_Gale, Richard - On the Nature and Existence of God.pdf

8/9/2019 _Gale, Richard - On the Nature and Existence of God.pdf

http://slidepdf.com/reader/full/gale-richard-on-the-nature-and-existence-of-godpdf 1/425



There has been in recent years a plethora of defenses of theismfrom analytical philosophers such as Plantinga, Swinburne, andAlston. Richard Gale's important book is a critical response tothese writings. New versions of cosmological, ontological, andreligious-experience arguments are critically evaluated, alongwith pragmatic arguments to justify faith on the grounds of itsprudential or moral benefits. A special feature of the book is thediscussion of the atheological argument, which attempts todeduce a contradiction from the theist 's way of conceiving ofGod's nature. In considering arguments for and against the existence of God, Gale is able to say interesting things about manyimportant philosophical concepts including explanation, t ime,free will, personhood, actuality, and the objectivity of experience.



On the natureand existence

of GodR I C H A R D M . G A L E

DEPARTMENT OF PHILO SOPH Y, U NIVERSITY OF PITTSBURGH

C A M B R I D G EUNIVERSITY PRESS



the Press Syndicate of the University of C am brid geiiilding, Trumpington Street, Cambridge CB2 1RP

40 West 20th Street, New York, NY 10011-4211, USA

10 Stamford Road, Oakleigh, M elbourne 3166, A ustralia© Cam bridge University Press 1991

First published 1991First paperback edition 1993

Reprinted 1996

Library o f Congress Cataloging-in-Publication Data is available.

A catalogue record for this book is available from the British Library.

ISBN 0-521-45723-8 paperback

Transferred to digital printing 1999



For the three wonderful wild onesAndy, Larry, and Julia



Contents

Introduction

AT H E O L O G I C A E A R G U M E N T SChapter 1 Atheology and the na tur e of GodChapter z " Jn e creation-imm utabili ty argum entChapter 3 "The om niscienc e-im m utability arg um en tChapter 4 fh e deductive argu m ent from evilChapter 5 t n e argument from world-relative actuality

T H E O L O G I C A L A R G U M E N T SChapter 6 Ontological argu m entsChapter 7 tosmological argum entsChapter 8 Religious-experience ar gu m en tsChapter 9 Pragmatic argum ents

NotesIndex

vii





Introduction

This book addresses the question whether there is rationaljustification to believe that God, as conceived of by traditional Western theism, exists. There are contemporary fide-ists who hold that there is no need to justify belief by appealto arguments, since the mere fact of belief is supposed to beself-justificatory. The re is no incon siste ncy in offering b oth afideistic and an argumentative support of belief, as do somecontemporaries. My concern is only with the latter part oftheir justification. The question, then, is whether there areany good arguments either for or against believing that Godexists.

I do not pr ete nd to answ er this que stion , since f completelyignore inductive arguments based on design, beauty, andlawlike regularity and simplicity for the existence of God, aswell as those based on evil to show the improbability of his

existence. A proper discussion of these arguments is thetopic for a separate book of considerable length, since itwould have to deal with the applicability of Bayesian modelsof probability to the aggregation of the premises of all thedifferent inductive arguments for and against God's existenc e. On e of the valuable lessons to be learne d from RichardSwinburne ' s The Existence of God, which makes out such aBayesian case for belief, is that the issues are exceedinglycomplex and need to be treated by those who are steeped inprobabili ty and confirmation theory, which eliminates me.

Our question whether there are any good arguments eitherfor or agains t belief ta ke s on special im po rta nc e in the light of



Introduction

the startling resurgence of theism within philosophy during the past thirty or so years. What might surprise some isthat the three leaders of this movement, William Alston,Alvin Plantinga, and Richard Swinburne, are themselvesanalytical ph ilos op he rs. Som e m istakenly see analytic ph ilosophy as the natural enemy of theism, no doubt because certain movements in twentieth-century analytic philosophy,such as logical atomism, logical positivism, and some versions of ordinary language philosophy, developed theoriesof meaning that were employed to slay the dragon of theismby showing that it did not measure up to certain minimalstandards of meaningfulness. But it is a mistake to identifyanalytic philoso phy w ith these m ov em ents an d their dog m as.

While these movements, along with their theories ofmeaning, have come and gone and the criticisms of theismbased on their theor ies of m ean ing ha ve becom e old hat, analytic philosophy has forged new weapons in the interim thathave been deployed by analytically trained philosophers on

behalf of theism: for instance, rational choice theory tobreathe new life into the perennial Rodney Dangerfield ofphilosophy, Pascal's wager; modal logic to reformulate amore powerful version of the ontological argument; language-game analysis for justifying the practice of religionand, in particular, the prima facie acceptability of existentialclaims based on religious experiences; and Bayesian modelsof probability for an inductive justification of belief. Philosophy of religion is to the core areas of philosophy - logic,scientific methodology, the philosophy of language, metaphysics, and epistemology - as Israel is to the Pentagon. Theformer are a proving ground for the weapons forged in thelatter. Wh ene ver there is a significant bre ak thr ou gh in on e ofthe core areas, it eventually finds a fruitful deployment in theperipheral areas, such as the philosophy of religion. And thisis what we have witnessed during the past thirty years.

Because theism has found such a powerful new formulation, due to the deployment of these new analytical weapons , there is a ne ed for a retu rn visit from H um e's P hilo. Butthe sceptical Philo w ho se spirit im bu es m y book is m or e th an



Introduction

just a crazed Charles Bronson, who is back again and reallyangry this t ime, even more so than on the previous thirty-six occasions; for my philosophical version of "Death WishXXXVII" has the positive upshot of helping us to command amore adequate conception of God - a God that will prove aworthy object of worship and obedience, even if the the casefor believing in his existence is shaky. My book, therefore,has both a negative and a positive pole.

There are two very different sorts of arguments to showthat belief (disbelief) is rationally justified. One is directed

toward establishing the truth (falsity) of the proposition thatGod exists. It will be called an "epistemological argument,"since it pu rp or ts to sup ply the sort of justification that w ou ldsupport a claim to know that God exists (does not exist). Theoth er, to be called a "pragm atic ar gu m en t," is directed tow ardshowing the prudential or moral benefits that result frombelieving (disbelieving) this proposition. Both ways of justifying the rationality of belief (disbelief) will be considered.

The epistemological arguments will be my first and forem ost con cern, with only the final Ch ap ter 9 dev oted to thepragm atic argu m en ts. I am going to reverse the usual o rderof presentation and begin with arguments against the existence of God, with special attention to so-called atheologicalarg um en ts that attem pt to dedu ce a contradiction from thetheist 's conception of God, with appeal to only necessarilytrue additional premises. My reason for doing so is that thefirst order of business should be to clarify the nature of the

God whose existence is in question. And that is just whatthese atheological arguments help us to do. They are thethought experiments that probe the internal consistency ofthe theist 's conception of God, often with the result thatthe theist must go back to the drawing board and redesignthe particular divine attribute(s) that is the focus of the argument. Their role in spurring consideration of the divinena tu re is similar to that of Zeno's parad ox es in forcing su bsequ en t philo sop hers to come to grips with the natur e of spaceand time. The idea of God in Western civilization has in factundergone just this sort of dialectical unfolding through

3



Introduction

the successive challenges posed by different atheologicalarguments .

The idea of redesigning our concept of God might strikesom e as bla sph em ou s. This becomes less shocking wh en it isrealized that the concept of God that is the target of an atheological arg um en t is that of the theolog ian, w hich is a high lytheoretical conce pt tha t is as distant from the so m ew ha tanthropomorphic concept of God in the Scriptures as is thephysicist's concept of a table from that of the ordinary person . The religiously available God - the one wh o com m un eswith men and intervenes in history - was metaphysicalizedby the great m ediev al theists so that he bega n to ha ve the sortof being enjoyed by a Platonic form. Our religious experiences and traditions serve as data for these metaphysicaltheories about God's nature, but in virtue of their underde-termining these theories, there is considerable room for conceptual maneuvering when a given theory of God's natureruns afoul of an atheological argument, just as there is when

a scientific theory faces anomalous facts. The basic problemthat a theological concept of God faces is that of over meta-physicalizing God so that he no longer is a person andthereby becomes religiously unavailable. This is a specialinstance of the problem faced by any theoretical or rationalreconstruction of an ordinary concept: Which features of theordinary concept must get retained? In Carnapian terms, thequestion concerns the conditions of material adequacy forthe analysis of our concept of God. This will be a recurringissue in this book.

Blasphemy aside, the idea of redesigning our concept ofGod raises the problem of how we can keep the referent ofthe word "God" constant amidst these conceptual reforms,w hic h can involve a cha nge in either wh at are taken to be th eessential or defining properties of God or how they areun de rsto od . Th ere m ust be some answ er to how this is possible, since this is in fact what has happened.

But just h ow is it possible? To an sw er this we m us t see ho w"God" refers. Some have claimed that "God" is a title thatapplies to an individual in virtue of his playing the role of the

4



Introiuciion

absolutely perfect sovereign being, which is the concept ofGod in traditional Western theism. If "God" is a title, it functions quite differently from the ordinary titles with which weare familiar, for instance, "the king of France," "the heavyweight champion of the world," and so on. Being God,unlike having one of these titles, is both essential to andconstitutive of the essence of its possessor, that is, this individual could not exist without being God, and no other individual could be God. The champion can lose his title toanother, which shows that this title is neither essential nor

necessarily possessed uniquely either within a single worldor across possible worlds. The character played by MarlonBrando in On the Waterfront said th at h e could ha ve be en acontender, even the champion; but it would be a violation ofthe meaning of God for him to have said that he could havebeen God or for God to say that he might have been a two-bitenforcer for the mob. No wonder there is no theologicalversion of the king-must-die legend.

Granted that any being who is God is God in every worldin which he exists and no other being is God in any world,there still remains the question of what qualifies an individual as the denotatum of "God." "God," no doubt, is a propername, but this is not sufficiently helpful, since there are suchwidely divergent views of how proper names refer. On theone hand, there is the so-called descriptivist theory accordingto which a prop er n am e has a sense that is expressed by somedefinite description or cluster of descriptions, it being both

sufficient and necessary for an individual to be its denotatumthat she satisfy this description or a good number of thedescriptions in the cluster. In regard to successive uses of aname, the cluster theory can require that they either haveenough members of the set in common or be connected by asequence of uses of the name, adjoining members of whichsatisfy this con ditio n. This wo uld allow successive coreferen-tial uses to have no properties in common, this parallelingthe bundle theories of the identity of material objects andpe rso ns over tim e. On t he basis of this analo gy, the latter willbe called the "bu nd le version of the cluster the or y." O pp os ed

5



Introduction

to this is the view of pr op er n am es as purely referential, theirreferent being de term ine d by various causal or historical factsconnecting the referent with the user of the name. "God"does not perfectly fit either theory, but by judiciously incorporating elements of both theories, along with some language-game analysis, an adequate answer can be given toour question as to how the referent of "God" can remainconstant amidst conceptual reform.

In recent years the descriptivist theory h as come un d er firefrom the likes of Kripke, Putnam, and Donnellan. The basis

of their attack is to take any description (or cluster of descriptions) that is offered as con stituting the sen se of a na m e an dshow that we can construct a counterfactual story in whichthe actual referent of the name does not satisfy the description (or a sufficient number of those in the cluster). In placeof this account, they suggest that typically, a name is osten-sively or indexically bestowed upon an individual and subsequent users of the name pick up the referent from theirpredecessors in an ongoing linguistic community, with thehistorical chain extending all the way back to the originalbap tizer. This secures con stancy of reference over time to th isindividu al and allows us in the interim to revise radically ou rviews of the essential nature of this being. It is this relianceon an ongoing linguistic community that will have a fruitfulapplication to the case of "G od ," in which the linguistic community is replaced by a religious community.

The historical-causal theory of reference also applies tonames of natural kinds. Consider "gold" in this connection.We begin by ostending a paradigmatic class of specimens ofgold. We then turn our scientists loose to investigate thenature or essence of these specimens. As time goes by werevise our definition of w ha t con stitutes th e essential p ro pe rties of gold, which is w ha t in fact ha p p en ed as the alch em ist 'stheory gave way to that of modern atomic theory. Thus, thedescriptive sense a name might have at some time is not

inviolable in that a later use of it can be coreferring eventhough it lacks this descriptive sense, due to a change in ourdefinition of this natural kind.

6



Introduction

If "God " we re to fit this simplified historical-cum-indexical-reference theory , our question wou ld be an sw ere d. But thereare two reasons for doubting that it does. First, because Godis a supernatural being, he seems to defy being indexicallypinned down or baptized. There are no lapels to be grabbedhold of by a use of "this." Some would contend that we canostensively pin do w n the nam e "G od " by saying "this" whe nhav ing or after jus t ha vin g a mystical or religious exp erience,in which "this" denotes the intentional accusative or contentof the experience. This would seem to require that these

experiences are cognitive and that their objective accusativeis a common object of the experiences of different persons aswell as of successive e xperien ces of a single pe rs on . These a revery controversial claims and must await a full discussion(and refutation) in Chapter 8.

A second disanalogy between "God" and the sort of ordinary proper names to which the historical-turn-indexical-reference theory applies is that whereas it is not an analytictruth or true by definition that the referent of an ordinaryproper name satisfy some description (e.g., we can imaginewhat it would be like to discover that the person whom webaptized as "Jones" and thought to be a human being is arobot), this does not appear to be so for "God." At any timeat which "God" is used, there will be some descriptive sensethat it has by definition. For example, at the present time it isanalytically true that God is a powerful, benevolent beingthat is eminently worthy of worship and obedience. To this

extent, "God" is not distinguishable from a natural-kindterm , which also can have at any time a desc riptive sense thatis definit ionally determined. But they part company becausesome of the descriptive properties that are definitionally tiedto God are hard core in that we wo uld not allow a use of " G od "to be coreferring with ours if these properties were not atleast partially constitutive of the sense of the name. Soft-coredescriptive properties, even if definitionally linked with"God," can alter over t ime without destroying sameness ofreference. Examples of such properties are being absolutelysimple, that is , adm itting no distinction betw een essence an d

7



Introduction

existence or betw een h is prop erties, and bein g unre strictedlyom nipo tent, both of which prop erties have come and go ne aspart of the sense of "God" without affecting its reference.

Examples of the hard-core descriptive properties of "God"are being a supremely great being, that is, as great as anybeing could possibly be, and being eminen tly w orth y of w orship and obedience. My reason for selecting these propertiesas hard core is that it is essential to our idea of God, to therole it plays in the form of life in which it is implicated, thatGod is eminently worthy of worship and obedience, and a

being could occupy this exalted position only if he is a beingthan w hich there could be non e greater. Th ese are high-leve l,emergent properties, since an individual can have them onlyin virtue of the possession of other, lower-level properties,such as omnipotence, benevolence, and so on.

The connect ion between these emergent propert ies andtheir lower-level determiners is very loose, and thereby permits there to be considerable conceptual reform withoutdestroying sameness of reference. First, we can change ourmind as to what the latter properties are without alteringreference, for insta nce , giving u p ab solute simplicity as beingone of these lower-level determiners. Second, we can reviseour analysis of those determiners that we take to be hardcore, such as being benevolent, powerful, or providential.The manner in which we account for the lower-level determiners of the hard-core emergent properties can, and has,varied greatly over time without causing any change in refer

ence. There is no analogue to this in respect to ordinaryproper nam es according to the standard descriptivist th eory ,including the bundle version variant of the cluster theory.

