Philosophical Review Freedom and Foreknowledge Author(s): John Martin Fischer Source: The Philosophical Review, Vol. 92, No. 1 (Jan., 1983), pp. 67-79 Published by: Duke University Presson behalf of Philosophical Review Stable URL: http://www.jstor.org/stable/2184522 . Accessed: 24/02/2011 13:15 Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use. Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at . http://www.jstor.org/action/showPublisher?publisherCode=duke .. Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. Duke University Press and Philosophical Revieware collaborating with JSTOR to digitize, preserve and extend access to The Philosophical Review. http://www.jstor.org
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Freedom and ForeknowledgeAuthor(s): John Martin FischerSource: The Philosophical Review, Vol. 92, No. 1 (Jan., 1983), pp. 67-79Published by: Duke University Press on behalf of Philosophical ReviewStable URL: http://www.jstor.org/stable/2184522 .
Accessed: 24/02/2011 13:15
Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless
you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you
may use content in the JSTOR archive only for your personal, non-commercial use.
Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at .http://www.jstor.org/action/showPublisher?publisherCode=duke. .
Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed
page of such transmission.
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 formsof scholarship. For more information about JSTOR, please contact [email protected].
Duke University Press and Philosophical Review are collaborating with JSTOR to digitize, preserve and extend
A owerfulrgumentanbe madethatGod'somnisciencesincompatiblewithhuman freedom.' If God is eternal and
omniscient, hen itmightseem thatmyfreedomnow to do otherthanwhat 'm doingmustbe thefreedom o to actthat fact bout
thepast God's priorbeliefaboutmypresent ctivity) ouldn'tbe afact bout thepast.But sincethepast s "fixed," tseems that fGodexists,then I am now not free to do other thanwhatI'm doing.
Many philosophers have been attracted to an Ockhamistre-sponse to this argument.2Both the Ockhamistand the incom-patibilist an distinguish etween "hard" and "soft"facts bout the
Review74 (January 1965), pp. 27-46; "Of God and Freedom: A Re-joinder," Philosophical eview75 (July1966), pp. 369-379; and "DivineForeknowledge,Human Freedom and PossibleWorlds,"Philosophical e-view 86 (April 1977), pp. 209-216. Pike also discusses the same basicargument n the fourthchapter of his book, God and Timelessness,NewYork: Schocken Books, Inc., 1970), pp. 52-86.
2Some examples are: MarilynAdams, "Is theExistenceofGod a 'Hard'Fact?"Philosophicaleview 6 (October 1967),pp. 492-503; and WilliamL.Rowe,PhilosophyfReligion,Encino, Dickenson, 1978), pp. 154-169.
The approach sketchedbelow is called "Ockhamist"because WilliamofOckhamdistinguished etweenpropositions bout thepastwhich re nec-essary nd those whichare notand argued that mong thosepropositionsabout the pastwhichare notnownecessary re certainpropositions boutGod. (WilliamOckham,Predestination,od'sForeknowledge,ndFutureCon-tingents. arilynMcCord Adams and Norman Kretzmann trans.), NewYork: Appleton-Century-Crofts,969), pp. 46-47; 92.) Roughly,Ockhamclaimsthatthose propositionsabout the past whichare trueby virtueofcontingentfuture events are not now necessary. Such propositions, tmight e said,express "softfacts" bout thepast. A usefuldiscussion fthe
Ockhamist approach can be found in: ArthurPrior,Past, Present, ndFuture,Oxford: Clarendon Press, 1967), pp. 121-127.John Turk Saunders agrees withthe Ockhamist thatcertainproposi-
tions about God express softfacts:John Turk Saunders, "Of God andFreedom," Philosophical eview 74 (April 1966), pp. 219-225. Saundersholds a positionwhich s even stronger hanOckhamism, incehe believesthatneither oftfactsnor hard factsneed be fixed:JohnTurk Saunders,"The Temptations of Powerlessness,"American hilosophical uarterly(April 1965), pp. 104- 107.
past; the hard facts re fixedwhilethe softfactsneed notbe fixed.But theOckhamistclaimsthatGod's priorbeliefaboutmypresent
activity s a softfactabout the past and hence notfixed;myfree-dom is thuspreserved.Some Ockhamists ven claimthat theveryexistence of God is also a softfact about the past.
