. They do so by employing particular concepts that activateparticular components of our background knowledge. Anticipatory rules that aremotivatedbythesecomponentsthenrestricthowtheensuingTEcanunfold.Butontheotherhand, andequally important,TE-narrativesalsogiveus informationaboutwhatwearenotsupposedto imagine(cf.Davies2007,35). Inpart this isdonethroughthetarget-thesis that automatically narrows our focus to certain aspects and leaves outothers.Butitisalsodonethroughthedeterminationofthequasi-worldinwhichtheTEmustbeembedded.Bybracketingcertainpartsofourbackgroundknowledgeinorderto immerse ourselves in a particular quasi-world, we prevent certain rules ofanticipationfromevenbecomingoperative.Ournormalanticipationsconcerningobjectsslidingdowninclinedplanes,forinstance,aredeterminedbythebackgroundknowledgeaccording towhich such objects are always subjected to kinetic friction and air drag.However, as reliable as this background knowledge may ever be under standardcircumstances, the resulting rules of anticipation are entirely impractical for theperformanceofTEslikeStevin’schain.Hence,inordertolettheimaginationinsuchTEsbe guided by the right kinds of anticipatory rules, parts of our common backgroundknowledgemust be bracketed already at the outset. Thus construed, bracketing is anessentialtoolforchannellingouranticipationsindesiredways.The previous remarks on the role of bracketing also allowme to finally explicatemythesis according to which TEs should be understood as anticipation pumps.13On theviewproposedhere,TE-narrativesarewelldesignediftheyaccomplishtwoobjectives:First,theymusttriggertherightkindofbackgroundknowledgeinordertomotivatetheright kinds of anticipatory rules. And secondly, they must ensure that these rules ofanticipation aren’t interfered by anticipations that are inessential or even animpedimentforreachingthedesiredoutcomeoftheimaginativeprocess.Hence,toputit ina slogan:ATE iswell-designed if theTE-narrative is successful in “pumping” therightrulesofanticipations.The“right”rulesarethosethat leadusfromtheinitialTE-setuptothedesiredendpointwithaminimumofunnecessarydistractions.5.3TheConclusionStageToperformaTEmeanstosuccessfully“livethrough”theanticipatoryhorizonsagainstthe background of which fulfilled intentions towards particular phases of theimaginativeprocessalwaysstandout.TheperformancestageofaTEiscompletedwhen

13Of course, in using this terminology, I take a cue from Daniel Dennett who famously dubbed TEs“intuitionpumps”.However,myagreementwithDennett’sviewsisrathersuperficial.WhileIagreethatthehallmarkofawell-functioningTE is that it is“cunninglydesignedto focus thereader’sattentionon‘theimportant’features,andtodeflectthereaderfromboggingdowninhard-to-followdetails”(Dennett1984,12),IdonotsharethedismissiveattitudethatischaracteristicforDennett’searlytakesonTEs.