Recent discussions of intentional identity can also illuminate some features of when two uses of "God" are co-referring.1 That they are coreferring should be independentof God's existence, and this should hold even if the referrerstake having necessary existence to be one of God's hard-core

properties, which would not be the case if the referent of"God" had to be pinned down indexically. Jones believesthat a certain witch poisoned his well, and Smith believes



Introduction

that the same witch killed his calf. Here is a case of intentional identity in which the existence of the referent is notrequired. Nor does their identity of reference require thatthey completely agree in their sortal characterization of theircommon referent. Smith could believe, pace Jones, that thepoisoner of Jones's well and the killer of his calf is not a witchbut a warlock or a vam pire . For Sm ith and Jones to be corefer-rers, it is not alone sufficient that Smith intend to refer to thesame individual as did Jones. What he says does not settlethe issue, unlike the case in which what an artist says his

pain ting repr esen ts settles the ma tter. It is also necessary tha tthere is sufficient similarity in their sortal characterizations oftheir common referent (e.g., they both thought of the referent as a being possessed of supernatural powers, in spite oftheir differences over just wha t sort of a su pe rn atu ra l be ing itis) an d they assign a similar exp lanatory role to it. W ere Sm ithto believe that some demon caused his rheumatism and Jonesto believe that the same individual caused his rheumatism,only disagreeing in his taking it to be germs rather than ade m on , we w ould n ot coun t his reference as being coreferen-tial with Smith's. Jones might say that demons are nothingbut germs, but this would be the eliminative use of "nothingbut" - the one that entails that there aren't any demons. It iscontrasted with the theoretical or reductive identity use of"nothing but" as in "water is nothing but a collection of H2 0m olecu les," which does not have any eliminative implication.

Similar considerations hold for intentional identities in

volving successive uses of "God." Abraham might believethat some supernatural being created the cosmos, and Isaacmight believe that the same divine person communed withhim. Again, their identity of reference does not require theexistence of a referent. Nor is it requ ired tha t the y conceive ofthe referent in exactly the same way (e.g., Isaac might differfrom Abraham in regard to what lower-level soft-core properties he takes to be essential to God or in how he understa nd s these divine attribu tes), for the re is bo th sufficientagreement in their sortal characterization of the referent andthe role they assign it as the explainer of the existence of the

9



Introduction

universe and various occurrences within it, such as numi

nous and other types of religious experiences. In spite oftheir differences in how they conceive of God, they agree inthinking of God as a supernatural being who is the creatorand sustainer of the universe and eminently worthy of worship and obedience. Here we see the importance of God'shard-core properties in securing coreference in successiveuses of "God."

The picture presented so far is overly intellectualized,stressing only the descriptive aspects of the name, both hardand soft core. No doubt these descriptive features are anessential part of the story, but they are not alone adequate toexplain how reference can remain constant amidst alterationsin the soft-core descriptive sense of the name, as well as inthe analyses given of the hard-core properties. It is here thatwe must avail ourselves of the historical-causal theory'snotion of a succession of referrers within an ongoing linguistic community who pick up the reference of a name fromtheir prede cess ors; only we m ust replace the linguistic with areligious community. The reason we think our use of "God"refers to the same God as was referred to by Abraham, Isaac,and Jacob, despite radical differences in our theories aboutthe nature of God, is that we are members of the same ongoing historical com m un ity of believers, sharing the sam e formof life.2

What is the form of life that is implicated in our commonreligious language game? It has to do with our having acommon historical root to our religious community, sharingsimilar attitudes toward the meaning and significance of life,common ethical beliefs and practices, and the subsequentpeo ple in this on go ing historical chain identifying them selveswith their predecessors in this chain. For their use of "God"to be coreferring with that of their predecessors, it is notenough that they intend it to be. What they say, thoughrelevant in this matter, is not alone decisive. They must alsoshare this common form of life with their predecessors andthink of themselves as continuing their traditions and aspirations. How we individuate our religious community will

1 0



Introduction

depend upon how ecumenical we want to be, which in turn

will depend on a complex of political, economic, and psychological factors. Jus t look at the sign ificant shift in the CatholicChurch's att i tude toward those they formerly took to be heathens in Third World countries. What consti tutes samenessof a religious community over time is a deep issue thatdeserv es a separa te volum e, just as doe s the problem of w hatconsti tutes sa m en ess of a l inguistic com m un ity over t ime forthe historical-causal theory of reference. Suffice to say that aproper account of how the reference of "God" can remainconstant over time despite significant conceptual revision iscrucially dependent upon how we fill in the details of whatmakes for sameness of a religious community over t ime.

Hopefully, this rough sketch, combining elements of boththe historical-causal and descriptivist theo ries, su pp lem en tedby some language-game analysis, makes intell igible how wecan redesign our concept of God without change of referenceand opens the way for our excursion into atheological argu

m en ts. Chap ter 1 will prese nt a bro ad o verview of the searguments, four of which will be singled out for in-depthtreatment in the following chapters.

1 1



Atheological arguments



Chapter 1

Atheology and the nature of God

In this chapter a broad overview will be given of atheologicalarguments , without any at tempt to determine whether theyultimately succeed. The aim is to give us a feeling for thesearg um en ts and a n appreciation for their positive role in bringing about important conceptual reforms in the way in whichGod's nature has been conceived of over time.

The aim of an atheological argument is to reveal a logicalinconsistency in the theist's concept of God. Accordingly, itbegins with an initial set of propositions each of which isaccepted by the theist as a necessary conceptual truth. Forexample, from the initial set containing the sole premise thatnecessarily God is om nipo tent, an attemp t is m ade to d edu cea contradiction, namely, that there is some task that anomnipotent being cannot do. Or, from the initial set containing the conceptual truth that God is benevolent in the sense

of always choosing the best alternative, it is deduced thatGod both does and does not actualize some possible world,with appeal being m ade to the additional prem ise that necessarily there is no best of all possible worlds. In both of theseatheological arguments, the initial set contains only propositions that the theist takes to be necessary truths.

It would seem that my definition of an "atheological argument" as a deduction of a contradiction from the theist 'sconcept of God, with appeal to only necessarily true additional premises, is unduly restrictive; for some of the mostimportant attempts to demonstrate the logical inconsistencyof theism begin with an initial set of propositions that contain

*5



Atheology and the nature of Cod

at least one proposition that is taken by the theist to be onlycontingently true, for instance, that there exists evil, createdfree persons, and so on. For example, the deductive argument from evil, in one of its versions, begins with an initialset containing two propositions:

1. There exists an omnipotent and benevolent God; and2. Evil exists,

from which it is deduced that evil does not exist, with appealbeing made to some extra premises that articulate conceptual

truths about omnipotence and benevolence. Herein the initial set is not comprised solely of conceptual truths, the theisttaking 2 to be only contingently true, and also 1, if God'sexistence is taken to be contingent. Similar considerationshold for the argument that deduces from the initial set comprised of

3. There exists an omniscient God; and4. There exist free created people

the contradiction that these created free pe op le are n ot reallyfree because their actions and choices are foreknown by thisom niscient G od. Proposition 4 certainly is no t take n by thetheist to be necessary, it not being taken to be an essentialprop erty of God that he create free p ers on s. Sim ilarly, th e argu m ent that atte m pts to de du ce from th e initial set co ntain ing

5. There exists an immutable and omniscient God; and6. There are temporal indexical facts, such as what time it

is right now

the contradiction that this omniscient (or immutable) God isnot omniscient (or immutable) contains an initial set thatseem s to contain at least one contingent m em be r - 5 and /or 6.

We might call these arg um en ts w hos e initial set conta ins atleast one proposition that is taken to be only contingentlytrue by the theist an "impure atheological argument," incontrast with the "pure atheological arguments" of my definition. This distinction turn s out to be un im po rta nt , sinceevery successful impu re atheological ar gu m en t ha s an eq ually

16



Atheotogy andthe *»f Mr* of God

successful counterpart pure atheological argument. The former can be recast as a pure argument by possibilizing theconjunction of the propositions in its initial set. Let us suppose that the initial set of the impure argument contains theproposi t ions p and q, in which p and/or q is taken to be onlycontingently true by the theist. It is possible that (p and q) isweaker than (p and q) in that it is entailed by but does notentail the latter. Therefore, if (p and q) is accepted by thetheist, so must it is possible that (p and q). For example, if thetheist grants that both God and evil exist, it is thereby grantedthat it is conceptually possible that they coexist. Now, if (pand q) entail a con tradiction, so do es the we ake r it is possiblethat (p and q). Given that (p and q) entail a con tradiction r, itfollows tha t it is no t possible tha t (pandq). If it is no t possiblethat (p and q), it is nec essa ry th at it is not possible that (p andq), that is, it is impossible that it is possible that (p and q),given that a proposition's modal status is invariant acrosspossible w or lds . Since an impossible proposition entails every

proposition, it follows that it is possible that (p and q), beingimpossible, entails this contradiction r. Another way of putting this point is that if (p and q) entail a contradiction, itfollows that not only (p and q) is impossible but also it ispossible that (p and q), given that a proposition has its modalstatus necessarily. Thus, the atheologian, by the trick of possibilizing the conjunction of the propositions in the initial setof an im pu re atheological argu m en t, can conv ert it into a pu reone. In the following discussion no attempt will be made toconvert impure atheological arguments into pure ones, sinceit is obvious that this can be done by the simple expedient ofpo ssibilizin g its initial set. With this in m ind w e can beg in ou rsurvey of the atheological arguments.

O u r brief o verv iew of atheological arg um en ts will be ba sedupon the divine attribute against which they are directed.Some of the arguments are not directed against just oneattribute but instead attempt to demonstrate a logical clashbetween two or more of them. As a result, there will besome arbitrariness under which divine attribute they areconsidered.

17




O M N I P O T E N C E

The first divine attribute to be taken up is omnipotence. Godis supposed to have this attribute, like all of his other attribu tes, e ssentially, m ean ing th at it is logically or conc eptuallyimpossible that God exist without having this attribute. Hispossessio n of om nip ote nc e is an exa m ple of a de re necessity,since it follows from his nature, not our way of referring toGod, that he has it. This contrasts with de dicto necessityclaims, the truth of which depend upon the mode of reference; for instance, it is true that it is necessary that God isomnipotent but not that Augustine's favorite individual isomnipotent, even though God is in fact Augustine's favoriteindividual.

Probably the most famous atheological argument is the onethat deduces a contradiction from the initial set containingonly the proposition that God is omnipotent. The trick in theargument is to require of an omnipotent being that he can do

anything without restriction. The following is my version ofthe paradox of the stone, which differs from the standardversion in that it is not mounted as a dilemma argument,thereby avoiding needless complications. It begins with aninitial set containing only

7. It is necessary that God is omnipotent;

and then proceeds as follows:

8. It is necessary that an omnipotent being can do anything [by definition];9. It is ne ces sary that God ca n do an y th in g [from 7 an d 8];10. It is nec essary that God can create a stone so hea vy that

God ca nn ot lift it [from 8 an d 9];11. It is ne ce ssa ry that (if G od can create a ston e so heavy

that God cannot lift it, then it is possible that there is something God cannot do) [conceptual truth];

12. It is possible that that there is something God cannotdo [from 10 a nd 11];

18



GmnipQten&e

13. It is no t necessary that God can do an yt hi ng [from 12];and

14. It is nece ssary tha t God can do any thi ng a nd it is notnecessary that God can do anything [from 9 and 13].

Since a contradiction has been deduced from 7, with appealonly to necessarily true additional premises, it follows that 7is necessarily false.

There is a variant of the "stone" argument in which "commit suicide," rather than "create a stone so heavy that Godcannot lift it," is substituted for "anything" in 9. If God cancommit suicide, it is possible that an omnipotent being ceaseto exist. If you find nothing conceptually absurd about this,try an absolutely perfect being, which is what God essentiallyis, ceasing to exist. This, certainly, is not possible. This variant on the suicide argument is not directed exclusively atGod's om nipo tence but at his being essentially both om nipotent and absolutely perfect. Accordingly, we will call it the

"omnipotence-perfect ion argument." Another mixed-bagatheological argument involving omnipotence is the "omnipotence-benevolence argument" in which "commit an immoral action " is sub stituted for "an yth ing " in 9. Because Go dis omnipotent, he can commit an immoral action; but, sincehe is essentially benevolent, he cannot. Thus, one of the twoproperties must be jett isoned.

What is the theist to do in response to these atheological arguments? Plainly, unless we are Kierkegaards, someresponse is necessary, since we do not want to accept alogically inconsistent creed. There is considerable room forconceptual reform here, since the theologian's notion ofGo d's om nip oten ce is a theoretical reconstruc tion of the biblical notion of God Almighty. From the idea of a God who issovereign over all there is comes that of a God whose powerknows no limitations, not even of a logical or conceptualsort.1 To escap e the ston e and suicide ar gu m en ts, a reli

giously acceptable restriction must be placed upon God'som nipotence. The mixed-bag argu m ents - the om nip ote nc e-

19





Omnipotence

There are many possible actions that God, in virtue of beingabsolutely perfect, cannot perform, such as playing football.Only a being possessed of a body could do this, but God"ain' t got no bo dy ," since to hav e a body m akes on e subject tothe possibilities of corruption and death, which certainlywould disqualify one from being an absolutely perfect being.Notice that some properties are great making in a sortallyrelative, but not absolute fashion, such as being well coordinated, which is great making relative to an animal but notabsolutely, since to have it one must be possessed of a body

and thus not be absolutely perfect.God's actions are confined to bringing things about in his

ow n inimitable sup ern atu ral fashion by simply willing th em .Thus , a more perspicuous way of rendering 02 is as

0 3 . For any proposition p, if it is logically possible thatGod bring it about that p, then God can bring it about that p,

in which "can bring it about" is to be understood in the full-

bloode d se nse of hav ing the pow er, abili ty, opportun ity, andso on of bringing about the state of affairs in question. The"can" must not be understood in the weak, logical-possibilitysense, for this would render 03 vacuously tautological.Again, there are logically possible propositions, for instance,that there exists an uncreated stone (assuming that it is notnecessary that God exists) that God cannot bring about, sinceit would be logically inconsistent for him to do so. Anotherexample is the proposition that God exists, and this he cannot bring about even if his existence is necessary becauseentailed by his essence.

While Oz and 03 are an improvement over O,, they stillface problems. The first difficulty is a superficial technicalproblem that is due to the fact that a given action or state ofaffairs ad m its of alternative descrip tions s uch that it could b elogically possible under some, but not all, of them. Imaginethat Igor's favorite object is the square circle. That Igor's

favorite object exists and that the square circle exists reportone and the same state of affairs, yet the former is logically

2 1



Atheology and the Mtureof God

poss ible an d t he latter is no t. There fore, it is logically po ssible(in the de dicto sense) that God bring it about that Igor'sfavorite object exists, and yet he cannot bring this about,pace 03 .

Clement Dore has suggested a way around this difficultyin which God "can perform any action such that the statem en t that He can perform it is logically con sistent an d w ou ldturn out to be logically consistent on any true description ofthe action."2 Using this strategy, 03 can be revised as

0 4 . For any proposition p, if it is logically possible thatGod bring i t about that p as well as any proposition that iscoreporting with p, then God can bring i t about that p,

in which two propositions are coreporting if the participialnominalizations of the sentences that express them are core-ferring. The participial nominalizations of "Igor's favoriteobject exists" an d "Th e squ are circle exists" are, respectively,

"Igor's favorite object existing" and "the square circle existing "; an d th ey are coreferring be caus e of the identity be tw ee nIgor's favorite object and the square circle.

