A Program for Naturalizing Metaphysics, with Application to the Ontology of Events by Alvin I. Goldman, The Monist, vol. 90, no. 3, pp. . Copyright 2007, THE MONIST, La Salle, Illinois 61301. A PROGRAM FOR NATURALIZING METAPHYSICS, WITH APPLICATION TO THE ONTOLOGY OF EVENTS 1. Metaphysics and Cognitive Science I wish to advance a certain program for doing metaphysics, a program in which cognitivesciencewouldplayanimportantrole.1Thisproposedingredientis absent from most contemporary metaphysics. There are one or two local parts of metaphysics where a role for cognitive science is commonly accepted, but I ad-vocateawiderrangeofapplication.Ibeginbylayingoutthegeneralprogram and its rationale, with selected illustrations. Then I explore in some detail a sin-gle application: the ontology of events. I do not push hard for any particular on-tologicalconclusion,abouteithereventsoranyothertopic.Thefocusismeth-odology, not a particular output of the methodology. Here is a recently published characterization of the metaphysical enterprise, one that probablycaptures orthodox practice pretty well and to which I take no exception. Metaphysicalinvestigationsbeginwithinitialappearances.Ineverydaylife, theseappearancesareseldomquestioned.Inmetaphysics,weinvestigatefurther. Aswepursueametaphysicaltopic,weseektogetbeyondappearances.Wecon-siderargumentsabouthowthingsreallyare.Weseektolearntherealityofthe situation.Realitymayconfirminitialappearancesoritmayundercutthem.Either way,ourgoalistofindtheultimatereality.Thissuggeststhatthesubjectmatter distinctive of metaphysics is ultimate reality. (Conee and Sider 2005: 200) Theauthorofthispassage,EarlConee,2doesnotendorsethestatedviewun-qualifiedly. He cavils at the claim that metaphysics is concerned with ultimate reality,suggestingthatultimateaddsnothingofimportance.Healsoworries that if metaphysics is simply concerned with reality, it doesnt differ from other factualinvestigations.Leavingthesequibblesaside,Coneeseemstofindthis characterizationofmetaphysicsfairlysatisfactory,aswouldothermetaphysi-cians, I suspect.Although Conee doesnt explicitly say this, reality is normally understood to refer to what exists (in the broadest sense) in a mind-independent way. Meta-physics seeks to understand the nature of the world as it is independently of how we think of it. The suggestion that we should study the mind to understand real-2ALVIN I. GOLDMAN itywouldthereforestrikemanymetaphysiciansaswrong-headed.Theywould readilyconcedethataportionofmetaphysicsthemetaphysicsofmindhas mind and mental states as its proper province. But the mind is just a fragment of reality.Mostsectorsofmetaphysicsareconcernedwithextra-mentalreality.It would be grossly misdirected for those other sectors of metaphysics to aim their inquiriesatthemind.Nonetheless,thisiswhatIpropose.Idontmeanthata study of the mind isthe final aim of all metaphysical inquiries; but it should be studiedinthecourseofothermetaphysicalinvestigations.Itshouldbeacon-tributing part of metaphysical investigation, even those investigations for which the mind is not the primary target.To clarify the proposal, it must be explained that the referent of the mind is(orincludes)theaggregateoforgansormechanismsofcognition.Cognitive organsormechanismsplayacriticalroleinthecausalproductionofappear-ances,includingmetaphysicalappearances(whateverexactlywetakethatto connote).Inconsideringwhethersuchmetaphysicalappearancesshouldbeac-cepted at face value or, alternatively, should be superseded through some sort of metaphysical reflection,it obviously makessense to beasinformed as possible about how these mechanisms of cognition work. That is why cognitive science is relevant. Letmeexpandonthisproposal.AsConeeindicates,metaphysicalinquir-iesusuallystartwithdefaultmetaphysicalassumptions,i.e.,nave,intuitive,or unreflectivejudgments.Thesecorrespondtowhathecallsappearances.We intuitivelyjudgethatobjectsarecolored,thatpeoplehavefreewill,thatsome events cause others, that time passes (always flowing in the same direction), and thatsomepossibilitiesareunactualized.Metaphysicalinquirystartsfromsuch defaultjudgments,butitispreparedtoanalyzeorinterprettheminalternative ways,oreventoabandonthemaltogether.Theyareallupforcriticalscrutiny, ofonesortoranother.Howshouldweproceedinthiscritical,reflectiveactiv-ity?To what degree should precedence, or priority, be given to our nave meta-physical convictions? Virtuallyallmetaphysiciansagreethatourdefaultmetaphysicalviewsare subject to philosophical refinement. If there are inconsistencies among our nave metaphysicalviews,somemustbeabandoned.Inaddition,mostcontemporary metaphysicianswouldagreethatscienceshouldsometimesoverrideournave metaphysics. Physics might give us reason toconcludethattime doesnt pass atall; thatit has noasymmetrical directedness; or, indeed, thatthere is no such thing astime, only space-time. Again, physics might give us reason toabandon certainassumptionsaboutcausalrelations.Mostexistingappealstosciencein defenseofmetaphysicalrefinements(orrevolutions)areappealstophysical science.Thisisunderstandable,giventhatmostofmetaphysicsisconcerned withostensiblynon-mentaltargets(e.g.,color,causation,time,possibilia).I A PROGRAM FOR NATURALIZING METAPHYSICS3 argue, however, that even in these sectors of metaphysics, evidence from mental science,thatis,cognitivescience,canandshouldbepartofmetaphysicalin-quiry.There are at least two ways in which a metaphysician might try to revise naveexperience or nave beliefin matters metaphysical. First, ametaphysician mightadvanceaneliminativethesis.Thisorthatontologicalphenomenon,as-sumed to exist on the basis of common sense or nave experience, might be de-nied any sort of existence at all. This transpires when a metaphysician denies the existence of time or free will, for example, yet common sense affirms their exis-tence.Second,ametaphysicianmightproposeamorenuancedkindofontologi-cal revision of common sense or nave experience. Starting with a nave concep-tionofacertainproperty,ametaphysicianmightsuggestthatthepropertyis really different in crucial ways from the way common sense or experience repre-sentsit.Theproposaldoesnotdenythephenomenonsexistence(somephe-nomenon worthy of the name). It merely suggests that the propertys ontological status is importantly different from the way it is ordinarily represented. One such moveistoclaimthat,althoughthepropertyisordinarilyrepresentedasbeing intrinsictoitsbearers,itisinfactarelationbetweenmembersoftheclassof bearersandanotherclassofobjectsorevents.Forexample,itsaresponse-dependent property, a relation between the presumed bearers of the property and someclassofmentalresponsesthatoccurinthesubject(perceptualresponses, cognitive responses, or emotional responses). Although we think of ourselves as detectinganon-relationalproperty,onethatresidesintheobservedobjects themselves,theclaimisthatthereisnosuchwell-behavedintrinsicproperty ofthebearers.