20

we reach that state of the TE-scenario which the presenter deems relevant for theprojected target-thesis. Like in the case of physical experiments, we can now askwhether the outcomeof theTE supports the conclusion that its presenter draws. Forinstance: Do the changing states of a water-bucket-system in an otherwise emptyuniversereallyforceustoaccepttheexistenceofabsolutespace?Doesalightraythatentersanelevatorhorizontally really forceus toaccept that theeffectsof gravityandinertialaccelerationareindistinguishable?But scrutinizing the relationbetween theoutcomeof the imaginativeprocess and theprojectedtargetthesisisnotallthathappensattheconclusionstage.Inmanycasesitisalso natural to reflect on whether the course of the imaginative process itself wasinevitable, independently from its purported impact on the target thesis. Here is anexamplethatmightillustratethepoint:ConsideramodernversionofGalileo’stower-TEinwhichwe imagineacannonball that is tied toamusketball.Weare thenasked tothrowthecombinedsystemfromatowerandtoseewhathappens.Ifthespeedoffallofbodies isreallyproportional to theirweights,as theAristotelian theorysuggests, thenthecombinedsystemwillfallfasterandslower.ThisoutcomeisusuallytakentoshowthattheAristotelianlawoffallingbodiesisfalseandmustbereplacedwithGalileo’s.Now,criticalreflectiononthisTEcanoccurontwodifferentlevels:14Ontheonehand,itisnaturaltowonderwhetherbothconclusions–therejectionoftheAristotelianlawandthe proclamation of the Galilean law – are equally supported by the outcome of theimaginative process. Yet, on the other hand, it is also possible to ask whether theoutcome itself is inevitable. Is it really the case that the anticipations in this TEunavoidablyleadustotwoconflictingscenariosinwhichthecombinedsystemfallsbothfasterandslower?Ordoour anticipations varydependingonhow tightly themusketballandthecannonballareconnected?Doouranticipationsvarydependingonwhetherthetwoobjectsareconnectedwitharopeorwitharubberband?Canacannonballthatis connected with a musket ball really be treated as one object, as the TE seems topresuppose (cf., for versions of these and similar other objections, Koyré 1968, 51;Gendler1998,404-406)?Suchquestionsobviouslydonotconcerntherelationbetweenthe outcome of TE-process and the projected target thesis. Rather, such questionsconcerntheTE-processitself.How can questions concerning the inevitability of the TE-process be resolved? Givenwhat has been said so far, an answer to this question goes along the following lines:SincetheevolutionofaTE-setupisdeterminedbytherulesofanticipationthatgovernthe TE, scrutinizing the outcome of the TE-process can only proceed by way ofexplicating the background knowledge that initially motivated the relevant rules ofanticipation.Duringtheperformancestage,i.e.duringtheprocessof“livingthrough”thehorizons inwhich thephasesof theTEaregiven, thisbackgroundknowledge is tacit.However, ifwewishtocriticallyreflectonthestrengthwithwhichtheoutcomeoftheTE-processimposesitselfonus,therelevantpartsofourbackgroundknowledgemust

14The two levels Iamreferring tohere resembleBrown’sdistinctionbetweenexperiment in thebroadsenseandexperimentinthenarrowsense:“Inthenarrowsense,anexperimentincludesthesetupandthe observation […]. In the broad sense, the experiment includes background assumptions and initialtheorizing, the setup, observation, additional theorizing, calculating, and drawing the final conclusion.”(Brown2007,157-158)WhatIcalltheleveloftheTE-processisroughlysimilartoBrown’sexperimentinthenarrowsense.

21

be made explicit.15The strength with which the outcome of a particular TE-processimposes itself on us will depend, among other factors, on the type of backgroundknowledgethatmotivatestherelevantrulesofanticipation:If,forinstance,aparticularset of anticipations ismotivatedby a theory that operates in thebackgroundof aTE,then the degree towhichwe accept the outcomeof theTEdepends on the degree ofbelief in the underlying theory. If, on the other hand, a particular set of anticipationsturns out to be determined by an essential law, then the outcome of the imaginativeprocesswillberegardedasnecessary.Let me conclude this section by briefly commenting on one last issue: On the viewproposedhere,toperformaTEisto“livethrough”therulesofanticipationthatareco-given at each stage of the imaginative process. These rules are motivated by thebackgroundknowledge that is triggeredduring thepreparatory stage, i.e. through theconcepts of which the TE-narrative is composed. Hence, the course as well as theoutcome of a TE is ultimately determined by chunks of pre-established backgroundknowledge.Butthisraisesanobviousquestion:HowcanwelearnsomethingnewfromTEs if their outcome isdeterminedbybackgroundknowledge thatmust alreadyhavebeen acquired before the TE can even be set up? Isn’t it a consequence of the viewdefendedhere that theperformanceofaTE is themererecallingofwhatwasalreadyknown?SincescientificTEsareaverydiverse lot, it is, I think, impossible togiveaprincipledanswertohowTEsmanagetogobeyondtheknowledgeonealreadyhastopossessinordertoperformtheTE.ItisthusmainlyforthesakeofbrevitythatIwillreducemyselftoonesingleaspectthatseemstobeofparticularimportanceinthiscontext:AsIhavepointedout,TEsrequireustoimmerseourselvesintoquasi-worldsthatdifferfromtheactual world in more or less drastic ways. One of the reasons why immersion isepistemically significant is that it gives us the opportunity to test our backgroundknowledgeunderconditionsthatcouldnotbereplicatedotherwise.Ofcourse,thereisalsoadangerinprojectingquasi-worldsthatarefarremovedfromtheactualworld:Themoreoutlandish thequasi-world, thehigher the risk thatour anticipations eventuallybecomeungrounded.Butoneofthethingsawell-designedTEaccomplishesbywayofprojecting idealized quasi-worlds is to provoke collisions between parts of ourbackgroundknowledgethatwouldnotevencomecloseundernormalcircumstances.16ItisinthiswaythatTEsallowustodetectandgetridofinconsistenciesinourexistingbackgroundknowledge, toexplicateandscrutinizebackgroundassumptionsthatwerepreviously left unquestioned and, in some particularly impressive cases, even to