There are more substantive objections to 02 and 03 , however, that are not circumvented by 04 . The difficulties resultfrom their sortal relativization of om nip ote nc e. First, it mig htbe objected that this makes it too easy to qualify as omnipotent. Let a pirod by definition be a bein g that can do no thin gsave for singing "Dixie." With no insult to the Confederacyinte nd ed , there is no do ub t that this bein g is not om nipo tent,and yet it qualifies as omnipotent in the sortally relativizedsense. That a pirod cannot create a pencil does not countaga inst its be ing om nip ot en t, since it is no t logically possib lethat a pirod bring it about that a pencil exists, this being dueto ho w we defined it. The resp on se to this coun terexam ple isthat there is a world of difference between relativizingomnipotence to God - an absolutely perfect being - and a

pirod. And this seems fair enough.But there is a more serious difficulty with a sortally relativized account of God's omnipotence; it creates a paradox of

2 2



Omnipotence

perfection co nsistin g in such a being h aving a lesser degree of

power than that possessed by some possible nonperfect being. Let a Pinrod be a possible being who has only one of thedivine perfections - omnipotence - and thereby does notqualify as an absolutely perfect being, Anything that God cando or bring ab ou t, ou r Pinrod can do or bring abou t, and the nsome. Not being burdened with being absolutely perfect, ourPinrod, in addition to having all ot God's powers, for instance, being able to create a universe ex nihilo and performot he r parlor tricks, can be poss esse d ()f a bod y and thereby beable to play football. Since he is not absolutely perfect, he isnot barred from committing suicide. Furthermore, since helacks benevolence, he is able to perform an immoral action.T hu s, this lesser bein g ha s a greater deg ree of freedom thanthat possessed by God, an absolutely perfect being. It looksas if no one is perfect, not even an absolutely perfect being

The great medieval theists would have had a ready response to the paradox of perfection based on the doctrine of

the divine simplicity. It will be worthwhile to consider thisdoctrine, not only as offering a way out of this paradox, butfor its own intrinsic interest. God's status as an absolutelyperfect being precluded either there being any distinctionsw ithin his na tur e or his being dep en den t upo n anything. Theformer entails that there is no distinction between God'sproperties: His omnipotence is identical with his omniscience, w hich is identical with his om nibenev olence, a nd soon. Were there to be any compositeness in his nature, hewould face the possibility of destruction through decomposition. (Think of Plato's argument for the immortality of thesoul based on the noncompositeness of the soul.)3 The latterentails that God does not instantiate any property, for werehe to do so he would be distinct from and dependent on thisproperty, thereby violating his aseity. Since God is identicalwith omnipotence, as well as all his other properties, noindividual could have only one of God's properties, for thisproperty is identical with all the other divine properties; andno individual distinct from God could have any one of theseproperties, since then he would be identical with God. As a

23




result, there can be no Pinrod, which eliminates this versionof the paradox of perfection, as well as versions based uponany other divine attribute.

There are two different versions of the doctrine of thedivine simplicity - the property and instance identity versions. W he n it is said tha t God is identica l with his pro pe rties ,say omnipotence, it could mean that he is identical eitherwith th e prope rty of om nipo tence itself or with his instancingof omnipotence. Both versions face formidable objections.

One objection to the property identity version is that by

making God identical with an abstract property - omnipotence itself - it renders him conceptually unfit to be a personal creator of the universe. An omnipotent being can be acausal factor but not omnipotence itself, to hark back to oneof Aristotle's objections to Plato's doctrine of the forms.Another objection to this version is that it has the unwantedconsequence that no individual other than God could haveany one of his properties, since if it were to instantiate oneof these properties it would be instantiating God himselfand thereby be God. But certainly individuals other thanGod can have som e of his intrinsic or non-C am bridge pro perties, for instance, being a person, self-identical, an entity,and so on. Furthermore, the property of omnipotence quaproperty seems to be a different property than that of, say,omnibenevolence. One is being able to do anything (withinsome limits, such as those imposed already), the other being disposed to perform only good actions. Another way of

making this point is that since "omnipotence" and "omniben evo lence " obviously differ in sens e and the sense of eachis the property it expresses, the properties of omnipotencean d om niben evo lence are different. In case this isn't obv iousto you, the following might be of some help. Since ordinarypower and benevolence obviously differ, there is all them ore reason to ho ld that increas ing deg ree s thereof differ.A nd , likewise, there is all the m ore re ason to hold that unlimited degrees thereof - omnipotence and omnibenevolence -also differ.

24



Omnipotence

The instance identity version would seem more plausible,

since it is by now a familiar story that two referring expressions can be coreferential though differing in sense, forinstance, the morning star" and the evening star." Th us, itmight be possible that an instancing of omnipotence" and"an instancing of omnibenevolence" are coreferential, although differing in sense, in that Go d's instancin g of the oneproperty is identical with his instancing of the other. Beforewe consider the possibility of this being so, it mus t bestressed that this instance identity version of the divine simplicity violates God's absolute aseity or inde pen den ce, s inceit conceives of God as instantiating properties and thus asdependent upon them. Stil l , it is interesting to consider thisversion, since, if it proves viable, the theist might want to gowith it and accept a restricted version of God's aseity.

Eleonore Stump and Norman Kretzmann develop a version of the instance identity theory that makes use of theaforementioned sense-reference distinction:

'Perfect power' and 'perfect knowledge' are precise analoguesfor 'the morning star' and 'the evening star'; non-synonymous expressions designating quite distinct manifestations ofone and the same thing.4

Before critically evaluating their claim, it is necessary torework their analogy so as to eliminate two ser ious blun der s.

In the first place, the mo rning star" and the evening star,"do not designate different manifestations of one and thesame planet but one and the same planet via different w ay sit manifests itself. Second, "perfect power" and "perfectknowledge" do not refer to an instance of a property but tothe property itself, and thus must be changed to an instancing of perfect power" and an instancing of perfect knowledge." To avoid these two difficulties we would do well torevise their claim as follows: An instance of perfect po w er "and an instance of perfect knowledge" are precise analogues for the morning star" and the evening star"; non-

25



Atheology and the nature of G@i

syn on ym ou s expressions designating on e and the same thin gvia different ways in which it manifests itself.

It would seem that we have the same intuitive grounds fordenying the identity of an instance of perfect power with aninstance of perfect know ledge as we previo usly h ad for de nying the identity of the property of om nip ote nc e with tha t ofomnibenevolence. Because an instance of ordinary power isobviously not identical with an instance of ordinary benevolence, there is all the more reason to hold that instances ofincreasing degre es thereof are not identical, and th us to d en ythat an instan ce of an unlimited de gree of the one is identicalwith an unlimited degree of the other.

In opposition to this line of reasoning, Stump and Kretz-mann offer an analogy to show how (an instance of) perfectpower can be identical with (an instance of) perfect knowledge although no instance of ordinary finite power is identical with a n instance of ordinary finite kno w led ge . The formeridentity "does not entail that power is identical with knowledge any m ore than the fact that the sum m it of a m ou ntain 's

east slope is identical w ith the sum m it of its w est slo pe e nta ilsthe iden tity of the slopes. 5 This spatial ana log y is m ost unfortunate for their purpose, since, when it is pushed, it resultsin a view of Go d's omnipro perties tha t is the o pp osite of w ha tthey inten d. Th e com mo n sum m it limits each of the slope s byserving as their common point of termination. Analogously,God's (instancing of) perfect power and perfect knowledgewould also represent a limitation on ordinary power andknowledge. But this is the opposite of what they mean by

these divine perfections, which are supposed to be ideal orunlimited instancings of power and knowledge. "The singlesummit is indeed the perfection of all the slopes" and "suggests that the idea that perfect phi and perfect psi might beidentical despite the plain difference between phi and psi."6

It might be objected that I have leaned too heavily on thespatial aspect of their analogy in drawing out the absurditythat God's omnipotence, and so on, represents a l imitat ionrather than a perfection of ordinary power. But if we are not

allowed to press their analogy in this way, it has no value.

26



Omnipotence

Even i f the ins tance iden t i ty vers ion of the doc t r ine of the

d i v i n e s i m p l i ci t y c o u l d s o m e h o w b e m a d e to w o r k , it w o u l ds t i l l no t succeed in neut ra l iz ing the paradox of per fec t ion ; fori t would fa i l to show tha t one could no t have one of God ' spe r fec t ions wi thou t hav ing a l l o f t hem, a s was the case wi thour P in rod . Le t u s a s sume fo r t he sake o f a rgumen t t ha tG o d ' s i n s t a n c i n g o m n i p o t e n c e is i d en t ic a l w i t h h i s i n s t a n c i n gbenevo lence . Th i s does no t en t a i l t ha t any in s t anc ing o fo m n i p o t e n c e i s i d e n t i c a l w i t h s o m e i n s t a n c i n g o f b e n e v o

lence . A token iden t i t y p ropos i t i on does no t en t a i l a t ypei d e n t i t y p r o p o s i t i o n . T h a t m y i n s t a n c i n g p a i n o n s o m e o c c asio n is ide nt ic al w i t h a specif ic f i r ing of m y C-f iber s do es n o ten ta i l th a t ev ery in s ta nc in g of pa in is iden t ica l w i th s o m efir ing of C-f ibers , no less a f i r ing of my C-f ibers .

As we have a l r eady seen , t he doc t r ine o f t he d iv ine s impl ic i ty faces a d i lemma in tha t i t s p roper ty iden t i ty vers ionr e n d e r s G o d c a u s a l ly i m p o t e n t , a n d i ts i n s t a n c e id e n t i t y v e r

s ion de s t r oy s h i s a se i ty. Wi l l iam Ma nn has a t t em pt ed to g rabth i s d i l emma by bo th ho rns by iden t i fy ing God ' s p rope r t i e sw i th h i s i n s t an c in g of t h em . Th i s i den t i fi ca t ion is su pp o se dto escape the ob jec t ion tha t God cannot be causa l ly e ff icacious because of being an abstract proper ty; but i t s t i l l facesth e a se i ty ob jec t ion ba se d on G od ins t anc ing a p ro pe r ty. H epo in t s ou t t ha t t he a se i ty ob jec t ion a s sumes tha t p rope r t i e sare abs t rac t en t i t i es :

I do n ot accept this assu m pti on . Properties, I am inclined tobelieve, are causal powers. P is a property of an object, onlyif P's presence in x confers some causal power(s) on x. Pan d Q are the same property if and only if (1) P and Q conferthe same causal powers on their objects and (2) whatever issufficient to bring about an instance of P in an object, x,is sufficient to brin g ab ou t an in stance of Q in x, and viceversa .7

The idea i s tha t God ' s power i s a cer ta in causa l capac i ty, andth i s is t he s am e caus a l capac i ty a s i s h i s ben evo lenc e , an dthe l ike .

27





Benevolence

Davidson criterion of event identity that omits the require

ment of spatiotemporal coincidence (since God isn't spa-tiotemporal for Mann) and goes only with the requirement ofhaving the same causes and effects of the same causal powers and dispositions. But this, as already indicated, does notseem to be true of God's instancing different properties.There are different theories of theoretical or transcategorealidentities based upon empirical correlations cum considerations of simplicity or scientific redactions of less-to-morebasic laws; but none of them could apply to God for obviousreasons.

I have not tried to show that there could not be an acceptable version of the doctrine of the divine simplicity, only thatnone of the leading contemporary accounts are adequate.Thus, the paradox of perfection still awaits a resolution.Maybe the best strategy for the theist is to bite the bullet andtake back the requirement that an absolutely perfect beinghave every perfection to an unlimited or unsurpassabledegree. That such a being, unlike a Pinrod, cannot commitsuicide or an immoral action does not dislodge it from itsprivileged place in the religious language game and form oflife that is involved in it; for it still retains its status as thebeing most eminently worthy of worship and obedience andso on. Giving up the requirement that the Deity be absolutelysimple in no way destroys the identity of the religious community over t ime and causes subsequent uses of "God" not

to be coreferring with earlier uses. Not only is no real harmdone, but it helps us to escape from a devastating atheologi-cal argument.

BENEVOLENCE

We have already considered one atheological argument in

volving God's benevolence - the mixed-bag omnipotence-benevolence argument - and it was seen that it is better forthe theist to reconstruct omnipotence, since essential benevolence seems to be required of a being who is going to play

29




the distinctive role of being eminently worshipable in a way

in which being unrestrictedly omnipotent is not. God is notsomeone who as a matter of contingent fact never failsmo rally but a being w h o cann ot fail because it is an essentialproperty of his nature that he cannot. Similar considerationsapply to the famous argument from evil, which also is amixed one since it app eals to God's om nipoten ce in additionto his benevolence (and also his omniscience in some versions). Again, it is one of the divine perfections other thanessential benevolence that gets reconstructed. There areother atheological arguments directed against benevolencethat concern the nature of God's benevolence rather than itsessentiality.

The point of this atheological argument, which will becalled the "b enev olenc e-creation arg um ent," is to force us toexamine our conception of God's benevolence. The theist ,unless she is willing to follow the dubious path of radicalmultivocalism and deny any similarity or analogy between

the properties of God and other beings, will conceive ofGod's goodness as being importantly similar to that of finitebeings. Thus, if the theist holds to a utilitarian conception ofth e goo d, G od w ill be conceived of as a utility max imizer. Thebenevolence-creation argument is directed against one suchutilitarian conception of God's benevolence, the Leibnizianversion, according to which he essentially follows the principle of perfection, tha t is, always choose s the best alternativeavailable.8 As Leibniz makes clear in his Theodicy, this principle enjoins God to make no choice at all when there is nouniquely best alternative, either because there are two ormore that are equally good or for each one there is an evenbetter one.

The benevolence-creation argument 's init ial set contains

15. Go d is ben evo lent; and16. A ben evo lent bein g always chooses in accordance with

the principle of perfection.It then adds this necessary premiss

17. Th ere is no w orl d th at is the best of all possib le w or ld s.

30



Benemkme

This could be true if there were two or more worlds of commensurate value or an infinite series of increasingly goodworlds, due to there being no upper bound on goodness. AsLeibniz himself said in the supplement to his Theodicy, "Thegoo d m ay an d do es go on to infinity." We shall wo rk with thelatter case of there being no uniquely best world, since it ismore intuitively obvious:

18. Go d doe s not mak e an y choice in regard to actualizing

a wo rld [from 16 and 17];19. G od 's optio n to actualize a world is a forced on e in tha tif he makes no decision, it is the same as if he actuallychooses not to actualize any world [necessary truth};

20. G od kn ow s that 19 is tru e [based on his om niscience ];21. God chooses not to actualize any world [from 17-20];

and22. God does make a choice in regard to actualizing a

w orld [from 21 , w hich explicitly contra dicts 18].The reason for 20 is that if God were not aware that his

choice to actualize a possible world is a forced one, it mightbe disputable whether his not making any choice is a choiceon his part not to actualize any world. That his choice in thismatter is forced results from the theist 's claim that a possiblew orld gets actua lized if an d o nly if Go d wills that it do es , thisbeing a special instance of his sovereignty. The theist could

escape the argument by denying that this falls within thepurview of God's sovereignty, but this would represent aradical departure from the traditional conception of God asalmighty.

Even if the Leibnizian can find some way out of this argum ent, a nd I do n' t think he can,9 there still is the problem ofwhether God can actualize any world other than the best ofall possible worlds. If he can't, then he is a Spinozistic type

creator who has no choice in the matter. And, what is evenmore unpalatable, it follows that the only possible world isthe best of all possible worlds, on the theistic assumptionsthat a possible world is one that could be actualized and that

31



Atheology and the nature ®f God

a world can get actualized only through God's creative

choice, this being a special case of his sovereignty.Most theists are not touched by the creation-benevolenceargument, since they have distinctly antiutilitarian moralintuitions. Rather, they conceive of God as giving greatestvalue to goodness of character developed through the person's own free endeavorings. Soul building is God's primarypurpose in creating a world. Or they might go the Robert M.Adams route and say that when God actualizes less perfectpossible persons than he could have, he is bestowing the

blessing of grace on the im perfect p ro du cts of his creative act;and since grace is con sidered a virtue by theists, Go d's be ne volence is not impugned by his not actualizing the best persons he could. At the opposite end of the spectrum from autilitarian concept of God's goodness is a purely formalisticone in which God's goodness consists in being morallyupright, alway s hon oring his moral du ties. How G od's goodness is conceived will have important repercussions for theconstruction of theodicies and defenses of God in the faceof evil. Whether a morally exonerating excuse can be constructed for God's permitting certain types of evils willdep end crucially on wha t consti tutes his goo dne ss. A p urelyformalistic conception of God's goodness makes the task ofconstructing a theodicy or defense easier, since it could besaid that God had no moral duty to eliminate or preventcertain kin ds of evils; to do so wo uld be an act of su pe rero gation, thu s the po int of the bum pe r sticker that said "G od do esexist: He just do es n 't w an t to get inv olv ed." But m ost theistswant to believe in a God who is more of a do-gooder than isthis coldly upright "he keeps his nose clean"-type deity.