(Thecriticsconceptionofwell-behavednessmayvaryacross revisionarytheories.Somemightrequirethepropertytobeanaturalkind, othersaphysicalproperty,etc.)Instead,whatunitestheobjectsoreventsin questionisonlyarelationbetweenthemandaclassofsubjectiveresponsesto them. This yields an anthropocentric property, rather than a well-behaved one. If it is to be revisionary, a response-dependence thesis must also show that there is adivergencebetweenitstrueresponse-dependentcharacteranditsnaveap-pearance.Cognitivescience-influencedmetaphysicsneednotberestrictedtorevi-sionarytheories.Cognitivesciencemightalsointroduceevidencethatfavors commonsenseviews.Iconcentrate,however,onhowcognitivesciencemight support revisionary claims. Here is an example.In speech perception, we seem to hear boundaries of words. It sounds as if therearesilencesbetweenwordsthatenableustodistinguishtheirbeginnings and ends. But this is an illusion. Steven Pinker (1994), a leading cognitive scien-tist, describes the matter as follows: 4ALVIN I. GOLDMAN Allspeechisanillusion.Wehearspeechasastringofseparatewords,butun-like thetreefallingintheforestwithnoonetohearit,awordboundarywithno one to hear it has no sound. In the speech sound wave, one word runs into the next seamlessly;therearenolittlesilencesbetweenspokenwordsthewaythereare white spaces between written words. We simply hallucinate word boundaries when wereachtheedgeofastretchofsoundthatmatchessomeentryinourmental dictionary.Thisbecomesapparentwhenwelistentospeechinaforeignlan-guage: it is impossible to tell where one word ends and the next begins. (1994: 159160) Althoughweseemtodetectwordboundariesinthespeechofournativelan-guage,therearenosuchboundariesintheacousticstimulus.3Theexperienced boundaries are supplied or constructed by the hearers speech perception system, building on its knowledge of words in the language that are likely to be spoken. Moreover,eventhesequencesofsoundswethinkwehearwithinawordare an illusion.Informationabouteachcomponentofawordissmearedoverthe entireword,andthisinformationisreconstructedorreassembledbythelis-teners system. Sowhatshallwesayaboutthemetaphysicsofwordboundaries?One could adopt eliminativism. That is the view suggested by Pinkers term halluci-nate.Butthereisalsoanotheroption:aresponse-dependencemaneuver.4Per-hapswordboundariesconsistinrelationsbetweenthespokensequencesof sounds and theauditory experiences of hearers who understand thelanguagein question. Word boundaries aretendencies forsuchsound sequences to produce boundary experiences in (suitable) hearers. These tendencies are genuine rela-tional properties, actually instantiated.So the response-dependencemoveis not equivalent to eliminativism. Nonetheless, it is clearly a revisionary view relative to naveexperience, becauseauditory experience represents word boundaries as being features intrinsic to the speech signal. The evidence that might support such a response-dependence maneuver in-cludes the evidence Pinker adduces from cognitive science. The acoustic proper-ties of speech signals are studied in the laboratory. These are the kinds of studies onwhichPinkerbaseshisconclusions.Thesameholdsforotherstudiesthat support the notion of a specialized sense for speech, which is responsible for the mental creation of our phonetic experience. Conjectures onthese matters might bemadeonthebasisofcasualobservationratherthanscientificexperimenta-tion,especiallyoffeltdifferenceswhenhearingonesnativelanguageversusa foreignlanguage.Butreallystrongevidenceagainstnaverealismandtoward eithereliminativismorresponse-dependencecomesfromresearchincognitive science. Cognitive science cannot adjudicate between the two revisionary theses. Philosophicalreflectionmustassesstheadvisabilityofeliminativismversus respondencedependence.Butevidencefromcognitivesciencetiltsstrongly against realism (or intrinsicalism) about audible word boundaries.A PROGRAM FOR NATURALIZING METAPHYSICS5 Theontologicalstatusofwordboundariesisnotaverysignificantmeta-physicaltopic.Soletmeturntoamoresalientexampleinmetaphysics:color. Here I focus on a slightly different issue than response-dependence, but one that again shows how cognitive science can be relevant.Sinceancienttimes,philosophersandphysicalscientistshaveoffered secondary quality accounts of color. These are response-dependence or dispo-sitional theories, which treat colors as (merely) dispositions to produce certain visual experiences in normal perceivers. On this type of approach, redness is the dispositiontolookredtostandardperceiversinstandardconditionsofvisual observation. This is a revisionary approach because color doesnt present itself asa dispositiontolook acertain way.Redness presentsitself (in visualexperi-ence)asbeinginobjectsthemselves,orontheirsurfaces.Thus,inplaceof dispositionaltheories,severalcontemporarytheoristspropoundtheviewthat colors are the categorical bases of such dispositions. Brian McLaughlin charac-terizes redness (in part) as that property which disposesits bearers to look red tostandardvisualperceiversinstandardconditionsofvisualobservation (2003: 100). As McLaughlin explains, this is nota dispositional theory of color properties. A dispositional propertyisa functional-role property. OnMcLaugh-linstheory,bycontrast,acolorpropertyisthebasisorgroundofafunctional role.Rednessistheoccupantoftherednessrole.Morefully,McLaughlinsays thatbeingabasisforthedispositionofobjectstolookreddoesntsufficefor being redness; to be redness, a property must be a basis common to all things so disposed(2003:102,emphasisadded).Healsoimpliesthat,toberedness,a property must be the unique property that occupies the redness role.McLaughlin proceeds to pose three problems for such an occupant, or role-player, theory, two of which Ill mention here. First is the problem of the com-mon ground. Is there a (physical) categorical property that all red things have in common?Thephenomenonofmetamerism,identifiedbycolorscience,seems to stand inthe way. Whatis seenasa single color can be produced by a single light-wave frequency or by many different combinations oflight-wave frequen-cies.AsC. L. Hardin notes:if an observers unique yellow [yellow that is not at all greenish or reddish] were at 575 nm on the spectrum [in a certain observa-tionalsetting],anappropriatemixtureofspectral550nmand650nmlight would match it exactly (1988/1993: 198). So there doesnt seem to be any com-mon basis of the disposition to produce an experience of unique yellow. Second istheproblemofmultiplegrounds.Ifthereismorethanonecommonbasisof the disposition to look red in circumstances C, then its not the case that there is somepropertythatisthepropertythatdisposesitsbearerstolookredtostan-dard perceivers in circumstances C.Theproblemofthecommongroundandtheproblemofmultiplegrounds flow directly from the demands imposed by the role-occupant theory. The ques-6ALVIN I. GOLDMAN tionishowserioustheseproblemsareandwhethertheycanbesatisfactorily solved.Cognitivescienceisrelevanthere,firstincreatingatleastoneofthe problems and second in helping to solve it. The problem of the common ground ismadeparticularlysalientbymetamerism,andmetamerismisaphenome-non discoveredbycognitivescience.AlthoughMcLaughlinthinksthatthe problemofmultiplegroundscanbeadequatelysolvedbypurelyphilosophical maneuvers(2003:125-126),theproblemofthecommongroundisviewedas one that can only be solved with the help of vision science. McLaughlin appeals specificallytotheopponent-processingtheoryofcolorvision(Hurvichand Jameson1957)(seeMcLaughlin2003:128-131).Opponentprocessingtheory is also the product of color science (indeed, the cognitive science branch of color science).WhetherMcLaughlinsconsiderationssucceedinsolvingtheprob-lem ofthecommongroundIshallnottrytosettle.Thepointissimplytoil-lustratehowdetailedtheoriesofcolorpropertiescanreadilyencounterprob-lems whenconfrontedwithempiricalfactsofthesortcognitivescienceun-covers.Atthesametime,cognitivesciencecanoftenhelpsolvesomeofthose problems.As things go with color, so they go with other properties that invite either a dispositional treatment or a categorical basis treatment. Another example of this typeisdisgustingness.Nowitsnotentirelyclearhowour ordinaryconception ofdisgustingnessshouldbeelucidated.Butletusmainlyfocusonaproperty that,onreflection,seemsworthyofbeingthereferentofthetermdisgusting (just as redness is the referent of red).Ifdisgustingnessisthegroundofthedispositiontoelicitfeelingsofdis-gust, mustntit bea property common to all disgusting things, asin thecase of