I shallargue that very ttractive resentation ftheOckhamistapproach, one explicitly ormulatedby MarilynAdams, is inade-quate.3 There are significantproblemswithAdams' attempt tocharacterize hehard fact/softactdistinction. urther, shallpre-
Nelson Pike claimsto exhibit heincompatibilityfhuman free-dom and divine foreknowledge,relative to certainplausible as-sumptions bout God's
nature.4These assumptionsreflect entralfeaturesof the standardJudeo-Christian onception of God. Pikeexplicitly dopts the assumptionthat if God exists,then God isessentially mniscient nd God iseternal.On Pike'saccount,God isomniscient f and only ifGod believes all and onlytrueproposi-tions, and we mightsay thatGod is essentially mniscient fandonly fGod isomniscientn all possible worlds nwhichGod exists.Pike saysthat God is eternal fand only fGod has alwaysexisted
and alwayswill.5FollowingPike'spresentation n a differentrticle, assumethat
theterm"God" is a descriptive xpressionused tomark a certainrole,ratherthan a proper name.6 Whoever occupies the role ofGod is omniscient, mnipotent, ternal, etc. In contrast, he term"Yahweh" is a proper name; it refersto the person who actuallyoccupies the role of God (if God exists). It is not necessarily rue
3Adams,op. cit.4Pike,"Divine Omniscience,"pp. 26-31.5Thus Pike conceivesofGod's eternality s sempiternality-existence t
all times.This conception s sharedbytheOckhamist; tcan be contrastedwith heatemporalconceptionofeternality eldbyBoethius and Aquinas.
6Pikemakesthis ssumptionexplicit n "Omnipotenceand God's Abilityto Sin," American hilosophical uarterly (1969), pp. 208-216, esp. pp.208-209.
thatYahweh is omniscient, mnipotent, ternal,etc.; itis logicallypossible thatsome otherperson has been God.7
Since "God" isbeingused here as a nonrigiddesignator, here ssome ambiguityn theassumptions bout God's attributes.God isessentially mniscient" oes notmean that hepersonwho is infactGod is essentially mniscient, utrather, hatnecessarily,whoeverisGod is omniscient. n terms f possibleworlds,God is essentiallyomniscientust in case forany possibleworld in whichthereis apersonwho isGod, thatpersonisomniscient.One can assume that
ifGod is eternalin a particularworld,thenitfollows hatthere sone and thesame personwho isGod at all times n thatworld.Pikeneed notaccept thisparticular ssumption, s it s notcrucial tohisargument.)
Though this s theapproach totheterm"God" thatPikeappearstoadopt, itmight eem tobe an unusual and unappealing position.I shall followPike in adopting this nterpretation,ut itis impor-tant to note-thatPike could just as easily embrace the strongerinterpretationccordingtowhichthepersonwho is in factGod isessentiallyGod. Nothing in Pike's proof, or in my criticism fAdams' Ockhamism,restson adoptingthe weakerrather hanthestronger nterpretation f God's attributes.
In effect,Pike also appears to adopt whatmightbe called the"fixedpast" constraint n power attributions:
(FPC) It isnevernanyperson's ower t a timeT so toactthat hepast (relative to T) would have been different romwhat itactuallywas.
Pike's view about the fixity f the past impliesnot onlythatonecannotcausally nfluencethepast; it mpliesthatno personis freeto do somethingwhich s suchthat,werehe todo it,thepastwouldhave been different romwhatit actuallywas.