SOVEREIGNTY

A totally unrestricted version of God's sovereignty requiresthat every thing be de term ined by his will, an d it is subject tothe same sort of atheological arguments as is a totally unrestricted version of his omnipotence. It seems to many phi-

32



Soveteigftty

losophers that abstract entities, such as nu m be rs an d pro per

ties, are not the sort of things that could be created. Thetheist is com mitted to there being objective mo ral tru ths , b u ta similar point could be made about them. This is the point ofthe following "sovereignty-benevolence argument," whoseinitial set contains these two propositions:

23. God determines the truth-values of ethical propositions; and

24. God is essentially benevolent.

Taking G od's eternality to involve beg innin gles s an d en dle ssdur ation in time, that is, om nitem pora lity, it is de du ce d from24 that

25. God is benevolent prior to his decision as to whattruth-values ethical propositions will have.

But from 23 it follows that

26. Nothing is good or bad, right or wrong, prior to God'sdecision as to what truth-values ethical propositions willhave.

But from 24 it follows that

27. Go d is benev olent prior to his d ecision as to what truth*values ethical propositions will have;

which contradicts 26.One might try to escape this argument by conceiving ofGod's decisions as to what truth-values ethical propositionsare to have as an abiding one that endures throughout aninfinite past and future, so that there is no time prior to thatat which he makes these decisions. This is not an ad hocmove, since God's immutability requires that he not changehis m ind from one time to an othe r. Th e sam e result w ou ld beachieved if the theist were to conceive of God as timelesslyeternal, that ;s , not subject to any temporal determinations ordistinctions. Again, the consequence will be that there is notime at which God has not yet made his ethical decisions.

33




A different version of the sovereignty-benevolence argu

ment must be deployed against these two ways of conceivingof God 's ethical decisions as being im mu table. An add itionalpremise must be added to the previous initial set - that Godm ust h ave a reaso n for any choice he m akes that is based onsome moral good that is realized by his choice. But thisrequires that there is something morally good that is prior inthe order of determination or explanation, rather than time,to God's choice as to what truth-values ethical propositionswill have in the sense that it helps to explain or determineGod's choice in this matter. Thus, there both is and is notsomething that is good prior in the order of explanation ordependency to God's choice as to what is good. This, I takeit, is the thrust of the argument of Plato's Euthyphro. And itseems right to me. Ethical propositions are not of the rightcategoreal sort to be made true by anyone's decision, evenGod's.

Another problem for God's sovereignty is whether it is

consistent for him to predetermine the free actions andchoices of created pe rso ns. T his will be addre ssed at leng th inChapter 4 when we take up the free will defense of God inthe face of m oral evil. The pred estina tion-fre edo m atheolog -ical arg um en t carries considerably more wallop than do es theom niscien ce-freed om argu m ent, since it is far more difficultto reconcile an act being free with God's predetermining itthan with h is only foreknow ing it. If one is convinced b y thesovereignty-benevolence and/or predestination-freedom arguments, it will be necessary to give an Gytype account ofGod's sovereignty, namely, as

S. God completely and solely determines everything thatit is logically consistent that an absolutely perfect being completely and solely determine.

God's sovereignty, like his omnipotence, must be relativized

no t just to w hat is logically possible but to w ha t it is logicallypossible for an absolutely perfect being to do or determine.

34



Absoluteness

I M M U TA B I L I T Y

God is su pp os ed to be imm utable in that he doe s not changeover time in respect to any of his intrinsic properties, asopposed to relational properties, such as being thoughtabout by Jones. According to the omniscience-immutabili tyargument, which will be our concern in Chapter 3, this getsin the way of his being completely omniscient; for there aresome tru ths that can be kn ow n only by a tempo ral being that

continually changes its beliefs, such as what the date is. Thetheist can respond to this argument by limiting either God'simm utability or his omnisc ience, as we will see in Cha pter 3.

Another mixed atheological argument involving immutability is the o ne tha t will be the topic of Ch ap ter 2 - thecreation-immutabili ty argument. God is claimed by theismto have created the world a finite number of years ago, butthis seems to require that he changed his mind in the past in

regard to creating the world, thereby compromising hisimmutability. This atheological argument, as we shall see,has played a very significant role in prodding the theologianto reexamine God's relation to creation, often with the resultthat God, along with his creative acts, is placed "outside" oft ime. This doctrine of timeless creation will figure prominently in the next chapter. The theologian is not forced tofollow this path . Instead it m ight be de nie d th at an absolutelyperfect bein g m us t be strictly im m uta ble. To sup po rt this itmust be shown that the reasons for the immutability of aperfect being, for instance, the Aristotelian argument thata perfect being is in a state of complete actuality and therebycannot change, are bogus.

A B S O L U T E N E S S

This divine property is not made explicit by traditional Western theism but is implicit in the way in which it conceives of

35



Atheology and the tmture of God

God's existence and actions. God's existence is absolute and

in no way world relative. Likewise, each of his actions, suchas his creative choice as to which world to actualize, is absolute. God actually makes only one creative choice, though hemight have made others, that is, in certain possible worlds inwhich he exists he makes these alternative creative choices.But only one of his many possible creative choices is actualsimpliciter or absolutely, just as only on e am on g all the po ssible worlds is actual simpliciter. Traditional theism picturesGod as being "outside" of possible worlds in that he initiallycontemplates them and then makes a unique and absolutechoice as to which on e of them is to becom e actualize d. Th atone among these worlds is actual simpliciter or absolutely isdue to the absoluteness of God's existence and his creativeactions. If actuality, along with existence and truth, are notabsolute, but only world relative, as David Lewis has argued,this fundamental tenet of traditional theism is undercut.Ch apter 5 will consider an atheological argum ent based on

the doctrine of world-relative actuality. It will be found to bewanting.I trust that the foregoing atheological arguments have cap

tured your fancy and that you are anxious to pursue some ofthem in further detail , hoping thereby to command a moreadequate conception of God's nature. And we now turn tothis task.

36



Chapter 2The creation-immutability argument

This argument is a dramatic case in point of the thesis ofC ha pte r 1 conce rning the positive role of the atheologicalarg um en ts; for Saint Au gu stine's famed theo ry of the relationbetween God's eternality and the temporality of the createdworld, which was accepted by all of the subsequent greatmedieval theists and became official church doctrine, wasdev elope d in resp on se to it. Fu rther m ore, t he account of timethat falls out of his theory h as gained w ide sp rea d acceptancedown through the present day. It will be shown that thisaccount of time itself forms the basis of an atheological argument. Had its import been understood by his Christian followers, they would have tar and feathered rather thansainted him. The notion of God's eternality then will be subject to critical scrutiny and will be shown to render him anonperson and thereby not religiously available.

In Book 11 of his Confessions, in wh ich he d evelop s this

theory of time and eternity, Au gustin e begins with this Man -ichaean formulation of the creation-immutabili ty argument:

Lo are they not full of their old leaven, w ho say to us, "Whatwas God doing before He made heaven and earth?" "For if (saythey) He were unemployed and wrought not, why does Henot also henceforth, and for ever, as He did heretofore? For didany new motion arise in God, and a new will to make a creature, which He had never before made, how then would thatbe a true eternity, where there ariseth a will, which was no t?"1

The following is an explicit mounting of this argument,with 1-3 comprising the initial set:

37



The creation-immutability argument

1. The world cam e into existence at som e past time, say t0;2. The world exists at some time if and only if God wills

that it does;3. God is immutable;4. At f0 God wills that th e wor ld exist [from 1 and 2];5. There are times earlier than f0, say (_„ at which the

world does not exist [from 1];6. At f_, God does no t will tha t the wor ld exist [from 2 and

51;7. God has a pro pe rty at one time that he lacks at ano ther

[from 4 and 6];8. God is not immutable [from 7]; and9. God is immutable, and God is not immutable [from 3

and 8].

Augustine's initial response is the old knee slapper thatprior to t0 God was preparing the fires of hell for those whowould give the creation-immutabili ty argument (and who

said tha t the good saint did n't hav e any sens e of hum or). Buthe immediately points out that this humorous excursionhardly serves to neutralize the argument (12). His first serious response is to deny that 5 follows from 1. His exactreason is not completely evident.

He asks "whence could innumerable ages pass by, whichThou madest not, Thou the Author and Creator of all ages?"and the n adds that before heaven and earth there wa s no time,"since God created the world and time together (13). Thenegation of 5 follows from time being ontologically dependent upon God's will and God's willing that the world andtime come into bein g togeth er. This w ay out of the argu m entseems to permit God to create an empty time, that is, absolute times. Being all-wise, God will make sure that in anyworld in which he avails himself of this creative option, hedoes not compromise his immutability by creating the cosmos at one of these times.

Augustine, however, has another way of blocking theinference of 5 from 1. In the sum m ation at the end of Book 11,he say s, "Let the m see therefore, that time cannot be witho ut

38




created being" (30). Earlier he had said something that also

could be construed as affirming the ontological dependenceof time upon a world of created beings: "If nothing passedaway, t ime past w ere not; and if noth ing w ere com ing, a t imeto come were not; and if nothing were, t ime present werenot" (14). Plainly, 5 is at variance with time's ontologicaldep end enc y u po n created beings. Time, i t m ight be ad de d, isstill ontologically dependent upon the will of God, since,given God's sovereignty, no beings exist without his willingthat they do.

Ne ither of these tw o ways out of the creation -imm utabili tyargument is satisfactory. First, they are not sufficiently perspicuous in that they leave us with an unresolved mysteryconcerning the manner in which God and his acts of willingare related to time. Second, a variant of the argument can beconstructed that does not require 5's com m itme nt to a bsolutetimes. It goes this way:

10. Some object, say O, within a world of objects, exists atf, but not at f2;11. An object exists at a time if and only if God wills that

it does;12. At tt God wills that O exist [from 10 and 11];13. At t2 God does not will that O exist [from 10 and 11];7. God ha s a pro pe rty at one time that he doe s not hav e at

an oth er [from 12 an d 13];8. God is not immutable [from 7];3. God is immutable;9. God is immutable and God is not immutable [from 8

and 3].

Herein there is no need to bring in absolute time. The argument is perfectly compatible with the theses that time isontologically dependent upon both God's will and the existence of created beings.

That these two theses do not succeed in neutralizing thecreat ion-immutabil i ty argument hardly is news to Augustine; for if he thought that they did, he would not have felt itnecessary to contrast God's eternal mode of existence with

39



The creation-immMfy$$i$$ argument

the temporal mode of existence of created beings. The pointof this contra st is to show th at neithe r God no r his actions aresubject to any temporal distinctions or determinations. Theup sh ot is that thes e two argum ents are to be rejected becau sein step s 4 and 6 of the first a nd 12 and 13 of the s eco nd , theylocate God's acts of willing in time. This is Augustine's deepobjection to the creation-immutability arguments. HarryWolfson once said that if we are to contrast two things, weshou ld k no w a little bit about at least one of them . A ug ust inesets a higher standard and thinks that we should know a

good deal about both of them. It is for this reason that heundertakes an analysis of both time and eternity, though hisexten ded analysis of time goes far bey ond w hat is ne ed ed forthe pres ent p urp os e of neutralizing the creation-imm utabili tyarg um en t. A nd it is this act of philosophical sup erero gatio n,as we shall see, that makes him a suitable candidate for tarand feathering. Atheological arguments are lurking everywhere

Augustine's notion of God's eternity closely follows thenotion of eternity articulated by Plotinus. An eternal beingenjoys a n illimitable life that adm its of no distinction be tw eenearlier and later phases, or past, present, and future, eventho ug h this life is durational in som e sens e. This conceptionof eternity has a striking similarity to Whitehead's conceptionof an actual occasion, and even Bergson's cotton-candyishduree, bo th of which are durational b ut no t internally structured successions. But there is a crucial disanalogy. Whereas

an actual occasion bears temporal relations to other actualoccasions and admits of temporal subdivision after it hasbecom e actual, God 's eternality adm its of neither these determinations. God's eternity is a totum simul, a nunc stans, inwhich all of his "years" are had at once. He exists in aneternal present - a nontemporal present because it admits ofno co ntra st w ith a past and future. All of God 's acts of w illingoccur within this eternal present and bear no temporal relations to any times or created being s and , according to the doctrine of the divine simplicity, are one and the same act. Andthese timeless acts of willing bring about the existence and

40



The creation-immutability argum ent

occurrence of things in time. Thus, God timelessly wills that

a certain sequence of events occur successively in time. Thisis a case of timeless causation in which the cause is not intime w hile its effects are . It is inte restin g to note that A ug ustine's doctrine of timeless causation is compatible with thecreated world being infinite in its past, in which case Godwould timelessly will that there be an infinite succession ofpa st eve nts . There is no p roblem of ho w G od 's act of will getsinto the act, at w ha t po int in t ime it is perfo rm ed. Those w h othink that an infinite past would preclude God's will beingthe cause of the world are assuming that God's creative actmust occur at some time, an infinitely extended past affording no time at which he can get into the act. Both versionsof the creation-immutabili ty argument make this mistake oftemporally relating God's creative act to its worldly effects.

God's eternality sharply contrasts with the temporallyincomplete mode of existence of created beings. Augustinewrites:

Thy years neither come nor go; whereas ours both come andgo, that they all may come. Thy years stand together, becausethey do not stand; nor are departing thrust out by comingyears, for they pass not away; but ours shall all be, w hen theyshall no more be. (13)

I assu m e that wh en A ugu st ine speaks of our years as comingand going he was not using this expression in the same waymy landlady at college did when she warned me that shedidn' t want any coming and going.

It is this coming and going - this negativity that infectstemporal beings - that is of the very essence of time forAugustine. Time, he says, is composed of past, present, andfuture. The past and future are types of nonbeing, the pastbeing the no longer and the future the not yet. The presentalone is real, things having being only when present. But

even it is shot through with a kind of negativity:

If time present (if it is to be time) only com eth into existence,because it passeth into time past, how can we say that either

4 i



The creation-immteitikiUty argument

this is, whose cause of being is, that it shall not be; so, namely,that we cannot truly say that time is, bu t because it is tendingnot to be ? (14)

Unfortunately, the pre se nt tu rns ou t to be of zero dur ation;for if we were to im pu te a ny finite d ur atio n to it, it w ou ld besubdivisible, if only theoretically or in our imagination, andthereby consist of successive phases. But successive phasescannot be copresent, for whatever is copresent is simultaneou s and thus not successive. The pres en t thereby turns o ut

to be a mere mathematical knife edge at which the not yetturns into the no longer.

This view, not surprisingly, is a breeding ground of paradox; for if the present is of zero duration and entities havereality only when present, it seems that it is impossible foranything to happen or come to pass, since i t doesn't haveeno ugh time in which to do s o.2 A ugu stine considers a particular case in point - the intervals of time itself - but, as we

shall see, it can be generalized to every type of temporalentity.It would appear that it is impossible for any time to be

long, that is, to be of no n ze ro du ratio n, since it m us t be lon gwhen present but the present "hath no space." He asks,"W here then is the t ime w hic h we m ay call long? " con tinuinghis use of spatial metaphors by the use of "where" (21). Thepresent does not have sufficient "space" within which tohouse a finite interval of time, or a finite event for that matter. It is on this ground that he finds it misleading to say, aswe ordinarily do, that it is a long time since some eventoccurred. The use of the present tense makes it appear as ifthe entire lengthy interval of time is present as a whole. Noris it correct even to say that it was a long time, since nointerval of time can be lon g w h e n pas t since the pa st is a kindof nonbeing. R ather, w e o ug ht to say that a pres ent t ime w aslong, that is, was long when it was present, since it can be a

bearer of properties only w h e n it is pre sen t. Unfortunately,since the present is of zero duration, it is impossible that itwas long.

