Salmon 1987

Ridgeview Publishing Company
Existence
Author(s): Nathan Salmon
Source: Philosophical Perspectives, Vol. 1, Metaphysics (1987), pp. 49-108
Published by: Ridgeview Publishing Company
Stable URL: http://www.jstor.org/stable/2214143 .
Accessed: 28/04/2013 16:07
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
Ridgeview Publishing Company is collaborating with JSTOR to digitize, preserve and extend access to
Philosophical Perspectives.
http://www.jstor.org
This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

All use subject to JSTOR Terms and Conditions
PhilosophicalPerspectives,1, Metaphysics,1987
EXISTENCE*
NathanSalmon
ofCalifornia
University
SantaBarbara
I shalldiscussherethetopicsofexistenceand nonexistence,ofwhat
it is foran individualto be actual and what it is foran individual
notto be actual. WhatI shall have to say about thesemattersoffers
littletoward our primordialneed to discoverthe Meaningof Ex-
istence,butI hope tosay some thingsthatwillsatisfy themoremodest
ambitionofthoseofus who wishto knowthemeaningof'existence'.
I shall also say some thingsthatbear on issuesin the grandesttradi-
tions of Philosophy.
The questionsI shalladdressherecan be approachedthroughthe

followingthought-exercise: For every one of us, priorto our con-
ception,the odds againstthe verygametesfromwhichwe in fact
developed comingtogetherto develop intoa particularhumanin-
dividualare astronomical.There are countlessbillionsof potential
pairingsof a humanspermcell witha humanovum thatare never
realized. Everyoneof us is among the elite groupof Elect whose
gametesdid manage, againstall odds, to unitein the normalman-
ner and develop intoa humanindividual.LetS be a particularmale
sperm cell of my father'sand let E be a particularovum of my
mother'ssuch thatneithergamete ever uniteswith any otherto
develop intoa humanzygote.Let us name the (possible)individual

50 / NathanSalmon
who would have developed fromtheunionofS and E, ifS had fer-

tilizedE in thenormalmanner,'Noman'.1Nomandoes not exist in
the actual world,but thereare many possibleworldsin which he
(it?)does exist.Thisis justto say thatNomandoes notactuallyexist
but he mighthave existed. Noman is, like all of us, a possible in-
dividual;it is trueof him,and it is likewisetrueof each of us, that
we mighthave existed.But somethingmore can be said about us
thatcannotbe said about Noman.There is a seeminglyimportant
differencebetween Noman and us. We are actual, Noman is not.
Nomanis merelypossible.Whatdoes thisdifference betweenNoman
and us consistin?Whatis it about us in virtueofwhichwe, butnot
Noman,may be said to be 'actual'?Whatis it forsomethingto have
the ontological status of being actual, and is there any special
metaphysicalsignificance attachedto somethingsolelyby virtueof
itsbeingactual? Is theresuch a thingas thepropertyof existence,
or thepropertyofactuality-a propertythatNomanlacks,and that
somethinghas solelyby virtueof thefactthatitexistsor is actual?
Whateveractualityis,we seem to matterin a way thatNomandoes
notseem to matterat all. (Nomandoes not mattereven to me, and
we are brothers!Well, at least we are brothersacross possible
worlds.2)Does thisrepresentan objectivefactabout us vis a vis
Noman and his kind,or is it ultimatelya formof prejudiceand
discrimination on our part?Are we objectivelybetterthan,or ob-
jectivelybetteroffthan,Nomanby virtueofthe factthatwe have
actuality,or solelyby virtueof the factthatwe exist,whereas he
does not?Is it objectivelybetterto have thisontologicalstatuscall-
thanto lackit?Ifso, whatis itaboutactualitythatmakes
ed 'actuality'
us countforso muchmorethanNoman?Is actualitysomethingwe
mighthave lacked? Specifically,in those possibleworldsin which
we do not exist,are we not actual? Conversely,in those possible
worldsin whichNoman exists,is he actual? In a possibleworldin
whichNoman existsand I do not,whichone of us inhabitsthe ac-
tual world?Does Noman have any properties?Does he lack every
property?Do we have any propertiesin those possible worldsin
whichwe do not exist?
In a sense, the question'What is it forsomethingto be actual?'
has one simple,correctanswer: For somethingto be actual is for
itactuallyto be-that is,foritactuallyto exist.Butthisansweronly
trades one ontologicalquestion fortwo new ones. What is it for
somethingto be, or to exist,and what is it forsomethingactually

Existence/ 51
to be the case? Ifwe can answerthese two questionssatisfactorily,

we will therebyhave an answerto the questionof what it is to be
actual.
Let us beginwiththequestionofexistence.Considerfirsta slight-
ly differentquestion:Whatexists?Quine pointedout thatthistime-
honoredontologicalquestionhas itscorrectanswerin a singleword:
Everything.3 Does thisobservationhelp us withour slightlymore
questionofwhat existenceis? It seems so. Iftheanswer to
difficult
the questionof whatexistsis the universalquantifier'everything',
thenforsomethingto existis foritto be one ofeverything. Butdoes
thisconstituteany sortof progresswithrespectto our questionof
what existenceis? Whatdoes it mean to say thatsomethingis "one
of everything"?
Modifying Berkeley'sfamousslogan,Quine gave substanceto the
idea thatwhatexistsis what is covered by theuniversalquantifier
withhisequallyfamousslogan'To be is to be a value ofa variable'.4
Taken as a responseto the question'What is existence?',Quine's
sloganseems at leastextensionally correct.Everyexistingindividual
is indeed the value of some variable or other, under some co-
operativeassignmentofvalues to variables,and itwouldseem that
everythingthat is assigned to a variable as its value is "one of
everything," i.e., itexists.Butitcannotbe seriouslymaintainedthat
being,in thesense of 'existence',simplyis thestateor conditionof
being the value of a variable,undersome assignmentof values to
variables. When Hamlet (pretendingthe play were nonfictitious)
agonized over the questionof whetherto be or not to be, he was
preoccupiedwithweightiermattersthanthequestionofwhetheror
not to be the value of a variable.Iftherewere no variables,would
therebe nothing?The dinosaurshad existence,buttheydidn'thave
variables. Perhaps there were no variables at the time of the
dinosaursforthemto be the values of. To be sure, the geometric
shapes and patternsthatformthe lower case italic'x', 'y', and 'z'
existedeven then,butwere theyvariables,and were functions from
themto objectsassignmentsof values to variables?Ifit is supposed
thattheywere, on the groundsthatin some futurelanguage they
are, thenit probablyshould also be said thatanythingthatmight
conceivablybe used as a variablein a possiblelanguageis a variable
(on the groundsthatany such objectis a variablein some possible
language),and any singularyfunction fromsuchobjectsis an assign-
mentofvalues to variables.IfQuine'ssloganis understoodto mean

52 / NathanSalmon
thatforsomethingto existis forit to be in the range of a function

whose domain is a set of objects that mightsomeday serve as
variables, one mightas well skip the variables and theirvalue-
assignmentsaltogetherand say thatto be is to be an elementof a
set. But thenwhynotsimplysay thatto be is to be the elementof
a singleton,or unitset?As explicationsofexistence,thesesomehow
fall flat. But I believe we have strayed fromQuine's intended
meaning.5
Taken literally,it is doubtfulthatQuine's slogan is even exten-
sionallycorrect.The dinosaursmaybe thevalues ofsome oftoday's
variables,undersome assignments, but none exist.(The dinosaurs
once existed,of course,but sadly none existtoday.)Assignmentof
past dinosaursto some presentvariablesis requiredto give thecor-
rectsemanticsfora suitableformalization ofsuchsentencesas 'There
was a dinosaurthatthisis a fossilof'6or the precedingsentence.
IfHamlet(pretendingthe play were nonfictitious) had decided not
to be, he wouldnothave ceased to be thevalue ofa variable.Quine's
slogan mightbe understoodinsteadas the claim thatto be (or to
exist)at a timet in a possibleworldw is to be thevalue ofa variable
undersome assignmentof values to variableswithrespectto t and
w. At least thisis extensionallycorrect.But it puts the cart before
the horse.The notionof a functionbeing an assignmentof values
to variableswithrespectto a timet and a possibleworldw is defin-
ed in termsof the notionof existence:an assignmentof values to
variables is an assignmentwithrespectto t and w if and only if
everythingit assigns existsat t in w.
Myclaimthatpast individualsare the presentvalues ofvariables
even thoughthesepast individualsno longerexistmayconflict with
the doctrinethatto be is to be a value ofa variable,but it does not
conflictwiththealternativedoctrine(extractedfromQuine'sobser-
vation thatthe universalquantifiercorrectlyanswersthe question
'Whatexists?')thatto be is to be "one ofeverything" (whateverthat
means). The universaland existentialquantifiers mustnot be con-
fusedwiththevariablestheybind.7A (typical)universalgeneraliza-
tion r(Va)(Pa' is trueunderan assignmentofvalues to variabless,
withrespectto a given timet, if and only ifevery (past,present,
or future)individuali thatexistsat t is such that jt is trueunder
theassignment s', withrespectto t,wheres' is theassignmentthat
assignsi to a and is otherwiseexactlythe same as s. The assign-
ments may already assignindividualsthatdo not existat t to cer-

Existence/ 53
tainvariables;hence,theassignments s' mayalso assignindividuals

thatdo not existat t to some variables,butnot to a. The universal
quantifier restrictsitsattention(typically)to assignmentsthatassign
existingindividualsto thevariableitbinds.Existenceper se matters
nothingto thevariablesthemselvesor theirvalue-assignments. Not
onlyare past individualsthe presentvalues of variables,butfuture
individualsare as well. Some possible assignmentsof values to
variableseven assignNomanas value to some variables.In fact,some
modal constructions requiresuchassignments, e.g., 'The gametesS
and E mighthave been unitedinthenormalmannerto develop into
an individual'.8It is thequantifier,
and notthevariableitbinds,that
insistson nothingbuttheexistent.And it insistson nothingbut the
existentonly as values forits adjacent variable,not as values for
othervariables in its less immediatevicinity.
It is a mistakein any case to attemptto explicatea metaphysical
notionby means of essentiallysemanticnotions.(Again,I believe
thisis not Quine's intent.See note 5.) One could say thatto be is
to be an elementofthe unionofthe extensionof an Englishquan-
tifier.Whynot?One mightas well say thatto be a personis to be
an elementof the extensionin Englishof 'person',to knowa given
proposition is to standintherelationexpressedby'knows'in English
to thatproposition,and so on. For one thing,we are wrong:the at-
tributesofexistence,beinga person,and knowingare notessential-
ly semanticin nature.For another,we are stillleftwonderinghow
theextensionofa quantifier in Englishis secured.Whatfeaturemust
individualspossess ifa class ofthemis to be an elementof the ex-
tensionof the Englishquantifier'there is'? Itneed notbe a mistake,
however,to use (ratherthanto mention)an Englishquantifier in at-
temptingto explicateexistence.We need to findan adequate way
to understandthe slogan 'To be is to be one of everything'.Some
progressis made towardansweringthequestionWhat is existence?'
ifour conceptof existencecan be definedin termsof our concept
of everything.
II
Philosopherswho address the questionsof what it is foran in-

dividualto exist,or what it is foran individualto be actual,often
do so with referenceto the fallacythey have uncovered in the

54 / NathanSalmon
classicalOntologicalArgumentforGod's existence.Indeed,the On-

tologicalArgumentis usefulas a vehicle by whichthisand other
issues in ontologyand the philosophyof logic may be introduced
and sharpened.Inwhatis perhapsitssimplestform,theOntological
Argumentis the following:
(la) The divineindividualis divine.
(lb) Any individualthatis divine exists.
Therefore,
(1c) The divine individualexists.
Let us call this 'Version 1'. The term 'divine' serves here as a
schematicterm,whichis to be interpreted relativeto a contextin
whichan argumentofa particularOntologicalArgueris in question.
If our concernis withDescartes'sargumentfromhis fifth Medita-
tion,'divineindividual'is to be interpretedto mean individualthat
has every perfection(with 'perfection'interpretedin Descartes's
sense). If our concern is withAnselm'sinstanceof the argument
schema (or at least the best knownof Anselm'sinstances,as given
in ChapterII ofhisProslogion),'divineindividual'is to be interpreted
to mean individualwhose magnitudeofgreatnessexceeds anyother
possible magnitude of greatness (with 'possible' interpretedin
Anselm'ssense of 'conceivable'and 'great'interpreted in his sense
of 'great').For presentpurposes,we may assume that,in each case,
the relevantconceptof divinityis such thatit is provable,or other-
wise manifesta priori,thattherecannotbe two or moredivinein-
dividuals.The OntologicalArguerassumespremise(la) as a logical
or manifesttruth,and contendsthat premise(lb) is likewise an
analyticor conceptualor demonstrabletruth.
Since the firstpremiseis sometimesregardedby the Ontological
Argueras a logicaltruth,theargumentmay be formulated withthe
firstpremiselefttacit,notincludedas an explicitpremiseoftheargue-
mentand notsupportedby further, additionalargument. Anselmpro-
vided an explicitargumentin supportof the premise,althoughhis
argumentclearlyindicatesthathe regardedthetruthofthepremise
as manifest, a truththateven theatheist'fool'is convincedof.When
the premiseis made explicithowever,it shouldimmediatelystrike
thereaderthatthereis a problemwithit.The atheistand theagnostic
doubt(by disbelievingand by suspendingjudgement,respectively)
thatthereexistsany divineindividualin the first place. Whyshould

Existence/ 55
theybe expected to acquiesce in the assertionthatthe divinein-

dividualis divine?
The intendedimportof Version1 is apt to be lost on the reader
unless he or she understandsa criticalfeatureof the argument:it
purportsto involvequantificationover morethingsthanare dreamt
of in Quine's philosophyof what exists.A more explicitand more
sophisticatedversionofthe OntologicalArgumentis the following:
(2a) The divine possibleindividualis a divinepossible
individual.
(2b) Any possible individualthatis divineexists.
Therefore,
(2c) The divinepossible individualexists.
Let us call this'Version2'. The difference between the two ver-
sions is thatVersion2 explicitlyinvolves,or at least explicitlyat-
temptsto involve,so-calledpossibilistquantification ratherthanso-
called actualistquantification. That is, version2 explicitlypurports
to employquantification over all thatmighthave existed,including
whatdoes notexist,ratherthanmerelyoverall thatdoes exist.The
readershouldtake specialnoteoftheimportofthepremisesofVer-
sion 2. The conclusionoftheargumentis supposedto be notmerely
thatthedivinepossibleindividual mighthave been an existentdivine
individual, butthatitactuallydoes existandactuallyis divine.Hence
premise(2a) mustbe read in sucha way thatitassertsthatthepossi-
ble individualthatactuallyis divineactuallyis divine,and premise
(2b) mustbe read in such a way thatit assertsthatany possiblein-
dividualthatactuallyis divineactuallyexists.Bothpremisesare in-
tendedto be taken in such a way as not to presupposethe real ex-
istenceofany divineindividual.So understood,whateverelse may
be problematicwiththeargument'sfirstpremise,itis notclear that
itsimplybegs thequestionagainsttheatheistor theagnostic.In fact,
one of thephilosophicalissues raisedby theOntologicalArgument
is preciselywhetherone can predicatea property (inthiscase, divini-
ty)of a possibleindividualwithoutpresupposingthe real existence
of the possible individual.The argumentcannot be summarily
dismissedon the groundsof an uncontroversial prohibition against
predicatingpropertiesofpossibleindividualswhose existenceis not
to be presupposed.Indeed,thereare some propertiesthatcan be
predicatedof possibleindividuals(such as Noman)withoutpresup-

56 / NathanSalmon
posingtheexistenceoftheseindividuals-forexampletheproperty
of not existing,and its entailments.
Ofcourse,in attributing Version2 to a particularhistoricalfigure,
such as Anselmor Descartes,some charitymay be requiredin in-
terpreting the modal locutionsinvolved;the term'possible'in the
phrase 'possible individual'need not be interpretedto mean the
modallogician'smetaphysical possibility
(although, itprobablyshould
be so interpreted fora contemporary OntologicalArguer).Anslem's
instanceofVersion2 is obtainedby interpreting the phrase'possi-
ble individual'in Anselm'ssense of 'thingthatexistsin intellectu'
(and by interpreting 'divine individual'to mean individualwhose
magnitudeof greatnessexceeds any otherconceivablemagnitude
of greatness).We may assume here thatthe conceptof divinityis
such thatit is provableor somehow manifesta priorithatno two
possibleindividualsare actuallydivine.We willreturnto the ques-
tionof whetherthe atheistor the agnosticneed denythatthereis
one possible individualwho is actuallydivine.
Once possibilistquantification is admitted,we may pose Quine's
ontologicalquestionin a new light:Whatpossibleindividualsexist?
Quine's simpleand correctanswerto the question'Whatexists?',if
resubmitted, apparentlybecomessimplyincorrect-provided itis in-
terpreted(contraryto Quine's intent)as the possibilistratherthan
theactualistuniversalquantifier. Noteverypossiblethingexists.Or
so it would seem. In any case, it is not necessarythateverything
actuallyexists;theremighthavebeen individuals thatdo notactually
exist. Noman,forinstance.
Is theEnglishword'everything' theactualistuniversalquantifier,
or is itthepossibilistuniversalquantifier? Is our ordinary,everyday
conceptofeverythingtheconceptofeverything thatexists,or is it
the conceptof everything thatmighthave existed,includingwhat
does not actuallyexist?Is it somehow(ambiguously)both?Or is it
none of the above? The doctrinethatthe standardquantifiersof
naturallanguage (the Englishwords'everything', 'something',etc.)
are possibilistquantifiers is sometimescalled 'possibilism',and the
doctrinethattheyare actualistquantifiers is sometimescalled 'ac-
tualism'.9In observingthatthestandardEnglishuniversalquantifier
is thecorrectanswerto thequestion'Whatexists?',Quine proclaims
his endorsementof actualism,and assumes his readers agree. I
believe that actualism is indeed the predominantview among
philosophersof logic and philosophersof language. My own view

Existence/ 57
is thatthequantifiers ofEnglishare typically actualist(andpresentist,

i.e., rangingwithrespectto a timet over onlythosethingsthatexist
at t)-that amongpotentialrestrictions on our use ofquantification,
restriction to existingthingsis,so to speak,the"defaultvalue"-but
thatthedomainofquantification maybe, and veryoftenis,adjusted
eitherupwardor downwardin variousways, at the drop of a hat.
(Considerour readinessto quantifyover no longerexistingobjects
in discourseabout the past, and forinstancein 'This is a fossilof
some dinosaur'.)Still,I believe thatour ordinary,everydayconcept
ofeverything is theconceptofeverything
(simpliciter) thatexists-
no more and no less-and I shall assume thisconstrualthroughout
mostof thisessay. In particular,then,I assume thatit is legitimate
to relyon theconceptofactualistuniversalquantification in attempt-
ingto explicatewhatexistenceis, forwe are merelyrelyingon our
ordinaryconceptofeverything. (Indeed,unlesswe mayrelyon our
priorgraspofactualistquantification, I doubtthata philosophically
satisfactory definition
or analysisofexistencecan be given.See note
16.) I shall notassume,however,thatthereis anythingillegitimate
aboutpossibilist perse or abouttheconceptofeverypossi-
quantifiers
ble individual.KitFine has shownthatthe possibilistuniversaland
existentialquantifiers are fullydefinableusingthe standardmodal
operatorsin tandemwithactualistquantifiers over bothindividuals
and "propositions"qua sets ofpossibleworlds,or alternatively, us-
ing standardmodal operatorsin tandemwithactualistquantifica-
tion over both individualsand possible worlds togetherwith a
predicatefora possibleworld'sbeingrealized.10In fact,the import
ofa possibilistquantificational assertioncan oftenbe easilyexpress-
ed usingonlyfirst-order machinery through thejudicioususe ofmodal
operators(includingan operatorforsomething'sactuallybeingthe
case) in tandemwithactualistquantifiers onlyover individuals(ex-
cludingpossibleworlds).Occurrencesofpossibilist quantifiers in this
essay are indicatedthroughout by modaladornment,in themanner
of'everypossibleindividual', and so on. (Do notread 'thereis a possi-
ble individualthatis such-and-such' as meaningthatthereexistsan
individualthatis bothpossibleand such-and-such. Insteadit means
thattheremighthave existedan individualthatactuallyis such-and-
such,etc.) Unadornedoccurrencesof Englishquantificational locu-
tionsare to be read actualistically (exceptin certainpassages in Sec-
tionVI below, wherethe phrase'everyobject'takeson a distinctly
Meinongianair).