Pike's argument s essentially s follows.Suppose Jonesdid X attime T2 and God exists. Since God exists, t follows fromGod'seternalityhat He existed at T1 (a timepriorto T2). Let us call thepersonwho was God at T1, "Y." SinceJonesdid X at Ti2,tfollows
7C. B. Martin argues for this approach in the fourthchapter of hisReligious elief, Ithaca: Cornell Press, 1964).
fromGod's omniscience hatHe believedatT1 thatJoneswoulddoX at T2. Now if it was withinJones' power at T2 to refrainfrom
doing X, then (1) itwas in Jones' power at T2 to act in such a waythatY would have been God and would have held a falsebeliefat
TI, or (2) it was inJones' power at T2 to act in such a waythatYwould have been God but wouldn'thave held thebeliefHe held at
TI or (3) it was inJones' power at T2 to act in such a waythatYwouldn'thave been God at T1.
But (1) is ruled out byGod's essentialomniscience, nd (2) and
(3) are ruledoutby FPC). Hence itwasnot nJones'poweratT2torefrainfromdoing X. If the argumentis sound, it can easilybegeneralizedto showthatGod's eternalitynd essentialomniscienceare incompatiblewithany human agent'sbeing freeat any time.
It should be pointedout that ncompatibilismbout divinefore-knowledge and human freedom needn't entail incompatibilismabouthuman foreknowledge nd human freedom.The problem sdeeper with-divine oreknowledgebecause of God's essentialom-niscience;perhaps it was inJones' powerat T2 so to act that Smith(who actuallyheld only correctbeliefs)would have held a falsebelief t T1. Pike wantsto insist n an asymmetryetweendivine andhuman foreknowledge.8
II. HARD AND SOFT FACTS
It is sometimes n one's power so to act thatfacts bout thepastwouldn't e facts.JohnTurk Saunders discussessuch a fact:
Althought strue hat f hadrefrainedromwritinghispaper n1965,Caesar's ssassinationouldhavebeenother han t s nthatt
8Pike says in his original paper: "The important hing to be learnedfrom hestudyof Smith'sforeknowledge f Jones'action s thattheprob-lemofdivineforeknowledgehas as one of its pillarstheclaimthattruth sanalyticallyonnectedwithGod's beliefs. o problemof determinismriseswhen dealing withhuman foreknowledgeof futureactions.This is be-cause truth s not analyticallyonnectedwithhuman beliefeven when asin thecase of humanknowledge)truthscontingentlyonjoinedtobelief."Pike, "Divine Omniscience,"p. 43. Thus it is clear thatPike as wellas theOckhamistneeds he distinction etweenhard and softfacts.
wouldnothave precededby 2009 yearsmywritinghispaper, twouldbeabsurd oargue hat thereforeidnothave t n my ower
to refrain romwritinghispaper n 1965.9
It isobviousthatthemere fact hat f Saundershad refrained romwritinghis paper, then Caesar's assassinationwouldn't have pre-ceded Saunders' writinghis paper by 2009 yearsdid not renderSaunders incapable of refraining;relativeto 1965, "Caesar died2009 years prior to Saunders' writinghis paper" expresses a softfact bout thepast. Of course, it was not in Saunders' power so toact that Caesar would not have died on the steps of the Senate.Relativeto Saunders' lifetime,he fact hatCaesar died on thestepsof the Senate is a hard factabout the past.
Pike agreeswith heOckhamist hatthereare bothhardand softfacts bout thepast.'0 It is noteasy to providea precisecharacter-izationofthe hard fact/softactdistinction. ikehimeselfprovidesno suchaccount,thoughhe claims wecan recognizeclear examples
ofeach sort. I1The disagreementbetweenPike and theOckhamistis about where to draw the line. Pike's position s that ftheordi-narynotionsof beliefand existence re applied to God, thenGod'sbelief t T1 and God's existenceat T1 (includingthefact hatY wasGod at T1) are hard facts bout thepastrelative oT2. And iftheyweresoftfacts bout the pastrelative o T2,thiswould show thatwewere ascribingbeliefsand existenceto God in a special,nonstan-
dard way.Given the hard fact/softact distinction, he appropriate in-
terpretation f Pike's claim about the fixity f the past should bemade explicit:
(FPC*) It is nevernanyperson's ower t a timeT soto actthat ny
9Saunders,op. cit.,p. 224. Unfortunately, aunders' arithmetics wrongsince there s no 0 B.C. or 0 A.D. Hence, Caesar's death preceded Saun-ders' writing is paper by 2008 years For simplicity'sake, however, shallignore thisand proceed withSaunders, Pike, and Adams in adding a yearto history.