redness?Isthereanysuchcommonground?Cognitivesciencefindingsmake suchathesisproblematic.Theemotionofdisgusthasaremarkablydiverse range of elicitors, which range from the concrete to the abstract. These are sum-marized by Daniel Kelly (2007). At the concreteend of the spectrum of disgust elicitorsarefeces,vomit,urine,andsexualfluids(Rozinetal.2000).Other likelycandidatesarecorpsesandsignsoforganicdecay(Haidtetal.1994). Bodily orifices and, via contamination, things that come in contact with them are also powerful elicitors (Rozin et al. 1995). All of these elicitors are disgust uni-versals.Finally,diseaseandparasiticinfectionprovideanothersetofdisgust universals (Curtis et al. 2004). At the abstract end of the spectrum, violations of normshavingnothingtodowiththetypesofelicitorsjustmentionedcanalso triggerdisgust.Floutinganormcentraltoaculturalingroup,forexample,can produce an emotion of disgust.What could possibly be a common ground among these diverse elicitors? It seemsunlikelythatthereisanysuchproperty.Ifthereisone,itsurelyisnta naturalkind,atmost,awildlyheterogeneousdisjunctionofproperties.Ifdis-A PROGRAM FOR NATURALIZING METAPHYSICS7 gustingness is either an intrinsic property shared by all disgust elicitors or noth-ing at all, eliminativism looms. Soletusconsiderthealternativeapproach:dispositionalismorresponse-dependence. On this view, disgustingness is simply the disposition to produce a disgustresponseinhumans.Butwhat,precisely,istheresponse?Kelly(2007) dividesthedisgustresponseintothreeparts,ofwhichIllemphasizetwo.The firstpartconcernstheso-calledaffectprogramfordisgust,whereaffectpro-gramisatheoreticaltermforreflex-likeemotionalresponses,asetofcoordi-natedphysiologicalandbehavioralelements.Inthecaseofdisgust,there-sponses include (1) a behavioral response involving immediate withdrawal, (2) a distinctivefacialexpression,knownasthegapeface,and(3)qualitativere-sponsesincludingrevulsionandthephysiologicalconcomitantsofnausea(Ek-man1992).ThesecondpartiswhatKellycalls,followingRozinetal.(2000), core disgust. Its features are distinct from those of the disgust affect program; theyareless reflexive andmorecognitive in nature. Two of itscentral features areasenseofoffensivenessandcontaminationsensitivity.Offensivenessisa type of aversionthat views the disgustingentities as somehow dirty, tainted, or impure.Contamination sensitivity refersto the fact that once an item is marked as offensive by the disgust system, the item can contaminate otherwise pure and un-disgustingentities.Entitiessocontaminatedarethentreatedasdisgusting. Moreover,thereisanasymmetricrelationshipbetweendisgustingnessandpu-rity.Itisfareasierforsomethingpuretobecontaminatedthanitistopurify something contaminated. A single drop of sewage can spoil a jug of wine, but a single drop of wine doesnt much help in purifying a jug of sewage (Hejmadi et al. 2004).Returningnowtothedispositionalapproachtodisgustingness,exactly which responses are constitutive of this disposition? Are they all constitutive of it? Here is a different way to approach the problem. Since we are already dealing with revisionary accounts of disgustingness, why assumethere is a single prop-erty? Why not distinguish different disgustingness properties, so as to cut nature better at its joints? One way to do thisis to bifurcate disgustingness in terms of thetwofamiliesofresponses:theaffectprogramresponsesandthecorere-sponses.WewouldthenhaveDISGUSTINGNESS1andDISGUSTINGNESS2, wheretheformerisadispositiontoproduceaffect-programresponsesandthe latter is a disposition to produce core-disgust responses.No matter which choice we make in this territory, the choice should reflect knowledgeofthetypesoffunctional(andevolutionary)relationsthatIhave beenpresentingfromcognitivescienceresearch.Anynaturalslicingofthe terrain should be informed by the relevant empirical facts, not restricted to what canbeknownaboutdisgustanddisgustingnessaprioriorviacommonsense alone.8ALVIN I. GOLDMAN Finally,Iturntothecaseofmoralconcepts.Thereispowerfulevidence from recentcognitive sciencethat moral judgmentisintimately associated with emotion.Thismightnaturallyleadtoaresponse-dependenceaccountofour moralconcepts,asrecentlyarguedbyJessePrinz(2006).Ofcourse,anumber of moral philosophers in recent decades have advanced sensibility accounts of moralconcepts,inthegeneraltraditionofHume.TheseincludeMcDowell (1985),McNaughton(1988),Johnston(1989),Dreier(1990),Wiggins(1991), andWright(1992).Thepointisnotthatittakescognitivesciencetothinkup such a position in metaphysics; rather, cognitive science is uniquely able to con-tributecertainkindsofcompellingevidence,evidencethatmighttiltmeta-ethicalormetaphysicalargumentsinonewayoranother.Hereissomeofthe relevantpsychologicalevidencethatsupportsaresponse-dependenceviewin meta-ethics.5 Moll et al. (2003) measured brain activity as subjects evaluated moral sen-tences.Theyfoundthatwhensubjectsmademoraljudgments(butnotwhen they made factual judgments) brain areas associated with emotions were active. Sanfeyetal.(2003)measuredbrainactivityassubjectsplayedanultimatum game. When a division of a monetary sum was deemed inequitable, a player had brain activity in areas associated with emotion. Berthoz et al. (2002) found simi-larengagementofbrainareaswhensubjectsconsideredviolationsofsocial rules.Again,Greeneetal.(2001)foundemotionactivationassubjectsconsid-eredmoraldilemmas.Inadifferentkindofstudy,Schnalletal.(forthcoming) found that negative emotion can lead people to make a more negative moral ap-praisalthantheywouldotherwise.Subjectsweregivenaquestionnairewitha series of vignettes and asked to rate the wrongness of the actions described. Half of the subjects who read the vignettes were seated at a nice clean desk; the other halfwereseatedatafilthydesk.Thelatterratedthevignettesasmorewrong thantheformer.Finally,thereisconsiderableevidencethatpsychopaths,who seem incapable of making genuine moral judgments, are profoundly deficient in negative emotions (Blair et al. 2001).Again, there is room for debate over which meta-ethical/metaphysical theo-ries of moral concepts or moral properties should be inferred from these studies. Thepresentpointisthefairlymildonethatempiricalstudiesareprofoundly importantinhelpingus(astheorists)understandtheordinarygraspofmoral matters,whichispartofthetaskofdoingmetaphysicsintheethicalrealm. Whatisespeciallynoteworthyisthataffectiveentanglementwithmoraljudg-mentis not evidentin consciousexperience. It presumablyoccurs, for the most part,atasub-thresholdlevel,whichiswhycognitivescienceandneuroscience are needed to tease out what takes place. Astute philosophers (such as Hume and his intellectual descendents) can tender such hypotheses, but their suspicions are not conclusive, and indeed are hotly disputed by other philosophers. One needs a A PROGRAM FOR NATURALIZING METAPHYSICS9 cleanermethodology,onenotbeholdentoconsciousawareness,togetsounder and more decisive evidence on the matter. For evidence on the role ofemotionevenin everyday factual judgment, consideranotherphenomenondiscoveredbycognitivescience.Thisconcerns thefascinatingdelusionalsymptomcalledCapgrassyndrome.IntheCapgras syndrome, people report that their acquaintances (spouse, family, friends, and so on) have been replaced by body doubles (Capgras and Reboul-Lachaux 1923). Theyacknowledgethattheirhusbandorwifelooksliketheirhusbandorwife (respectively).Theycanpickthepersonoutfromaline-upasresemblingtheir husband or wife, but they steadfastly maintain that he or she is an impostor. Ap-parently,theyconsciouslyrecognizethepersonsface,buttheylackanemo-tionalresponsetoit.6Theythereforeconcludethatthepersonobservedisntin fact the husband, wife, friend,etc.7 Of special interest hereis what theCapgras syndrome tells usabout normal people. Although weare quite unaware of this, we apparently use emotional associations, at a sub-threshold level, in identifying aperson.Iamnotgearingup,onthisbasis,todefendaresponse-dependence accountofpersonalidentity.Iaddthisitemtothemixonlytounderlinethe pointthatwhattranspiresinjudgmentmakingandconceptpossessionisnot introspectively transparent. This is why metaphysics needs cognitive scienceat least insofar as metaphysics proceeds by starting from our nave world-view and proceeding outwards from there.Intheremainderofthepaper,Iillustratethecognitivistprogramwitha ratherdifferentexample:themetaphysicsofevents.TheangleIshallhighlight again turns on the fact that the ways we conceptualize many types of entities of metaphysicalinterestarenotavailabletointrospection.Weneedthehelpof cognitivesciencetoilluminatetheseconceptualizations.Thisisespeciallytrue in the case of minor entities, the focus of the present journal issue. 