58 / NathanSalmon
The actualistuniversalquantifier'everything'remainsa correct

answer to the question'What possible individualsexist?',but it is
nota veryusefulresponse.It does nottellus, forexample,whether
Nomanis one ofthepossibleindividualsthatexists-except by tell-
ing us thathe existsifand onlyifhe is one of everything. It is not
yet clear what it means to say thata possibleindividualis "one of
everything."Descartes'scogitoergo sum may be takenas specify-
ing one possibleindividualthatenjoysthe ontologicalstatusof real
existence (viz., oneself).The OntologicalArgumentpurportsto
specifyanother,as do existence proofs in mathematics.Unlike
Quine's actualist universal quantifierresponse, however, these
responsesofferonlyparticular instances,notan exhaustivespecifica-
tion.The question'Whatpossibleindividualsexist?'may be posed
as a requestfora philosophicalanalysisoftheconceptofexistence,
in thesense ofan illuminating specification ofa necessaryand suffi-
cientconditionC such that,necessarily,a possibleindividualexists
ifand only if it satisfiesC. This request is inextricablytied to our
questionof what existenceis. Whatis it fora possibleindividualto
be "one of everything"?
The explicituse ofpossibilistquantification inVersion2 oftheOn-
tologicalArgumentmayshiftthecritic'sfocusfromthefirst premise
ofthe argumentto the second. Once possibilistquantification is ad-
mitted, itmightbe objectedthatpremise(2b)is nota conceptualtruth,
on the groundsthatit is logicallypossibleforthereto be a merely
possibleindividualthatis divinebut does not exist. Ifpremise(2a)
is to be taken as manifesteven to the atheist,and ifpremise(2b)
is to be takenas strongenoughto ensurevalidityforVersion2, then
surelysomething needsto be saidbytheOntologicalArguerto assure
the reader thatpremise(2b) is indeed a conceptualtruth.In fact,
historically,Ontological Arguers have offeredsupport for their
premise(2b) by means ofanothera prioriargument.Sometimesthis
supporting argumentis verybriefand mentioned onlyinpassing("Ex-
istenceis a perfection").Sometimesit is the veryheartof the On-
tologicalArguer'smoregeneralargument.Anselm'ssupportforhis
premise (2b) came in the formof the notoriousargumentthat,
necessarily,the magnitudeof greatnessof any possibleindividual
thatexistsexceeds its actual magnitudeof greatnessifit does not
actuallyexist;hence,anypossibleindividualwhose actualmagnitude
ofgreatnessexceeds anyotherpossiblemagnitude ofgreatness, since
itcouldexist,mustexist-otherwise, itsactualmagnitude ofgreatness

Existence/ 59
would notexceed itsown possiblemagnitudeof greatness.(Gotit?)

That shouldsatisfythefoolwho doubtsthatpremise(2b) is concep-
tuallyor demonstrablytrue.
It would be in the spiritof the OntologicalArgument,however,
to replyto theobjectionthatpremise(2b) is nota conceptualtruth
by pointingout that,ifthe objector'sconceptof divinitydoes not
alreadyincludethe conceptof existenceas a necessaryor entailed
condition(so thatthe objectordoes not read premise(2b) as a con-
ceptualor logicallydemonstrabletruth), we may formthenew con-
ceptofexidivinity,definedin termsoftheobjector'sconceptofdivini-
ty thus:
exidivine=def. divineand existent.
Now we replace the word'divine'by 'exidivine'throughout
Ver-
sion 2, to obtainVersion3:
(3a) The exidivinepossible individualis an exidivinepossible
individual.
(3b) Any possible individualthatis exidivineexists.
Therefore,
(3c) The exidivinepossible individualexists.
Usingthissimplestrategy,the OntologicalArguercan removeany
need to supportthe second premiseby further a prioriargument.
Even thefoolis convincedofthetruth of(3b);thenew secondpremise
is beyond all reasonable doubt.
Well,premise(3b) is beyondall reasonabledoubtprovidedwe can
be persuaded that the concept of exidivinityis a genuine and
legitimateconcept. Ay, there'sthe rub. In fact,in attemptingto
trivializetheargument'ssecondpremisein thisway,theOntological
Arguershiftsthe criticalfocusfromthe second premiseback again
to the first.Version3 apparentlyattemptsto treatexistencein such
a way thatrealexistencemaybe provedofa possibleindividual simp-
ly by conceivingof it as existent.
The OntologicalArgumentmaybe usefulas a device forintroduc-
ingand discussingvariousphilosophicalissues,buttakenpolemical-
ly as a contribution to the debate over God's existenceit is surely
worthless.It is appropriatethatAnselmshouldlabel his opponent
'thefool',since itis difficult
to imaginea genuineatheistor agnostic
who is not also a foolbeing convertedto theismon thestrengthof

60 / NathanSalmon
thispiece of sophistry.That the OntologicalArguement(takenas

a purportedproofofGod'sexistence)involvessome error,therecan
be no doubt. This was conclusivelyestablishedduringAnslem's
lifetimein a reductioad absurdumby his formidablecriticGaunilo,
who firstobserved that if the OntologicalArgumentsucceeds in
demonstrating itsconclusion,thenone can also prove theexistence
ofa fantasyislandbyan exactlyanalogousargument.11 Unfortunate-
ly,thisreductiodoes notpinpointtheerrorin theOntologicalArgu-
ment.The creditforhavinglocated the fallacyin the argumentis
oftenattributed to Kant,who purported to debunktheargument with
his observationthatexistenceis not a predicatethatcan be legiti-
matelyincludedin the definition or concept of something.Kant's
refutationis widelyregardedas conclusive,or at least sound, as
regardsthe versionsof the OntologicalArgumentdiscussedhere.
One exceedinglyplausibleidea thatlends supportto thisrefutation
is thatone cannotcreatenew entitiessimplyby definingtheminto
existence.Ifexistencewereregardedas an admissibledefining prop-
ertyor concept,exactlyon a par withsuch mundaneconceptsas
being green-eyedor being an island,then it would be possible to
initivemerelypossibleindividualssuchas Nomanintotheeliteclub
of Existence,simplyby definingthemas existing.Your next stop:
the TwilightZone.
Kant's observationthat existenceis not an admissibledefining
predicatewas echoed by bothofthetwo greatestfiguresofcontem-
poraryanalyticphilosophy,GottlobFrege and BertrandRussell-
baskingin theglowoftheirpowerful, new quantification
theory,with
itspreciseand mathematically respectablenotionofexistential quan-
tification.
In thefinalfootnoteto "FunctionundBegriff'(1891)Frege
wrote: "The ontologicalproofof God's existencesuffersfromthe
fallacyof treatingexistenceas a first-levelconcept."This is essen-
tially the same idea he advanced seven years earlier in his
Grundlagender Arithmetik, wherehe wrote:"Because existenceis
a propertyof conceptsthe ontologicalargumentforthe existence
ofGod breaksdown" (section53). Russellwas even moreemphatic
in his lectureson logical atomism:
When you take any propositionalfunction and assertof it ...
thatit is sometimestrue,that gives you the fundamental
meaningof 'existence'.You may express it by sayingthat
thereis at least one value of x forwhich the propositional

Existence/ 61
functionis true.. . . Existenceis essentiallya propertyof a

propositionalfunction.It means thatthe propositional
functionis truein at least one instance.("The Philosophyof
Logical Atomism,"in Russell'sLogic and Knowledge,ed., R.
C. Marsh,at p. 232.)
. . . As regardsthe actual thingsthereare in the world,
there is nothingat all you can say about themthatin any
way correspondsto thisnotionof existence. It is a sheer
mistaketo say thatthereis anythinganalogous to existence
thatyou can say about them.... There is no sortof point
in a predicatewhich could not conceivablybe false. I mean,
it is perfectlyclear that,iftherewere such a thingas this
existence of individualsthatwe talk of,it would be
absolutelyimpossibleforit not to apply, and thatis the
characteristicof a mistake.(ibid.,241.)
... there is a vast amountof philosophythatrestsupon the
notionthatexistence is, so to speak, a propertythatyou can
attributeto things,and thatthe thingsthatexist have the
propertyof existenceand the thingsthatdo not exist do
not. That is rubbish... (ibid.,p. 252.) See also Russell'sA
CriticalExpositionof the Philosophyof Leibniz, London:
George Allen & UnwinLtd, 1971, at pp. 174-175;and his
Introductionto MathematicalPhilosophy,London: George
Allen and UnwinLtd, 1953, at pp. 203-204.)
The problemwiththeOntologicalArgument-accordingto Kant,
Frege, and Russell-is thatby invokingthe alleged concept of ex-
idivinityin talking about an exidivinepossible individual,it il-
legitimately treatsexistenceas an admissibleconceptor property
of individuals,on a par withsuch mundaneconceptsor properties
as beinggreen-eyedor beingan island,therebyviolatingitsproper
status(as a second-levelpropertyor conceptof first-levelconcepts
or ofpropositional or as a pre-requisite
functions, forhavingany pro-
pertiesat all, or as somethingof the sort).Alas, the foundersof
mathematicallogic wouldapparentlycast out Descartes'slovelylit-
tle cogitoergo sum alongsidehis OntologicalArgument.The quest
foran answerto thequestion'Whatpossibleindividuals exist?'begins
to look more and more quixotic.

62 / NathanSalmon
III
Schopenhauergave expressionto a verycommonreactionto the

OntologicalArgumentwhen he called it "a charmingjoke."'2 The
argument'spropounders,however,do not offerthe argumentas a
curiousphilosophicalparlortrickor riddle;they advance it in all
seriousnessas a deductiveproofofa thesisthatmostofus had been
trainedto believe since childhood,withverylittlein theway of ra-
The thesis it purportsto prove is extremely
tional justification.
implausible-atleast for(i.e., withrespectto theepistemicsituation
of)thosewho are able to breakfreeoftheirchildhoodreligioustrain-
ingand forthosewho neverhad any-and forthatreason alone the
thesisneeds somethinglike evidence or argumentforitsepistemic
Ifthereis any area in whichphilosophersare to be held
justification.
to a higherstandardthannonphilosophers, itis in providing
justifica-
tionfortheirotherwiseimplausiblereligiousbeliefs.'3Whereasthe
OntologicalArgument(taken polemicallyas a purportedproofof
God's existence) has always struckme as philosophyat its least
dignified,I have never seen any meritwhatsoeverto the Kantian
sort of reply recounted in the preceding section. Furthermore
Descartes'scogito has always struckme as an excellentexample of
philosophyat itsshiningbest.Letus distinguish threeseparateKan-
tian theses about existence:
(i) The Englishverb 'exist' (and its cognates)represents,
fromthe pointof view of logic, not a first-order
predicateof English,but a logical quantifier;
(ii) There is no propertyor conceptof existencefor
individuals;
and
(iii) It is illegitimateto invoke the term'exist' or the alleged
propertyor concept of existencein formingthe concept
of somethingor in specifyingone of the necessarycon-
ditionsin the definitionof something-so thatone can-
not legitimatelydefinesomethingas the existentsuch-
and-such,or as a such-and-suchthatexists.14
Insofaras Kant,Frege,or Russell,or theirfollowers, have heldany
or all oftheses(i),(ii),and (iii),it is virtuallyprovablethattheyare
completelymistaken.

Existence/ 63
It is widelyrecognizedthatthesis() is false.Anynumberofcom-
mentatorshave noted thatthe term'exists'is fullyand completely
definablein formallogicas a first-orderpredicateofindividuals,us-
ingstandard,actualist,Frege-Russellianexistentialquantification.
Its
definition(whichalso employsthe logical notionsof identityand
abstractionbut nothingmore) is the following:
(Ax)(3y)[x= y].
the Englishword 'exists'may be regardedas being
Less formally,
definedby thephrase'is identicalwithsomething',or moresimply,
'is something'.This yieldsan aternativeway to give substanceto
theidea thatto be is to be one ofeverything: To be one ofeverything
is to be something.The phrase'is something', inthesense of'is iden-
ticalwithsomething',is paradigmaticofthesortofexpressionthat,
fromthe pointof view of logic,would ordinarilybe regardedas a
first-orderpredicateofindividuals. (Ofcourse,itwouldnotbe regard-
ed as a simple first-order predicate;it is a compoundexpression.)
It satisfiesevery reasonable logical,grammatical,or semantictest
or criterionforfirst-order predicatehood.In any case, the expres-
siondisplayedabove is unquestionably a logicalfirst-order
predicate.
The factthatit (correctly) applies to everyexistingindividualwhat-
soever, and does so by the rules of semanticsalone, does nothing
to threatenits statusas a full-fledged predicateof individuals.On
thecontrary, the factthattheprinciplesofclassicalsemanticsassign
a class ofindividualsas an extensionto thisexpressionconfirms that
itis indeeda first-order predicate,and one ofpurelogicat that.Fur-
thermore,the factthatitsextensionin any modelis justthe domain
ofindividualsin thatmodel confirms thatitis the verypredicatewe
want. Ifany individualin the domainof any modelwere leftout of
the predicate'sextension,then whateverpropertyor concept the
predicatewouldbe an expressionfor,itwould notbe an expression
forthe existenceof individuals.
Althoughithas been less oftennoted,itshouldbe equallyobvious
thatthereis a concept of existenceforindividuals,and thatthere
is a special property-thepropertyof existing-thatan individual
has onlyby virtueofthe factthatit exists.Each of the notionsin-
volved in the definitionof the predicate 'exists' is precise and
mathematically respectable;each ofthe expressionsmakingup the
definienshas a definitesense or content.In fact,each of the three
notions involved-existential quantification, identity, and

64 / NathanSalmon
preciseina waythatmanyeveryday
abstraction-is arenot.
notions
is fully
quantification
Existential intermsofthelogicalno-
definable
tionsofnot and everything,as follows:
(AP)[ - (vx) -Fx].
theoccurrence
(Moreaccurately, oftheexistential inany
quantifier
existential
generalization
r(3a),palmay be by
defined
contextually
r - (vJ - q)a- .) Identityis justthe binaryequivalencerelationthat
eachindividual andtonootherindividual.
standsintoitself Abstrac-
tionis justtheformaloperation bywhicha compoundfirst-order
predicateis formedfroman open sentenceof formallogic.The
English expressions'something'and 'is identicalwith' are
paradigmatic ofthesortofexpression thatisordinarilyregarded as
expressing an attribute
(propertyorrelation)orconceptas itssense
orcontent. Ifanyexpressions expressconcepts as their
orattributes
sense or content,thesedo. Theirsensesor contentsare easily
specified.Thesenseorcontent ofthesecond-order predicate(quan-
'something'
tifier) is theproperty ofclassesofindividualsofnotbe-
ingempty, theproperty of havingat leastone element.Moreac-
curately, thesenseorcontent of'something',withrespecttoa given
timet, is thetemporally indexedproperty or conceptofnotbeing
empty att.Thesenseorcontent ofthephrase'isidenticalwith',with
respectto a giventimet, is thetemporally indexedbinaryrelation
ofbeingone and theverysamethingat t, or thecorresponding
concept. 15
Ifa setofexpressions thatexpressconceptsorattributes as their
senseorcontent areappropriately combined toform a newexpres-
sion,thecompound expressionthusformed hasa senseorcontent
thatis determined ina certinwaybythesensesor contents ofthe
combined component Hencethephrase'isidentical
expressions. with
something', andthedisplayed expression,expressa definiteproper-
tyorconceptas their(shared)senseorcontent. Thisis theproperty
or conceptofbeingidenticalwithsomething(ormoresimply, the
property or conceptofbeingsomething).Itis thisproperty or con-
ceptthatis thesenseor contentofthepredicate'exists'.Anditis
thisproperty or conceptthatwe call'existence'.We havehereour
answertothequestionofwhatitisforsomething tobe, ortoexist.
To be is to be identicalwithsomething.
I do notmean,ofcourse,thatthepredicate'exists'expressesthe
property withsomeparticularthing,
or conceptofbeingidentical

Existence/ 65
suchas Socratesor Russell.Suchproperties as theseare nowadays

called'haecceities' or 'thisnesses'(Robert Adams),and are express-
ed bysuchphrasesas 'isidentical withSocrates' and'isidentical with
Russell'.The property or conceptof existenceexpressedby the
predicate 'exists'involvesexistential quantification. Itistheproper-
tyorconceptofbeingidentical withsomethingor other,thefeature
thatan individual hasonlyinvirtueofthefactthatnoteverything
is distinctfromit.Moreaccurately, thesenseorcontent oftheterm
'exists',withrespecttoa giventimet, is theproperty ofconceptof
beingsomething at t,theproperty thatan individual hasonlyinvir-
tue ofthatfactthat,at t, noteverything fromit.
is distinct
Itstandstoreasonthatthefirst-order conceptofexistenceforin-
dividuals shouldinvolvetheFrege-Russellian higher-level
logicalno-
tionofsomethingor other.To be is to be identical withsomething.
Notto be is to be distinct fromeverything. Moresuccinctly, to be
is to be something, notto be is to be nothing. To be andnotto be:
theseare theanswers.'6
As faras I cansee,thereisnothing at alltobe saidforeither thesis
(i) or thesis(ii). In anycase, despitetheirimpressive credentials,
neither KantnorFregenorRussellhasanypersuasive argument to
offer foreitherofthesetheses.
Thesis(iii)is no betteroff.Thereis nota singleplausiblereason
whythepredicate'exists',or theproperty orconceptofexistence,
shouldbe precludedfromthedefinition ofsomething or fromthe
formation ofsomeinclusive concept, suchas theconceptofan exis-
tentfantasy islandorthatofan existent lion.Whyshouldany con-
ceptbe precluded fromtheformation ofmorecomplexones?The
conceptofan existent fantasyislandistheconcept ofa fantasy island
thatis notdistinct fromeverything, and theconceptofan existent
lionisthatofa lionthatis notdistinct from everything. Theconcept
ofan existent lioniseverybitas legitimate, qua concept, as thecon-
ceptofa green-eyed lion.Similarly, we maydefinetheterm'ex-
iunicorn' as follows:
thatexists.
exiunicorn=dd& unicorn
Letus calltheprocedure offorming suchconceptsor definitions
as these'existential
definition'.
Whatcanpossiblybe wrongwithex-
istential Ifthereis anything
definition? inourdefinition
illegitimate
of an exiunicorn, it comes from'unicorn',not from'exists'.'7
Philosophers oftenformor invokecomplexconceptsthatinclude