10Pike,"Of God and Freedom," pp. 369-370; Rowe makes a similardistinction etween factswhich re "simply bout the past" and factswhichare not, in WilliamRowe, op. cit.,pp. 162-165.
MarilynAdams presents n accountofthedistinctionwhichshe
believessupportscompatibilismgainstPike'sattack. t willbe use-
fultoconsiderMarilynAdams' attempt t giving n accountofthe
distinction:
(B) "Statement is at least n = df. "The happening r nothap-part bouttimeT' pening,actuality r nonac-
tualityf somethingt T is anecessaryconditionof thetruth f P."
l2Put in termsof possibleworlds,the fixedpast constraints:
(FPC*) A possibleworldW* (in whichan agentdoes other thanwhathe does in W at T) can establishthatthe agent had it in his
power at T in W to do otherwiseonly fW and W* have thesame hard facts bout thepast relative o T.
In "Pike on PossibleWorlds,Divine Foreknowledgeand Human Free-dom,"Philosophical eview88 (July1979), pp. 433-442, Joshua Hoffmancriticizes ike's interpretationf thefixityf thepast.HoffmanconstruesPike as claimingthatthepossessionof everypowerentails heoccurrenceor nonoccurrenceof past circumstances.That is, Hoffmanattributes oPiketheclaimthatthetruth f a statement scribing particularpowertoan agent at a time in a world W entails hat the past be as it is in W
(Hoffman,pp. 441-442).Pike himselfputs the constraint n a misleadingway, saying: "If weassumethatwhat swithinmypowerata givenmomentdetermines setofpossibleworlds,all of the membersof thatsetwillhave to be worlds nwhich what has happened in the past relativeto the given momentispreciselywhat has happened in the past relativeto thatmoment n theactualworld."Pike,"DivineForeknowledge,"p. 215. But nothing nPike'spositionrequires acceptance of the radical doctrineattributed o himbyHoffman.Pike's fixedpast constraint ommitshimto theclaimthat fanagentperforms n act inworldW,thenanypossibleworldW*inwhichhe
refrains romperforming heactmusthave the same pastas W, fW* s toestablishthatthe agent can in W refrainfromperforming he act. Buttheremaybe possibleworlds includingW) in which he agentcan performthe act and doesperform heact) inwhichthe pasthistoriesrelative othetimeof the act) are all different romone another; hence, the truth f apower-ascription eed not entailthe past history.Hoffman'scriticism fPike missesthemarkand leaves the fixedpast constraint nscathed.
Thusthe tatement,Caesardied2009 years efore aunderswrotehispaper" s at east npart bout44 B.C.,sinceCaesar's eath tthat
time s a necessaryonditionfthe ruthfthat tatement.t s alsoatleastnpart bout1965A.D.since aunders'writingispaper n1965A.D. is also a necessaryondition fthe ruthfthat tatement.iven(B) thenotion f a "hard"factmaybe explained s follows.
(C) "Statement expresses = df. "P isnot t least npart bout'hard'fact bout time " any time future relativeto
T. ts13
Adams uses this account to present an Ockhamist response toPike's argument. On her account, God's belief at T, and the factthatY was God at T1 are deemed soft facts bout T,.