2. Events and the Problem of Event Individuation Letusapproachthemetaphysicsofeventswhilelookingoverourshouldersat the metaphysics of time and disgustingness. In the case of time, a variety of sci-entifictheorizing,especiallyinphysics,suggeststhatthereisnosuchthingas time,atleastasnaivelyexperienced.Theflowoftime,theexistenceofsome-thing called the present, and the directionality of time are all problematic from thestandpointofphysics.Butiftimeisanythingatall,itcannotbesomething merelyinourminds.Sowhatexternalthingcanbenominatedbytheword time? Metaphysicians struggle with this problem. It will not do to pick out any old metric or dimension that is countenanced by contemporary space-time phys-icsandcallittime.Metaphysicianslookforapropertyordimensionofthe space-timemanifoldthatcomportsascloselyaspossibletoournaveunder-standing or grip on time, that is, our time experience. But when we try to look at 10ALVIN I. GOLDMAN whatthat understanding orexperienceconsistsin, we needhelp fromcognitive science. To identify a suitable physical dimension to identify with time, we need a better grip on our anthropocentric conception of time; and that may require an understanding of temporal psychology. Thecaseofdisgustingnesshasparallels.Ifwewanttounderstandwhat disgustingnessis,weneedtounderstandthepsychologyofdisgust(theemo-tion). Disgustingnessmay prove to have an additional parallel with events, as I shalllatersuggest.Aswehaveseen,oncetherelevantfactsarelearnedabout disgust and its family of elicitors, we see that a good case can be made for bifur-catingthephenomenonofdisgustingness.Insteadofpositingasinglekindof state,itmaymakebettersensetoposittwotypesofdisgustingness.Astrong case for such a proposal, however, hinges partly on empirical findings about the internalresponsesassociatedwithdisgust.Similarly,Ishallargue,certainphi-losophicalpreconceptionsabouteventsshouldgivewayoncewehaveabetter grip on the way we (psychologically) conceptualize items that can be considered events. Instead of positing a single ontologicalcategory of events,itis better to posit a pair of such categories. Theprecedingparagraphdescribeswhereweareheaded.Nowletmeset the stage for the discussion by making some preliminary clarifications. There are narrower and wider uses of the term event,colloquial uses and technical uses. Thesedifferentusesindicatethatdifferentwritersabouteventshavedifferent targetsintheircross-hairs.Colloquially,eventsareequatedwithhappenings: things that occur, or take place, which cant be said of objects like books or stat-ues. A variety of theories conceive of events rather differently, however (Casati and Varzi 2007). There are scientific theories that treat events as qualified points inspace-time(GeneralRelativity),orassetsofoutcomes(probabilitytheory); and some philosophical theories treat events as properties of cross-world classes of individuals (Lewis 1986). These technical uses are putaside for present pur-poses.Anotherdistinctionisbetweenevent-typesandevent-tokens.Theterm crashcanbeusedtorefereithertoageneraltypeofeventortoindividual events that occur on particular occasions. Our principal interest is event tokens.In the 1960s and 1970s a lively debate emerged about the problem of event individuationandthenatureofevents.Somephilosophers,mostprominently Donald Davidson, championed a view that came to be called the coarse-grained, or unifier, approach. This was initially motivated by the theory of action, where actions,ofcourse,areasubcategoryofevents.Davidsonsunifierapproachis succinctly illustrated by the following passage about actions: [I]tishardtoimagineasatisfactorytheoryofactionifwecannottalkliterallyof the same action under different descriptions. Jones managed to apologize by saying Iapologize; but only because, underthecircumstances,saying I apologize was apologizing. (1969/1980: 164-165) A PROGRAM FOR NATURALIZING METAPHYSICS11 TherearenttwodistinctactionsofJones,hissayingIapologizeandhis apologizing;theseactionsareoneandthesame.Similarly,ifOlivermoveshis finger, flips a switch,turns on a light,and alerts a prowler,all theseactionsare identical with one another.Multipliers contend that Olivers actions are four distinct actions. The mul-tiplier position was developed (semi-)independently by Jaegwon Kim (for events generally)andme(foractions)(Kim1966,1969,1973,1976;Goldman1965, 1970, 1971). We each defended a general account of events or actions as exem-plifications of properties. In Kims well-known formulation, event E is identical with event E only if they have as constituents the same substance(s), the same propertyorrelation,andthesametime(Kim1976).Forexample,thedyadic eventconsistingofthecollisionoftheTitanicwiththeiceberghasasitstwo substances the ship and the iceberg, as its relation colliding with, and as its time a period of around ten seconds on the fateful night (Simons, 2003). In the Oliver case,theswitch-flippingactionisnotidenticaltotheprowler-alertingaction becausetheyinvolvebothdifferentsubstances(theswitchversustheprowler) anddifferentrelations(flippingversusalerting).Nocomparablyagreedupon principle of individuation is associated with the unifier theory.The unifier approach commanded a majority of adherents in the 1960s and 1970s,andthisprobablyremainstruetoday.Becausetheproperty-exemplificationapproachhasbeenaminorityview,Ibeginthepresentdiscus-sionwithalimiteddefenseofit.Thepurposeisnottodefenditasthecorrect theory, but only as a serious contender. So let me review some considerations in its favor and some of the dialectic that has characterized the debate. Here are some arguments from causal relations that I advanced in the early period of debate (Goldman 1970, 1971). Consider these actions:(B1)Boriss pulling (of) the trigger(B2)Boriss firing (of) the gun (B3)Boriss killing (of) Pierre.According to the unifier approach, B1 = B2 = B3. If this is correct, every effect ofB1mustequallybeaneffectofB2andB3.Nowconsiderthefollowing event:(F)the guns firingIs F an effect of B1? Surely it is. The pulling of the trigger causes the gun to fire. IsFaneffectofB3?Apparentlynot;itcertainlysoundswrongtosaythatBo-riss killing (of) Pierre caused the gun to fire. Hence, some effect of B1 is not an effect of B3, a counterexample to the unifier approach. This assumes, of course, that if X and Y are identical, then there are no properties possessed by X that Y lacks. This principle holds for causal-relational properties among others.12ALVIN I. GOLDMAN Davidson tried to ward off this kind of difficulty by arguing that the singu-lartermBorisskilling(of)Pierreshouldbetranslated(roughly)asthedefi-nite description The action of Boris that caused Pierres death. The denotation ofthisexpressionissaidtobeB1,andsinceFisaneffectofB1,itisalsoan