66 / NathanSalmon
existenceas a necessary condition. We havetheconcept,forexam-

ple,ofa temporary existent,i.e.,an individual thatexistsbutdoes
notalwaysexist.We alsohavetheconceptofa contingent existent,
i.e.,an individual thatexistsbutdoesnothavenecessary existence.
Theseconcepts areperfectlylegitimate,andindeed,extremely useful
forcertainpurposes.
SaulKripke's powerful 'schmidentity' form ofargument canbe ap-
pliedhere.'8Supposemyclaimsthatexistenceis theproperty or
conceptofbeingidentical withsomething andthattheEnglish word
'exists'is a first-order
predicate forthisconceptare mistaken. Then
takeinsteadtheexpression '(Ax)(3y)[x = y]'.As I havealreadysaid,
thisisunquestionably a logicalfirst-orderpredicate.Bytheprinciples
Qfsemantics alone,thispredicate (correctly)applies,withrespectto
anytimet and possiblew, to everything thatexistsat t in w, and
tonothing else.Following Kripke, we mayabbreviate thispredicate
bytheword'schmexists', andwe maycalltheproperty orconcept
thatis thesenseorcontent ofthispredicate 'schmexistence'.There
isabsolutely noreasonintheworldwhywe cannotusethispredicate
indefining newexpressions, orwhywe cannotinvoketheconcept
ofschmexistence informing morecomplexconcepts, as oftenas we
like.Thefollowing expression, forexample,is perfectly well-formed
and meaningful:
(Az)[(Ax)(3y)[x= y](z) & = y](z)].
LI1(Ax)(3y)[x
Thisis a predicatefortheconceptofcontingent existence. Exactly
analogously, we maysay thatsomething is schmexidivine ifit is
divineand identical withsomething. We are perfectly freeto use
eitherofthesedefined notionsinourreasoning. Whoisto stopus?
IfthebestthatKantandthefounders ofmathematical logiccan
do toblocktheOntological Argument istoprohibit existential
defini-
tion,theirresponseto theargument constitutes
nothing morethan
an especiallyarrogant formofreligious persecution. Let theKan-
tiansscream'Bluemurder!' as oftentheyplease,existential defini-
tionsare perfectlylegitimate.
It maybe plausibly arguedthatthereis no pointin performing
an existential Itis true,forexample,thattheconceptof
definition.
an existentlionisinsomesensenotverydifferent from theconcept
ofa lion,andas Kantpointedout,a hundred existent dollarsis not
worthone pennymorethana hundred dollars.(I finditimpossible
to agreewithhim,however, thata hundred existentdollarsis not

Exstence/ 67
worthone pennymore than a hundredmerelypossibledollars.If

all mydollar billswere merelypossible,I would gladlytradethem
forjust one existentdollar bill.) How does this show anythingil-
legitimateabout the conceptof an existentlionor about thatof an
existent dollar? At most, it only shows that such concepts are
superfluous, thattheylacka raisond'etre,notthattheyare somehow
illegitimately formed.I doubtthatiteven showsthismuch:Consider
any class thathas me as an element-{Nathan Salmon}, forexam-
ple (theunitclass thathas me as itsonlyelement).WhenI am dead
and gone, thisclass will no longerexist. It will not be an existent
class. It is farfromclear, however,thatit willnot be a class of any
kind.I believe thatit will stillbe a class afterI am gone, and that
I will stillbe an elementof it (although,of course, since I will no
longerexist,therewillnot be anythingthatis an elementofit and
it will not have any elements).Irrespectiveof one's philosophical
dispositiontowardthe (admittedlysomewhatbizarre)questionof
whetheran existentclass can become an nonexistentone, thevery
fact that we can raise a substantive(albeit bizarre) question of
whethera given such-and-such remainsa such-and-such in certain
circumstances in whichitdoes notexistindicatesthatthereis perfect
legitimacyto theconceptofan existentsuch-and-such, qua concept.
My view that,at some timein the future,singletonme will stillbe
a class butno longeran existentclass,whethercorrector incorrect,
involvesjustsucha concept,as does theview ofanyonewho denies
thatsingletonme will ever be a nonexistentclass.
Furthermore, iftheOntologicalArguerswere correct,therewould
be yet another,and no less significant, purposethatmay be served
by forming complexconceptsthatincludetheconceptofexistence.
The charge thatexistentialdefinition is pointless,in a sense, begs
the question againstAnselm,Descartes,et. al.
IV
Version3 oftheOntologicalArgument is unscathedby Kant'salleg-

ed refutation. Moreover,althoughitssecond premiseis notbeyond
all possible doubt,it is beyond all reasonable doubt.
Well,itssecond premiseis beyondall reasonabledoubtprovided
we can makesense ofthepossibilist phrase'anypossibleindividual'.
The use ofpossibilistquantification offendsthe sensibilities
ofsome

68 / NathanSalmon
actualists,and mayobfuscatetheevaluationoftheargumentas valid

or invalid.Itwouldbe desirableto eliminatesomehowthepossibilist
quantification of Version3.
That is somethingwe can do (at least to theextentdemandedfor
presentpurposes),throughthejudicioususe ofmodal operatorsand
standardactualistquantification over individuals(and by assuming
Russell'sTheory of Descriptionsin order to explicatethe descrip-
tion'theexidivinepossibleindividual'by meansofquantification and
identity).'9 fromthe argument,
In removingpossibilistquantification
one mustbe sensitiveto possiblemisinterpretations ofthepremises
and conclusion.Recall theimportofthe premisesofVersion2. The
same is trueof Version3, replacing'divine'by 'exidivine':conclu-
sion (3c) is supposedto be notmerelythatthe exidivinepossiblein-
dividualmighthave been an existentdivineindividual,but thatit
actuallydoes existand actuallyis divine.Hence premise(3a) must
be read in such a way thatitassertsthatthepossibleindividualthat
actuallyis exidivineactuallyis exidivine,and premise(2b) mustbe
read in such a way thatit assertsthatany possible individualthat
actuallyis exidivineactuallyexists.An actualistrenderingof Ver-
sion 3, then,is the following:
(4a) There mighthave been an individualx such thatactual-
ly:x, and necessarilyonlyx, is exidivine.
(4b) Necessarily,every individualx is such that,actually,ifx
is exidivinethenx exists.
Therefore,
(4c) There mighthave been an individualx such thatactual-
ly: (i) x, and necessarilyonlyx, is exidivine;and (ii) x
exists.
Let us call this'Version4'. It is what Version1 becomes when it
is submitted to regimentation in accordancewithcontemporary stan-
dards of rigor,withan aim to satisfying certainreasonable formal
desiderata.Version4 is a valid modal argument,one thatinvolves
actualistquantification. And itssecond premiseis now beyond all
reasonable doubt.
Yet Gaunilodemonstratedthatthe argumentmustinvolvesome
error.Ifexistentialdefinition is notthe sourceofthe error,whatis?
Here again, the correctanswer has been noted by a numberof
commentators. As Kanthimself pointedout (and othersbeforehim),

Existence/ 69
theonlyconceptualtruths thatfollowdirectly fromconceptsor defini-

tionsare hypothetical conclusionsof the form"Ifanythingsatisfies
the conceptor definition, thenithas such-and-such properties"and
"Anythinghavingsuch-and-such propertiessatisfiesthe conceptor
definition." Itis a conceptuala prioritruth,forexample,thatifthere
is a green-eyedlion,thenita lion and has greeneyes, and thatany
lion thathas green eyes is a green-eyedlion. Exactlyanalogously,
it is a conceptuala prioritruththatifthereis an existentlion,then
itis a lionand exists,and thatanylionthatexistsis an existentlion.
It is trueby definition, ifyou will,thatall and onlyexilionsare lions
thatexist.20Analogouslyagain, it is a conceptuala prioritruththat
all exiunicomsare unicornsand exist,and onlyunicornsthatexist
are exiunicorns.
The importantpointis thatin thisrespectnothingchangeswhen
we move fromactualistto possibilistquantifierlogic. Even in
possibilistquantifier logic,the conclusionsthatfollowdirectlyfrom
definitions are alwayshypothetical in form.Itis a conceptuala priori
truth,forexample,thatifa possibleindividualis a green-eyedlion,
thenit is a possibleindividualthatis a lionand has greeneyes, and
any possibleindividualthatis a lion and has greeneyes is a possible
individualthatis a green-eyed lion.Itis similarly
a conceptuala priori
truththatifthereis an exidivinepossibleindividualthenitis a divine
possibleindividualand exists,and ifthereis exactlyone divinepossi-
ble individualthatexiststhenit is theexidivinepossibleindividual.
Itis trueby definition thata possibleindividualis theexidivinepossi-
ble individualifand onlyifit,and (amongpossibleindividuals)only
it,is bothdivineand existent.Itcertainlyis nota conceptuala priori
truth,or trueby definition, thatsome possibleindividualis an exis-
tentlion,thatsome possibleindividualis an exiunicorn, orthatsome
possible individualis exidivine.For all thatcan be knownmerely
by reflection on theconceptofa such-and-such (or on thatofan ex-
istentsuch-and-such), theremay not be anythingthatfitsthe con-
cept,noteven a possiblething.Even ifitcan be knowna priorithat
thereis a possiblethingthatmightfitthe concept,theremay not
be any possiblethingthatactuallyfitstheconcept.Thereis noteven
a kernelof truthto the idea thatifexistencewere treatedas an ad-
missibledefiningpropertyor concept,thenit would be possible to
createentitiesby definingthemintoexistence.The mostwe obtain
directlyfromthe existentialdefinition of an existentsuch-and-such
is that,if a possibleindividualis an existentsuch-and-such, thenit

70 / NathanSalmon
exists and is a such-and-such.

The problemwiththe OntologicalArgument(as it has been for-
mulatedhere) is not thatit involvesexistentialdefinition, but that
itsexpounderscommitMeinong'sfallacyofassumingthat,forany
formula(p,it is logicallyor triviallyor manifestly truethatthe q) is
a (P. (To put the matterless anachronistically, Meinongcommits
Anselm'sfallacy.)Far frombeinga logicalschema,thisassumption
is actuallycontradictory. In fact,theassumption is contradictory even
ifthe rangeof the formula-variable (P is restricted to consistent for-
mulas thatmay apply to possibleindividuals.To see this,let (4, be
'F(x) & p'. Now let (4, be 'G(x) & -p'. Logic gives us only thatif
thereis exactlyone (, thenthe (P is a (4.21 Even in thestrangeland
of possibilist quantification, logic-through its Law of
Noncontradiction-rejects the claimthat,forevery(consistent)for-
mula (P,the possibleindividualthatis a (P is a (P. Possibilistquan-
tifierlogic gives us onlythat,ifexactlyone possibleindividualis a
(P,thenthe possibleindividualthatis a (P is a (P. Premise(2a) is no
truthoflogic.Itis no piece oftriviaeither.Likewiseforpremise(3a).
This muchis obviousfromitstranslation intoactualistdiscoursevia
(4a).
The considerations raisedin theprecedingtwoparagraphsare both
necessary and sufficientto expose the fallacy in ontological
argumentsin the styleof Versions1through4, whetherforthe ex-
istenceof lions,fantasyislands,unicorns,or divineindividuals.No
doubt more is required to debunk more sophisticatedversions,
althoughI believe notmuchmore.Certainly, one need nottake the
drasticmeasure of retreatingto theses about existence that are
demonstrablyfalse (or nearlyso).
In fact,Kant,Frege,and Russellall recognizedexplicitlyin their
writings truthsare alwayshypothetical
thatdefinitional in form.Why,
then,did each thinkitnecessaryto insistalso on one or all oftheses
(i)-(iii)?I do not know. It is possible thattheyreasoned along the
following lines:Conceptualor definitional truthsconcerning thesuch-
and-suchalways have thehypothetical form'Ifthesuch-and-such ex-
ists,thenit is thusand so'. Ifit were legitimateto includethe very
conceptofexistenceitselfin thedefinition ofthesuch-and-such, then
we couldsatisfy theantecedentofthishypothetical bytheverydefini-
tion,therebysecuringtheconsequentcategorically, withouttheex-
istentialproviso,via modusponens.And theconsequentin thiscase
includesexistenceas one oftheconditions itascribes.We wouldhave

Existence/ 71
obtainedan analyticexistential;we would have definedsomething

intoexistence.Butwe cannotcreatenew entitiessimplybe defining
themintoexistence.Itis illegitimate, therefore, to includeexistence
itself(or anythingthat entails existence,such as the concept of
necessaryexistenceor thatofexidivinity) in forming a complexcon-
ceptor in defining a term(thesis(iii)).Iftherewerea conceptor prop-
ertyofexistenceforindividuals, or iftherewerea first-order predicate
of individualsthat correctlyapplied to all and only existingin-
dividuals,thenit would be perfectly legitimateto includethiscon-
cept or propertyin formingmore complexconcepts,and it would
be perfectly legitimateto use thispredicatein definingotherterms.
Consequently,thereis no conceptor propertyof existenceforin-
dividuals(thesis(ii)),and theword'exists'is nota first-order predicate
(thesisi)). In fact,it is clear on independentgroundsthatexistence
is a second-levelconceptratherthana first-level concept.This fur-
therconfirmsthesis(iii).
The reasoninghere is fallacious.The mistakeoccurs when it is
argued thatby buildingexistenceintothe conceptof the such-and-
such, one would make it trueby definition thatthe such-and-such
exists.This is just Meinong'sfallacyagain. Whereas the Version3
Arguercommittedthisfallacyis assertingpremise(3a) as a logical
or manifesttruth,Kantand his followersmay have committedthe
verysame fallacyinattempting to locatethesourceoftheillegitimcy.
One may defendthe Kantianrefutation of the OntologicalArgu-
mentby claimingthatwhatKant and his followershave in mindis
thatitis illegitimate to includeexistenceinthedefinition ofsomething
in such a way thatit followsjust fromthe definition thatthe thing
categoricallyexists.Ifexistenceis to be legitimately includedin the
very definition of the such-and-such, it mustbe includedin such a
way thatonly hypotheticalconclusionsfollowfromthe definition.
Ifthisis what theyhave in mind,thenat least theydo not commit
theverysame fallacyas theiropponents.Theyare, however,guilty
ofsomething,even ifonly grossunderstatement. It is not thatit is
merelysomehow illegitimate to includeexistencein the definition
ofsomethingin such a way thatitsexistencefollowsfromthe very
definition. Thoughitis perfectly possibleto includeexistencein the
definition ofsomething, itis quiteliterally impossibleto do so in such
a way thatexistencefollowsfromthe definition. If 'oughtnot' im-
plies 'can', then it is false thatone oughtnot to includeexistence
in the definitionof somethingin thisway.

72 / NathanSalmon
The factthatthe firstpremiseofthe OntologicalArgumentis not

a truthoflogicand notmanifestly truedoes notentailthatit is not
trueat all. Butdo atheistsand agnosticshave anyreasonto suppose
ittrue?Indeed,does theOntologicalArguerhave any reasonto sup-
pose ittrue?Whyshouldanyonebelieve thatthereis a divinepossi-
ble individual?
One mightinferthatthereis a divinepossibleindividualfromthe
observationthatit is perfectly possibleforthereto be a divinein-
dividual.Thisis essentially Anselm'sargumentforhispremise(2a).22
It is fallacious,but it is understandablewhyso manywritershave
been convincedbyit.In orderto steerclear ofthefallacy,one must
distinguish sharplybetweentheassertionthatitis possibleforthere
to be something thatis such-and-such,and theseparateassertionthat
it is possiblethatthereis somethingthatactuallyis such-and-such.
In possibilistdiscourse,we mustdistinguish theassertionthatthere
is a possibleindividualthatmighthave existedhavinga certainpro-
pertyfromthestrongerassertionthatthereis a possibleindividual
thatactuallyhas the property.Thattheseare different assertionsis
confirmedby consideringthe (yetto be analyzed) propertyof ex-
istingwithoutactuallyexisting:There mighthave been something
thatdoes notactuallyexist(thereis a possibleindividualthatmight
have had theproperty ofexistingwithoutactuallyexisting),
butthere
could notbe somethingthatactuallyexistsbutdoes notactuallyex-
ist (thereis no possibleindividualthatactuallyhas the propertyof
existingwithoutactuallyexisting).Let us suppose forthe moment
thatit is somehow manifestor knowablea priorithatit is possible
thatthereis divineindividual.In possibilist logic,itfollows
quantifier
thatthereis a possibleindividualthatmighthave been bothdivine
and existent.Itdoes notfollowthatthereis a possibleindividualthat
actuallyis bothdivineand existent.The factthattheremighthave
been an individualthatwould have been divinedoes notentailthat
theremighthave been an individualthatactuallyis divine.The lat-
ter is what is needed to legitimizepremise(2a).
I am notmakingthecommonobjectionthattheOntologicalArguer
begs thequestionsince one mustestablishor assumethatthedivine
possibleindividualexistsbeforeit can be concludedthatthedivine
possible individualis divine.23That may be true,but as faras the
presentobjectiongoes, it need not be. It is open to theOntological
Arguer to attemptsome neutral,non-question-begging way of
establishing thatthereis a possibleindividualthatis actuallydivine,