15It shouldbenoted,however, that theexplicationof tacitbackgroundknowledgeusuallycomesat thepriceofdiminishingthecognitiveefficacyandeleganceofagivenTE.ThispointhasbeenemphasizedbyDavidGooding(1992,286).16Anexamplemighthelpmake thispointmorevivid.Asnotedearlier (cf. footnote6), ourbackgroundimage of the world is in part composed of pieces of non-propositional, practical and sensorimotorknowledgeabouthowthingsaredoneandhowweuseourbodiestodothem.At firstblush,onemightthinkthat,sinceTEsareperformedinthelaboratoryofthemind,implicitsensorimotorandkinaestheticknowledgeisirrelevantforthepracticeofTE-reasoning.However,asGooding(1992)andYiftachFehigeand I (2013)haveargued, this is far frombeing thecase.WhileGoodingemploysabroadlynaturalisticframework to make this point, Fehige and I have approached the issue from a phenomenologicalperspective. The aim of our paper was to a) show that implicit body knowledge is operative inmanyinstancesofTE-reasoningandtob)illustratethisclaimbymeansofananalysisofNewton’sbucket-TE.

22

reconfigure theconceptualapparatuswithwhichweapproachtheworld(Kuhn1977;Gendler1998).

6. ConcludingRemarksThegoalofthispaperwastosetoutsomebasicdirectionsinwhichaphenomenologicalframework for the analysis of scientific TEs should be developed. On my view,phenomenologydiffersfromtheexistingapproaches,amongotherthings,initsstartingpoint: Themain objective, at least initially, is not to construct a theory that seeks tobring TEs in line with certain pre-established epistemological, ontological andmetaphysicalviews.Themainobjectiveofaphenomenologicalaccountisrathertogivea faithfuldescriptionof theactualperformance ofTEs froma first-personperspectiveandtogoonfromthere.Oneoftheaimsintheprecedingsectionswastoindicatehowsuch a descriptionmight actually look like. But, of course,muchwork remains to bedoneinordertopositionphenomenologyasaseriouscontenderintheongoingdebateonTEs.17

7. LiteratureArabatzis,Theodore(2012):“HiddenEntitiesandExperimentalPractice:RenewingtheDialogueBetween History and Philosophy of Science”, in: Mauskopf, Seymour & Schmalz, Tad (eds.):IntegratingHistoryandPhilosophyofScience.ProblemsandProspects.Dordrecht:Springer,125-140.Brendel, Elke (2004): “Intuition Pumps and the Proper Use of Thought Experiments”, in:Dialectica58/1,89-108.Brown,JamesRobert(2004):“WhyThoughtExperimentsTranscendEmpiricism”,in:Hitchcock,Christopher(ed.):ContemporaryDebatesinPhilosophyofScience,Malden:Blackwell,24-43.Brown, JamesRobert (2007): “CounterThoughtExperiments”, in:RoyalInstituteofPhilosophySupplement61/82,155-177.Brown, JamesRobert (22011):TheLaboratoryoftheMind.ThoughtExperiments intheNaturalSciences,London&NewYork:Routledge.Brown, James Robert (2013): “What Do We See in a Thought Experiment?”, in: Frappier,Melanie, Meynell, Letitia & Brown, James Robert (eds.): Thought Experiments in Science,Philosophy,andtheArts,London&NewYork:Routledge,53-68.Chudnoff,Elijah(2013):Intuition,Oxford:OxfordUniversityPress.Crowell,StevenGalt(2006):“HusserlianPhenomenology”,in:Dreyfus,Hubert&Wrathall,Mark(eds.):ACompaniontoPhenomenologyandExistentialism,Malden:Blackwell,9-30.Davies, David (2007): “Thought Experiments and Fictional Narratives”, in:Croatian JournalofPhilosophy19,29-45.Dennett, Daniel C. (1984):TheElbowRoom.TheVarietiesofFreeWillWorthWanting, Oxford:ClarendonPress.Fehige, Yiftach & Wiltsche, Harald (2013): “The Body, Thought Experiments, andPhenomenology”,in:Frappier,Melanie,Meynell,Letitia&Brown,JamesRobert(eds.):ThoughtExperimentsinScience,Philosophy,andtheArts,London&NewYork:Routledge,69-89.Faraoni,Valerio(2013):SpecialRelativity,Springer:Dordrecht.Foster,John(2000):TheNatureofPerception,Oxford:OxfordUniversityPress.