effect of B3. But the proposed translation is contentious. If the translation is cor-rect, why does it seem intuitively incorrect to say that Boriss killing (of) Pierre causes the gun to fire?Hereisanotherexample.Johnsings;indeed,hesingsloudlyandoff-key. Are the followingidentity statements true: Johns singing (J1) = Johns singing loudly(J2)=Johnshissingingoff-key(J3)?Theunifiertheoryaffirmsthese identitystatements.ButconsiderJohnsbeingangry(A)asacandidatecause. SupposethatstateAwasacontributingcauseofJ2,becausetheloudnessof Johns singing was an inadvertent effect of his anger. Although state A is a con-tributing cause of J2, its implausible that its a contributing cause of J1, because the decision to sing was independent of the angry mood. Since one of the causes of J2 is not a cause of J1, we must conclude that J2 ! J1, and another counterex-ample confronts the unifier theory (Goldman 1971).Thiskindofdifficultyhasresurfacedinmorerecentliteratureoncausa-tion. L. A. Paul (2000) gives the example of Suzy who falls and breaks her right wristwhileskiing.Thenextdayshewritesaphilosophypaper.Simplifying Paulsexample,notethatSuzysbreakingherrightwristisplausiblyacon-tributingcauseofherwritingthepaperwithherlefthand,butitisntacon-tributingcauseofherwritingthepaper.NowPauldoesnotregardthispoint (or therestofherexample)asgroundsforendorsingthepropertyexemplifica-tionviewofevents.Sheintroducesthetermaspecttorefertopropertyin-stances,butdoesntidentifythemwithevents.Anaspectcanbeapropertyin-stanceofanevent;itdoesnthavetobeanevent.Paulproceedstoarguethat aspects,ratherthanevents,arecausesandeffects.Thispositionhassubstantial affinity with the multiplier theory because it views property-individuated entities (viz.,aspects)asthebearersofcausalproperties,aroleusuallyassignedto events. Noticethatthepartiestothedebateagreeontwothings.First,theyagree thattherearesuchthingsasevents.Second,theyagreethatexactlyoneofthe twoapproachesiscorrect;thetaskistodeterminewhichonethatis.8Forpur-posesofthisdebate,then,eliminativismisntanoption.9It alsoisntanoption to consider both theories to be correct, since one view explicitly denies what the otherasserts,viz.,theidentityofseveralenumerateditems.Itispremature, however,toexcludethispossibility.Perhapstherearetwoviableconceptsor conceptions of events: under one the unifier view is correct, and under the other the multiplier (property-exemplification) view is correct. I shall now try to show how cognitive science lends support to this compromise position.A PROGRAM FOR NATURALIZING METAPHYSICS13 3. Two Systems of Representation and Their Relevance to the Ontology of Events ThethesisIwishtoadvanceisthatthemindhastwowaysofrepresenting events, or two cognitive formats for event representation. I dont mean that there are two representational systems dedicated exclusively to events, but two differ-entsystems that can be used to represent events among otherthings.Crucially, theyrepresenteventsusingdifferentrepresentationalprinciples,andthisgives rise to different intuitive principles of identity or individuation.ItisarguablethatJonathanBennetts(1988)theoryofeventsrecognizes this kind of point. His position might be encapsulated by saying that people get ustofeelmultiplierintuitionsbyusingonesetofeventnames(thekillingof theman)andtheunifiersuseanothersetofeventnames(themurder).If BennettstheoryapproximatestheoneIlloffer,thatsfinefromthepresent perspective.ButtheoneIshallpresenthasamoreexplicitbasisincognitive science, at a level below the level of language.10

Thegeneralideaofmultiplesystemsorformatsofmentalrepresentation for (roughly) the same target domain is extremely common in cognitive science. One example is the two systems of visual representation:the ventraland dorsal systems (Milner and Goodale 1995). Conscious visual experience of the world is subserved by the ventral visual stream, which generates conscious visual percep-tion of objects. In addition, the recently discovered dorsal visual system uncon-sciouslymediatesthevisualcontrolofmotorbehavior(e.g.,reachingforacup that is in view). That there are two distinct systems is evident from neurological dissociationsincertainpatients,inwhichonesystem,butnottheother,isim-paired. These systems deploy differenttypes of representation, although each is activated by visual inputs.The study of cognitive processes directed at events and actions is fairly re-cent. Some of this research is specifically focused on the individuation of actions and events, although the particular individuation problem psychologists study is abitdifferentfromtheonemetaphysiciansconcernthemselveswith.Wewill have to tread carefully. Nonetheless, here is a passage by the psychologist Karen Wynn that indicates what motivates her interest in the subject.[V]irtually nothing is known about how people individuateactions. The individua-tionofactionsislikelytobecomplexbecauseactionsarenotdefinablepurelyin terms of objective properties. Freges observation that number is not an inherent property of portions oftheworldappliesas muchtoactions asitdoesto physical matter.Justasagivenphysicalaggregatemaybeaninstanceofmanydifferent numbers(threeforksisalsoaninstanceof12tines,2decksisaninstanceof104 cards), so may a given portion of activity (one dance may equally be an instance of 437steps,onespeechmaybeaninstanceof72utterances).Thereisnoobjective fact of the matter as to where, in the continuously evolving scene, one action ends andthenextactionbegins.Theindividuationofdiscreteactionsfromthiscon-tinuous scene is a cognitive imposition. (Wynn 1996: 165, emphasis added) 14ALVIN I. GOLDMAN Ofspecialinteresttodevelopmentalpsychologists(includingWynn)is howinfantsmanagetoparsetheongoingactivityoftheworldintodistinctac-tions (or events). Infants parsing abilities should interest us as well, sincethey bearonthequestionofwhatrepresentationalcapacitiesarealsoavailableto adults for individuating actions and events.SharonandWynn(1998:357)proceedtoenumerateseveralreasonswhy an ability to individuate actionsand events is essential to theinfant. First,such parsing is required to perceive causality. For example, to recognize causality in a launchingepisodesay,whereabilliardballcausesmovementinanother billiard ball by striking itit wouldnt be possible to recover the causality if the wholemotionofthetwoballswereperceivedasoneundifferentiatedswoosh. Second,individuatingactionsisanecessaryprecursoroflearningtoproduce them.Achildwhowantstothrowaballshouldrealizethatpullingbackones arm is a necessary component of the throw, but jumping up and down in excite-ment afterward is not. A third major motivation for parsing motion into discrete actionsliesin the need to interpret the behaviors of others.Parsingmotion into discreteactsprovidesthebehavioralunitsthatareexplainableintermsofde-siresand beliefs. Finally, an ability to parseactions into the same units as other people do is necessary for verb acquisition and hence linguistic communication. Forlanguagetobecomprehensible,twopeoplemusthaveinmindthesame bounded pattern of motion when they refer to a jump, a hug, or a hit. What cues might infants use to individuate actions? Ideas are suggested by studies of individuation in other domains. Spelke (1988) found that infants con-struespatiallyseparatedsurfacesinthevisuallayoutasbelongingtodistinct objects.Similarly,Bregman(1990)demonstratedthatportionsofacousticen-ergy that are temporally separated by silence are interpreted by adults as arising from independent acoustic events, whereas energy that is temporally continuous and judged to originate from the same location is commonly interpreted as aris-ingfromthesameacousticevent.Thesefindingssuggestthatinfantsmayuse