Adams claims that her account shows why "Caesar died 2009yearsbefore Saunders wrotehis paper" does not express a hardfact bout 44 B.C. But her account does not xplain thisunlessit s
interpreted o imply hatno sentence expresses a hard fact.Adamssaysthat"Caesar died 2009 yearsbefore Saunders wrotehis paper"is at least in part about 1965, since Saunders' writinghis paper in1965 isa necessary onditionof the truth f that tatement. ut thisseems plainly false; the statement ntailsthat Caesar's death andSaunders' writing ispaper be separated by 2009 years,but tdoesnot entail any two particular dates for the two events. The state-
mententailsthatthe two events tand n a certain emporalrelation,but itdoes notentail that theyoccur on any specificdates. Hence,Saunders' writinghis paper in 1965 is not necessary onditionofCaesar's death being 2009 years priorto Saunders' writing is pa-per,ifwe interpret Q is a necessary onditionforP" as "P entails
Q"~
One might reply that since it is true that Saunders wrotehispaper in 1965, "Saunders wrote his paper in 1965" is materiallyimplied y "Caesar died 2009 years prior to Saunders' writinghispaper." So ifwe interpret Q is a necessary onditionforP" as "Pmaterially mplies Q," Saunders' writinghis paper in 1965 is a
dition of "Caesar died 2009 years prior to Saunders' writinghispaper." Also, it is perhaps reasonable to say (thoughI'm notsure)
that f Oswald had not shotKennedy n 1963, then Kennedywouldnot have been assassinated. f this s so, then Adams could saythat"Oswald shotKennedy in 1963" is a necessary onditionof "JohnF. Kennedywas assassinated."Similarly, fJones hadn'tdone X at
T2,then twould have been false thateitherSmithknew at T1 thatJoneswould do X at T2orJonesbelievedat T1 thatJoneswoulddoX at T2. Thus, Adams could say that "Jones did X at T2" is a
necessarycondition of the disjunction.But there s anothersortofproblemwhich fflicts othplausible
accounts-both the counterfactual nd entailment nterpretationsof"necessary ondition."Suppose "Smith xistedatT1" istrue. t isa necessarycondition of the truthof this statement on both thecounterfactual nd entailment ccounts)that t s not the case thatSmithexisted for the first ime at T2. It is obvious that Smith's
existing t T1- ntailsthat he doesn't existfor the first ime atT2*
And if Smith had existed for thefirst imeat T2,thenhe wouldn'thave existed at T1, so the counterfactual ccount fares no betterthan the entailment ccount. Thus, by (B), the statement Smithexisted at T1" is at least in part about T2; by C) the statementfails
to expressa hard fact bout T1. But sinceSmithneed notbe eternal(or essentiallyomniscient),this is a disastrous resultforAdams'account. The same sort of argumentshows that Adams mustsay
that "Jonesbelieved at T1 thatJoneswould do X at T2" does notexpressa hard fact boutT1. This is because "It isnot the case thatJonesbelievedforthefirst imeatT2 thathe woulddo X atT2" is anecessary onditionof Jones believed at T, thathe would do X at
T2"
Also, it is a necessarycondition on both interpretations)f thetruth f the statement, Piece of salt S dissolvedat T1j," thatS did
notdissolveat T2. One wantstosaythatthis tatement xpressesahard fact bout T , but Adams' account does notcapturethis ntui-tion (since the statement s at least in part about T2).
It is not easy to see how Adams could provide an account of'necessary condition"whichwould avoid all the problemsraisedabove. Without uch an account,she hasn'tpresentedan adequateexplanationof the distinction etween hard and soft facts.
Various contemporaryOckhamistshave argued thaton anyac-ceptable account of the distinctionbetween hard and softfacts,
God's priorbeliefwill be a softfact bout the past. I shallnotherefurther iscussparticular ompatibilistccountsof thedistinction;rather, shall sketch constraint n the accountof the distinctionwhich n incompatibilistmightuse todefeatany ompatibilisthar-
acterization f the distinction. hat is, I shall develop an explana-
tionoftheclaimthatGod's priorbelief sa hardfact bout thepast;this explanationwillnot mplythathumanforeknowledges also ahard factabout the past. This mightprovide a wayin whichPikecould defend both his incompatibilitylaim and the asymmetrythesis-the thesisthat God's foreknowledgeundermineshumanfreedom n a wayin whichhuman foreknowledgedoes not.