Existence/ 73
withoutassumingthatone actuallyexists.What I am pointingout

is that,even ifthe atheistand agnostichave been persuadedon a
priorigroundsthattheremighthave been a divineindividual,the
OntologicalArguerstillowes themsome furtherargumentto con-
vince themthattheremighthave been an individualthatactually
is divine(beforeit can be concludedvia premise(2b) thatthereac-
tuallyexistsa divineindividual).Merelyestablishingthatthereis a
possible individualthat mighthave been divinedoesn'tcut it.
For the purposesof the OntologicalArguer,the additionalargu-
mentmustproceed frompremiseswhose truthis a priori,or other-
wise manifest.The additionalargumentmustnotdependin anyway
on the assumptionthata divinepossibleindividualactuallyexists,
since thisis whatis supposedto be ultimately
proved.Furthermore,
theadditionalargumentmustbe sound;itmustbe an argumentthat
cannotbe extendedto fantasyislandsand the like. Surelyno such
argumentexists.Theredoes noteven possiblyexistanypossiblesuch
argument.
In a penetratingcritiqueoftheOntologicalArgument, David Lewis

suggestedone reason on behalfof the OntologicalArguerforsup-
posingthatsome possibleindividualnot only mighthave been but
actuallyis divine.24We were supposing,forthe sake of argument
(althoughit has not yetbeen established),thatit is manifestthatin
some possibleworldsthereexistssomethingthatis divine.Hence,
in some possibleworldstheproperty ofdivinityis exemplified.Now,
the actual worldis generallythoughtto be a special possibleworld
in that,unlikeanyotherpossibleworld,italone is actual.Ifthespecial
propertyof divinityis exemplifiedin any possibleworld,it seems
only fittingthatit shouldbe exemplifiedin the mostspecial of the
worlds,the actual world.Lewis writes:
This reason seems prima facie to have some force:whatever
actualitymay be, it is somethingwe deem tremendouslyim-
portant,and thereis only one worldthat has it . ..
Thereforeit may well seem plausiblethatthe actual world,
being special by its unique actuality,mightalso be special
by being a [worldin whichdivinityis exemplified].This
does not pretendto be a proofof [premise(2a)], but [I] do

74 / NathanSalmon
notdemandproof; I] wishto knowiftheontologicalarguerhas

anyreasonat all to accept[(2a)],even a reasonthatdoes no more
than appeal to his sense of fitness.(p. 184)
Lewis's suggestionis a Trojan'sgift.For he goes on to argue that
actualityis nota specialproperty at all. Accordingto Lewis,theword
'actual' is,in itsprimarysense,an indexicaltermanalogousto 'here'
or 'now':itsreferencevarieswiththecontextin whichitis uttered-
treatingpossibleworldsalong withtimesand locationsas relevant
featuresof contextsof utterances.
The fixed meaningwe give to 'actual' is such that,at any
world w, ..., in ourlanguage ...'the actual world' denotesor
names w; the predicate'is actual' designatesor is trueof w
and whateverexistsin w; the operator'actually'is trueof
propositionstrueat w, and so on forcognate termsof other
categories ....
A complication:we can distinguish primaryand secondary
senses of 'actual' by askingwhat world 'actual' refersto at a
world w in a contextin whichsome otherworldv is under
consideration.In the primarysense, it stillrefersto w, as in
'If Max ate less, he would be thinnerthan he actuallyis'. In
the secondarysense it shiftsits referenceto the world v
under consideration,as in 'If Max ate less, he would actually
enjoy himselfmore'. (p. 185)
Lewis extractsfromhistheorythat'actual'is indexicalthe conse-
quence thatactualityis not a special propertyof the actual world,
or equallyinsignifi-
and thatall possibleworldsare equallysignificant,
an
cant, from enlarged and objective modal pointof view:
If I am right,the ontologicalarguerwho says that[theac-
tual] worldis special because [it]alone is the actual world is
as foolishas a man who boasts thathe has the special for-
tune to be alive at a unique momentin history:the present.
The actual worldis not special in itself,but only in the
special relationit bears to the ontologicalarguer.Other
worldsbear the same relationto [their]ontologicalarguers
We should conclude,therefore,that [Version2 of the On-
tologicalArgument]is a valid argumentfroma premise[(2a)]
we have no non-circularreason to accept ..... [Premise(2a)]
derives its credibilityentirelyfromthe illusionthatbecause

Existence/ 75
[theactual] world alone is actual, therefore[theactual] world

is radicallydifferent fromall otherworlds-specialin a way that
makes it a fitting place of [divinity].
But once we recognizethe
indexical nature of actuality,the illusionis broken and the
credibilityof[(2a)]evaporates.Itis trueofany world,at thatworld
but not elsewhere, that that world alone is actual. (pp.
187-188)25
We have arrivedat lastat the questionofwhatit is fora possible
individualto be actual. IfI am rightthatto be is to be identicalwith
something, thento be actualis actuallyto be identicalwithsomething.
Lewis has providedus withan answerto the questionof whatit is
forsomethingactuallyto be the case.
Unfortunately, thereis muchin Lewis's analysisof actualitythat
commentators have takenexceptionto. Yet thereis muchin theac-
countthathas the ringoftruth.It is important to sortthesematters
out ifwe are to be clear about what it is fora possibleworldor a
possible individualto be "actual," properlyso-called.
One immediatedifficulty withLewis's theoryof actualityis that
hisstatement ofthetheorypresupposeshishighlycontroversial view
thatthe (standard)inhabitantsof possibleworldsare world-bound
individuals,i.e., thateach possibleindividualexistsin one and only
one possible world. This view, in turn,is connectedwithLewis's
idiosyncraticview that possible worlds are physical systems.
Nowadays,philosophersmorecommonlyregardpossibleworldsas
abstractentitiesof a certainsort,such as maximal consistentsets
of propositionsthatmighthave been jointlytrue (RobertAdams),
maximalsituationsthatmighthave obtained(Saul Kripke),maximal
historiesthecosmosmighthave had (Kripke),totalstatesthecosmos
mighthave been in (Kripke,RobertStalnaker),maximalstatesofaf-
fairsthatmighthave obtained(AlvinPlantinga), or maximalscenarios
thatmighthave been realized(myself). As withmostofLewis'scom-
mentators,I regardLewis's presuppositions concerningthe nature
ofmodalityas inessentialto themainidea ofhistheorythattheterm
'actual' is indexical,and I propose to considerinsteada versionof
the indexical theorythat makes the considerablymore plausible
assumptionthatpossibleworldsare maximalabstractentitiesofone
sortor another.(IfLewis'spresuppositions concerningthenatureof
modalityare essentialin some sense to the whole of his indexical
theory,thenmy concernhere is witha properpartof thattheory,

76 / NathanSalmon
especiallywiththoseaspectsofthetheorythathave theringoftruth,
supplementedwithan abstract-entity conceptionofpossibleworlds.)
On the abstract-entity conceptionof possible worlds,possible in-
dividualsneed notbe world-bound-although, forall thatis demand-
ed by the conceptionitself,it may turnout thatan extremeversion
of the doctrineof essentialismis true,makingevery possible in-
dividualworld-bound insome sense.(Logically,itcould turnoutthat
the actual worldis the onlypossibleworld-extrememetaphysical
determinism-sothatthe onlypossible individualsare both actual
and world-bound.)
On theabstract-entityconceptionofpossibleworlds,ineverypossi-
ble worldtherewillexistalternativepossibleworlds(unlessextreme
metaphysicaldeterminism is true),but in any singlepossibleworld
w, everyworldotherthanw itselfis merelypossible.Ifworldsare
maximal compossiblesets of propositions,then accordingto any
singlepossibleworld,itis theonlyworldwhose elementsare all true,
and everyotherworldis a set of propositionsthatare not all true.
Ifworldsare maximalstatesthe cosmos mighthave been in, then
accordingto any singlepossibleworld,it is the onlyworldthatthe
cosmos is in, and everyotherworldis a maximalstatethe cosmos
is not in,and so on. We may abbreviatethisby sayingthatin any
singleworldw, one and onlyone possibleworldis realized,and that
is w itself.The exactmeaningof'realized'dependson whichabstract-
entityconceptionofpossibleworldsone adopts.Ifworldsare max-
imalpropositions, then'realized'simplymeans'a truemaximalpro-
position'.Ifworldsare maximalstatesofaffairs, then'realized'simply
means 'a maximalstateofaffairsthatobtains',and so on. Whatever
particularabstract-entity conceptionis decided upon, it will be an
analyticor conceptualtruththatone and only one possibleworld
is realized.One thingthatshouldemergefromany properaccount
ofindexicalityis thattheterm'realized'justintroducedis not index-
ical, even if the term'actual' is indexical.26
The notionsofa worldbeingactualand ofa propositionbeingac-
tual (i.e., of somethingactuallybeing the case) are interdefinable.
For the purposesof thisinvestigation, it will be convenientto take
thepropositional operator'actually'to be thefundamental termand
varioususes oftheterm'actual'tobe derivative.We maymarkthese
variouscognatesof 'actually'by way of superscripts indicatingthe
type of entityto whichthe termis applicable.A possibleworld is
said to be actuapw(or an actual world)if it is actuallyrealized. A

Existence/ 77
possibleindividual is saidto be actuaP(oran actualindividual) ifit

actuallyexists,i.e.,ifitexistsin theactualwworld.An individual
is said to be an actualP? ifit is actually? (i.e.,iftheproposition
thatitis ? is actuallythecase),andso on. (Exactlyanalogously, a
worldispossiblew ifitis possiblyrealized;an individualisnecessary
ifitnecessarily exists,etc.)Sincetheindexicaltheoryof'actually',
as propounded byLewis,admitsa secondary, nonindexical senseof
'actually'anditscognates,forcompleteprecision a subscriptof'1'
or '2' shouldbe addedto indicatetheprimary orsecondary sense,
e.g., 'actualil'.
Whereasa numberof objectionshave been raisedagainstthe
theorythat'actually'(initsprimary sense)and itscognatesare in-
dexical,everyobjection thatI am awareofis basedstraightfoward-
lyon one ormoreconfusions. Insomecases,theconfusion belongs
to someoftheadherents and defenders oftheindexicaltheoryas
wellas to thetheory's critics.
Perhapsthesimplest confusion is theideathat,ifonetreatspossi-
ble worldsalongwithtimesandlocationsas features ofcontexts of
utterance, thenanynonrigid definite description willemergeas an
indexical expression,sincethereferent (inactualEnglish) ofanysuch
description withrespecttoa worldinwhichitisuttered varieswith
the world.27 Thisconfusion betweenindexicality and nonrigidity
stems from a common misdescription, and a concomitant
misunderstanding, ofthesemantic term'indexical'. Anindexicalex-
pression is usuallydefinedas an expression whosereferent (denota-
tion)orextension withrespecttoa context varieswiththecontext,
so thatthereare possiblecontexts c and c' suchthatthereferent
or extension oftheexpression withrespecttoc is notthesameas
itsreferent withrespecttoc'. (Lewismisdefines indexicality inex-
actlythiswayinthepassagequotedinthepreceding section, thereby
helpingto foster theconfusion I am discussing, althoughhe is not
guiltyof thisconfusion.) The definition is too general;it failsto
discriminate betweengenuinely indexicalexpressions, suchas 'the
presentU.S. president', and certainnonindexical expressions, such
as 'theU.S.president'. A moreaccuratedefinition is this:Anexpres-
sionisindexicalifitsreferent orextension withrespecttoa context
ofutterance andwithrespect toothersemantic parameters ofevalua-
tion,suchas a timeand a possibleworld,varieswiththecontext
(holding theotherparameters fixed).28 Anexpression is nonindex-
ical ifitsextension withrespectto a contextand withrespectto a

78 / NathanSalmon
set of additionalsemanticparametersdoes not vary withthe con-

text.On theotherhand,a singulartermis (intensionally) rigid,with
respectto a given context,ifits referentwithrespectto the given
contextand withrespectto othersemanticparametersdoes notvary
with the additionalparameters.A singulartermis (intensionally)
nonrigidif its referentwithrespectto a contextand withrespect
to a set of parametersvaries withthe set of semanticparameters.
It is commonto distinguish betweenthe extension and theinten-
sionof an expression.The intension of an expression,withrespect
to a given contextof utterance,is the functionthatassignsto any
time and possible world (and perhaps some furthersemantic
parametersotherthana contextofutterance)theextensionthatthe
expressiontakes on with respectto the given context and with
respectto those parameters.An alternativedefinition of the term
'indexical',then,is the following:An indexical expressionis one
whoseintension withrespectto a givencontextvarieswiththecon-
text,so thattherewillbe possiblecontextsc and c' such thatthe
intensionof the expressionwithrespectto c is not the same as its
intensionwithrespectto c' .29 A singulartermis (intensionally)rigid,
withrespectto a givencontext,ifitsintensionwithrespectto that
contextis a constantfunction.
For example, the referentof 'the U.S. president',withrespectto
my presentcontextand with respectto the year 1978, is Jimmy
Carter,since he was presidentin 1978. Carterremainsthe referent
ifone changes the contextwhileretainingthe year 1978 as the se-
cond parameter,althoughthe referent changesas one changes the
second parameter.The referent of'thepresentU.S. president',with
respectto the same two parameters,is Ronald Reagan, since he is
presidentat thetimeofthecontext.The referent variesifone changes
the context,even ifthetimeparameteris heldconstantat theyear
1978. Withrespectto any actual context,the formerexpressionis
(temporally)nonrigid;yetitis also nonindexical,since itretainsthe
same intensionwithrespectto everycontextofutterance.The lat-
terexpression,beinga trueindexical,takeson different intensions
withrespectto different contexts,butwithrespectto anyactualcon-
text it is (temporally)rigid.
The referentof a singular term with respect to a context c
simpliciter (thatis,withrespecttoc butnotwithrespectto any other
parameters)maybe definedas thereferent ofthetermwithrespect
to c and withrespectto various featuresof c (such as the timeof

Existence/ 79
c) to act as theneededextraparameters.30 In thegeneralcase, if

onespeaksofthereferent ofa term(orthetruth-value ofa sentence,
etc.)withrespectto a certaindiminished or incomplete setoffur-
therparameters, itis understoodthatthediminished orincomplete
set is to be augmentedor completedby drawingadditional
parameters fromthegivencontext.Sincethereferent of'theU.S.
president' withrespecttoa context anda timevarieswiththetime,
so doesitsreferent withrespecttoa context simpliciter
(sincevary-
ingthecontextin thiscase involvesvarying thetimeparameter).
Theexpression is nonindexicalnevertheless.A singulartermwhose
referent withrespectto a contextsimpliciter varieswiththecon-
textis eitherindexicalor (intensionally) nonrigid.It mayeven be
both(e.g.,'thepoliticalleaderofthiscountry'), butit can be one
withouttheother.
The theorythat'actually'(initsprimary sense)and itscognates
are indexicalclaimsthatthereis a similardifference betweenthe
expressions 'theU.S.president in 1985'(nonindexical)and'theper-
sonwhoactually,is U.S.president in 1985'(indexical),
bytreating
possibleworldsalongwithtimesand locationsas features ofcon-
textsof utterance. As such,thetheoryinstructs us to index(i.e.,
relativize) theextensions ofexpressions bothtoa context, whichis
to includea possibleworldas one ofitsvariousfeatures, andto an
additional possibleworld,whichis tobe treatedas an independent
parameter ofsemanticevaluation.
Itmayseemthatoncepossibleworldsareincludedas features of
contexts, thereis no purposetobe servedby doublyindexing ex-
tensions to bothcontexts and possibleworlds,treating each as in-
dependent semantic parameters.Weshouldbe abletomakedo with
thepossibleworldsofthecontexts. We maysay,forexample,that
a sentenceoftheformrItis possiblethatS' is truewithrespect
toa context c ifandonlyifS itself istruewithrespecttosomecon-
textthatisjustlikec ineveryrespect otherthaninitspossibleworld
and whosepossibleworldis accessibleto thatofc. Thissinglyin-
dexedaccountseemsto yieldthecorrectresults untilwe consider
sentencesthatembedone modaloperatorwithinthe scope of
another.Considerthefollowing sentence:
(5) It is possiblethattheactualP,U.S. president
be a woman.
According to thesinglyrelativized
account,thissentenceis true
withrespectto a contextofutterancec ifandonlyifthereis some

80 / NathanSalmon
worldw' accessibleto theworldofc suchthattheU.S.president

in w' is a womanin w'. Butthisis thewrongtruth-condition
for
the sentence.In fact,it is correcttruth-condition
forthe wrong
sentence,to wit,thenonindexical sentence.
(6) It is possiblethattheU.S. president
be a woman,
or moreidiomatically,
The U.S. president
mighthavebeen a woman,
on oneofitsreadings (theRusselliansecondary occurrence orsmall
scopereading). Sentences (5) and(6) differ
intheirtruth-conditions;
ifbothsentences areuttered ina worldinwhichthepersonoccupy-
ingthepresidency isessentiallya man,sentence(5) is falsewhereas
sentence(6) is true.Sentence(5) is truewithrespectto a context
ofutterance c (roughly) ifand onlyifthereis someworldw' ac-
cessibleto theworldofthecontextc suchthattheU.S. president
in the worldof the contextof utterancec-rather thanin w' -is a
womanin w', ratherthanin theworldofc. The modaloperator
'itis possiblethat'directs us to evaluateitsoperandsentence'The
actualP, U.S. president is a woman'withrespectto worldsw' ac-
cessibleto thatofthecontextofutterance c. Thissentenceis true
withrespecttothesamecontext c anda worldw' accessibletothat
ofc ifand onlyifthedescription 'theactualP" U.S.president' refers
withrespectto c andw' to something to whichthepredicate 'is a
woman'applieswithrespecttoc andw'. Incomputing thereferent
ofthedescription withrespecttoc and w', theindexicaloperator
'actualPl'directsus to seekan objectto whichitsoperandphrase
'U.S.president' applieswithrespect totheveryworldofthecontext
ofutterance c itself,
forgetting abouttheworldw'. Thusinevaluating
sentence(5)withrespecttoa worldofutterance w (theworldofits
contextofutterance c),we are concerned simultaneously withthe
extension of'U.S.president' withrespecttow and theextension of
'is a woman'withrespectto someworldw' accessibleto w. The
truth-value ofthewholedependsentirely andsolelyonwhether the
uniqueobjecttowhichthephrase'U.S.president' applieswithrespect
to w is something towhichthepredicate'isa woman'applieswith
respect toanaccessible worldw'. Itisforthisreasonthata systematic
theory oftheextensions oftheexpressions ofa languagecontaining
indexicalmodaloperators requiresdoubleindexing, i.e.,ingeneral
thenotionoftheextension ofan expression (e.g.,thetruth-value

Existence/ 81
of a sentence)is relativizedto both a contextand a world,treated

as independentsemanticparameters.The notionof the extension
of an expression with respect to a contextc simpliciteris then
definableas theextensionoftheexpressionwithrespectto thecon-
textc and the world of c.
A commonobjectionto the indexicaltheoryof 'actually'is that
it requires a commitmentto utterancesor theirproducersbeing
world-bound (existinginonlyone world),and therebyto Lewis'sun-
popularmetaphysicalview thatindividualsare world-bound.3' The
reasoninggoes as follows:When we say of an expressionthatit is
indexical,whatwe are sayingis thatdifferent utterancesofthe ex-
pressionmaytakeon different semanticvalues (referent, truth-value,
intension,etc.),so thatitis nottheexpressiontypebutitsutterance
(inscription, token)thatis theproperobjectofthesesemanticvalues.
An utteranceoftheindexical'now'refersto thetimeoftheutterance,
an utteranceof 'here' to the place of the utterance,an utteranceof
'I' to the producerof the utterance,and so on. To say, then,that
'actual' (in itsprimarysense) is indexicalis to say thatan utterance
ofitdesignatesthepossibleworldinwhichtheutterancetakesplace,
or the possibleworldin whichtheproduceroftheutteranceexists,
or somethinglike that.Butwhereasit is perfectly legitimateto talk
about the timeor place of an utterance(in a given world),it is il-
legitimateto talkaboutthepossibleworldofan utteranceor itspro-
ducer,since one and the verysame utteranceis producedby one
and theverysame speaker in indefinitely manydifferent worlds.If
I utterthe sentence
(7) Actually,,a Republicanwill be elected U.S. presidentin
2100 A.D.,
I would have made the same utteranceregardlessof whichparty
controlsthepresidencya hundredand fourteen yearsfromnow.The
same utterancebyme occursindifferent possibleworlds,trueinsome
and falsein others.No worldmaybe singledout as theworldofmy
utterance-unless(contrary to whathas been said)utterances or their
producersare world-bound. Yet theutteranceis eithertrueor false,
and not both.
This piece of reasoninggoes wrongwhen it is arguedthatto say
thatan expressionis indexicalis to say thatitsutteranceis the pro-
per object ofsemanticvalues. The properobjectofsemanticvalues
is the expression(type)itself;but thesemanticvalues are had only