17Indevelopingmyviewsoverthelastcoupleofyears,IprofitedfromtheconversationsIhadwithJimBrown, Keizo Matsubara, Danny Goldstick, Sonja Rinofner-Kreidl, Walter Hopp, Mike Stuart, YiftachFehige,GeordieMcComb,HanneJacobs,SebastianLuft,JeffHilderley,GeorgeHeffernan,MichaelWallnerand James McGuirk. I would also like to say “thanks” to the Austrian Science Fund for the generoussupportofmyongoingresearchonTEs(projectnumber:J3114).

23

Froese, Tom & Gallagher, Shaun (2010): “Phenomenology and Artifical Life: Toward aTechnological Supplementation of PhenomenologicalMethodology”, in:HusserlStudies26, 83-106.Gendler, Tamar Szabó (1998): “Galileo and the Indispensability of Scientific ThoughtExperiment”,in:BritishJournalforthePhilosophyofScience49,397-424.Gooding,DavidC.(1992):“WhatisExperimentalaboutThoughtExperiments?”,in:ProceedingsoftheBiennialMeetingofthePhilosophyofScienceAssociation.Volume2,Chicago:UniversityofChicagoPress,280-290.Gunstone, Richard & White, Richard (1981): “Understanding Gravity”, in: Science Education65/3,291-299.Hopp, Walter (2011): Perception and Knowledge. A Phenomenological Account, Cambridge:CambridgeUniversityPress.Hopp,Walter(2014):“ExperimentsinThought”,in:PerspectivesonScience22,76-97.Horowitz, Tamara & Massey, Gerald J. (eds.): Thought Experiments in Science and Philosophy,Lanham:Rowman&Littlefield.Husserl,Edmund(1960):CartesianMeditations.AnIntroductiontoPhenomenology,translatedbyDorionCairns,TheHague:MartinusNijhoff.Husserl, Edmund (1973): Experience and Judgement. Investigations in a Genealogy of Logic,translatedbyJamesS.ChurchillandKarlAmeriks,Evanston:NortwesternUniversityPress.Husserl,Edmund (1983), IdeasPertainingtoaPurePhenomenologyandtoaPhenomenologicalPhilosophy.FirstBook,translatedbyFredKersten,TheHague,MartinusNijhoff.Husserl, Edmund (1997):ThingandSpace:Lecturesof1907, translated by Richard Rojcewicz,Dordrechtetal.:Kluwer.Husserl,Edmund(2001a):LogicalInvestigations.Volume1,translatedbyJ.N.Findlay,London&NewYork:Routledge.Husserl,Edmund(2001b):LogicalInvestigations.Volume2,translatedbyJ.N.Findlay,London&NewYork:Routledge.Husserl, Edmund (2001c): Analyses Concerning Active and Passive Synthesis. Lectures onTranscendentalLogic,translatedbyAnthonySteinbock,Dordrechtetal.:Kluwer.Husserl,Edmund(2005):Phantasy,ImageConsciousnessandMemory(1898-1925),translatedbyJohnB.Brough,Dordrecht:Springer.Kind,Amy(2013):“TheHeterogeneityofImagination”,in:Erkenntnis78/1,141-159.Koyré,Alexandre(1968):“Galileo’sTreatise‘DeMotuGravium’:TheUseandAbuseofImaginaryExperiment”, in:Metaphysics andMeasurement: Essays in the Scientific Revolution, Cambridge:HarvardUniversityPress,44-88.Kuhn, Thomas S. (1977): “A Function for Thought Experiments”, in: The Essential Tension.SelectedStudiesinScientificTraditionandChange,Chicago:UniversityofChicagoPress,240-265.Kujundzic,Nebosja (1995): “ThoughtExperiments:ArchitectureandEconomyofThought”, in:JournaloftheBritishSocietyofPhenomenology26/1,86-93.Lampa, Anton (1924): “Wie erscheint nach der Relativitätstheorie ein bewegter Stab einemruhendenBeobachter?”,in:ZeitschriftfürPhysik27/1,138-148.McAllister,JamesW.(1996):“TheEvidentialSignificanceofThoughtExperimentinScience”,in:StudiesinHistoryandPhilosophyofScience27/2,233-250.Miščević,Nenad(1992): “Mentalmodelsand thoughtexperiments”, in: InternationalStudiesinthePhilosophyofScience6/3,215-226.Mohanty, JitendraNath (1991): “TheMethod of Imaginative Variation in Phenomenology”, in:Horowitz, Tamara &Massey, Gerald, J. (Eds.):ThoughtExperiments in Science andPhilosophy,Savage:RowmanandLittlefield,261-272.Nersessian,Nancy(1993):“IntheTheoretician’sLaboratory”,in:PSA:ProceedingsoftheBiennialMeetingofthePhilosophyofScienceAssociation,Vol.1992,VolumeTwo,291-301.Noë,Alva(2004):ActioninPerception,Cambridgeetal.:MITPress.Norton,John(1996):“AreThoughtExperimentsJustWhatYouThought?”,in:CanadianJournalofPhilosophy26/3,333-366.Norton,John(2004):“WhyThoughtExperimentsdonotTranscendEmpiricism”,in:Hitchcock,Christopher(ed.):ContemporaryDebatesinPhilosophyofScience,Malden:Blackwell,44-66.