42




This creates the paradox of the impossibility of measuringtime; for we can measure an interval of time only when it ispre sen t, but w ha t is pre sen t is no thin g, tha t is, a time of ze roduration. Notice that the paradox of the seeming impossibility of measu ring time could equally well be directed at the ac tof measuring time rather than at the interval of time that ismeasured: An act of measuring time takes a finite time andoccurs only when present, but the present "hath no space"within which to house it. No one is fast enough to measurea time in no time at all.

Augustine has a way out of this paradox, but one that,unfortunately, is an occasion for a good tar and feathering.His solution consists in questioning the assumption that inorder to measure t ime we must measure something that hasa mind-independent being in rerum. It was because the pastand future did not enjoy such a being that time could not bem eas ure d, given the punctua l nature of the pres ent. Instead ,Augustine proposes that what we measure has only a being-for-thought, an intentional inexistence to use a later phrase.What we measure is not a past interval of time itself but onlywhat we represent through our act of recollection, namely,the intentional accusative of this act. Consciousness, whilenot fully able to resurrect the past and future, can bestow onthem a second-class being-for-thought sort of existence. Thisis the only sort of being these temporal ekstacies can enjoy.

Nor is it properly said, "there be three times, past, present,and to come:" yet perchance it might be properly said, "therebe three times; a present of things past, a present of thingspresent, and a present of things future." For these three doexist in some sort, in the soul, but otherwise do I not see them;present of things past, memory; present of things present,sight; present of things future, expectation. (20)

Not only the past and future, but even the present, is rele

gated to this shadowy being-for-thought type of ontologicalsta tus . To be prese nt is to be repres ente d as pr ese nt by an actof sight. As we shall shortly see, there is an outstanding

43




problem concerning the temporality of this act of sight, aswell as those of memory and expectation. They take a finitetime in which to occur, but the present hath not sufficient"space" in which to house them.

The doctrine of the ideality of time is immediately put touse in explaining ho w w e can me asu re time. We kn ow thatwe "measure not time to come, for it is not yet; nor past,because it now is not"; thus, we must measure "times passing." We perceive things while they are passing and thisleaves us with memory impressions or contents - the beings-

for-thought - that can be compared with each other, one ofthem can be selected as a time unit and another can be estimated to contain so many of these units. This will yield ameasure of time past. Augustine speaks of time as a protraction of the m ind ("Time is no thing else than pro traction . . .of the mind itself ; 26), m eaning a mental re presen tation of aprotracted temporal interval, rather than a temporally protracted mental state or process. He gives an example of howwe mea sure so un ds . The sound, while passing, is "extendedinto some space of time so that it might be measured, sincethe present hath no space."

Both (sounds) have sounded, have flown, passed away, areno more; and yet I measure, and confidently answer . . . thatas to space of time this syllable is but single, that double. Andyet I could not do this, unless they were already past andended. It is not then themselves, which now are not, that I

measure, but something in my memory, which there remainsfixed. It is in thee, my mind, that I measure times.(27)

The impression, which things as they pass by cause in thee,remains even when they are gone; this it is which, stillpresent, I measure , not the things which pass by to make thisimpression. (27)

Augustine even goes so far as to say that a "long past" is "a

long memory of the past" (28). The latter refers not to thelength of time that the act of remembering takes but to the

44



The crmtion-immittaMlity argument

m an ne r in wh ich it rep res en ts. It is a recollecting som eth ing -

as - being - long - in - t im e - in - com parison - with - some - shor ter-time-unit.Wherein lies the heresy of this theory of time and the man

ner in which it is m easure d? To be su re it is ultra-subjectivisitic,bestowing a m ind -de pe nd en t existence on the past , pres ent ,and future; and it is no easy task to square such an idealistictheory of time with a fairly literal reading of Genesis. Moreseriously, it com pletely d en ies the reality of time, if w e acc eptA ugu stine's claim that t ime is com pose d of past, p rese nt, an dfuture. A nd this, as we shall see, can't be squa red with esse ntial tenets of theism. How does it wind up denying the veryreality of time? It was previously claimed that Augustine'sparadox of the impossibility of measuring time can be fullygeneralized. When it is, the result is tha t there are no tem po ralentities, an d th us time itself is un re al. It is as su m ed by A ug ustine that only the present is real and moreover is of zeroduration. Given, furthermore, that something is real, that is,occurs or exists, only w h en it is pr es en t, it resu lts in th ere no tbeing any intervals of nonzero duration; there is nothing tobe m easu red if w ha t is to be m eas ure d is som ethin g in rerum.But by the same reason it follows that there aren't any mentalacts of representation, such as memories, perceptions, andexpectations; for these acts, like intervals of time, require apresent of nonzero duration in which to occur or exist. Theseacts of representing things as past, present, and future are

not themselves in time. Their being consists in their beingrepresented, but what represents them is not i tself temporal.This sounds like Kant, or one legitimate way of interpretinghim, according to which things are not really in time but areonly represe nted as in time by a no um en al self w ho se acts arenot themselves in time. Nothing really is in time

What/Augustine has in effect done is to create the following "disappearance of t ime argument":

14. Only the present is real;15. The present is of zero duration;

45



The creation-imm utability argum ent

16. Temporal entities (including intervals of time) occur or

exist only when present [from 1];17. Temporal entities require a finite time within which tooccur or exist;3

18. No temporal entity occurs or exists [from 15-17];19. Time is unreal [from 18].

A contemporary philosopher might try to escape this argument by espousing a B-theory of time according to whichtime, pace Augustine, does not consist of past, present, andfuture - the A-series - but only a series of events runningfrom earlier to later - the B-series. Not only would it beunacceptably anachronistic to read a B-theory of time backinto A ugu stine or his pred ece ssors, b ut this theo ry, as will beargued in Ch apte r 3, is unacc eptable in its ow n rig ht be cau sea proposition reporting an event's position in the A-series isnot reducible to one that reports its position in the B-series,pace the B-theory of time.

The disappearance of t ime argument turns into an atheo-logical argument when it is combined with the theist 's commitment to the reality of time. If time is unreal, certainessential tenets of theism must be rejected. There really is nopast Creation and Incarnation, nor a future Resurrection andJudgment Day. The very ontological distinction between thecreator and his creation dissolves. Christianity and Judaismcontrast sharply with mystically based religions in their viewof time, for not only do they hold time to be real but impor

tant as well. In contrast to the cyclical views of time andhistory that were common to all archaic civilizations, Christianity and Judaism have a linear view of time in which eachevent is unique and imbued with significance. There are noinstant replays: Each of us has only one chance to make it.These biblically based religions, certainly, cannot be reconciled with the unreality of time.

Obviously, Saint Augustine has lost his way, unwittinglygiving us the premises of the disappearance of time argum ent, in violation of his very ow n theistic com m itm ent to th ereality of time. Fortunately, his solution to the creation-

46



The creMim-4mmMbiUty argument

immutabili ty argument, based on God being an immutably

eternal b eing w ho se creative act bears no tem poral relation toany w orldly time or even t, such as the universe's coming intoexistence, does not require him to go on and commit this actof philosoph ical hara-kiri. He just got carried aw ay, m esm erized by surface grammatical analogies and spatial metaphorsinto thinking that time is something very queer indeed. It isnot my purpose to find a way out of the disappearance oftime argu m ent. There mus t be a way out, since we kno w thattime is real. Even a mystic should be unhappy with the argum en t, since the denial of time's ultimate reality, based on w ha tis revealed thro ug h m ystical experience, doe s no t require o neto have a muddled conception of what we take this mereappearance, t ime, to be. My purpose, instead, is to probeAugustine's account of God's eternity and manner of connection with worldly time, and, in particular, to see whether itleaves us with a religiously available God - the problemadumbrated in Chapter 1. The discussion of God's eternality

will be concerned, first, with the intrinsic features of thiseternity and, second, with its extrinsic relations to the temporal world. It will be found inadequate in both areas.

God's eternity is not to be confused with the sort of timeless existence enjoyed by such denizens of the Platonicheaven as numbers, properties, and propositions. While theformer is like the latter in not being subject to any temporaldeterminations or distinctions - that is, does not not have apast, present, or future, stand in temporal relations, or endure in time - it involves more than mere timelessness.Augustine says that "in the Eternal nothing passeth," so farimputing to it nothing more than the sort of timeless existence of the aforementioned abstracta; but he then goes onto say that "the whole is present" (11). "Thy years neithercom e n or go. . . . Thy years stand togeth er, because theydo stand" (13). God has an illimitable life that is comprisedof something like years, but, unlike ordinary temporal years

that tem pora lly succeed each other, these "y ear s" are hadall at once in a single eternal or timeless present. The sameview of eternity is expressed b y Boethius in the Consolation of

47



The creaUonîmmuttiMlity argument

Philosophy (Book 5, Prose 6): "Eternity . . . is the com pletepossession all at once of illimitable life." In explaining thisdoctrine, Stump and Kretzmann say that i t involves an"infinite du ratio n" th at is non tem po ral because i t involves nointernal succession.4 At other places they speak of it as an"atemporal du rat ion."5 Just as there is a non tem po ral sort ofduration, there is a nontemporal sort of presentness that ishad by God, a present that has no contrast with a past andfuture.

The first thing that must be pointed out is that these

descriptions of God's eternality, if taken literally, involvestraight-out contradictions. The notion of a "timeless duration," judged by the stan dar ds of ordinary languag e, is acontradiction in terms. The OED says that "duration" is"lasting, continuance in time; the continuance or length oftime; the time during which a thing, action, or state continues ." It says no thi ng abo ut a timeless du ration . Similarly, theclaim that God has all his years at once is contradictoryaccording to our ordinary ways of using language. By definition, numerically distinct years are successive. There areother contradictions that will be explored later. Stump andKretzmann seem oblivious to these contradictions, and thusma ke no effort to find so m e way to sur m ou nt th em . And thisis what I will now try to do.

W hile such ph ras es m ake n o sen se according to the logic ofordinary discourse, they do make sense when interpretedwithin the mystical tradition, with its grounding in unitivemystical experiences. The doctrine of the timeless or eternalpresent bears a striking similarity to the Zen doctrine of theeternal now according to which all times are present, therebeing no past or future distinct from the present moment.During the summer of i960, I was privileged to spend anafternoon with Daisetsu Suzuki - Mr. Zen Buddhist to theWest at that time - and in the course of the conversation Ichallenged the Zen doctrine of the eternal now with absur

dity. "Present" is used without the needed contrast with"past" and "future." This causes language to idle, just as itwould if we were to continue to use first-person discourse

48



Tile cfi0iion-itnmut»bility argument

but eliminated from our language the needed contrast with

second- and third-person discourse. He listened intently tothis polem ic an d th en replied that I w as ind ee d right that theZen doctrine of the eterna l now is ab su rd from the stan dpoint of ordinary language or common sense but that, nevertheless, it is a fact

Under fear of being swatted with a broom, I venture thisanalysis of what he might have meant, judged by the rulesfor the use of ordinary language, the Zen doctrine is meaningless; but, if I w ere to have a satori (a mystical exp erienc e),I would both understand what i t meant and know that i t istrue. Leaving aside the issue of the cognitivity or truth ofsuch a my stical experience, w hich will be the topic of C ha pte r8, the following reply could be m ad e to his claim th at only themystic will be able to understand the doctrine. I, along witha host of other nonmystics, understand this doctrine alongwith other mystical claims, such as those concerning timelessduration and an illimitable life that is had all at once. Ifnonmystics did not understand the writings of mystics, therew ou ld n't be such a hefty sale of mystical literature w ithin the"straight" community.

This belies the claim of mystics that their experience isuniquely ineffable, being beyond any conceptualization ordescription. No doubt, i t cannot be described while one ishaving it, but that is also true of nonmystical experiences,such as wrestling with an alligator, that admit of subsequent

description. Nor is the description of a mystical experience asubstitute for it in that it is qualitatively isomorphic with it,but, again, this is true ot nonmystical experiences, such as asense experience of a red patch. In both cases, the description, though not a phenomenological replica of the experience it describes, is such that it enables us to identify thisexperience were we to have it.

I believe that th e real reaso n for the mystic's claiming som e

sort of unique ineffability for their experience is to be foundin the inestimable significance and value it has for them. Itseems that the more highly we prize some experience, themore we shun applying concepts to i t . Like the composer

49




who shu ns writing program n otes for her sym ph on y because

she fears, and rightly so, that eventually the reading of theprogram notes will take the place of actually listening to themusic, the mystic is afraid tha t the concep ts by wh ich the experience is described will become surrogates for this experience itself. Both persons are telling us by their refusal toconceptualize their experience that it is the direct experienceitself that counts and that language is a very poor substitute.

But how is it that we can understand these claims withouthaving had mystical experiences? I don't know, nor do Iknow of any satisfactory account, but I do know that weunderstand what the mystics are saying, in spite of theirlanguage appe aring, on purely syntactical gr ou nd s, to violatethe ordinary canons for informative discourse. If a sports-caster were to say, "Joe Montana is dropping back to passwhile remaining in place," my reaction would be, "What aschmuck, he is contradicting himself, but when it is said inthe Isa Upanishad, in reporting the pheno m enological conten t

of the mystical experience, that it stirs without moving, Iun de rst an d. I apply a different logic to the mystical claimthan to the empirical one; for instance, I do not ask, as Iwould in the Joe Montana case, whether it stirs and remainsat rest at one and the same time, for unitive mystical experiencesare phenomenologically timeless, containing no distinctionsbetween different times within their content.

Similar considerations hold for sayings like "It's full and it'sem pty" and "The stone an d the tree are o n e ." If the former issaid by someone peering into the fridg, he's a candidate forBellevue; but if it is said by someone who has just finisheddoing zazen, we might consider boo king him on a lecture tou rof college campuses. If the gardener says that the stone andtree are one, we consider replacing him; but if it is said byMeister Eckhardt in a sermon describing a "nature" type ofmystical state, we say, "H ow pro foun d " Ob viously , the contexts in which these thing s are said is crucial in sup ply ing thekey for their interpretations, including the sort of logicalstandards we appeal to in understanding and evaluatingthem . I do n't kno w w hat th e "logic" of m ystical discou rse is,

50




but I do know that it is radically different from that of

"straight" discourse.I venture to speculate that the reason nonmystics canunderstand these "contradictory" notions is that they havehad experiences that are steps along the way to fully developed unitive mystical experiences: altered states of consciousness, such as drunkenness and drug-induced highs,the em path etic identification with an oth er in a love relationship , especially when it involves the joining of the flesh( I've got you under my skin"), aesthetic experiences of reconciliation between opposites, for instance, the union of joyand tragedy in the allegretto movement of Beethoveen's Seven th S ym ph on y (an d do n't ask if the joy and tragedy occur atone and the same time, since they permeate the entire experience of the movement), and last, but not least, the poor-man's "oceanic experience" upon sinking into a hot tub.

Too often the rationalistic theist muddies the waters bytrying to show that the notions of a timeless present and

timeless duration make sense according to the canons ofordinary discourse. This is clearly evident in the Kretzmann-Stump exposition. At the heart of their account is the claimthat the timeless and temporal present, along with timelessand temporal duration, are species of the same genericnotions of present and duration respectively. (As we shallsee in the next chapter, they also claim, with equal implausi-bility, that there is a generic notion of simultaneity that hasboth a temporal and nontemporal species.) What is basicallywrong with this account is that it runs together incommensurable languages. If the mystical notions of a timeless presen t an d a timeless d uratio n w ere species of the same gen us asare the temporal present and temporal duration respectively,there should be considerable overlap in their logics. Theysh ou ld at least ob ey the same laws of logic, such as the law ofnoncontradiction. But, as seen above, they don't. There isn'tany ordinary sense of duration other than the temporal one.