82 / NathanSalmon
relative toa context,andmayvaryaccordingly. To saythat'actual'

(initsprimary sense)is indexicalis notto saythatan utterance of
itdesignates theworldoftheutterance; rather, itis tosaythat,with
respectto anycontext, itdesignates theworldofthecontext. This
requires seeingthecontexts ofutterances, rather thantheutterances
themselves, as world-bound. The notionofcontextthatis relevant
hereis suchthat,foranyparticular actualutterance ofanyexpres-
sionbyanyone,ifanyfactshadbeendifferent in anyway,evenif
onlyfactsentirely independent ofand isolatedfromtheutterance
itself,thenthecontextoftheutterance wouldipsofactobe a dif-
ferentcontext-evenifthe utterance is made by theverysame
speakerintheverysamewaytotheverysameaudienceat thevery
sametimeintheverysameplace.To putitanotherway,whereas
a singleutteranceoccursin indefinitely manydifferent possible
worlds, anyparticular possiblecontext ofan utterance occursinone
andonlyonepossibleworld, so thatineverypossible worldinwhich
thesameutterance occurs, itoccursina newanddifferent context-
even ifthespeaker,hisor hermannerofuttering, thetimeofthe
utterance, thelocationofthespeaker,theaudiencebeingaddress-
ed,andall othersuchfeatures andaspectsoftheutterance remain
exactlythesame.A singleutterance occursin manydifferent con-
texts,eachofwhichoccursina different possibleworld.Thisiswhat
it meansto includea possibleworldas one of thefeatures of a
context.
Whereasutterances arenotworld-bound entities,itisnevertheless
perfectly reasonabletotreattheircontexts as world-bound entities.
Indeed,notdoingso wouldbe unreasonable. Suppose,forexample,
thatitwillcometo passthata Democratis electedpresident inthe
year2100,and considera worldW thatis exactlyliketheactual
worldineverydetailuptoJanuary 1,2099,butinwhicha Republican
is electedpresident in2100.SupposeI hereandnowuttersentence
(7).Intheactualworld, I thereby asserta proposition thatisnecessari-
lyfalse.In W,on theotherhand,I thereby asserta proposition that
is necessarily true.32I uttertheverysame sequenceof wordsof
English withtheverysameEnglish meanings inthetwoworlds, yet
I assertdifferentpropositions, oneproposition beingnecessarily false
and theotherbeingnecessarily true.Ifwe refuseto treatcontexts
as world-bound we are forcedto say,-quitemysteriously,-that I
uttertheverysamesentencewiththeverysame meaningin the
verysamecontext inthetwoworlds, yetassertdifferent things!The

Existence/ 83
informationcontent ofsentence (7)wouldemergeas a semanticfunc-

tionnotonlyofthemeaningofthesentenceandthecontextofut-
terance,butalso of theapparently irrelevantquestionof which
politicalpartywinstheU.S. presidency intheyear2100.Treating
contextsas world-bound, we may say insteadthatthe adverb
is indexical,
'actually1' and thatthesame utterance takesplace in
differentcontexts, resultingin different
propositionsasserted.We
thereby assimilate thisphenomenon tothesortofcontext-sensitivity
thatisfamiliarincasesofsuchsentences as 'A Republican
ispresently
U.S. president'.
The centralthesisof the indexicaltheoryof 'actually'and its
cognatesmaythusbe statedbysayingthattheextensional seman-
ticsgoverning 'actually'initsprimary senseis givenbythefollow-
ingrecursion rule:33
A1: A formula oftheformFActually1 l (where? is any
formula) is truewithrespectto a contextc, a possible
worldw, and othersemanticparameters (suchas a time
t, and an assignmentofvaluesto variabless) ifand only
if? is itselftruewithrespectto thecontextc, thepossi-
ble worldofc ratherthantheworldw, and theother
semanticparameters.
The extensionalsemanticrules governingthe cognatesof
'actually1'(in theirprimarysenses)are easilyderivedfromthis
clause governing 'actually1'
together withthe definitions
of the
cognatesintermsof'actually1' andsomeelementarymodalseman-
tics.For example,we thusobtain:
Awi:The predicate'actualwl' applieswithrespect
(correctly)
to a contextc, a possibleworldw, and othersemantic
parameters,to theworldofc, ratherthanto theworld
w, and to nothing else.
and
Ai1:The predicate'actuali1'
(correctly)applieswithrespectto
a contextc, a possibleworldw, and othersemantic
parameters i if
(suchas a timet), to a possibleindividual
and onlyifi existsin theworldofc (at t).
Anothercommonobjectionto thistheoryof 'actually'and its
cognatesis thatis clasheswithourunderstanding
ofwhatitmeans

84 / NathanSalmon
to say thata world(or individual,or proposition)is "possible."For

to say thatworldsotherthan the actual worldare possible is just
to say thattheworldthathappensto be actual mightnothave been
actual and thatsome otherworldmighthave been actual instead.
But if'actualwl'is indexical,such a claimis ruledout as semantical-
ly incoherent.On the indexicaltheory,in any actual contextof ut-
terance,onlytheactualworldmaybe properlycalled'actualwl'with
respectto any world,and everyworldotherthantheactual world
is properlycalled 'nonactualwl'even withrespectto itself.On the
indexicaltheory,then,itis a necessarytruthabouttheactualworld
thatit is theactual world.In whatsense are theotherworldspossi-
ble iftheycould not have been actual?34
A closely relatedobjectionraised by RobertAdams is this:One
may easilyglean fromthe indexicaltheory'ssemanticrulesA1,Ai1,
and Awl thatactuality(thepropertyof being actual),on the index-
ical theory,is of no special metaphysicalsignificance.Specifically,
thefactthatsomethingis actually1thecase, on theindexicaltheory,
does not make it ontologicallyor metaphysically more substantial
or importantthanifit were possiblythe case but not actually1the
case. For thefactthata certainpropositionis actually1the case, on
the indexicaltheory,is justthefactthatitis the case in a particular
possibleworld(whichhappensto be theworldoftheactualcontext
of utterance)-injustthe same way thatthe factthe some (recurr-
ing) event is occurringnow is justthe factthatit is occurringat a
particulartime(whichhappensto be the timeofthe contextof ut-
terance).Froman objectivepointofview,the factthata giventime
is the presenttimedoes not make itspecial in any way: it is just a
timelike any othertime,one thathappensto be the timeofa par-
ticularutterance.The factthatthe timein questionhappensto be
the timeof a particularutterance,by itself,is of no consequence.
Any timeis properlycalled 'thepresent'at thattimeand no other.
Similarly,on the indexicaltheory,to call our world 'actualwl',per
se, is notto attribute
to itany metaphysically significantdistinction.
The factthata givenpossible worldis actualwlis just the factthat
itis thisworldratherthansome otherworld.Thisdoes notconstitute
any special status;every world is the worldit is and not another
world.Indeed,thisfeatureoftheindexicaltheoryis preciselywhat
gives the point to Lewis's response to his envisaged Ontological
Arguer.But it is greatlyat odds withour ordinarythinkingabout
actualityand mere possibility especiallyas reflected
(nonactuality),

Existence/ 85
in our ordinaryvalue judgementsin connectionwith actual and

nonactualevents.We judge it good thata cure forsome terrible
disease is actuallydiscovered.We do notjudge it good (indeedwe
probablyjudge it bad) thata cure mighthave been discoveredbut
is notactuallydiscovered.We condemnsomeone foractuallycom-
mitting assault.We do notcondemnsomeonemerelyon thegrounds
thathe or she mighthave committedassault in radicallydifferent
circumstances. We mighteven applaudsomeoneforactuallyresisting
provocationto assault(unlessit is ClintEastwood).We feelpityfor
the victimsofactual disasters.We do not feelpityforthewould-be
victimsof disastersthatmighthave occurredbut did not actually
occur.To quote Adams:"ifwe ask, 'Whatis wrongwithactualizing
evils,since theywill occur in some otherpossibleworldanywayif
theydon'toccur in thisone?', I doubtthatthe indexicaltheorycan
providean answerwhichwillbe completelysatisfying ethically."35
These objectionshave considerableforce.But theycan be com-
pletelymetwhileaccommodatingwhat truththeymay containby
invokingLewis's secondary,nonindexicalsense of'actually'and its
cognates.36The secondary-senseanalogues to the threesemantic
rules given above are the following:
A2: A formulaof the form rActually2 1 (where ? is any
formula)is truewithrespectto a contextc, a possible
world w, and othersemanticparameters(such as a time
t, and an assignmentof values to variabless) ifand only
if ? is itselftruewithrespectto the contextc, the possi-
ble world w (ratherthan the world of c), and the other
semanticparameters.
AW2:Thepredicate'actualW2' (correctly)applies withrespect
to a contextc, a possibleworld w, and othersemantic
parameters,to the world w (ratherthan to the world of
c) and to nothingelse.
Ai2: The predicate'actualY2' (correctly)applies withrespectto
a contextc, a possible world w, and othersemantic
parameters(such as a timet), to a possible individuali if
and only ifi exists in the world w (at t).
It is immediatelyapparentfromthese semanticrules governing
the secondarysense of 'actually'and its cognatesthatthe expres-
sionsin questionare nonindexicalin theirsecondarysenses.By con-
trastwiththesemanticrulesgoverningtheprimarysenses,thecon-

86 / NathanSalmon
textc playsno significant rolein thesemanticrulesgoverning the

secondarysenses. More interesting, the propositional operator
'actually2'itselfplaysno significant semanticrole.It is completely
superfluous, in thatthetruth-conditions (withrespectto semantic
parameters) ofanyformula oftheformFActually2 q]l, as givenby
A2,are exactlythoseoftheimmediate subformula ? itself.To say
thata proposition is actuallythecase in thesecondary senseis just
to say thatitis true,no moreand no less.Thisfullyaccordswith
Lewis'sexampleofthesecondary sense:'IfMaxate less,he would
actuallyenjoyhimself more'.Italsohelpstoexplainwhytheadverb
is often
'actually' usedas a deviceforemphasis orforindicating con-
trastbetweenbeliefor expectation andreality, as in Lewis'sexam-
ple,ratherthanas a modalauxiliary. Therecanbe littledoubtthat
theadverb'actually'has thesetwodistinct uses.37The contextof
use-thepointoftheutterance-will generally favoronereading over
theother,although itneednotineverycase.Anexactlyanalogous
ambiguity arisesinthetemporal modewiththeworld'current'. Con-
sider:'In 1950thecurrent U.S. president was a Democrat'.
Sincea possibleworldis saidtobe actualifandonlyifitisactual-
lyrealizedanda possibleindividual is saidto be actualifandonly
ifitactuallyexists,tosaythata possibleworldisactualinthesecon-
darysenseisjustto saythatitis realizedandtosaythata possible
individualisactualinthesecondary senseisjusttosaythatitexists.
ThusactualityW2 is justtheproperty ofbeingrealized.Thisproper-
tywas explainedabovein termsoftheabstract-entity conception
ofpossibleworlds:Ifa possibleworldis a maximalcompossible set
ofpropositions, theproperty ofbeingrealizedistheproperty ofbe-
inga maximalsetoftruepropositions; ifa possibleworldis a max-
imalstateofaffairs thatmighthaveobtained, theproperty ofbeing
realizedis theproperty ofbeinga maximalstateofaffairs thatob-
tains,and so on. Likewise,actualityi2 is justtheproperty of ex-
istence.Thisproperty wasanalyzedinSectionIIIaboveintermsof
thelogicalnotions ofabstraction, negation, universal quantification,
andidentity. Theseconstruals of'actual'inthesecondary senseare
complemented bythesemantic rulesAW2andAi2,whichimputethe
verysame nonindexical extensionalsemanticsto 'actualW2' and
as wouldbe correct
'actualY2' for'realized'and 'exists', respectively.
Thesecondary senseof'actualw' is evidently theappropriate sense
forunderstanding thetruism thatsomepossibleworldotherthan
theactualworldmighthavebeen"actual"instead.Itis also in the

Existence/ 87
secondary senserather thanintheprimary sensethatcallinga world

possibleistosaythatitmight havebeen"actual." Eachpossible world
is realizedinitself andinno otherworld;henceeveryworldis the
actualW2 worldinitself. Butonlythisworld-thewaythings happen
tobe-is actuallyrealizedintheprimary sense,andineveryworld
it is theone and onlyactualwlworld.38
Bythesametoken,itmustbe saidthatactuality inthesecondary
sense,bycontrast withactuality intheprimary sense,isinsomesense
a specialstatus.The feature ofa proposition thatitis true,andthe
feature ofa stateofaffairs thatitobtains, andthefeature ofa possi-
ble individual thatitexists,are unlikethefeatures ofbeingtrueor
obtaining or existingina particular world(actualityintheprimary
sense)in thatthelatterare all more-or-less ordinary extra-world
features havingno metaphysically specialentailments whereasthe
former areallspecialintra-world featuresthataffordtheirpossessors
ina givenworlda metaphysically statusin thatworld.39
significant
Thatactuality inthesecondary senseis insomesensean objective-
lyspecialsortofstatusis notthesortoffactthatwouldordinarily
requirea substantiating argument. In thesensein whichitis true,
itis alsoperfectly obviousandcompletely trivial.40
Amongproposi-
tionsin a givenworld,thosethatare trueare obviously specialin
a certainway.Likewise, ifyouwerea stateofaffairs, wouldyou
ratherobtainor notobtain?Thatexistenceis metaphysically more
significant thannonexistence is hardly thesortoffactthatcouldbe
opentoquestion. Anyonewhodoubtsorseriously questions whether
existence ismetaphysically moresignificant thannonexistence simply
does notunderstand thephrase'metaphysically as itis
significant',
usedin thepresentcontext, or else misunderstands theword'ex-
istence',or else is takingone or theotheroftheseexpressions in
somenonstandard sense.
Onefinalpointaboutthistheory of'actually'anditscognates must
be stressed. Itis oftenclaimed,byproponents andcriticsalike,that
on theindexicaltheory, to say thata possibleworld(ora possible
individual, ora proposition) is actual(intheprimary sense)is tosay
merelythatitis (orexistsin,or is truein)theworldofthecontext
ofutterance. Similarly,itisoften saidthattheactuality(intheprimary
sense)oftheactualworld(orofan actualindividual, orofa proposi-
tionthatis actuallythecase) on theindexicaltheoryis a property
thatis possessedonlyin relationto a speakerand hisor hercon-
text.Forexample,in hisoriginalpaperLewiswrites:"Theactual

88 / NathanSalmon
worldis notspecialinitself,
butonlyinthespecialrelation
itbears
to the ontologicalarguer.... It is trueof any world,at thatworld
butnotelsewhere, Morerecently,
thatthatworldaloneis actual."'41
hesaysthathisindexical theoryof'actual'"makesactuality
a relative
matter: everyworldisactualat itself, andthereby all worldsare on
a par.... The 'actualat' relation
betweenworldsis simply identity.
... Surelyit is a contingentmatterwhichworldis actual ... at one
world,one worldis actual;at another,
another."42 in his
Similarly,
critiqueofLewis'stheoryAdamswrites:"According to theindex-
icaltheoryofactuality,
theactualityoftheactualworldconsistsin
itsbeing... theworldin whichthisact oflinguisticutteranceoccurs.
... According to theindexicaltheory, actuality is a propertywhich
theactualworldpossesses, notabsolutely, butonlyinrelation tous,
itsinhabitants."43Theseclaimsinvolvea confusion aboutthenature
ofindexicality in general,and maybe traceableto a use-mention
confusion. The claimsare moreappropriate fortheproperty ofbe-
ingcorrectly called'actual'in English, thanfortheproperty ofac-
tualitythereby attributed.
Indexicality is a featureofexpressions,
notoftheproperties designated bytheseexpressions. Forthisreason,
itis betterto speaknotoftheindexicaltheoryofactuality, butof
theindexicaltheoryof'actuality' in English.Thatactuality in the
primary senseis neither context-relative norcontingent on thein-
dexicaltheory caneasilybe seenfrom thesemantic rulesgoverning
'actually1'and itscognates.On theindexicaltheory,to say that
something is actually1thecase is to say thatitis thecase in a par-
ticularpossibleworld.Theparticular worldinquestionis,ofcourse,
theworldofthecontextofutterance, butthatthisis so is notpart
ofwhatis asserted.Exactlyanalogously, theproperty ofoccurring
nowisnottheproperty ofoccurring simultaneously withanyspeech
acttoken,buttheproperty ofoccurring at a particulartimet. That
timet is theverytimeatwhichI wrotethepreceding sentence,but
theproperty ofoccurring at t is notthesamethingas theproperty
ofoccurring whenI wrotethepreceding sentence. On anygivenoc-
casionofutterance of'occurring now',theproperty designated will
be indexedtotheverytimeoftheutterance, so thatwhatproperty
is designated willvaryfrom utterance toutterance. Alsoanalogous-
ly,theproperty ofbeingmeisnottheproperty ofbeingthespeaker
orproducer ofa particular utterance. Rather, itis NathanSalmon's
haecceity, theproperty ofbeingtheveryindividual NS.Theproper-
tiesdesignated bysuchindexical expressions as 'occurringnow'and