Consider thefactthatCaesar died 2009 years priortoSaunders'writinghis paper. What lies behind our viewthat this fact s not ahard fact bout44 B.C.? We might aythat t s a softfact bout44
B.C. because one and thesame physicalprocesswould havecount-ed as Caesar's dying2009 yearspriorto Saunders' writinghis pa-per, if Saunders wrote his paper in 1965, and would nothavecountedas Caesar's dying2009 yearspriorto Saunders'writing ispaper, if Saunders hadn'twritten is paper in 1965. This capturesthe "futuredependence" of softfacts; softfact s a fact n virtuef
eventswhichoccur in the future.Similarly,uppose that Smithknew at T1 thatJoneswould do X
atT2. Smith'sknowledge sa softfact boutT1 because one and thesame stateof Smith'smind (at T1) would count as knowledgeifJonesdidX atT2,and wouldnotcountas knowledge fJonesdidn'tdo X at T2. Exactlythe same sort of futuredependence explainswhyboth facts-the fact about Caesar's death and the fact about
Smith's knowledge-are soft facts.Thus, an incompatibilistmightinsist on the following ortofconstraint n an accountof the hard fact/softactdistinction: heonlywayin whichGod's beliefat T1 about Jonesat T2 could be asoftfact bout thepastrelative o T2 wouldbe ifone and thesamestateof themindof thepersonwhowas God at T1 would countasone belief fJonesdid X at T2, but different elief ornot a beliefat all) ifJonesdid not do X at T2. But it is implausibleto supposethatone and thesame stateof themind of the personwhowasGod
at T, would count as different eliefsgivendifferent ehaviorbyJones at T2.
Suppose again thatJonesdid X at T2. Y (beingGod) believed atT, thatJoneswould do X at T2. Let's saythatY'smindwas instateat TI; thisconstitutedHis believingthatJoneswould do X at T2.Now ifY's mind were in state s and Jonesdid notdo X, Ys mindbeing n would still ountas a beliefthatJoneswoulddo X. (In thiscase, Y wouldn'tbe God, since he wouldhave a falsebelief.)Hence,Ys mindbeingins at T, wouldnot ountas one belief fJones did X
at T2 and anotherbelief or not a beliefat all) ifJones did not do Xat T2.Someone mightagree thatthe incompatibilist'sonstraint s ap-
propriate but disagreewithwhat I have said about itsapplication.That is,one might rgue that fJoneshadn'tdone X atT2,thenthestate of God's mind that actuallyconstitutedHis believingthatJones would do X would not have constitutedthatbelief. Thispositionmight be supported by extending Putnam's point thatmeanings and beliefs in't n the head.'4 AccordingtoPutnam,mybeliefthatwater is wet-the stateof my mind thatconstitutesnfact,mybelievingthat-would have been a different elief-thebeliefthatXYZ iswet-if lakes and oceans on earthhad been filledwithXYZ ratherthanwater. On thisapproach, the state of God'smind at T, that counts as His belief thatJones will do X at T2countsas thatbeliefpartly n virtueof the factthatJones does in
fact do X at T2'But thispictureofGod's omniscience shighlymplausible.God's
omniscience would be seriouslyattenuated if the same state ofGod's mind at T, would constitutedifferent eliefs about Jones,depending on Jones' behaviorat T2. The following s a more ap-pealing pictureof God's omniscience.An Ockhamistmightdenythe appropriateness of the constraint, laimingthatwhile it's nottrue that one and
the same stateof God's mind at T, would con-stitutedifferent eliefs,depending on Jones' behavior at T2, it istruethatGod's mindwouldhavebeen ina differenttate tT, (fromtheone itwas actually n), ifJoneshad notdone X at T2.Whereas
'4HilaryPutnam,"The Meaningof Meaning"' reprintednHilaryPut-nam,Mind,Language,andReality, London: CambridgeUniversity ress,1975), pp. 215-271, esp. pp. 223-227. RobertStalnakersuggestedtometheidea for the incompatibilist'sonstraint nd pointed out therelevanceof Putnam'spointto it.