By cutting the notion of the eternal present loose from itsmystical roots, they block the only way in which sense can bemade of it and thereby do a great disservice to the doctrine

51



The creation~immuitiMltt§ mgwnmt

that they attempt to defend against the charge of meaning-lessness.

Their attempt to make rational sense of the notion of atimeless mind also commits the sin of cutting a mystical doctrine loose from its mystical roots. God has an illimitable lifeand thus must be alive and possessed of a mind; but he hasthis "life" all at once. Thus, he must enjoy a mental life ofsome sort that does not involve any kind of temporal succession or duration. The proper thing to say about this is that itis a mystically ro oted notion that m ake s no sense in terms of

what these expressions mean in ordinary language. Maybethe quickest and m ost direct way of sho w ing the absurdity ofa timeless mind is as follows: A mind is conscious, and consciousness is a temporally elongated process.

Obviously, not everyone agrees with this. Stump andKretzmann attempt to make good ordinary sense of a t imeless mind by claiming that there are many psychological ormental states that do not require temporal duration. Inductive support can be given for my claim that consciousness isessentially temporal by meeting any alleged counterexamples, such as those given by them. They give knowing as acounterexample:

Knowing seems to be the paradigm case; learning, reasoning,inferring take time, as knowing does not. In reply to thequestion 'What have you been doing for the past two hours?'it makes sense to say 'Studying logic' . . . but not 'Knowing

logic'.6

Willing and awareness are other examples they give of mental states that on e can hav e a tem pora lly.

While they are right that it makes no sense to ask "Howlong have y ou b een kno w ing logic?" since know ing logic is astate, not a process, they fail to see that it does make sense toask "How long have you known logic?" and "When did you

begin (cease) to know logic?" Furthermore, one could not bein the state of knowing that p w itho ut h aving disposit ions toeng age in vario us sorts of epi sod es, such as having occu rrent

52



The creation-imm utability argument

beliefs that p employing that p in one's reaso ning , an d so on,

just as one could not be in love without having dispositionsto perform various actions, such as caring for, protecting,cherishing, and so on, the loved one. While these states fillt ime in a homogeneous manner, they require having dispositions to engage in temporal episodes, occurrences, orprocesses. This raises an especially acute problem for thetimelessly eternal Deity of the great medieval theists, who issupposed to have pure actuality and therefore no unrealizeddispositions. Thus, God's being in one of these states, suchas loving or knowing, would require that he not have anyunrealized dispositions but instead be engaging in all of theactions that one in such a state is disposed to do. But theseactions take time to perform.

The same points hold with respect to the extrinsic aspectsof God's timeless eternality. Our ordinary concept of causation involves some sort of temporal relation, which can bethat of simultaneity, between cause and effect. This holdseven for the notion of agent causation in which the cause isnot an event but a person. God's timeless causation is aspecies of such agent causation but one that has no temporalrelation to its temporal effect. Our ordinary concept of causation does not make room for timeless causation; but itmight make sense within the mystical tradition, with its ideaof the eternal intersecting time at every moment. Mysticsclaim to directly experience this intersection between time

and eternity - that all times are one, namely, the present.And we might be able to make sense of the idea of timelesscausation on such an experiential basis. Again, we seemforced to give a mystical interpretation of God's timelesseternity.

So far it has been argued that the intrinsic and extrinsicfeatures of God's timeless eternity can be understood only inmystical terms. But the theist must pay a significant price forgoing this mystical route, namely, he winds up with a Godwh o is a non person. A person has a mind an d thus en du resin time. It has purposes, intentions, and goals and interactswith other perso ns, wh ich, again, require it to have a tem po -

53





The ermtion-immutabitity argument

seemed to require that he be timeless, for only such a being's

existence can be deduced from its essence.The underlying problem of theism is that it wants its Godto play contradictory roles - to be both a perso n an d a mystical reality that is beyond being, time, and distinctions. Thedoctrine of the Trinity is an attempt to resolve this contradiction by giving God, through the different members of theTrinity, both of these aspects, Jesus the Son being the personal God to whom we can relate in a personal way and Godthe Father being the mystical reality - the absolutely simplebeing who is beyond being and distinctions - to whom wehave access through unitive mystical experiences. This doctrine is not all that different from Eastern forms of mysticismin which an eternal one or undifferentiated unity somehowundergoes emanation into lesser deities of a personal sort,which em ana tions even tually flow back into its on en es s. I amconvinced that the doctrine of the Trinity makes sense onlyon mystical grounds, for, on the surface, it appears contra

dictory by its identifying the three members of the Trinitywith each other but refusing to allow them to have all theirproperties in common, thereby violating Leibniz's law.

Not surprisingly, Stump and Kretzmann also try to makegood rational sense of this doctrine. Their way out of theproblem of how we can identify the persons of the Trinitywitho ut hav ing to apply L eibniz's law to them is to hold thateach of them has its properties qua that member of the Trinity. Th us, the propo sition that Christ died is to be un de rsto odas "Christ with respect to his human nature (or qua man)died."8 If ever a pred icate w ere referentially tra ns pa ren t, it is"dies" or "gets crucified," pace what they claim. Being crucified or dying, unlike being loved, does not occur under aconcept or desc ription. It wo uld hav e been a bad joke for t heRoman executioners to have said to Jesus, "Don't feel bad.We're crucifying you only qua Son, not qua Father."

Concerning the extrinsic features of God's eternality, onlythat of timeless causation has been considered. An evenmore serious problem concerns the epistemological relations

55




that such a being has to the temporal universe. It appearsthat his timeless immutability precludes his knowing certainfacts about the temporal world, such as what is happeningright no w . This is one of the reasons that it se em s impossib leto have personal interactions with such a God. The nextchapter will consider the omniscience-immutability argument against such a God, and it will be argued that it succeeds. And, as a result, we will have to make due with atemporal God, the nature of which will partially emerge asa result of this argument, to which we now turn.

56



Chapter 3

The omniscience-immutability argument

This argument 's initial set is composed of these three propo*sitions:

1. Go d essentially kn ow s all an d believes only true pr op ositions;

2. God essentially does not change from one time to another in respect to any nonrelational property; and

3. There are true temporal indexical propositions (to becalled "A-propositions") to the effect that certain events and/or times are now past, present, or future.

The ism's co m m itm en t to 1 and 2 resu lts from its conceivingGod to be essentially both omniscient and immutable; and itis committed to 3 because, as pointed out in the previouschapter, theism (at least as represented by Judaism andChristianity) is a historically rooted creed that accords notonly reality but significance to the sort of A-facts over which3 quantifies, such as that the world came into being in thepast, has reached a certain stage in its spiritual developmentat present, and awaits a future day of judgment.

The qualification on God's omniscience, that he believesonly true propositions, which will be dropped in the futurefor the sake of brevity, is due to our not wanting to allow anomniscient being to go around having any false beliefs, even

if he does know every true proposition. Notice that an omniscient being must not only know of every true propositionthat it is true but must also know every true proposition. Aswe shall see later in this chapter, these are not the same,

57



The omniscienee-immmtabUity argument

since one could know that some proposition is true withoutknowing it.

The reason for the restriction of God 's imm utability to no nrelational pro pe rties , in w hich a no nrela tiona l prop erty is on ethat an indiv idua l could po ssess eve n if there were no t to existany other non un ive rsal pa rticulars, save for spaces an d tim es,is to escape such "Cambridge" changes in God as beingtho ug ht abo ut by Jones at on e time bu t not at ano ther, hereinthe change being in Jones rather than in God. How the argument proceeds from here depends on whether i t has as i ts

target the timelessly ete rnal God of St. Aug ustine or the omni-temporally eternal God of the Bible, that is, a God that endures throughout a beginningless and endless t ime and logically could not begin or cease to exist. If it is the former, itargues that a t imeless being cannot know an A-proposition,and therefore God isn't omniscient; and, if the latter, that inorder literally to keep up to date, God would have to changein his temporal indexical beliefs from one time to another,and thereby is no t strictly im m uta ble . A contradiction resultsin either case. Since we are intent on following u p the prece ding chapter's concern with the viability of the conception ofa timelessly eternal God, we shall first consider the versionof the argument that is directed against this sort of Deityand then go on to see how it fares w h e n it is dep loyed againstthe omnitemporal God. It will turn out that only the latter escapes the argument, and only if we give up his strictimmutability.

No contradiction can be deduced from the initial set 1-3. Itis necessary to add the following supposedly necessary additional premise:

4. It is conceptually impossible for a timeless being toknow an A-proposition,

from which it follows, in conjunction with 3, that

5. God (conceived of as timelessly eternal) does not knowevery true proposition,

wh ich contradicts 1. Th us, God bo th is an d is not om niscient.

58



The omniscience-immutability arguntettt

Faced with this contradiction, our theist must deny at least

one of the prem ises of the abov e arg um en t, either 1, 2, 3, or 4.Premise 2 cannot consistently be den ied, since God's im m utability is a logical consequence of his being timelessly eternal: A being that is not in time at all cannot change in thenon-Cambridge properties it has from one time to another.Denial of 3 is not an attractive optio n, for the reaso ns alread ygiven, nor is the denial of 1, since omniscience is usuallytaken by theists to be at least a soft- if not h ard -co re pr op er tyof God. It is the added premise 4 that the theist will want todispute, many of the great medieval theists, for example,arguing that God could timelessly know A-propositions.Obviously, the defender of the omniscience-immutabili tyargument owes us an explanation and defense of 4, since it isnot clear, much less acceptable, without further ado.

The first thing that must be done toward this end is toclarify the very troublesome notion of an A-proposition. Wewill restrict our treatment to contingent propositions. An

A-proposition is expressible by (but not necessarily only by)the use or intentional tokening of an A-sentence, that is, asentence that is such that successive tokenings of it by thesame person, at the same place, and with the same meaninglogically could express propositions that differ in truth-value.1 It must also be stipulated that the sentence is usedwith the same m ean ing on each of these occasions. The grammatical device that is used in Indo-European languages to

render sentences nonfreely repeatable in time is the tense ofa verb, while in Ch ines e a nd Eskimo it results from the inflection of a noun. The meaning of an A-sentence is given by its"tokening rule," which specifies the conditions under whichan intentional tokening or use of it expresses a true proposition or, as some wo uld pu t it toda y, a rule that takes us froma context of tokening, of which the relevant feature is thetime of tokening, to a truth-value. (Throughout this book Iam adop ting the con ven tion that a token ing of "no w " or othertemporal indexical term refers to the time of tokening ratherthan the time of its being perceived, which is our usual, butnot universal convention.) For instance, the tokening rule for

59



The omniscknce-immutaMMty argument

the English sen tence "Event E is now pr es en t (past, future )"requires of a true tokening of it that it be simultaneous with(later than, earlier than) event £.

A B-proposition, however, is one that is expressiblethrough the use of a freely repeatable sentence and describesa temporal relation between two events and/or times, forinstance, "Event £ is (timelessly) earlier than event E'. " It isfreely repeatable in that it is logically impossible that successive tokenings of it, with the same meaning, can expresspropositions that differ in truth-value. A freely repeatable

sentence will be called a "B-sentence."Care has been taken to give only a sufficient condition forbeing an A-proposition, namely, being expressible by the useof an A-sentence, thereby leaving it open whether an A-proposition also is expressible by the use of a B-sentence. Ifthe latter were to be the case, it could be shown that everyA-proposition is identical with some B-proposition, tha t is, aproposition expressible by the use of a B-sentence. Thiswould afford the theist a way out of the omniscience-immutable argument, since God could timelessly know anA-proposition via timelessly knowing the B-propositionwith which it is identical, which certainly is possible, if timeless knowledge itself is possible. The previous chapter castdoubt on the possibility of the latter, and this chapter willargue against the possibility of a B-reduction, that is, thedoctrine that every A-proposition is identical with some B-proposition, thereby presenting a two-pronged attack

against this strategy for escaping the argument. Appeal to aB-reduction, however, is only one of many ways of rebuttingit, as we shall see.

Given this account of an A-proposition, why should weaccept 4's contention that only a being in time can know anA-p roposition? A. N . Prior has given both a negativ e andpositive defen se of 4 / The former consists in an argu m en tagainst the chief gro und for deny ing 4 - the B-reduction.

Since, as we shall see w he n w e consider Cas taned a's rebuttalbased on a timeless way of knowing an A-proposition in

60



The otnrikmence-immutabUity argument

oratio obliaua, there are oth er gro un ds for deny ing 4 than that

supplied by the B-reduction, Prior's negative defense suffersfrom incompleteness. We will consider i t anyway and thengo on to criticize his positive argument for 4.

Against the B-reduction's claim that every A-proposition isidentical with some B-proposition, Prior repeats his earlierobjection from his article "Thank Goodness That's Over."3

When we say of some particularly unpleasant experience -the final examinations - "Thank Goodness that the finalexaminations are over (past)," we are not expressing thesame proposition as we would were we to say "Thank goodne ss that the final exa m inatio ns occur (timelessly) at t7 (or a re(timelessly) 2000 years later than the birth of Christ)"; forwhy should we be glad about the latter? In brief, Prior'sargument is that the A- and B-propositions in question arenot identical, since one could be glad about the former butnot the latter. Prior's point can be generalized to all propositional attitudes, since one could believe, know, and the like

that the final examinations are now past without having thesame propositional attitude to the B-proposition that theseexams occur (timelessly) at t?.

Herein Prior is implicitly appealing to (at least the only-ifpart of) the following propositional-attitude-based criterionfor propositional identity:

PA. Proposition p is identical with prop osition a if an d on lyif it is not possible that som eon e w ho u nd ers tan ds both p andq have a propositional att i tude at some time T toward one ofthem that she does not have toward the other.

A similar type of propositional-attitude-based criterion canbe given for determining when two sentences express thesame proposi t ion:

PA'. T he sente nce s "s" and " s '" express the same proposition at some time T if and only if it is not possible thatsomeon e w ho un ders tand s the meaning of both "s" and "s' "hav e a different propo sitional a tt i tude at T toward wh at "s"expresses than toward what "s' " expresses.

61





The emn isdence-imnttitabtiity argum ent

mind , and more o v e r d i d so w i thou t a ccep t i ng e i t he r t he B-

reduct ion (which he re jec ts ) or the poss ib i l i ty of t imeless lyknowing an A - p ropos i t i on i n oratio obliqua. I n o t h e r words ,Geach c la ims tha t God can t ime less ly know an A-propos i t ionin oratio recta, even th ou gh n o A -prop os i t io n i s iden t i ca l w i than y B-propos it ion . Geac h beg ins by g iv ing th e fo l lowing s imple ru le fo r God ' s omnisc ience : " 'God knows tha t p' is true ifand only if the plain 'p ' i s t rue ." He goes on to say:

We need not lose our he ads in dea ling with te nse d prop ositions; we need only stick to the simple rule I have just given.In 1939 it was true to say 'H itler is alive '; it was th erefo re trueto say in 1939 'God knows that Hitler is alive'. In 1970 it wastrue to say 'Hitler is dead'; it was therefore true to say in 1970'God knows that Hitler is dead'.7

No t ice tha t no t emp ora l qua l i f i ca t ion i s m a d e on w h e n G odknows . Wh a t eve r he k now s , he knows t ime l e s s ly. Wha t i st emp ora l l y qua li fi ed is t he p r opo s i t i on h e k n o w s and w he nwe t ime-bound c rea tu res can t ru ly say tha t God t imeless lyknows t h i s p ropos i t i o n .