Exstence / 89
'beingme' are notthemselvescontext-relative in anystraightforward

sense. Quite the contrary;the propertyof occurringnow is a tem-
porallyindexed property,and hence it is not the sortof property
thatsomething(a recurringevent) has relativeto some timesand
notto others.In thesame way, the propertydesignatedby 'actuali1'
is an extra-world property;ifa possibleindividualhas thisproperty
at all, it has thispropertyrelativeto everyworld,and ifa possible
individuallacks theproperty,itlacks thepropertyrelativeto every
world.In fact,thepropertydesignatedby thenonindexical'actual'2'
may be said to be context-relative in a way thatactualityi1 cannot.
The formerpropertyis just existence,whichis an intra-world pro-
pertythata possibleindividualhas relativeto any worldin which
it exists.The temporalanalogue of thisis equally trueof 'current'
in itsnonindexicalsense. Similarly,the propertydesignatedby the
nonindexicalphrase 'being the speaker' mightbe called 'context-
relative'in thatan individualhas thispropertyrelativeto any con-
text in which he or she is the one doing the talking.By contrast,
theproperty designatedinthepresentcontextbytheindexicalphrase
'being me' is such thatan individualhas it relativeto a givencon-
text if and only if he or she is NathanSalmon,regardlessof how
much talkinghe or she may be doing in the context.
Actualityin theprimarysenseper se is ofno special metaphysical
significance;actualityin the secondary sense is in some sense
metaphysically Lewis'scriticism
significant. oftheOntologicalArgu-
mentis that,since actualitywl is no specialdistinction,it is a mistake
to arguethatifany possibleindividualis divineinanypossibleworld,
itis onlyfitting thatsome possibleindividualshouldbe divinein the
one and onlypossibleworldthatis specialby virtueofitsactuality.
We have justarguedthateven iftheactual worldis nothingspecial
just forbeing uniquelyactualwl,neverthelessit is triviallyspecial
and metaphysically distinguished by virtueof being actualW2, that
is,byvirtueofbeingrealized.Lewis'sacknowledgedsecondarysense
of 'actual' thus seems to undercuthis criticismof his suggested
groundsor basisforpremise(2a) ofVersion2. (Butsee note40 above.)
Does thismean thatLewis's suggestedbasis for(2a) is adequate
afterall?Surelynot.Actuality in thesecondarysenseis metaphysical-
ly special in some sense, but it is notso special thatany otherpro-
perty(ofa givensort)thatis specialor important in some sense will
ipsofactohave some instancein actuality2.Considertheverypro-
pertyin question:divinity.For Descartesdivinityis thepropertyof

90 / NathanSalmon
having all perfections.For Anselm it is the propertyof havinga

magnitude of greatness that exceeds any other conceivable
magnitudeofgreatness.Whicheverconstrualone chooses, divinity
is no doubtin some way a veryspecial status,one thatenjoysvery
special religioussignificance.In the same way, the propertyof be-
ing the state of affairsof therebeing some possible individualthat
has divinityis itselfveryspecial,ofconsiderablereligioussignificance.
The propertyof beinga possiblestate of affairsthatobtainsis also
special,butin a verydifferent way. Itis special ina distinctly secular
and peculiarlymetaphysicalway. The factthatthe state of affairs
of therebeing some possible individualthatis divine is special in
thefirstway is no groundwhatsoeverforthe hypothesisthatitalso
has a propertythatis special in thesecond way,the metaphysically
specialproperty ofobtaining.Atmost,itsupportsonlythehypothesis
thatthisstate of affairsdeserves or oughtto obtain-in the sense
thatitwouldbe good of"fitting" ifitdid.Whatis wrongwithLewis's
suggestedbasis for(2a) is notthatactualityin any reasonablesense
is notspecial;itis thatthesuggestedbasisis no basisat all.One might
as well argue that,since being the best of all possibleworldsis in
some sense a veryspecial property,it is only fitting, and therefore
true,thattheworldthatis special foritsactualityW2 shouldalso en-
joy thisotherspecialproperty. Itwouldfollowfromthislineofreason-
ing(assumingthattheproperty ofbeingthebestofall possibleworlds
is necessarilyspecial)thateveryworldis,accordingto itself, thebest
ofall possibleworlds.The incurableoptimist, and themetaphysical-
ly deterministic pessimist,may be contentwiththisargument.The
restof us knowthat,fitting thoughitmay be, theactualW2 worldis
hardlythe best of all possibleworlds(even thoughit is indeed the
mostrealizedofall possibleworlds),and thattherefore, itis literally
impossiblefortheactualwlworldto be thebestofall possibleworlds.
VI
I suggestedin SectionII above thatnonexistent

possibleindividuals,
suchas Noman,have properties-for example,theproperty ofnonex-
istenceand itsentailments. These entailments
includesuch negative
propertiesas thatofnotbeinga philosopher.It does notfollowthat
ifyou are asked to count up everythingthatis not a philosopher,
Nomanis to be includedin the count.Norshould the dinosaursbe

Existence/ 91
includedin the count.Like the dinosaurs, Nomanin notone of

everything. Consequently, he is notone ofeverything thatis nota
philosopher. Indeed,notbeingoneofeverything is theveryproper-
tyofNomanwe startedwith.
By contrast withMeinongians, I am notclaimingthatthereare
individuals thatdo notexist.Ifthequantifier 'thereis' is actualist,
thatMeinongian claimis simply contradictory-and otherwise, itis
trivial.WhatI amclaiming isthattheremighthave beenindividuals
thatdo notactually,existand thatactually,have certainproper-
ties.AlvinPlantinga has giventhename'seriousactualism' to the
doctrine thatnecessarily, everyindividual issuchthatitmustexist
ifit is to have anyproperties at all.44In Plantinga's terminology I
am denying seriousactualism whilemaintaining (a versionof)ac-
tualism.ButI am dead serious.Myclaimis philosophically quite
moderate, notnearly as radicalas itmight seem.Exactly analogously,
therehavebeenindividuals thatdonotnowexistbutthatnowhave
certainproperties. Somepastdinosaurs nowhavetheproperty of
beingfossilized, andsuchimmortal artistsas MozartandJohnLen-
non are justlyadmiredby millionstoday.Not to mentionsuch
posthumously acquiredproperties as arisefrom posthumous awards
andthelike.Ifnothing else,therearealwayssuchproperties as hav-
ingonceexistedandhavingbeena musician. Thisis fundamentally
thesamephenomenon: Anindividual from onecircumstance hascer-
tainproperties in anothercircumstance in whichitdoes notexist,
as a resultoftheproperties it has in itsowncircumstance.
In fact,so-calledseriousactualismis reallyquitea radicaldoc-
trine.45 Thereis no groundforthisdoctrine thatwouldnotprovide
analogousgroundsfordenying presentproperties to suchpastin-
dividualsas JohnLennonand dinosaurs. It mightbe thought that
pastindividuals andpaststatesofaffairs areinsomewaymorereal
thanpossibleindividuals thatnevercomeintoexistenceandpossi-
ble statesof affairs thatneverobtain.46 We are concernedmuch
morewithindividuals andeventsfrom ourpastthanwithindividuals
and eventsthatnevercometo pass,andthisissometimes takenas
evidenceofthegreater degreeofreality we attribute tothepastover
thepossible-but-never. Thosewhosee thingsthisway usuallyat-
tribute an intermediate degreeofrealityto future individuals and
future statesofaffairs-more realthanneverexistent individuals and
neverobtaining statesofaffairs, butlessrealthanpastindividuals
andpaststatesofaffairs. Thisisall a mistake. Pastindividuals were

92 / NathanSalmon
morerealthanmerelypossibleindividuals are,andeventsthatoc-
curredinthepastweremorereal(insomesense)whentheyoccur-
redthaneventsthatneveroccurare now.Forthatmatter, future
individualswillbe morerealthanmerelypossibleindividuals are,
and future eventswillbe morerealwhentheyoccurthanevents
thatneveroccurarenow.Thepastreality ofan individual orevent
maygiveusa present reasonforconcerninregardtothatindividual
orevent.Contrary towhatonewouldexpectaccording tothecom-
parativerealityviewI amdisputing, we aretypicallyconcerned more
aboutfuture realitiesthanaboutpastrealities, at leastwithregard
to future we knowofor anticipate.
realities Thebondageofcausa-
tiontotime'sarrowgivesus a present andpressing reasonforcon-
cernaboutfuture generationsandfuture events.What'sdoneisdone.
We cannotchangethepast,butourpresentactionsand inactions
to a greatextentdetermine thefuture. As faras thepresent is con-
cerned,pastindividuals and statesofaffairs, futureindividuals and
statesofaffairs,andforever merelypossibleindividuals and states
ofaffairsare on a par:theyare nowequallyunreal.The future is
nevertheless a topicofspecialpresentconcern,becauseitwillbe
real,andwhatwe do nowdetermines whatitwillbe. Furthermore,
on a journeyin thedirection
we are all time-travellers, oftime's
arrow.
Ofcourse,sincesuchmerelypossibleindividuals as Nomanhave
propertieseventhoughtheydo notexist,ifourquantifiers are ac-
tualist,
thentheclassicallogicalrulesofuniversal instantiation and
existentialgeneralizationare fallacious.47 Insteadwe have free
logicalversions:FEverything is (P. Q exists.Therefore,Q is (' and
FQ is (. Q exists.Therefore,somethingis qP1 . In additionto these
we have the following possibilistvariations:rEverypossiblein-
dividualis (P. Q is a possibleindividual.
ThereforeQ is (- and rQ
is (. Q is a possibleindividual.Therefore,somepossibleindividual
is (P . Ifthesingular termQ is a simpleindividual
constant (proper
name)orvariableandthepossibilist aredefined
quantifiers interms
of the actualistquantifiers, thenthesepossibilistversionsof free
logicalUI and EG are tantamount to thefollowing:
Necessarily,
everythingactually,is (P.
Q mighthaveexisted.
Therefore,Q is (P.
and

Existence/ 93
Q is (P.
(P mighthave existed.
thatactually,is
theremighthave been something
Therefore,
qP.48
We couldhavesomething more.Theoriginal freelogicalversions

ofUIandEGarerequired bythepresence oftruesentences inwhich
singular termsthatdo notreferto (denote)existing individuals oc-
cur(outsideofnonextensional contexts,suchas thosecreatedby
quotations marks),whether or notthesetermsrefertopossiblein-
dividuals thatdo notexist.Ifwe requirethatall ourtermsreferto
possibleindividuals, we mayretaintheformofclassicalUI and EG
usingthepossibilist Ifthepossibilist
quantifiers. quantifiersaredefined
in termsoftheactualist quantifiers,thisis tantamount to deleting
themodalexistential secondpremise from thepossibilist freelogical
UIandEGrulesdisplayed above.Unfortunately, notallsingular terms
thatdo notrefer to existingindividualsrefertopossibleindividuals
thatdo notexist,as witnessQuine's'themerelypossiblefatman
in thatdoorway'and Meinong's'theroundsquare'.49 No merely
possiblemanis actually,fator actually1 in Quine'sdoorway,let
aloneboth,and no merelypossibleindividual is actually1roundor
actually1 square,letaloneboth.Andofcourse,therecouldnotbe
any impossibleindividuals. These descriptions are thusstrongly
nonreferring, inthattheynotonlydo notrefer toanyexisting thing,
theydo notevenrefer toanymerely possiblething. Yetthereseem
to be truesentences inwhichsuchstrongly nonreferring termsoc-
cur;forexample,thenegativeexistential 'Theroundsquaredoes
notexist'.We couldfollowFrege'sstrategy and stipulate thatall
strongly nonreferring termsshallhereafter refer toNoman.Wecould
thenhave ourclassicalUI and EG back,at leastin form, byinter-
pretingthequantifiers Thenegative
possibilistically. existential'The
roundsquaredoesnotexist'wouldstillbe true,as wouldthemodal
sentence'Itis possiblefortheroundsquaretoexist'.Indeed,thelat-
terwouldbe logically true.Butas Russellnotedindiscussing Frege's
strategy, "thisprocedure, though itmaynotlead to actualP2 logical
anddoesnotgivean exactanalysisofthe
error,is plainlyartificial,
matter." Darn!Russellis right.
Does Nomanhave any positiveproperties in additionto such
negative propertiesas notexisting andnotbeinga philosopher? Yes.
Forexample,he hasthemodalproperty ofpossibly existing andits

94 / NathanSalmon
entailments.He also has the dispositionalpropertythathe would

be male if he existed.
Does he have any nonnegativenonmodalproperties,then?Yes
he does. He has the propertyof being mentionedand discussedin
theseverypassages. In fact,as was intimatedtwoparagraphsback,
he has the more fundamentalsemanticpropertyof being referred
to by thename 'Noman'.Indeed,Nomanis rigidlydesignatedby the
name 'Noman'. Again, it does not followthatthe name refersto
something. Nomanisnotsomething, and hence,even though'Noman'
refersto him,thereis nothingthat'Noman' refersto. Still,Noman
mighthave been someone;he mighthave existed.Although'Noman'
does not referto any actuali' individual,it does referto a possible
individual.It is thus only a weakly nonreferring term. That is,
although'Noman'does not actually,referto anything, theremight
have been someone x such that 'Noman' actually1refersto x.50
Referenceprecedesexistence.This is not to say thatifNoman had
existed,the name 'Noman'would have referredto him.Indeed, if
he had existed,thename wouldnot have been conferredonto him.
The name only contingently refersto him.In fact,the name con-
tingentlyrigidlydesignateshim.
How does a name like'Noman'come to referto a merelypossible
individuallikeNoman?Throughfixingitsreferenceby description,
in a standardKripkeanstipulation.Ofcourse,thedescription operator
involvedmustincludemerelypossibleindividualsin its range,but
we have already seen thatthispresentsno problem.(See notes 10,
19 above.) The hard part is findinga propertythatuniquelyiden-
tifiesa particularmerelypossibleindividual.In Noman'scase, that
was not difficult:
Nomanis the onlypossibleindividualwho would
have developed fromthe unionof the particulargametesS and E
ifS had fertilizedE in the normalmanner.Notall merelypossible
individualsare so easily pinned down.51
Since 'Noman' refersto Nomaneven thoughhe does not exist,a
sentencecontaining'Noman' mightexpressa possibleproposition
about Noman even thoughthe possiblepropositiondoes not exist.
Considerthe following:
(8) Noman is a native Californian.
This sentenceexpressesthe possiblepropositionthatNomanis a
It is arguablethatthispropositionis a Russellian
nativeCalifornian.
singularproposition(David Kaplan)is whichNomanhimselfoccurs

Existence/ 95
as a constituent.52 In any event, by uttering(8) one asserts of

Noman,de re, that he is a native Californian.Many philosophers
would agree thatin assertingof an individual,de re, thatit has a
certainproperty, one therebyassertsa singularproposition inwhich
the individualin questionoccurs directlyas a constituent.53 Thus,
in uttering(8) one maybe regardedas assertingthepossiblesingular
propositionabout Nomanthathe is a nativeCalifornian.This pro-
positionis false.In fact,itdoes noteven exist.(Recalltherestriction
on EG.) Butitis possible,in twoimportant senses.First,itmighthave
existed.Second, it mighthave been true.(As a matterof fact,ifit
had existed,it verylikelywould have been true.)There is no pro-
positionthatsentence(8) actually1expresses,buttheremighthave
existeda propositionthatthe sentenceactually,does express.This
is the possiblepropositionabout Nomanthathe is a nativeCalifor-
nian. The factthatthispossible propositionmighthave been true
underliesthe factthatthe modal sentence
Noman mighthave been a native Californian
actually,is true.
In fact,some merelypossiblepropositions are truedespitethefact
thattheydo not exist,forexample, the possiblesingularproposi-
tionabout Nomanthathe does notexist,and itsentailments. Indeed,
forany possibleindividualx, thepossiblesingularproposition to the
effectthatx does notexistis necessarilysuchthatifitis true,itdoes
not exist. Its truthentailsits nonexistence.
There is an especiallyremarkableanomalythatarisesfromthese
considerations.Let ENS be the ovum from which I actually,
developed. Considernow the possibleindividualwho would have
developed fromtheunionofthespermcellS fromNoman'spossible
zygote withthe ovum ENS frommy actualPzygote,ifS (insteadof
the spermcell fromwhichI actually1developed) had fertilized ENS
in the normalmanner.Let us namethispossibleindividual'Nothan'.
It would seem thatitis literallyimpossibleforbothNothanand me
to existtogether.Ifone ofus exists,the othercannotalso exist.We
are incompossibleindividuals.Nevertheless, Nothanand I stand in
certaincross-world relationsto one another.(Infact,we are incom-
possible brothersacross possibleworlds.)IfNothanhad existedin-
steadofme,he wouldhave growntoreachsomedeterminate height.
It is eithertruethatNothanwould have been tallerthanI actually1
am, or else it is truethatNothanwould not have been tallerthan

96 / NathanSalmon
I actually,am. SupposeI utterthesentence

Nothanwouldhave beentallerthanI actually,am,
thereby assertingofNothanandmyself, de re,thathe wouldhave
beentallerthanI actually1 am,andsupposeSaulKripkedenieswhat
I assert.Hereagain,itseemsverylikely thatwhataretrueorasserted
arecertain singularpropositions inwhichNothan andI occurdirectly
as constituents,to wit,thesingular proposition thathe wouldhave
been tallerthanI actually1 am or thesingular propositionthathe
wouldnothavebeentallerthanI actually, am.54Although one of
thesesingular propositions is trueand theotherfalse,and one of
themassertedbyme and theotherbyKripke, ifNothanand I are
incompossible individuals,
neither singular proposition canpossibly
exist.Inanypossible worldinwhichoneofitsindividual constituents
exists,theotherindividual constituentdoesnot.Something exactly
analogousis trueofthecomplexdispositional statesofaffairsofit
beingthecase thatNothanwouldhavebeentallerthanI actually1
am,and itbeingthecase thatNothanwouldnothavebeentaller
thanI actually1 am.One ofthesestatesofaffairs obtains,yetneither
can exist.Orconsider insteadthede re modalproposition concern-
ing Nothanand me thatit is impossible forbothof us to exist
simultaneously. Thissingular proposition is no moreexistent than
thepossibleproposition thatNomanmighthave existed,and it is
no less true.But ifit is true,it cannotexist.Its truthentailsits
necessarynonexistence. Thus,therewouldseemto be a sensein
whichthereare'someimpossible objects(certainsingular proposi-
tionsorstatesofaffairs) thathavecertain properties (beingthecase,
obtaining, beingassertedor denied,etc.),eventhoughtheycannot
exist,and indeedinsomecases,theveryproperty in questionen-
tailstheimpossibility ofexistence.
Hereagain,I amnotmaking theMeinongian claimthatanydescrip-
tion,eveniflogically contradictory,refers to somepossibleor im-
possibleat object.Quine'sdescription 'themerelypossiblefatman
inthatdoorway'does notreferto anysortofobject,whether exis-
tent,merely possible,orimpossible. Itis a verystrongly nonreferr-
ingterm.Similarly, Meinong's roundsquareis notonlynota possi-
ble object,itis notevenan impossible object.Whatmakesan im-
possibleobjectimpossible is notthatithas contradictory or other-
wise incompatible properties. No object-whether existing,past,
future, forever merelypossible,or forever impossible-has incom-