spatiotemporaldiscontinuitiestosegmentscenesofmotionintodistinctunits. Portions of motion separated by gaps in space, time, or both may make differ-entactions.ThishypothesiswastestedandcorroboratedbyWynn(1996).Ina first test, infants were able to counthence must have individuatedthe jumps of a puppet when there were brief pauses betweenjumps. In a second test, mo-tionless pauses were replaced by having the puppet gently wag its head back and forth between jumps. Infants were still able to detect boundaries between jumps, thoughnotasstronglyasinthefirsttest.SharonandWynn(1998)extended these findings, buttheyalso showedthatinfants are notable to parse heteroge-neous actions from a continuous, nonpatterned stream of motion.Thepossibilityoftwodistinctsystemsforactionoreventindividuation wassuggestedbyfindingsontheindividuationofobjects.Anotherstudyby A PROGRAM FOR NATURALIZING METAPHYSICS15 Spelkeandcolleaguessupportedthenotionthatyounginfantsanalyzespatio-temporal continuity in establishing representations of one object or two. Spelke, Breinlinger,Macomber and Jacobson (1992) arranged a scene with two screens separatedbyagap.Younginfantswhosawobjectsemergeoneatatimefrom opposite sides of the two screens but never appear in the middle interpreted this as two objects, not one. This was another example of the significance of spatio-temporalcontinuity.(Iftherewereoneobject,continuousmotionwouldhave required it to appear in the gap between the screens.) However, by 12 months of age,infantsbringkindmembershiptobearontheproblemofobjectindividua-tion.XuandCarey(1996)foundthatwhen12-month-oldinfantswereshown objectsoftwodifferentkindsemergingoneatatimefromoppositesidesofa singlescreen,theyestablishedrepresentationsoftwonumericallydistinctob-jects. This couldnt have been inferred on the basis of spatio-temporal continuity considerationsalone,becausesuchconsiderationswouldbecompatiblewitha single object. Shipley and Shepperson (1990) also used counting as a wedge into studyingpreschoolchildrenscriteriaforobjectindividuation.Childrenwere showndisplaysofobjects,someofwhichwerebroken.Forexample,there mightbethreeintactcarsandonecarbrokenintotwodistinctpieces.Even whenaskedtocountthecars,childrenasoldas5yearswouldoftencount each separate piece of a car, along with each intact car (yielding a total count of five,intheforegoingexample).Thisisdespitethefactthat5-year-oldswould have known the kind word car for some time. Still, when considering a scene consistingofwholecarsandsplitcars,childrenanalyzedthesceneintermsof spatio-temporally discontinuous units, not units determined by the basic-level kind term car. This tends to show that although 5-year-olds have the capacity toindividuateonthebasisofkinds,orsortals,theyhaveapreferenceforthe apparently more primitive spatio-temporal criterion of object individuation. What about event individuation? Wagner and Carey (2003) compared chil-drenandadultsonindividuationtasksconcerningbothobjectsandevents. Shownaseriesofslidesonacomputerscreen,subjectswereaskedforeach slide, How many Xs are here? The X term was either an appropriate kind label for the object (e.g.,car, or fork) orelsea kind-neutralword (e.g., thing). Following the object task, children were told that the game was going to change andnowtheywouldcountwhathappenedinmovies(theeventindividuation experiment).HeretheanalogoustermXwasanappropriateeventdescription, targetingeither the goal oftheevent orits temporally discrete sub-actions. The latterwasconsideredaspatio-temporalparseoftheobservedactivity.Inboth the object and event cases, children showed a spatio-temporal bias in individua-tion.Theyoftenignoredthekindterm(orgoal-basedterm)thatwasspecified andcountedintermsofpurelyspatio-temporalcriteria.Nonetheless,evenby age 3 and definitely by age 5,they showed somesensitivity to kind-basedindi-16ALVIN I. GOLDMAN viduation. Adults, by contrast, were at or near ceiling in their use of kind-based (or goal-based) individuation.Allthestudiesandtheoreticalinterpretationswehavereportedarefrom developmental psychology, which focuses on the growth or unfolding of cogni-tivecompetences.Thedevelopmentangleisnotourprimaryinterest,however. Weareconcernedwiththecognitivecompetencesofadults,becausecommon sense ontological judgments issue mainly from adults (philosophers themselves, in most cases). The main point I wish to make is that adults apparently have two systems for individuating events, one system based on spatio-temporal disconti-nuities and a second system based on kinds. The first is a very early developing systemandthesecondisnotquitesoearlydeveloping.Ireportevidencefrom developmentalpsychologynotbecauseIamintrinsicallyinterested(here)in infants and children but because it happens to be the branch of cognitive science where the most relevant work on event individuation has been done.Thepsychologicalresearchjustreviewedisthemain,alreadyexisting, empiricalworkmostrelevanttothemetaphysicsofevents,butadmittedlyit isntasdirectlyrelevantasonemightwish.Itisntfullyandsquarelyrelevant becausetheindividuationissueaddressedincognitivescienceisdifferentfrom theindividuationissuediscussedinsection2.Thepsychologistsareconcerned witheventsegmentation,orparsing.Thisisdecomposingtheworldsspatio-temporalfluxintodisjointunits.Thephilosophersareconcernedwithevents occurring at one and the sametime and place.11 Arguably the two questions are connected, but they arent the same. Considertheexamplefromsection2inwhichBorispullsagunstrigger andtherebyfiresthegun.Thetrigger-pullingandthegun-firingbothoccurat thesametimeandplace.Butwecanstillaskwhetheradditionalontological slicingshouldbedone.Arethetrigger-pullingandthegun-firingtwodistinct actions,oraretheyoneandthesameaction,simplydescribedintermsofdif-ferent effects? Ill call this new question the slicing question. It isnt one that the psychologyliteratureaddresses.Althoughthepsychologicalworkdoesnot squarely address the slicing question,itmay beindirectly relevant.Recognition of the two disparate systems can help us diagnose the emergence and persistence of the dispute between unifiers and multipliers.Theproperty-exemplificationviewassignsapivotalroletothekindor propertyofwhichatokenactionoreventinstantiates.Accordingtothisap-proach, ifaction a is atoken of action-type A andaction b is a token of action-typeB,thenaisnotidenticaltob,evenifaandbareperformedbythesame agentatthesametime.Borisspullingthetrigger(att)andBorissfiringthe gun (at t) are not the same token action, because pulling the trigger and firing the gunarenotthesamepropertyorkind.Thisconceptualizationofactionsor eventscomportswiththesecond,late-developingsystemofeventrepresenta-A PROGRAM FOR NATURALIZING METAPHYSICS17 tion, in which kinds (or sortals) play a crucial role. By contrast, the unifier view ofactionsoreventsassignsnocomparablypivotalroletokind-differencesin action or event individuation. According to the coarse-grained approach, Boriss pulling of the trigger andBoriss firing of the gun are one and the sameaction. Isitplausiblethatthisconceptualizationofactionsisassociatedwiththefirst, early-developingsystemofeventrepresentation,asystemfocusingonspatio-temporal factors without attention to (other) kinds? I think it is plausible. If purely spatio-temporal representations are used to represent actions, it is natural to represent them as bodily movements only, not as entities with higher-order, non-geometric or non-kinematic properties, such as firing the gun, poison-ingtheking,orwritingacheque.Interestingly,theleadingproponentofthe coarse-grained theory, viz. Davidson, explicitly says that actions are, really, just bodily movements:We must conclude, perhaps with a shock of surprise, that our primitive actions, the ones we do not do by doingsomething else, [are] mere