24

Palmieri,Paolo(2005):“‘Spuntar loscogliopiùduro’:didGalileoeverthinkthemostbeautifulthoughtexperimentinthehistoryofscience?”,in:StudiesinHistoryandPhilosophyofScience36,223-240.Penrose, Roger (1959): “The Apparent Shape of a Relativistically Moving Sphere”, in:MathematicalProceedingsoftheCambridgePhilosophicalSociety55/1,137-139.Sartre, Jean-Paul (2004): The Imaginary. A phenomenological psychology of the imagination,London&NewYork:Routledge.Siewert, Charles (2007): “In favor of (plain) phenomenology”, in: Phenomenology and theCognitiveSciences6,201-220.Smith,DavidWoodruff&McIntyre,Ronald (1982):HusserlandIntentionality.AStudyofMind,Meaning,andLanguage,Dordrecht,Boston&London:Reidel.Smith,DavidWoodruff(2004):MindWorld.EssaysinPhenomenologyandOntology,Cambridge:CambridgeUniversityPress.Smith,DavidWoodruff(2007):Husserl,NewYork&London:Routledge.Sturgeon,Scott(2000):MattersofMind:Consciousness,reasonandnature,London:Routledge.Terrell,James(1959):“InvisibilityoftheLorentzContraction”,in:PhysicalReview116/4,1041-1045.White,AlanR.(1990):TheLanguageofImagination,Oxford:Blackwell.Wiltsche, Harald A. (2013): “How Essential are Essential Laws? A Thought Experiment onPhysicalThingsandTheirGivennessinAdumbrations”,in:Günzler,Ingo&Mertens,Karl(eds.):Wahrnehmen, Fühlen, Handeln. Phänomenologie imWettstreit der Methoden, Münster: Mentis,421-436.Wittgenstein,Ludwig(1980):RemarksonthePhilosophyofPsychology,Oxford:Blackwell.