Existence/ 97
patibleproperties. An impossible object,suchas thesingular pro-

positionthatNothanwouldhavebeentallerthanI actually,am,is
a complexconstructed outofpossibleobjects.Anysuchobjecthas
a perfectlyconsistent setofproperties; itisimpossible onlybecause
someofitsessentialconstituents are incompossible. Animpossible
objectcannotexist,butitcan anddoeshavetheproperties ithas.55
Presentexistence is nota pre-requisite forpresently haveproper-
ties.Noristhedisjunction ofpastandpresent existence, i.e.,thepro-
pertyofeitherexisting or have onceexisted.Noreven is thedis-
junctionofpast,present, and future existence, i.e.,theproperty of
existingat sometimeor other.Evenpossibleexistenceseemsnot
tobe a pre-requisite
forhavingproperties, sinceitseemsthatinsome
sense,some impossible thingshave properties! The moral:The
metaphysical condition ofhavingproperties is quiteseparablefrom
theontological condition ofexisting.Predication precedesexistence.
Ofcourse,anything thatexistshasproperties, butthisisbecausehav-
ingproperties is metaphysically utterlyunavoidable-ina waythat
even deathand taxesare not.Nomanis sparedthelatter, butno
object,notevenan impossible one,issparedtheformer. Suchisthe
negative-existentialpredicament.
Ifnonexistence, andevennecessary nonexistence, do notpreclude
havingproperties, whatcan be metaphysically so specialorimpor-
tantaboutexistence? How can actualityi2 be an important proper-
tywhenitis a necessary truth thateverything hasit,andeventhe
possibleindividualsthatdo nothaveit,andtheimpossible individuals
thatcouldnothaveit,nevertheless haveotherproperties? Whatis
itaboutactuality inthesecondary sensethatmakesitmetaphysically
important?
One reasonthatactuality2 is metaphysically important mightbe
thatso manyothersignificant properties dependuponit.Ifa possi-
ble stateofaffairs does notobtain,itcannotexplain,or cause,or
be theresultofanyotherstateofaffairs. Andunlessa particular
possibleindividual exists,it cannotbe anywhere or do anything.
Although Noman'sproperties arenotrestricted to negativeproper-
tiesandmodalproperties, theyareseverely restricted. Nomandoes
nothaveexperiences. A merely possibleindividual doesnotliveand
learn;itdoes notfeelpleasureand pain,or knowjoyand sorrow;
itdoes notlaughorcry;itdoesnotevenlie stillat rest.(Letalone
is any merelypossibleindividual divine,in anysignificant sense.)
The properties ofmerely possibleindividuals, andofimpossible in-

98 / NathanSalmon
dividuals, are inert; they include only such unimpressive

characteristicsas beingreferredto, not beinga nativeCalifornian,
and possiblyexistingor necessarilynot existing.Not an enviable
resume.The mere propertyof existing,once it is acquired,opens
up a galaxyofnew possibilities.The questionofwhetheran actual'2
individualis betteroffthana nonactuali2one probablydepends on
whichpropertiesthe actuali2individualhas. Existenceper se does
not make one well off,except insofaras it opens door to thepoten-
tial forbeingwell off.Unfortunately,it also opens the door to the
potentialforbeing badly off.
Notes
*Portionsof the presentpaper were presentedat a symposiumon pro-

blems of Existence and Identityat the Universityof NorthCarolina at
Greensboro(April1986); to the University of Padua, Italy;the University
of Belgrade,Yugoslavia; the AnalyticSectionof the PhilosophicalSociety
of Serbia,Yugoslavia; the UniversityofCalifornia,Santa Barbara; and the
1987 AlbertaPhilosophyConference.It has benefitted fromthe discussions
thatfollowed,fromcommentsby W. R. Carter,and fromfruitful discussions
withRobertAdams,AnthonyBrueckner,WilliamForgie,David Kaplan,Ali
Kazmi, and TimothyWilliamson.
1. I assumeherethatthereis onlyone possibleindividualwho wouldhave
resultedfromthe unionofS and E, ifS had fertilized E in the normal
manner.(This assumptioncan be expressedthroughthejudicioususe
of standardmodal operatorswithoutthe aid of a quantifierthatpur-
portsto quantify overmerelypossibleindividuals,as follows:Theremight
have existedan individualx such thatx and actuallynecessarilyonly
x actuallywould have developed fromthe union of S and E ifS had
fertilizedE in the normal manner. This alternativeformulationis
somewhatcumbersome,though,and more difficult to grasp than the
originalformulation.) The intuitionthatthisassumptionis trueis very
widelyshared.I am hererelyingon theassumptionmerelyas a device
to introducethe questionthatis the maintopic of thisessay. For fur-
therdiscussionof thisand relatedsufficiency principlesof cross-world
identitysee myReferenceand Essence(Princeton University Press,1981),
pp. 196-252,especiallyp. 209f;and "Modal Paradox:Partsand Counter-
parts,Pointsand Counterpoints," in P. French,T. Uehling,and H. Wett-
stein,eds.,MidwestStudiesin Philosophy XI:Studiesin Essentialism (Min-
neapolis: Universityof MinnesotaPress, 1986),pp. 75-120.
2. Cf.Salmon, 1981,pp. 116-133,on thisand othercross-world relations.
3. In thefirst paragraphof"On WhatThereIs,"in Quine'sFroma Logical
Point of View (New York: Harper & Row, 1961),pp. 1-19.
4. Ibid.,p. 15;andMethodsofLogic (NewYork:Holt,Rinehart and Winston,
Inc., 1972),p. 234.

Exstence / 99
5. Quine'smaximdoesnotdirectly concernthequestionofwhatthings
actuallyexist;itconcernstheontological commitments ofthisorthat
theoryor pieceofdiscourse (andby extension, theontological com-
mitments ofthisor thattheorist or speaker),irrespective ofwhether
thesortsofthings towhichthetheory ordiscourse isontologicallycom-
mittedactually exist.Quine'sthesisisthata theory orpieceofdiscourse
is committed totheexistence ofthings ofa givensort(and,as a special
case,totheexistence ofa givenpossible thing)ifandonlyifsomethings
ofthatsort(orthatpossiblething) mustbe countedamongthevalues
ofvariablesina suitablereformulation ofthetheory or thediscourse
ifthereformulated theoryordiscourseis to be true.The ontological
commitments ofa theory orpieceofdiscourse willthusindudeanything
whoseexistence affirmed,
is explicitly butI takeitthatthepointofthe
thesisis thatthe ontological commitments of a theoryor piece of
discoursemayoutrun theexplicitexistenceavowals.(Otherwise, itwould
be muchsimpler to say insteadthata theory or pieceofdiscourseis
ontologicallycommitted toallthings, andtoonlythosethings, thatare
explicitlysaid to existin a suitablereformulation of the theoryor
A number
discourse.) ofdifficulties
andproblems forQuine'sthesis could
be raised,thoughonlyfewwillbe mentioned here.
6. Cf.DavidKaplan,"BobandCarolandTedandAlice,"inK.J.J. Hintikka,
J.M.E.Moravcsik,and P. Suppes,eds.,ApproachestoNaturalLanguage
(Dordrecht: D. Reidel,1983):pp. 490-518,appendixX, at pp. 503-505
and especiallyp. 516,note15.
Quine'sthesismentioned suprain note5 appearsto havethefalse
consequencethatifthissentenceconcerning a particular fossilis true,
thenthingsthathavebeendinosaurs existtoday.(Immortal dinosaur
souls?)
It has beensuggested to me thatQuine'sactualproposedcriterion
ofontological commitment avoidsthisdifficulty sincethecriterion is
restrictedtoone'scommitments concerning existence at sometimeor
other,rather thantoone's(stronger)commitments concerning existence
i.e.,commitments
simpliciter, concerning whatsortsofthings areinthe
condition orstatethatsomething comesintowhenitbeginsto exist
and fallsoutofwhenitceasesto exist.Although I havebeenunable
tofindan explicit andclearformulation ofthisrestrictionamongQuine's
writings on hisproposedcriterion forontological commitment, this
tenselessconstrual seemstruerto thespiritof hisexplicit(and not
altogether independent) viewsconcerning canonicalnotation, verb
tenses,and theregimentation ofordinary language.See forexample
"Mr.Strawson onLogicalTheory," inQuine'sTheWaysofParadox(New
York:RandomHouse,1966),pp. 135-155, at pp. 143-146.(Thanksto
PetervanInwagen forprovidingthisreference.)Ifthisrestricted
criterion
accordsbetter withQuine'sactualintent, histhesiswouldbe lessdecep-
tively(albeitlessneatly)encapsulatedas follows:"To-be-or-to-have-been-
or-going-to-be is to be thevalueofa variable"'
Unfortunately, asidefrom uglinessofformulation, thisleavesus with
no criterionforone'scommitments concerning existence orbeingper

100 / Nathan Salmon
se, as opposedtoone'scommitments concerning existence-at-some-time-

or-other.Whatisdesiredisa tense-sensitive criterionthatcommits one
whoutters thepasttensedsentence'Thereusedto be (things thatat
sometimeorotherare)sea serpents' at a timet totheexistence prior
totof(things thatatsometimeorotherare)sea serpents, andonewho
utters thepresent tensed'Thereare (things thatat sometimeorother
are)sea serpents' at t to theexistence at t of(thingsthatat sometime
orotherare)sea serpents. Underthesuggested interpretation,Quine's
criterion is insensitive tothesedifferences intense,assigning toutterers
(att)ofeither tensedsentence theverysame(timeless) ontologicalcom-
mitment (att) to (things thatat sometimeor otherare)sea serpents.
7. Quineappearsto fallintojustthisconfusion, forexampleop. cit.,p.
13 (wherehe speaksof "thethingsover whichthe boundvariable
'something: ranges"),and elsewhere.
8. Quine'sthesismentioned suprain note5, ifI understand itcorrectly,
has thefalseconsequencethatifthissentenceconcerning S andE is
true,thensomeindividual whomighthavedevelopedfrom theunion
ofthesegametes actually exists(atsometimeorother). Cf.note6. These
apparentconsequencesof Quine'sthesismay demonstrate thathis
criterion ofontological commitment actually appliesnottothisorthat
theory, as Quineintends, buttothesemantic metatheory fora suitable
language inwhichthisorthattheory isformulated. Eventhusconstrued,
however, thecriterion givesat mostonlya sufficient condition foron-
tologicalcommitment ofthemetatheory as augmented withtheaffir-
mationofthetruth oftheobjecttheory; nonecessary condition isgiven.
In fact,I believethatthiscondition is notevena sufficient condition.
WhenI assignNomanto somevariableas itsvalueI commitmyself
toNoman's suitabilityas a valueforvariables, nottohisactualexistence.
9. Theseare nottheonlydoctrines thatgo bythese'ism"s;norarethese
theonly'ism"sthatthesedoctrines go by.
10. See KitFine,"Prior ontheConstruction ofPossible Worlds andInstants,"
postscript toA. N.PriorandK.Fine,Worlds, TimesandSelves(Amherst:
University ofMassachusetts Press,1977),pp.116-161. Forexample,the
locutionrSomepossible individual is(P1 maybe defined as FThepossi-
ble worldw thatis realizedis suchthattheremighthaveexistedan
individual that,in w,is (P1 .
11. "OnBehalfoftheFool,"inA. Plantinga ed.,TheOntological Argument
(GardenCity;Doubleday& Company, Inc.,1965),pp. 6-13,at p. 11-12.
12. InTheFourfold RootofthePrinciple ofSufficient Reason,inPlantinga,
1965,pp. 65-67.
13. Itis oftenargued(mostnotably byAlvinPlantinga) thatbeliefin God
is nolessrationally justified thatmanyotherunproved andcontestable
philosophical beliefsthatare widelysharedand usuallyregardedas
knowledge, suchas thebeliefin othermindsor thebeliefthatthere
is an external, material world. See forexampleA. Plantinga, TheNature
ofNecessity (Oxford: Oxford University Press,1974),atp.221.Theissue
oftherationality ofbeliefinGodcannotbe discussed adequately here,
ofcourse,butitshouldbe notedthathistorically, thefunction androle

Existence/ fOI
oftheOntological Argument inphilosophy is integrally relatedto the

viewthatthehypothesis ofGod'sexistence requires substantial justifica-
tion,intheform ofsomething likeproof, ifitistobe rationally adopted.
Theobservation thatmanyexternal existence beliefsusuallyregarded
as knowledge are basedonverylittleinthewayofdecisiveevidence
seemsbothcorrect and epistemologically significant. However,there
is an epistemologically important pointofdisanalogy betweenbeliefin
Godandbeliefinothermindsorintheexternal worldc Thehypotheses
ofothermindsandoftheexternal worldareextremely plausible(even
withrespecttotheepistemic situation ofsomeonewhohas notbeen
philosophically indoctrinated sincechildhood concerning otherminds
ortheexternal world), whereasthehypothesis ofGod'sexistence isfun-
damentally implausible (at leastforthosewhoare able to breakfree
oftheirchildhood religious indoctrination or whoneverhad any),or
atmost,notsignificantly moreplausible thanthehypothesis ofthereal
existenceof themythological Olympiangodsof old, or thanother
superstitious orocculthypotheses. Indeed, itisdifficult toimagine a non-
philosopher whodidnotbelieveinotherminds orintheexternal world,
yetthereare massesofnonphilosophers whodo notbelievein God.
It is notthecontestability orunprovability ofthehypothesis ofGod's
existenceas much as its intrinsic implausibility thatrendersthe
hypothesis in need ofevidenceor proofforitsjustification.
14. In callinga thesis'Kantian', I do notmeanthatitwas infactheldor
endorsedbyKant,onlythatitis inthespirit ofthesesoftenattributed
to Kant.
Thereis,ofcourse,thefourth Kantianthesisthattheallegedproper-
tyorconceptofexistenceis nota predicate ofGerman, or anyother
natural language, butitisdifficult tosee howthistruism couldbe thought
to offeranyfoodforthought to thelikesofAnselmand Descartes.
15. Foran argument thattheidentity predicate is notvague,see Salmon,
1981,pp. 243-245;and theappendixto Salmon,1986b.
16. I havesaidthatthequantifiers 'everything' and'something' ofstandard
Englishdo nothavea fixeddomain,andmaybe restricted invarious
waysaccording tothecontext ofuse,butthatthedefault valueisrestric-
tionto existing things.Thissuggests a treatment oftheEnglishquan-
tifiersonthemodeloftheindexical phrases'everything ofthatsortand
'something ofthatsort',tobe supplemented orcompleted bya contex-
tualindication or"demonstration" ofthesortinquestion, whereno ex-
plicitdemonstration constitutes bydefault a contextual indication ofthe
sortexisting thing.Onthispicture, the"definition" provided bytheslogan
'To be is to be identical withsomething' makesthe'is'ofbeingan in-
dexicalpredicate ofindividuals, shorthand for'isidentical withsomething
ofthatsort'.Indeed,the'is' ofbeingin Englishdoesseemto display
thesamesortofcontext-sensitivity as thequantifiers 'everything' and
'something'. Itisonlywhenthedemonstrative element takesitsdefault
valuethatthesloganbecomesa "definition" ofthe'is'ofexistence. The
result isa specialsortofostensive definition, rather thana nominal defini-
tion,one employing a peculiarsortofostension-by-default. Giventhis

102 / NathanSalmon
pictureoftheinter-relations amongthequantifiers, the'is' ofbeing,

andthe'is'ofexistence, itis doubtful thata philosophically satisfactory
nominaldefinition of the'is' of existencecan be given.(I have not
said howfarthispictureshouldbe maintained.)
17. See SaulKripke, Naming andNecessity (Harvard University Press,1980),
pp. 24, 156-158.
18. Kripke, 1980,p. 108;and"Speaker's Reference andSemantic Reference,"
inP. French, T. Uehling, andH.Wettstein, eds.,Contemporary Perspec-
tivesinthePhilosophy ofLanguage (Minneapolis: UniversityofMinnesota
Press,1979),pp.6-27, at p. 16.
19. See note10.Fine'sresults combined withRusselrs TheoryofDescrip-
tionsenableonetosecuretheeffect ofreferring bydefinite description
topossiblebutnonexistent individuals. ThelocutionrThepossiblein-
dividualthatis (Pis Ip1 maybe defined as rThepossibleworldw that
isrealizedissuchthattheremight havebeenan individual x suchthat,
in w,x and necessarily onlyx is (P,and in w,x is tl .
20. Actualists claimthatitis alsoa logicala priori truththatanything that
is a lionis a lionthatexists.Ifactualism is correct,itis a conceptual
truth thatall andonlylionsare exilions.
21. In "OnDenoting" Russellnotoriously raisesa number ofobjections to
Frege'stheory ofSinnandBedeutung thatareapparently basedonone
or moreconfusions or misunderstandings. One particular objection in
"On Denoting"' is quitepowerful, butis briefly statedamidtheother,
mistaken criticisms, and consequently hasbeen undulyneglected. In
connection withhis exampleof 'thepresentKingofFrance'Russell
writes:"Oragainconsider sucha proposition as thefollowing: 'Ifu is
a classwhichhasonlyonemember, thenthatonemember isa member
ofu', or,as we mightstateit,'Ifu is a unitclass,theu' is a u. This
proposition oughtto be alwaystrue..." OnFrege'theory, anyEnglish
sentencecontaining thephase'thepresentKingofFrance'(andfree
ofobliquedevices)is neither truenorfalse.Russellcorrectly pointsout
thatwhereasMeinong's theory errsinonewaybycounting thesentence
'The presentKingofFranceis a presentKingofFrance'as logically
truewheninreality itis noteventrue,Frege'stheory errsin another
butequallyobjectionable waybydiscounting theweaker,conditional
sentence'Ifthereis exactlyone present KingofFrance,thenthepre-
sentKingofFranceis a present KingofFrance',whichis logically true,
as noteventrue.Bycontrast withFrege,the"secondary occurrence"
or "narrowscope"readingof thelattersentenceis indeeda trivial
theoremofPrincipia Mathematica. (Thisis notto say,though,that
Russell'saccountof definite descriptions as contextually defined"in-
complete symbols" thataremoreanalogoustosecond-order predicates
thantosingular terms is superior toan accountofdefinite descriptions
as complete, genuinesingular terms.)
22. In Plantinga, 1965,at p. 4.
23. Thisobjection wasapparently firstraisedbyGaunilo. SeePlantinga, 1965,
at p. 11.

Existence/ 103
24. "Anselmand Actuality," Nous,4 (1970):pp. 175-188.