movements of the bodythese are all the actions there are. We never do more than move our bodies; the rest is up to nature. (1971/1980: 59) AccordingtoDavidson,whatmanyphilosophersofactioncallbasicactions are the only actions there are. He does not mean to deny, of course, that grander-soundingthingslikesendingspaceshipstoMarsalsoareincludedamongour actions. He certainly doesnt deny that these types of descriptions can be applied toouractions.Buttheseotheractionsarejustidenticaltothebasicones, which aremovements of the body (e.g., the finger movement that depresses the button thatlaunchesthe spaceship).This fits a purelyspatio-temporalconstrual ofaction,whichconformscomfortablytotheconstraintsoftheconjectured early-developing system of event representation. Adifferenttakeoneventindividuationemergesifonefocusesonevent-referringexpressionssuchasQuislingsbetrayalofNorwayorQuislings betrayingNorway,bothofwhicharederivedfromtheindicativesentence Quisling betrays Norway. Verbs pick out kinds or sortals, which typically ab-stract from the spatio-temporal properties displayed by their particular instances. If one mentally represents an event by concentrating on the abstract property by which it is specified, this conceptual mode of mental representation would natu-rally be processed by the later-developing system for event representation.My hypothesis, then, is that the two groups of partisans in the event-slicing debate are theorizing aboutthe referents oftwo differenttypes of mental repre-sentations,eithereventrepresentationsinthespatio-temporalsystemorevent representationsinthekind-basedsystem.Bothgroupsofpartisans,ofcourse, have access to both systems of representation, but when they engage in philoso-phical reflection abouteventsthey implicitly focus on representations from one system rather than the other; and they differ on which system of representations 18ALVIN I. GOLDMAN they select. Since Davidson did not dispute the existence of grander actionshedidnotdenythatbodilymovementshavehigher-orderpropertiesinaddi-tiontogeometricorkinematiconeshedidnotmentallyrepresentactionsex-clusivelybymeansofthespatio-temporalsystem.Heusedthekind-basedsys-tem as well. However, he might have privileged, or concentrated on, the deliver-ancesofthespatio-temporalsystemascomparedwiththekind-basedsystem. This is my conjecture. In other words, I dont suggest that unifiers turn off their kind-based system whenthinking (during everyday business) about events. Nor do I suggest that property-exemplification theorists turn off their spatio-temporal system when thinking about events. It is only a matter of which representational formatis given greater weight or preferencein ones theorizing moments. Each representationalformatcanbeasourceofintuitionsaboutevents.Atheorist merelychoosestoaccentuateonefamilyofintuitionsratherthananother.Nor doImeantodenythatthetwotypesoftheoristsareunmovedbythevarious philosophicalprinciplesthatareproposedanddebated.Obviouslytheyareso moved. Nonetheless, I suspect that theorists of different persuasions dwell more persistently(intheirminds)ononefamilyofintuitionsascomparedwiththe another,andthatsanon-negligiblefactorinwhatinfluencestheirtheoretical judgments.Letmereturnnowtothecomparisonsdrawnearlier(atthebeginningof section2)withthemetaphysicsoftimeanddisgustingness.Inbothcasesthe (initial) question is: what is the property in the world that makes the best fit with a term used uncritically and unreflectively to characterize objects or relations in theworld?Astheexampleoftimeillustrates,findingasuitablefitmaynotbe an easy matter. There may be less in the world than what our cognitive struc-turesinclineustosuppose.Asthedisgustingnessexampleillustrates,wemay alsofindthatthingsintheworldandinourcognitiveandemotionalreper-toireoffer a more complex array of alternativesthan offered by our language. (Thelatteroffersjustasinglepredicatedisgustinganditsapproximatesyno-nyms). In the case of disgustingness, I proposed that the best solution is to reject the assumption that there is a single property of disgustingness, even understood inresponse-dependentterms.Instead,thereisDISGUSTINGNESS1andDIS-GUSTINGNESS2.Similarly,themetaphysicalmoralthatmayemerge(inpart) fromthefindingsofcognitivescienceisthatweshouldabandontheoriginal assumptionofbothunifiersandmultipliersthatexactlyoneoftheirviewsis right (the only interesting question being which one?). Instead we should con-cludethatbothareright.Thebestsolutionistocountenancetwometaphysical categories of events, EVENTS1 and EVENTS2. This is how cognitive science can play a role in the conduct of metaphysicalizing.12

Alvin I. Goldman Rutgers University A PROGRAM FOR NATURALIZING METAPHYSICS19 NOTES 1. Thethemeofthepresentpaperisveryclosetothatofanearlierone:(Goldman 1989).Intermsoftheearlierpapersdivisionbetweendescriptiveandprescriptive metaphysics, the present paper emphasizes the prescriptive enterprise. 2. Although the passage appears in a co-authored book by Conee and Sider, Conee is listed as the lone author of the chapter containing this passage. 3. Insteadofdescribingourauditoryexperienceofwordboundariesasfeaturing silences (as Pinker puts it), we might describe the matter in the language of categorical perception.Theboundarybetweentwowordsisperceivedasbeingdiscreteand discontinuous,whereasinfactitisasnon-discreteandcontinuousaswhatisperceived within a spoken word. Even in this formulation of the phenomenon, there is an illusion or errorinournaverepresentationofwordboundariesthatisopentocorrectionby cognitive science. Thanks to Roberto Casati for emphasizing this alternative formulation. 4. Sarah-Jane Leslie suggested this as an example of response-dependence.5. I cull this summary from Prinz (2006). 6. EvidencesupportingthisexplanationwasfirstproducedbyEllisandYoung (1990).TheypredictedthatCapgraspatientswouldnotshowthenormallyappropriate skinconductanceresponsestofamiliarfaces,andthispredictionwasconfirmed. Subsequent research has added further confirmation. 7. For an instructive novelistic account of the syndrome, see Powers 2006. 8. Iassumethattheunifierandmultiplierapproachesexhausttheterritory.That assumptionmayappearproblematic,becausethereis,forexample,thepartor componential view of events (Thalberg 1971; Thomson 1971, 1977), which treats some members of a candidate list of identical events as spatiotemporal parts of other members. For present purposes, however, the part theory will be considered a species of the multi-plier view. It does, after all, answer the standard question of identity, in relevant types of cases, by saying different rather than same.9. For a defense of eliminativism about events, see Horgan 1978. 10. ThankstoDeanZimmermanforsuggestingtheparallelwithBennett.Iamnot convinced,however,thatthetwo-conceptions-of-eventsthesisquitefitsBennetts consideredview.Inchapter8ofhisbook,Bennettsaysthatagoodsemanticsofevent names should be intermediate between Kims view and Quines, where Quine defended a zonal view that identifies any two happenings in the same spatiotemporal zone. Bennett doesnotsaythatbothkindsofsemanticsareright.Onthecontrary,hesearchesfora single intermediate position between them. He despairs of finding one, saying The truth lies between Kim and the Quinean, but there is no precise point between them such that it lies there (188: 126). In searching for a single intermediate position, Bennett appears to disavow a two-conceptions-of-events view, in which both conceptions are correct. 