Th ere i s an ambig u i ty in Ge ach ' s p r es en ta t io n tha t m us t b eclari f ied before we can ev alu ate h is po s i t ion . W h e n I n o w say"G od k no w s tha t Hi t l e r i s no w d e a d ," am I a sc r ib in g to Godknowledge o f the ve ry same A-propos i t ion as the one Iexpress by my presen t use o f "Hi t l e r i s now dead" in the

oratio obliqua c l aus e? T h i s que s t i on ge t s do w n to w he t h e r t heoccur rence o f "now" in the oratio obliqua c l au se e x p re s se sGod ' s p resen t - t ense index ica l mode o f r e fe rence to thep re sen t a s w e l l a s my own . As He c to r-Ne r i C as t aneda ha s soab ly shown, an index ica l word , such as "now," a lwa ys ha sthe wides t scope in a sentence , even i f i t occurs in an oratioobliqua c l ause , beca use it exp ress es th e spe ak er ' s o w n ind ex ica l reference .8 The que s t ion i s w he th er in case s in w hic h anindexical word occurs in an oratio obliqua cons t ruc t i on a t t r i bu t i ng a belief, des i re , and the l ike to someone i t a lso exp re s se s t h i s pe r son ' s own index i ca l r e f e r ence . C as t anedade n i es tha t it do es . In m an y cases it do es no t , e spec ia l ly th os e

63



The omniscienceHtftmutabUity argument

involving personal indexicals. If I say "Jones knows that I amtir ed ," the occurrence of "I" expresses my own first-personalindexical reference, not also Jones's; for were Jones to haveemployed this mode of indexical reference, he would havereferred to himself rather than me. In this case, the oratioobliqua clause abstracts from ho w Jones refers to me andthereby underdetermines the proposition known by Jones.Bu t the issue is not so clear cut with "n ow ." Ord inar y usa ge ,in my opinion, is not sufficiently precise to determinewhether an occurrence of "now" in an oratio obliqua clauseat tr ibut ing a belief, and the like to a person does or does notindicate this person's indexical reference as well as thespeaker 's .

If we take such an occurrence of "n ow " as exp ressing onlythe speaker's indexical reference, we abstract from how theperson in question refers to the present moment and therebyunderdetermine the proposition he believes, and so on. Onth is construa l, Geach would ha ve failed to sho w tha t God can

know an A-proposition; for it could be true that God knowsthat Hitler is now dead without God making a present-tenseindexical reference to the present moment and thus withoutG od kno win g any A-proposition. In this case we und erde termine the proposition known by God.

We can, of course, simply decide to construe an oratio obliqua occurrence of "n ow " or any indexical wo rd as expressingboth the speaker's and the believer's indexical reference; and

we could adopt the convention that when and only when anoratio obliqua construction is represented in the explicit para-tactic manner are we so to construe it. If I say "God knowsthat (this). Hitler is now dead," the oratio obliqua occurrenceof "now" expresses both my and God's indexical reference,and, thereby, I do attribute to God knowledge of the verysame A-proposition as I express by my use of "Hitler is nowdead" in the oratio obliqua clause. Th us, Geach w inds u p withthe seemingly bizarre position that God is able to make apresent-tense indexical reference to the present momentwithout doing so at any time9 But many a bizarre claim is true,

64




a n d t h u s it b e h o o v e s u s t o f in d s o m e c o g e n t a rg u m e n ta g a i n s t G e a e h .

Such an a rgument i s to be found in Rober t Coburn ' s ve rs i o n of t h e o m n i s c i e n c e - i m m u t a b i l i t y a rg u m e n t , d i r e ct e daga ins t the t ime less ly e te rna l De i ty :

If a being is omniscient, then presumably it follows that thisbeing knows everything which (logically) can be known. Butit is easy to see that an eternal being could not know everything wh ich (logically) can be kn ow n, an d this is becau se som e

of the facts wh ich (logically) can be kno w n, are know able onlyby temporal beings, by beings who occupy some position (orsom e position s) in tim e. . . . ass um e that the idea of a non -temporal kno w er m akes sense. Then ask, could such a know erknow, e.g., that today is May 12, 1962? The obvious answer,I submit, is that it could only if it could use temporal indicatorwo rds. For other wise , i t could no t express and a fortiori couldnot entertain a truth such as the above. But a necessary condition of being able to use temporal indicator words is being anoccupant of time. Hence, God's alleged eternity is logicallyincompatible with his alleged omniscience.10

Before we can p roper ly unders tand and c r i t i c i ze th i s a rgum en t , it m us t be g iven an exp li c it m o u n t i n g . In the fo l lowing ,Coburn ' s use of an indica tor word wi l l be replaced by tha t oft h e t o k e n i n g o f a n A - s e n t e n c e i n w h i c h t h e t o k e n i n g is i n t e nt iona l . Fu r th e r, e ach of i ts p r em ise s i s a ne cess a ry con cep tua lt ru th , a s i s r equ i red , s ince the conc lus ion , 4 , has to be a necessary t ruth i f i t is to be permissible to add i t to the ini t ia l seto f an a theo log ica l a rgument :

4 b . A pe r s o n can kn o w a p r op os i t io n on ly if sh e can t ru lyexpress i t [p remise ] ;

4c. A pe r so n can tru ly ex pre ss a p r op os i t io n on ly if she c ant r u l y t o k e n a s e n t e n c e t h a t e x p r e s s e s t h i s p r o p o s i t i o n[p remise ] ;

4d . A p e r so n can t ru ly ex pre ss an A -pro po s i t ion o n ly ifshe can t ru ly token a sen tence tha t expresses th i s p ropos i t ion[from 4c];

65



The omniscience-mmutttkUity argument

4 c An A-proposition is expressible only by the tok en ingof an A-sentence [premise];

4f. A per son can truly express an A-pro pos ition o nly if s h ecan truly token an A-sentence [from 4d and 4e];

4g. A person can truly token an A-sentence only at a time[premise];

4h. A person who can truly token an A-sentence exists intime [from 4g]; and

4. It is conceptually impossible for a timeless being toknow an A-proposition [from 4b, 4f, and 4h].

This arg um en t is valid, but its pre m ises s tan d in ne ed of further elucidation and justification.

Premise 4b m ust not be taken as requ iring of som eo ne w h oknows a proposition that she can give overt expression to itin some public language. (Let us not forget Gunga Din andJames's deaf mute.) For the sake of ar gu m en t w e can as su m ethe possibility of an in-principle private lan gu ag e an d the re by

permit our Deity to express his thoughts in his own privateDeitese language.To establish 4c we m ust begin w ith th e fact that kn o w in g

entails believing, which in turn entails having either a certainoccurrent thought or the disposition to have this thought.Since God is pu re actuality, he has no unfulfilled disp os itions, although he does have counterfactual possibilities inthat he varies across possible worlds in which he exists. As aresult, all of his thoughts are occurrent. Now, to think something occurrently is to say something to oneself, to express aproposition ev en if only in one's mind or he ar t . H ere w e m us tappeal to an important thesis championed by Peter Geachand W ilfrid Sellars that such internal sa yi ng s are to be u n de rstood as an analogical extension of the overt use of language.This means that we are to construe a person's saying someproposition to herself as involving a mental tokening of asentence that ha s the same m ean ing or pl ay s the sam e role as

does a sentence by which her thought could be publicly expresse d. To sum m arize: Premise 4c req uire s of a kn ow er of aproposition that she can either publicly express it by token-

66





The omm seknce-mmwtflbitit^ argument

in which he exists. Therefore, in every world in which bothGod and A-facts exist, he brings about the occurrence of anA-sentence token for every A-proposition that is true in thatworld. The empirical evidence available to us in this worldindicates tha t this is no t the case The theist, of course , canhave faith that these tokens occur on some remote galaxy orthe like, bu t this is a pretty bizarre thing to have to accept onfaith - a blabber-mouth Deity who must publicly remark oneverything. There should be a new hymn, "We Believe InOur Heavenly Bringer About of A-sentence Tokens." Or he

might offer as an apologia that we wouldn't be able to recognize these tokens were we to come upon them, since theywould be in Deitese.

A far mo re serio us objection to the timeless causation rebu ttal of 4g is that it abuses the tokening rules for A-sentences.These rules specify what temporal relations must hold between an intentional tokening and the reported event, foronly an intention al tokening of an A -sentence expresses an A-proposition. Thus, an intentional tokening of an A- sentenceinvolves the thinking of the thought or proposition expressed by the tokening. This requires that the thinking ofthe proposition expressed by the tokening of the A-sentenceis simultaneous with the tokening. As a consequence, thethinking of the proposition must have the same temporalrelation to the reported event as does the tokening, if therelevant tokening rule is to be satisfied. But this is exactlywhat a timeless God cannot do. God's occurrent believing

and thinking of an A-proposition (remember that he has nodispositional potentialities), unlike the occurrence of the A-sentence token he causes, has no temporal relation to anyevent, and this results in a failure to satisfy the relevanttokening rule. Furthermore, God believes all at once, in asingle timeless p re se nt , every A-p roposition (e.g., that E isnow past, present, and future), which hardly satisfies thetokening rules.

There is another path that leads to 4g. It begins with theconceptual truth that a person cannot express a propositionunless she can understand the expression of this proposition

68



The omniseiittceîmmutabiUiy argument

by someone, in which the "someone" can be herself. Godcannot understand the expression of an A-proposition byanyone, including himself, since to do so he would have todetermine the temporal relation between the tokening of theA-sentence and his perceiving it; but his perceiving it standsin no temporal relations.

At this point, our proponent of the timeless causationrebuttal might take the desperation tack of charging us withbeing und uly ant hro po m orp hic in requiring the Deity to jum pthrou gh the same conceptual ho op s as w e mortals do, in particular, of having to satisfy our tokening rules if he is to haveknow ledge of, and thereby be able to express, A-propositions. Itis sufficient if his occurrent believings of A-propositionsstand in timeless relations to the reported events that arecounterparts to the temporal relations of precedence andsimultaneity that are required by the tokening rules. Let usfirst consider a nontemporal counterpart relation to simultaneity and then consider what such a counterpart would be

for precedence.Stump and Kretzmann supply us with a nontemporal

counterpart to temporal simultaneity that relates a timelesslyeternal Deity to every event and moment of time, and it willbe instructive to determine whether it can help us to makesense of this Deity's knowing A-propositions. They begin byclaiming, as is their wont, that our ordinary notion of temporal simultaneity, that is, existence or occurrence at one and

the same time, is a species of a generic notion of simultaneitythat means "existence or occurrence at once (i.e., together).", a

Another species of this genus is what they call "ET-simultaneity." This relation holds between the timelesslyeternal God and every moment of t ime and event becauseeach of the latter is "obs erved as tem porally p re se nt " by God,and God, in turn, is observable by an observer at each ofthese times as eternally present, supposedly through a mystical experience, though they do not commit themselves tothis interpretation. '3 Their account assumes that successivemystical experiences apprehend one and the same eternalpresent, an assumption that will be challenged in Chapter 8.

69



The omniseienee-mmmiabUity mgummt

Why couldn't God timelessly know an A-proposition, atleast a present-tense one, by standing in this relation of ET-simultaneity to the event reported by this proposition?

In the first place, there isn't any nontemporal genericnotion of simultaneity. The O.E.D. has this entry un de r"simultaneity": "from the Latin sitnul, at the same time." Ifthey were right that the re is a generic sense of "sim ultan eity"m eaning "existence toge the r," it w ou ld be correct to say thattwo fictional characters that exist together in the same novelare simultaneous and that two numbers that exist together in

the same number system are simultaneousThis departure from ordinary language is not decisive,

since Stump and Kretzmann could maintain that their ET-simultaneity relation, by wh atev er na m e w e decid e to call it ,is sufficiently analogous to ordinary simultaneity that itshould allow God to have knowledge of present-tense A-propos itions by stan din g in this relation to the ev en t re po rtedby such a proposition. Just how relevantly analo gou s is ET-

simultaneity to ordinary simultaneity? The relevant featuresof the analogy concern the respective roles these relationsplay as relations, that is, the manner in which they orderthings. Ordinary simultaneity is reflexive, symmetrical, andtransitive, but ET-simultaneity is only sym me trical. It cann otbe transitive; for, if it were, any two times would have tostand in the ET-simultaneity relation to each other since eachhas this relation to God, which would violate the requirement that the relata in an ET-simultaneity relation are a timelessly eternal being and a temporal entity. And for the samereason , the relation cann ot be reflexive. T hu s, it ha s only o neou t of the three logical featur es of ord inar y sim ultane ity. Th ismakes it look quite disanalogous to simultaneity, andthereby gives us no reason for thinking that God's standingin it to the event reported by a present-tense A-propositionsatisfies the tokening rule that requires of a true tokening ofa present-tense A-sentence that i t be simultaneous with the

reported event.Even if the Stump-Kretzmann strategy were to work for

God's knowledge of present-tense A-propositions, and it

70



The omrti$cienee~iMtnMctbilit$ti¥gument

doesn't, it would fail to show how he can know past- andfuture-tense A -pro po sition s. To accom plish this they wo uldhave to concoct a nontemporal counterpart relation to ordinary precedence, call it "ET-precedence," that again wouldhave an eternally present and temporal entity as its relata.Given the ontological disparity between its relata, it must beirreflexive and nontransitive, for the reasons already given.And, also like ET-simultaneity, it will be symmetrical. Bygiving the relation these logical properties, it avoids introducing any succession within God's timeless eternity or

requiring him to bear any temporal relations to temporalentities. And it could be claimed by Stump and Kretzmannthat God, in virtue of sta nd ing in ET -precedence to the ev entreported by a future-tense A-proposition, can know it,since this relation is sufficiently analogous to the ordinary-precedence relation between tokening and reported eventthat is required by the tokening rule.

As with ET-simultaneity, the analogy with its temporal

counterpart is very thin. Again, only one out of the threelogical properties of the latter is fulfilled. Ordinary precedence is irreflexive, asymmetric, and transitive while ET-precedence is irreflexive, symmetric, and nontransitive. Andone out of three just isn't good enough for it to count as arelevant ana logu e to the ord inary relation. To allow G od toknow A-propositions in virtue of standing in ET-simultaneityand ET-precedence relations to worldly events is on a parwith a radically multivocalist theory that permits God tocount as benevolent even though he does not satisfy ourord ina ry criteria for bein g benev ole nt. In bo th cases it is a dishonest piece of equivocation.

The B-reduction. This family of rebuttals argues that every A-prop osition is identical with so m e B-prop osition, and thu s byknowing the latter God also knows the former. The firstversion of this rebuttal to be considered was developed by

Nelson Pike for the expressed purpose of refuting Coburn'sargument .1 4 P ike be gin s by correctly po inting out that a timeless God, in virtue of his knowledge of B-propositions that

7»



The omniscience^mmutabitity argument

report a temporal relation between an A-tokening and theevent it repo rts, can kno w wh en time-bo und p erso ns e xpresstrue A-propositions. For instance, God can know the B-proposition that Smith's token ing of "E occurs no w " is (t imelessly) sim ultane ous w ith E an d there by kno w th at th eA-proposition expressed by Smith is (timelessly) true invirtue of satisfying the relevant tokening rule. If Smithshould be called up on to justify his A-a ssertion, he w ou ld doso by pointing ou t that it satisfied the releva nt tok enin g r ule,that is, he would assert this very B-proposition. And, fur

thermore, God also would offer this B-proposition to justifyhis knowledge of the B-proposition that the A-propositionexpressed by Sm ith is (timelessly) true . An d, P ike con troversially adds, this identity in justification shows that the latterB-proposition that the A-proposition expressed by Smith is(timelessly) true is identical w ith the A-p roposition expre sse dby Smith. And, in virtue of this identity, God timelesslyknows the A-proposition expressed by Smith.