25. Lewisisactually concerned withan alternative formulation ofAnselm's
OntologicalArgument, andmorespecifically withtheweakerpremise
thatsome possibleindividual is actuallydivine,in the particular
Anselmesque senseof'divine'. Hisarguments, though,extendstraightfor-
wardlytopremise (2a),whether 'divine'is understood intheAnselmes-
que senseor insomeothersense(suchas theCartesian), as wellas to
premise(3a).
26. Contraryto PetervanInwagen'sinterpretation oftheindexical theory
in"Indexicality
of'actuality' andActuality," ThePhilosophical Review,
89,3 (July1980):403-426, at p. 409 (andapparently contrary toLewis,
1970,at thefinalfootnote). Itshouldalso be remembered thatLewis
admitsa secondary, nonindexical senseof'actual'.
27. van Inwagen,1980,pp. 413-416.
28. See my"TenseandSingular Propositions," in J.Almog,J.Perry, and
H. Wettstein, eds.,ThemesFromDavidKaplan(forthcoming, 1987).
29. Thenotion ofthecharacter ofan expression introduced inDavidKaplan,
"On theLogicof Demonstratives," JournalofPhilosophical Logic,8
(1978):pp.81-98(alsoinP. French, T. Uehling, andH. Wettstein, eds.,
Contemporary Persecptives inthePhilosophy ofLanguage, Minneapolis:
UniversityofMinnesota Press,1979,pp.401-412, at p. 409)is defined
(roughly)as thefunction thatassigns toanycontext c theintension that
theexpression takeson withrespectto c. In thisterminology, an in-
dexicalexpression is one whosecharacter is nota constant function.
30. See Kaplan, 1978, especiallyat p. 408 of French,Uehling,and
Wettstein, 1979.
31. See forexampleAdams,1974,at pp. 195-199ofLoux,1979;and van
Inwagen,1980,at pp. 416-417.
32. See Kaplan,1978,andAllenHazen,"OneoftheTruths AboutActuali-
ty,"Analysis, 39, 1 (January 1979),pp. 1-3.
33. Cf.Kaplan,1978,recursive definition 10 at p. 407 ofFrench,Uehling
and Wettstein, 1979.
Inspeaking ofthe"extensional semantics" governing an expression,
I meanthesemantics oftheexpression atthelevelofextension (singular
termreference, sentence truth-value, predicate application), ratherthan
at somehigherlevel,suchas thelevelofcontentor proposition ex-
pressed.Formoreonthenotion ofdifferent "levels"ofsemantic values,
see myFrege'sPuzzle(Cambridge, Mass.:MITPress/Bradford Books,
1986),chapter2.
34. See Adams,1974,at pp.201-202ofLoux,1979;Plantinga, 1974,at pp.
48-51;van Inwagen,1980,at pp. 423-425.
35. Adams,1974,at pp. 194-195ofLoux,1979.
36. Cf.Lewis,"Postscripts to'Anselm andActuality'," Postscript B,inLewis,
1983,at p. 22.
37. Cf.Hazen,1979;and Lewis,ibid.
38. Hereis a simplequizquestion: On theindexicaltheory, is theexpres-
sion'theactualworld'a (modally) rigiddesignator?
39. An intra-world property is onethatsomething has,or lacks,in a pos-

104 / NathanSalmon
sibleworld(e.g.,beinga nativeCalifornian), whereassuchrelativiza-

tiontopossibleworlds is unnecessaryorsuperfluous inconnection with
extra-worldproperties (e.g.,beinga native-Californian-in-world-w).
For
moreon thedistinction between intra-worldandextra-world attributes,
see Salmon,1981,section13.2,pp. 118-120.
40. Nevertheless,I believeitis deniedbyLewis.Thisis duetothefactthat
Lewisdoesnotendorse theabstract-entity
conception ofpossibleworlds,
on whichactuality inthesecondary sensereducesto suchproperties
as thatofbeingtrueorthatofobtaining. Instead,Lewisadoptsa con-
crete or physicalistic conceptionof possibleworldsas maximal,
spatiotemporally self-contained,causallyisolatedphysicalsystems, on
whichactuality intheprimary sensereducestosomething liketheon-
tologically
unimportant property ofbeingpartofa particular maximal
causallyandspatiotemporally isolatedphysicalsystem andnotanother,
andactuality inthesecondary sense,ifitreducesto anything, reduces
totheequallyunimportant (inthepresent context)
binary relationbet-
weena partofsucha physical system andthesystem ofwhichitis a
part.Themaximal causallyandspatiotemporally isolated
physical system
ofwhichwe are a partis,froman objectivepointofview,no more
specialontologicallythananyothersuchphysical systemsthatmayexist.
41. Lewis,1970,at pp. 187-188.Ironically, justone page earlier(at pp.
186-187)Lewiscautions againsta common confusionthatisveryclose-
ly relatedto thesortofconfusion exhibitedin thequotedpassage.
42. OnthePlurality ofWorlds (Oxford: BasilBlackwell,1986),atpp.93-94.
43. Adams,1974,at pp. 193-194ofLoux,1979.
44. "De Essentia,"in E. Sosa, ed., Essayson thePhilosophyofRoderickM.
Chisholm (Rodopoi,Amsterdam, 1979):pp.101-121, at p. 109;and"On
Existentialism," Studies,44 (1983):PP. 1-20,at p. 11.
Philosophical
45. Cf.KitFine,"Plantingaon theReduction ofPossibilist
Discourse," and
JohnPollock,"Plantinga on PossibleWorlds,"inJ.Tomberlin and P.
van Inwagen,eds.,AlvinPlantinga (Dordrecht: D. Reidel,1985):pp.
121-186,at pp. 164-171and pp. 126-129,respectively.
46. See forexampleRobert Adams,"TimeandThisness," inP. French,T.
Uehling,and H. Wettstein,
eds.,MidwestStudiesinPhilosophyXI:Studies
inEssentialism (Minneapolis:UniversityofMnnesotaPress,1986),pp.
315-329.
47. In his"Repliesto myColleagues," sectionII.B,in Tomberlin and van
Inwagen, 1985,pp.316-323, Plantingaattempts a response toPollock's
denialofso-calledseriousactualism.SomeofPlantinga's arguments for
so-called
seriousactualismbegthequestion bycritically
relying
(pp.319,
322)on classicalexistential
generalization. Also,in defendinghimself
againstPollock'schargeoffallaciousmodalreasoning Plantingaappears
(atp. 319,first
complete paragraph) tocommit theveryfallacyPollock
attributesto him.(Specifically,he appearsto inferthe falsehood
'Necessarily,
everythingisnecessarilysuchifitexemplifiesnonexistence
thenitexists'fromthetruth 'Necessarily,everythingis suchthatifit
exemplifies nonexistence thenit exists')
48. Similarly,we also havesuchtemporal versions as FEverypresentor

Existence/ 105
pastindividual is (P.a is a presentor pastindividual. Therefore, a is

(PI and ra is<4a isa future individual.Therefore,somefuture individual
is (P , etc.(Moreaccurateversions oftheseruleswouldincludean ad-
ditional premise requiring theinter-substitutabilityofa andthevariable
ofgeneralization underanyassignment ofa valuetothevariableunder
whichit and a are co-referential.)
49. Quine,op. cit.,at p. 4; AlexiusMeinong, "TheTheoryofObjects,"in
R. Chisholm, ed.,RealismandtheBackground ofPhenomenology (New
York:FreePress,1960):pp. 76-117,at p. 82.
50. Contrary to MonteCook,"NamesandPossibleObjects," Philosophical
Quarterly, 35,140(July 1985):pp.303-310,atp. 309.See Kaplan,1973,
at pp. 506 and 517,note19. I once foundtheseclaimsbaffling. Cf.
Salmon,1981,p. 39n.I wasconfused. Onceitis admitted thatclassical
UI and EG are fallacious, andthatan additional existential
premiseis
all thatis requiredin each case to correctthefallacy, whatonceap-
pearedutterly mysterious becomesperfectly clearandstraightforward.
The claimthat'Noman'refersto Nomanand yetdoes notreferto
anything, properly understood, isreallynomorebaffling thantheclaim
that'Shakespeare' refers toShakespeare, whoislongdead.Whenrefer-
ringtomerely possible individuals,itissomewhat morenatural (although
by no meansmandatory) to allowone'squantifiers to go possibilist,
thereby preserving theform ofclassicalUIandEG.Likewise, whenrefer-
ringto pastorfuture individuals, itis naturaltoallowone'squantifers
to rangeoverall pastor all future individuals.
51. See Kaplan,1973,appendixXI,at pp. 505-508.
52. See Salmon,1986afora defenseofsingular propositions as thecon-
tentsofsentencescontaining propernames.
53. See Salmon,1986a,at p. 4-6,fora defense oftheclaimthattheobjects
ofde re propositional attitudes are singular propositions.
54. Thefirst ofthesepropositions maybe spelledoutmorefully as follows:
Theheight thatNothan wouldhavehadifhehadexistedisgreater than
theheight thatI actually, have.Thesecondproposition maybe regard-
ed as thenegationofthefirst. See note2 above.
55. A simplerexampleofan impossible objectthathas properties is the
pairset {Nothan,Nathan},i.e.,thesetthata possibleindividual is an
element ofifandonlyifthatpossibleindividual is eitherNothan orme.
Thisimpossible set has suchproperties as itsmembership, notbeing
empty, beingfinite, and so on,all ofwhichare perfectly compatible
withone another.The term'{Nothan,Nathan}'maybe regardedas
a strongly nonreferring termthatis notverystrongly nonreferring; it
doesnotrefer toanyexisting ormerely possiblething, yetitdoesrefer
toan impossible thing.Similar remarks maybe madeinconnection with
the 'that'-clause 'thatNothanwouldhave been tallerthanNathan
actually,is'.
Hereis a not-so-simple
quizproblem:Finda wayto makediscourse
involving overimpossible
quantification objectspossibilistically
accep-
table, by defining,analyzing,or somehow reconstructing the
superunrestricted
impossibilist possibleand every
quantifiers-'every

106 / NathanSalmon
impossibleindividual'and 'some possibleor some impossiblein-

dividual'-intermsof thepossibilist
quantifiersand standardmodal
(See note10 above.)Ifthiscannotbe done,howare we to
operators.
understandtheclaimthatitis true(orI assert,or Saul Kripkedenies)
ofNothanandmethathe wouldhavebeentallerthanI actually, am?
Whatis itthatis true(asserted,
denied)?
References
Adams,R. M.,"Theories ofActuality," Nous,8 (1974):pp.211-231;also in

Loux,pp. 190-209.
Adams,R. M.,"Actualism and Thisness," Synthese, 49 (1981):pp. 3-41.
Adams,R. M., "Timeand Thisness," in P. French,T. Uehling,and H.
; Wettstein, 1986,pp. 315-329.
Cook,M.,"Namesand PossibleObjects," Philosophical Quarterly,35, 140
(July1985):pp. 303-310.
Fine,K."Prior ontheConstruction ofPossible Worlds andInstants,"postscript
toA. N.PriorandK. Fine,Worlds, TimesandSelves(Amherst: Univer-
sityofMassachusetts Press,1977),pp. 116-161.
Fine,K.,"PlantingaontheReduction ofPossibilistDiscourse,"inJ.Tomberlin
and P. van Inwagen,1985,pp. 121-186.
Frege,G. "Function undBegriff'(1891),in Frege'sCollectedPaperson
Mathematics, Logic,andPhilosophy, ed.,B. McGuinness,trans.,M.Black,
V. H. Dudman,P. Geach,H. Kaal,E.-H.W. Kluge,B. McGuiness, and
R. H. Stoothoff (Oxford:BasilBlackwell, 1984):pp. 137-156.
Frege,G.,Die Grundlagen derArithmetik, trans,J.L. Austin (Evanston, Ill.:
Northwestern University Press,1968).
French, P.,T. Uehling,andH. Wettstein, eds.,Midwest Studies inPhilosophy
XI StudiesinEssentialism (Minneapolis, Minn.:University ofMinnesota
Press,1986).
Gaunilo,"On BehalfoftheFool,"in A Plantinga, ed., 1965,pp. 6-13.
Hazen,A., "One oftheTruths AboutActuality," Analysis,39, 1 (January
1979):pp. 1-3.
Kaplan,D., "Boband Caroland Ted andAlice,"in K.J.J. Hintikka,J.M.E.
Moravcsik, and P. Suppes,eds.,Approachesto NaturalLanguage
(Dordrecht: D. Reidel,1973):pp. 490-518.
Kaplan,D., "Howto Russella Frege-Church," TheJournal ofPhilosophy,
72 (November 6, 1975):pp. 716-729;also in Loux,pp. 210-224.
Kaplan,D.,"OntheLogicofDemonstratives," JournalofPhilosophicalLogic,
8 (1978):pp.81-98(alsoinP. French, T. Uehling,andH. Wettstein, eds.,
Contemporary PerspectivesinthePhilosophy ofLanguage, Minneapolis:
University ofMinnesota Press,1979,pp. 401-412).
Kripke,S., Namingand Necessity(HarvardUniversity Pressand Basil
Blackwell, 1972,1980).
Kripke, S.,"Speaker'sReference andSemantic Reference,"inP. French,r.
Uehling,and H. Wettstein, eds., Contemporary Perspectives in the
Philosophy ofLanguage(Minneapolis: University ofMinnesota Press

Existence/ 107
1979),pp. 6-27.
Lewis,D.,"Anselm andActuality," Nous,4 (1970):pp.175-188; alsoinLewis,
1983,pp. 10-20.
Lewis,D. "Postscriptsto 'Anselm andActuality',"inLewis,1983,pp.21-25.
Lewis,D., Philosophical PapersI (Oxford UniversityPress,1983).
Lewis,D., On thePlurality of Worlds(Oxford: BasilBlackwell, 1986).
Loux,M.,ed.,ThePossibleandtheActual(Ithaca:CornellUniversity Press,
1979).
Meinong, A.,"TheTheoryofObjects," inR. Chisholm ed.,Realismandthe
Background ofPhenomenology (Atascadero: Ridgeview PublishingCo.,
1960):pp. 76-117.
Parsons,C., "Objectsand Logic,"TheMonist, LXV(1982):pp. 491-516.
Plantinga,A.,ed.,TheOntological Argument (Garden City:Doubledayand
Company, Inc.,1965).
Plantinga, A.,TheNatureofNecessity (OxfordUniversity Press,1974).
Plantinga, A.,"De Essentia," in E. Sosa,ed.,Essayson thePhilosophy of
Roderick M. Chisholm (Rodopoi,Amsterdam, 1979):pp. 101-121.
Plantinga,A.,"OnExistentialism,"PhilosophicalStudies, 44 (1983):pp.1-20.
Plantinga, A., "Repliesto my Colleagues,"in J. Tomberlin and P. van
Inwagen,eds., 1985,pp. 313-396.
Pollock,J."Plantinga on PossibleWorlds," in J.Tomberlin and P. van
Inwagen,1985,pp. 126-129.
Quine,W. V. O., "On WhatThereIs,"in Quine'sFroma LogicalPointof
View(NewYork:Harper& Row,1961),pp. 1-19.
Quine,W. V. O.,Methods ofLogic(NewYork:Holt,Rinehart andWinston,
Inc.,1972).
Quine,W. V. O., "Mr.Strawson on LogicalTheory," in Quine'sTheWays
ofParadox(NewYork:RandamHouse,1966),pp. 135-155.
Russell,
B.,A CriticalExpositionofthePhilosophy ofLeibniz(London: George
Allenand Unwin,1971).
Russell,B., "On Denoting," in Russell,1956,pp. 39-56.
B.,Introduction
Russell, toMathematical Philosophy(LondonGeorgeAllen
and Unwin,1953).
Russell,B.,LogicandKnowledge, ed.,R. C. Marsh(London:GeorgeAllen
and Unwin,1956).
Russell,
B.,"ThePhilosophy ofLogicalAtomism," inRussell's
ThePhilosophy
ofLogicalAtomism, D. Pears,ed.(LaSalle:OpenCourt, 1985):pp.35-155;
also in Russell,1956,pp. 175-281.
Salmon,N.,Reference and Essence(Princeton University Pressand Basil
Blackwell,1981).
Salmon,N.,Frege'sPuzzle(Cambridge, Mass,:TheMITPress/Bradford Books,
1986).
Salmon,N.,"ModalParadox:Partsand Counterparts, Pointsand Counter-
points,"in French,Uehling, andWettstein, 1986,pp. 75-120.
Salmon,N.,"TenseandSingularPropositions," in J.Almog,J.Perry, and
H. Wettstein, eds.,ThemesFromDavidKaplan(forthcoming, 1987).
Salmon,N., "Reviewof Lewis,1986"The Philosophical Review(forth-
coming1987).

108 / NathanSalmon
Salmon,N.,"The LogicofWhatMightHave Been,"unpublished.

Schopenhauer,A., The FourfoldRoot of thePrincipleofSufficient
Reason,
trans.,Mme.KarlHillebrand (London: GeorgeBellandSons,1889);ex-
cerptin Plantinga, 1965,pp. 65-67.
Tomberlin,J.E.,"IsBeliefinGodJustified?" ofPhilosophy,
Journal 67(1970):
pp. 31-38.
Tomberlin,J.E.,"Plantinga andtheOntologicalArgument,"inJ.Tomberlin
and P. van Inwagen,1985,pp. 257-270.
Tomberlin,J.andP. vanInwagen, eds.,AlvinPlantinga D. Reidel,
(Dordrecht:
1985).
van Inwagen,P., "IndexicalityandActuality,"TIhePhilosophical
Review,
89, 3 (July1980):403-426.
Whitehead, A. N.,and B. Russell,PrincipiaMathematica,secondedition
(Cambridge University Press,1963).


Salmon 1987

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Salmon 1987

Uploaded by

Copyright:

Available Formats

Ridgeview Publishing Company

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

The questionsI shalladdressherecan be approachedthroughthe

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

who would have developed fromtheunionofS and E, ifS had fer-

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

to be the case? Ifwe can answerthese two questionssatisfactorily,

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

thatforsomethingto existis forit to be in the range of a function

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

tainvariables;hence,theassignments s' mayalso assignindividuals

Philosopherswho address the questionsof what it is foran in-

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

classicalOntologicalArgumentforGod's existence.Indeed,the On-

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

theybe expected to acquiesce in the assertionthatthe divinein-

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

is thatthequantifiers ofEnglishare typically actualist(andpresentist,

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

The actualistuniversalquantifier'everything'remainsa correct

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

would notexceed itsown possiblemagnitudeof greatness.(Gotit?)

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

thispiece of sophistry.That the OntologicalArguement(takenas

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

functionis true.. . . Existenceis essentiallya propertyof a

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

Schopenhauergave expressionto a verycommonreactionto the

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

suchas Socratesor Russell.Suchproperties as theseare nowadays

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

existenceas a necessary condition. We havetheconcept,forexam-

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

worthone pennymore than a hundredmerelypossibledollars.If

Version3 oftheOntologicalArgument is unscathedby Kant'salleg-

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

actualists,and mayobfuscatetheevaluationoftheargumentas valid

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

theonlyconceptualtruths thatfollowdirectly fromconceptsor defini-

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

exists and is a such-and-such.

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

obtainedan analyticexistential;we would have definedsomething

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

The factthatthe firstpremiseofthe OntologicalArgumentis not

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

withoutassumingthatone actuallyexists.What I am pointingout

In a penetratingcritiqueoftheOntologicalArgument, David Lewis

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

notdemandproof; I] wishto knowiftheontologicalarguerhas

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

[theactual] world alone is actual, therefore[theactual] world

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

possibleindividual is saidto be actuaP(oran actualindividual) ifit

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

set of additionalsemanticparametersdoes not vary withthe con-

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

c) to act as theneededextraparameters.30 In thegeneralcase, if

This content downloaded from 140.254.87.103 on Sun, 28 Apr 2013 16:07:15 PM

worldw' accessibleto theworldofc suchthattheU.S.president