11. Thischaracterizationisslightlycontentious,becausemanymultiplierswould deny that the events in question take placeat the sametime and place. Forexample,the timeofkillingapersonmightincludethetimeofthevictimsdeath,makingitlonger thanthetimeofpullingthetrigger.Theexacttimesandplacesofcomplexactionsand eventsisahighlyvexingissue.Nonetheless,theformulationinthetextisauseful approximation.12. Thanks to Roberto Casati, Holly Smith, and Dean Zimmerman for valuable com-ments on an early draft of this paper. Zimmerman was particularly helpful in persuading methatcertainmaterialcontainedinthatdraftmightprofitablybedeleted(although those were not his words).20ALVIN I. GOLDMAN REFERENCES Bennett,J.,1988,EventsandTheirNames,Indianapolis(IN):HackettPub-lishing. Berthoz, S., Armony, J. L., Blair,R. J., and Dolan,R..J., 2002, An fMRI Study ofIntentionalandUnintentional(Embarrassing)ViolationofSocial Norms, Brain 125: 16961708. Blair,R.J.,Colledge,E.,Murray,L.,andMitchell,D.G.,2001,ASelective ImpairmentintheProcessingofSadandFearfulExpressionsinChildren with Psychopathic Tendencies, Journal of Abnormal Child Psychology 29: 491498. Bregman,A.S.,1990,AuditorySceneAnalysis:ThePerceptualAnalysisof Sound, Cambridge (MA): MIT Press. Capgras, J., and Reboul-Lachaux, J., 1923, Lillusion des Sosies dans un Delire SystematiseChronique,BulletindelaSocietCliniquedeMedicineMen-tale 2: 616. Casati,R.,andVarzi,A.C.,2007,EventConcepts,inT.F.ShipleyandJ. Zacks (eds.), Understanding Events: How Humans See, Represent, and Act on Events, Oxford: Oxford University Press, in press.Conee,E.,andSider,T.,2005,RiddlesofExistence:AGuidedTourofMeta-physics, Oxford: Clarendon Press. Curtis,V.,Aunger,R.,andRabie,T.,2004,EvidencethatDisgustEvolvedto Protect from Risk of Disease, Proceedings of the Royal Society: Biological Science Series B, 271(4): S131S133. Davidson,D.,1967,TheLogicalFormofActionSentences,inN.Rescher (ed.),TheLogicofDecisionandAction,Pittsburgh:UniversityofPitts-burgh Press; reprinted in Davidsons Essays on Actions and Events, Oxford: Oxford University Press, 1980, pp. 105122. Davidson,D.,1969,TheIndividuationofEvents,inN.Rescher(ed.),Essays inHonorofCarlG.HempelDordrecht:Reidel,pp.216234;reprintedin Davidsons Essays on Actions and Events, Oxford: Oxford University Press, 1980, pp. 163180. Davidson,D.,1971,Agency,inR.Binkley,R.Bronaught,andA.Marras (eds.), Agent, Action, and Reason, Toronto: University of Toronto Press, pp. 337;reprintedinDavidsonsEssaysonActionsandEvents,Oxford:Ox-ford University Press, 1980, pp. 4361. Dreier, J., 1990, Internalism and Speaker Relativism, Ethics 101: 722748. Ekman, P., 1992, An Argument for Basic Emotions, Cognition and Emotion 6: 169200. Ellis, H. D., and Young, A. W., 1990, Accounting for Delusional Misidentifica-tions, British Journal of Psychiatry 157: 239248. A PROGRAM FOR NATURALIZING METAPHYSICS21 Goldman,A.I.,1965,Action,Ph.D.dissertation,PrincetonUniversity,Depart-ment of Philosophy. Goldman, A. I., 1970, A Theory of Human Action, Englewood-Cliffs (NJ): Pren-tice-Hall. Goldman, A. I., 1971, The Individuation of Action, Journal of Philosophy 68: 761774. Goldman, A. I., 1989, Metaphysics,Mind,andMental Science, Philosophical Topics17:131145;reprintedinGoldmansLiaisons:PhilosophyMeets theCognitive and Social Sciences,Cambridge (MA):MITPress, 1992, pp. 3548. Greene, J. D., Sommerville,R.B., Nystrom,L. E., Darley,J.M.,Cohen, J. D., 2001,AnfMRIInvestigationofEmotionalEngagementinMoralJudg-ment, Science 293: 21052108. Haidt,J.,McCauley,C.,andRozin,P.,1994,IndividualDifferencesinSensi-tivitytoDisgust:AScaleSamplingSevenDomainsofDisgustElicitors, Personality and Individual Differences 16: 701713. Hardin,C.L.,1988,ColorforPhilosophers:UnweavingtheRainbow,Indian-apolis (IN): Hackett Publishing. Hejmadi, A., Rozin, P., and Siegal, M., 2004, Once in Contact, Always in Con-tact:ContagiousEssenceandConceptionsofPurificationinAmericanand Hindu Indian Children, Developmental Psychology 40: 467476. Horgan, T., 1978, The Case Against Events, Philosophical Review 87: 2847. Hurwich, L. M., and Jameson, D., 1957, An Opponent-Process Theory of Color Vision, Psychological Review 64: 384408.Johnston, M., 1989, Dispositional Theories of Value, Proceedings of the Aris-totelian Society 63 (Supplement): 139174. Kelly, D., 2007, Projectivism Psychologized: The Philosophy and Psychology of Disgust, Ph.D. dissertation, Rutgers University, Department of Philosophy. Kim,J., 1966, On the Psycho-physical Identity Theory, American Philosophi-cal Quarterly 3: 227235. Kim, J., 1969, Events and Their Descriptions: Some Considerations, in N. Re-scher(ed.),EssaysinHonorofCarlG.Hempel,Dordrecht:Reidel,pp. 198215. Kim, J., 1973, Causation, Nomic Subsumption and the Concept of Event. Jour-nal of Philosophy 70: 217236. Kim,J.,1976,EventsasPropertyExemplifications,inM.BrandandD. Walton (eds.), Action Theory. Dordrecht: Reidel, pp. 159177. Lewis,D.K.,1986,Events,inhisPhilosophicalPapers,vol.2,NewYork: Oxford University Press, pp. 241269. McDowell,J.,1985,ValuesandSecondaryQualities,inT.Honderich(ed.), Morality and Objectivity, London: Routledge, pp. 110129. 22ALVIN I. GOLDMAN McNaughton, D., 1988, Moral Vision: An Introduction to Ethics, Oxford: Black-well.McLaughlin, B., 2003, Color,Consciousness,and ColorConsciousness,in Q. Smith and A. Jokic (eds.), Consciousness, Oxford: Oxford University Press, pp. 97154.Milner, A. D., and Goodale, M., 1995, The Visual Brain in Action, Oxford: Ox-ford University Press. Moll, J., De Oliveirra-Souda, R., and Eslinger, P. J., 2003, Morals and the Hu-man Brain: A Working Model, Neuroreport 14: 299305.Paul, L. A., 2000, Aspect Causation, Journal of Philosophy 97: 235256. Pinker, S., 1994, The Language Instinct: How the Mind Creates Language, New York: William Morrow. Powers, R., 2006, The Echo Maker, New York: Farrar, Strauss and Giroux. Prinz,J.,2006,TheEmotionalBasisofMoralJudgments,PhilosophicalEx-plorations 9: 2943. Rozin,P.,Nemeroff,C.,Horowitz,M.,Gordon,B.,andVoet,W.,1995,The BordersoftheSelf:ContaminationSensitivityandPotencyoftheBody AperturesandOtherBodyParts,JournalofResearchinPersonality29: 318340. Rozin,P.,Haidt,J.,andMcCauley,C.,2000,Disgust,inM.LewisandJ.M. Haviland-Jones(eds.),HandbookofEmotions,2ndedition,NewYork: Guilford, pp. 637653. Sanfey,A.G.,Rilling,J.K.,Aronson,J.A.,Nystrom,L.E.,andCohen,J.D., 2003,TheNeuralBasisofEconomicDecision-MakingintheUltimatum Game, Science 300: 17551758. Schnall,S.,Haidt,J.,andClore,G.,forthcoming,IrrelevantDisgustMakes Moral Judgment More Severe, for Those Who Listen to Their Bodies. Sharon,T.,andWynn,K.,1998,IndividuationofActionsfromContinuous Motion, Psychological Science 9: 357362. Shipley,E.F.,andShepperson,B.,1990,CountableEntities:Developmental Changes, Cognition 34: 109136. Simons,P.,2003,Events,inM.LouxandD.Zimmerman(eds.),TheOxford HandbookofMetaphysics,NewYork:OxfordUniversityPress,pp.357385. Spelke, E., 1988, Where Perceiving Ends and Thinking Begins: The Apprehen-sionofObjectsinInfancy,inA.Yonas(ed.),PerceptualDevelopmentin Infancy, vol. 2, Hillsdale (NJ): Erlbaum, pp. 219234. Spelke,E.,Breinlinger,K.,Macomber,J.,andJacobson,K.,1992,Originsof Knowledge, Cognition 99: 605632. Thalberg,I.,1971,SinglingOutActions,TheirPropertiesandComponents, Journal of Philosophy 68: 781786. A PROGRAM FOR NATURALIZING METAPHYSICS23 Thomson,J.J.,1971,IndividuatingActions,JournalofPhilosophy68:771781. Thomson,J.J.,1977,ActsandOtherEvents,Ithaca(NY):CornellUniversity Press. Wagner,L.,andCarey,S.,2003,IndividuationofObjectsandEvents:ADe-velopmental Study, Cognition 90: 163191. Wiggins, D., 1991, A Sensible Subjectivism,in hisNeeds, Values, Truth:Es-says in the Philosophy of Value, Oxford: Blackwell, 185211. Wright,C.,1992,TruthandObjectivity,Cambridge(MA):HarvardUniversity Press. Wynn,K.,1996,InfantsIndividuationandEnumerationofActions,Psycho-logical Science 7: 164169. Xu, F., and Carey, S., 1996, Infants Metaphysics: The Case of Numerical Iden-tity, Cognitive Psychology 30: 111153.