There are several flaws in this. In gene ral, it is no t true tha ta person cannot know that the proposition expressed byanother is true without kn ow ing this very proposition. Thereare at least two cases in which this does not hold. In one casea suitable justification cannot be given; for instance, I canknow that the mathematical proposition expressed by somenoted mathematician is true but fail to know this propositionbecause I cannot derive or prove it. A more relevant case isone in which I cannot express the proposition known byanother, thereby failing to satisfy requirement 4b for knowledge; for instance, I can know on the basis of reliable eyewitness reports that the first proposition expressed by Smithyesterday is true and yet not know this proposition becauseI cannot express or specify it. And this is exactly the predicament God seems to be in regarding the A-propositionexpressed by Smith: He cannot express it since he cannottruly token an A-sentence. That God can identify but not

express this A-p roposition is not paradox ical once w e realizethat there are two different w ay s of identifying a p rop ositio n.In one we use a nou n "thaf '-clau se that expresses the prop o-

72



The omniscknce--imMuiaît§WgUtnmt

sition, that is, has a sen se that is the prop os ition ; for ins tan ce ,one says "that snow is white." In the other we use a non-specifying singular identifying expression, for instance,"Smith's favorite proposition." Pike fails to show that Godcan identify an A-proposition in the specifying manner. Andif God cannot specify it, he cannot know it.

Another flaw in the argument from sameness of justificatory responses for two assertions to identity of propositionsasserted is the assumption that what is asserted is identicalwith how it is justified. Pike is right that we justify an assertion of an A-sentence by stating a B-proposition indicatingthat the relevant tokening rule is satisfied, that is, that thetokening bears the proper temporal relation to the eventreported. But obviously, the A-proposition asserted is notidentical with the justificatory B-proposition, since the latteralone entails that a token bears som e tem po ral relation to th ereported event. The A-proposition that it is now rainingcould be true even if no tokens were to occur now. The sen

tence, "It is now raining even though no tokens occur now,"is pragmatically self-refuting in that every tokening of itexpresses a contingently false proposition.

An interesting new version of a B-reduction can be develope d from Kripke's views on rigid design ator s an d a criterionfor propositional identity based on the semantics of possibleworlds, namely:

PS. The proposition p is identical w ith the pro pos ition q if

and only if they have the same truth-value in every possibleworld.

Consider any A-proposition, say the one expressed by nowtokening "£ occurs now." The temporal indexical word"no w " me an s the same as "this t im e" an d, in some con texts,"at this tim e," and thereby deno tes a m om en t of time (wha tever that might turn out to be). Therefore, there is a trueidentity proposition, say that t7 is now. Both the indexicalword "now" and the proper name "t7" are rigid designatorsin that they denote in every counterfactual proposition inwhich they occur the same individual as they do when used

73



The ommscietmîmfnutaMitf •argument

in a categorical proposition about the actual world. Therefore, if "d" is rig id, the p roposition ex pressed by "d might nothave been d" is false. Both "now" and "t7" are rigid designators by this test. A s a conseq uence , the proposition expressedby the identity sentence "t7 is no w " is necessarily tru e if tru eat all. Given that it does express a true proposition, it followsthat there is no possible world in which now - this verymoment of t ime, namely, t7 - is not identical with t7. And,since E occurs n o w , there is no po ssible world in which Eoccurs now b ut d o es not occur at t7. Therefore, that E occurs

now and that E occurs (tunelessly) at t7 have the same truth-value in every p oss ible w orld an d, th us , by PS, are identical.Th e only th in g su spect in this B-reduction is its criterion for

propositional identity. PS is a far weaker criterion than PA,and the question is which one should be operative in ourdiscussion of G o d 's om niscience. It seems that PS no t only isinadequate in general, since it yields the counterintuitiveresult that there is only one necessary (impossible) proposition, but demands too little of an omniscient being. Godwould not be disqualified from being omniscient accordingto PS we re he to b elieve tha t 2 + 2 = 4 bu t fail to believe th at2 + 3 = 5. Thus, PS permits God to have a second-class sortof omniscience, which is definitively not the sort we expectof the being than which none greater can be conceived.

The co-reporting thesis. That w e can no t perm it God to have asecond-class sort of omniscience also undermines this way

ou t of the Co bu rn argu m ent. It is bas ed o n a far more m ode stclaim than that made by the B-reduction - that for everyA-proposition there is some B-proposition that reports oneand the same ev en t, in which the event reported by a prop osition is the event referred to by its participial nominaliza-tion. That S is Fin g rep orts the e ven t of S's Fing. Th us , th eco-reporting thesis claims that for every A-proposition thereis a B-proposition such that their participial nominalizationsare co-referring. This seems correct, given that temporalindexical terms like "now" denote a moment of t ime.

74



The omniscience-immutability a rgument

Consider the A-proposition expressed by now tokening "S

is Fing now." As already seen, "now" denotes a moment oftime, and thereb y there is a true prop osition identifying no wwith a moment of t ime, say that t7 is now. As a result, theB-proposition that S is (timelessly) Fing at t7 is co-reportingwith that S is Fing now, since S's Fing now is one and thesame event or state of affairs as S's Fing at f7, the reasonbeing that these ev ents involve the sam e subject instantiatingthe same property at the same time. While our t imelessDeity, for the reasons given by Coburn, cannot know theA-p ropo sition that S is Fing no w , he can tim elessly kn ow theB-p roposition tha t S is (timelessly) Fing at t7. Th us, a t imelessGod can know of every event, including events such as S'sFing now.

But this way out attributes a second-class omniscience toGo d. God, on this accoun t, does kn ow of every event but notun de r every, true description. H e kno w s w hat is going onno w , but he doe sn't kn ow of thes e events that they are occurring now . The co-reporting way ou t receives a spu rious plausibility from an ambiguity in "God knows of S's Fing now"that exactly parallels the aforementioned ambiguity in "Godkn ow s that S is Fing no w " in regard to w he the r the oratioobliqua occurrence of "now" indicates both the speaker 's andGod's indexical reference or only the speaker's. Since theco-reporting thesis agrees that it indicates only the speaker's,i t must say the same about the occurrence of "now" in "God

knows of S's Fing now." And this gives the game away.15

The elimination thesis. Yet an oth er p ossible escape from Co-burn's argument is supplied by a radical theory of indexicalsthat has been developed by John Perry in two articles ofunsurpassed clarity and brill iance.16 It is radical because itdenies the very existence of the sort of A-propositions whoseexistence is asserted by the claim of premise 3 that there are

true A-proposi t ions. These proposi t ions are supposed toresemble nonindexical ones in that both are a complete, unified Fregean sense compounded of the senses of the subject

75




and predicate of a sentence that expresses this unified sense.Such a Fregean proposition or "th ou gh t," in addition to serving as the sense of this sentence, also serves as its indirectreference (i.e., its direct reference when the sentence occursin oratio obliqua) and most importantly the object of propo-sitional attitudes. Let us follow Perry and call such propositions "traditional" or "de dicto" propositions. Perry has anargument for the elimination of de dicto A-propositions, aswell as other de dicto indexical pro po sitio ns, in favor of wh athe calls "de re" propositions; and it will turn out that a time

less God can know these de re propositions.Perry begins by clearly demonstrating that the traditional

Fregean theory does not fit indexical propositions. The A-sentence "Now is w he n E occurs" obviously has the samesense or meaning from one time to another, since the senseor meaning of "now" is invariant over time, although itsreferent is not. And yet one could believe what this sentenceexpresses at one time but not believe what it expresses w hentokened at another time. This shows that the sense of thisA-sentence canno t be identified with w ha t it is that som eon ebelieves w ho believes the proposition or thou gh t it expresseswhen tokened at some time. The source of the difficulty isthat the predicate of a subject-predicate sentence suppliesan incomp lete sens e (e.g., the sen se of " is w he n Eoccurs") that requires a sense completer su pplied by the subject expression so as to form a complete, unified sense, thatis , a de dicto proposition. But the subject of an indexical

sentence (and for the sake of simplicity indexical sentenceswhose subject alone is indexical are to be considered) is notfitted to serve as a sens e com pleter, for the rea son ju st give n.Plainly, if we are to go on believing in the existence of dedicto indexical propositions, we must modify this traditionaltheory.

The most likely way to do this, and one that Frege himselfseemed to pu rsu e, is to contin ue to hold that indexical w ord sare sense completers but impute to them a sense of limitedaccess that varies w ith the contex t, as op po se d to the context-invariant senses that other words have. Every time "now" is

76




tok en ed, it ha s a different sense tha t is ap pre he nd ab le on ly at

that time, and similarly for other indexical words. As a result,the proposition or complete, unified sense expressed by atokening at t7 of "N ow is w hen £ occurs" can be appr ehended or thought only at t7; an d only Jones can gras p orapprehend the proposit ion he expresses by tokening "I amF. " The sense that "n ow " has at t7 cannot be conveyed by an ynonindexical or fully general description, since the replacem en t of "n ow " by such a description results, according to PA,in a different proposition being expressed at t7.

The notion of a sense that is private or of limited access isan ath em a to Perry, as it is to a host of con tem pora ry defen ders of de re propositions, and thereby he cannot accept theexistence of de dicto indexical propositions that involve sucha sense:

Such a theory of propositions of limited accessibility seemsacceptable, even attractive, to some philosophers. Its accept

ability or attractiveness will depend on other parts of one'smetaphysics; if one finds plausible reasons elsewhere forbelieving in a universe that has, in addition to our commonworld, myriads of private perspectives, the idea of propositions of limited accessibility will fit right in. I have noknockdow n argument against such propositions, or the metaphysical schemes that find room for them. But I believe onlyin a common actual world.17

Perry de velo ps his de re theory of indexical propo sitions soas to avoid countenancing these mysterious private sensesand perspectives. Instead of thinking of the sense of anindexical sentence as the proposition or thought it expressesin some context, he proposes that we think of it as the "role"of the se nten ce, a role being specified by wh at w e hav e calleda tokening rule. And instead of thinking of the propositionexpressed by an indexical sentence in some context (i.e., as

tokened by some person, at some time, and place, etc.) as acomplete, unified sense - an abstract entity composed ofabstrac ta - w e are to thin k of it as an ontological mixed b ag -

77



The omniscience-immutability ttfpmmt

a pair comprised of the value or referent of the indexicalsubject expression in that context and the incomplete senseof its predicate. For relational sentences with more than oneindexical subject expression, for instance, "This hit that," thede re proposition expressed in some context will be anordered n- tup le co m pris ed of a sequence of the differentvalues of its indexical words in that context and the incomplete sense of its relational predicate. This ontological mixedbag is called a "de re" proposition. It is true just in case thevalue or referent of its indexical subject expression in that

context falls u nd er or is an instance of the concept den oted bythe inc om plete s en se of its pred icate. For instance, m y tokening of "I swim every day" expresses this de re proposition:the pair made up of Richard Gale the man (which is thereferent of my use of "I") and the incomplete sense of "sw im s every da y ." A nd it is true because Richard Gale - thevalue of "I" in that context - falls under the concept being-someone-who-swims-every-day, which is denoted by the

incom plete sense of " swim s every day ."The next step is to identify a de re proposition with theinformation conveyed by the tokening of an indexical sentencein a certain context. Perry says of the tokening on August 1,1976, of "Russia and Canada quarrelled today" that it

seems to yield just this information:(i) an incomplete sense, that of 'Russia and Canada quarrelled';(ii) an object, the day August 1st, 1976.(i) and (ii) do not uniquely determine a thought, but only anequivalence class of thoughts. Belonging to this equivalenceclass will be just those thoughts obtainable by completing thesense of 'Russia and Canada quarrelled' with a sense completer which determ ines, as reference, August 1st, 1976.1 shallcall thoughts related in this manner informationally equivalent. l8

Herein we have the key to escaping Coburn's argument.God can timelessly know a B-proposition, for instance, thatRussia and Canada quarrel (timelessly) on August 1, 1976,

78





The omniscimce~4i®imtMU ty argument

is F " The app rop riate tok ening of differently index ed cou nterparts can express differently indexed counterpart de dictoprop ositions th at a re at least logically equivalen t if not id entical. The propositions expressed respectively by tokening at t7

"S is now F at t7" a nd at a later time t9 "S w as F at t7" mutuallyentail each other, on the previous assumption that thesetimes exist, just as do the propositions expressed, respectively, by my to ke nin g of "I am F" and your tokening of "He(pointing to m e) is F." T hu s, eve n if it w er e true that each dedicto indexical proposition is apprehendable or expressible

only in a single context, we can still apprehend or express inother contexts a differently indexed counterpart propositionthat entails this proposition. And this should go a long wayto allaying our worries about the "limited accessibility" of dedicto indexical propositions.

Furthermore, by appeal to the intuitively plausible principle that if X kn ow s bo th p an d that p entails q then X kno wsq, it can be shown that any indexical proposition can beknown in a different context than that in which alone it isexpressible; for, in general, people know that differentlyindexed cou nte rpa rt p ropositions entail each othe r. This indirect way of knowing an A-proposition is of no help to a timeless God, since he is unable to express, and a fortiori unableto know, a differently indexed counterpart proposition toany A-proposition.

Perry's worries about indexical words as sense completers also applies to proper names if Kripke is right that they

aren't replaceable by a description. Just as we had to settlewith something like being-now as the current sense of"now," we must accept being-Socrates as the sense of "Socrates," the only difference being that the latter has a context-independent sense; but the problem posed by senses of l imited accessibility, supposedly, has already been solved. Ofcourse, Perry might find being-Socrates just as objectionableas being-now and extend the scope of his de re theory tocover propositions expressed by sentences containing suchproper names, which is commonly done by defenders of dere propositions.

80



T te &mmscience-immutabUity argument

Not only is the de re elimination theory not properly moti

vated, it faces serious conceptual problems of its own. It willbe argued that the de re theory fails to offer a suitable objectfor an indexical belief as well as a suitable bearer of a truth-value.

According to Perry, when someone believes an indexicalproposition, the object of belief is not a de dicto propositionbut one of his ontological mixed-bag de re propositions consisting of an inc om plete sense and the object that is the valueor referent of the indexical expression in that context. This isa variant on Russell 's multiple relation theory of judgment,according to which belief is an ordered n + l-tuple relationwhose relata are the believer and the n number of objects referred to by the subject(s) and predicate of the sentence expressing her belief: And it faces the same serious problems.'9

There has been a standard objection to the multiple relation theory that applies with equal force against the de retheory, namely, that it fails to account for the unity of whatis believed, said, and the like, since it gives us a mere heap,

reported by a laundry list, as the accusative. To say or believesomething is not to refer successively to different items. Atip-off that something is amiss surfaces with Perry's misuseof "inform ation " w he n he tells us that the toke ning of "Russiaand Canada quarrelled today" on August 1, 1976, "seems toyield just this information (i) an incom plete sense, tha t of'Russia a n d Ca na da qua rrelled'; (ii) an object, the day A ug ust1st, 1976"; for, if I w ere w itho ut further a do to pr ese nt yo u

with Jone s the m an as I say "the incom plete sense of' isa fireman'/ ' I would not be said to have informed you ofanything nor to have expressed a belief. Of course, were I tohav e do ne this in respon se to the request "Present m e with aman and tell me of an incomplete sense that refers to a concept un d er w hich this man falls," I wo uld be informing youof something, because, in effect, I would be saying "Jones(this man here) falls under the concept referred to by theincom plete sens e of ' is a firem an'." But this blo w n-u pPlatonic proposition, like the proposition that Jones instantiates f ire m an ne ss , is a de dicto prop osition , not on e of P erry'sde re propositions.

81



The omnisdenceîfûtabUity argument

Because a de re proposition lacks an internal articulationand unity, it is categorically unsuited to be the bearer of atruth-value, and thereby also unfit to serve as a relatum inan entailment relation. Were I, without any special stagesetting, to present you with the man Jones as I say "is theincom plete sen se of' is a fire m an '," it w ould be conceptually absurd for you to reply "That's true," for I have notpresented you with a suitable subject of a truth-value. Andsince a de re proposition categorically cannot have a truth-value, it is unable to entail or be entailed by anything. But as

seen in the discussion of differently indexed counterpartpropositions, indexical propositions do enter into entailmentrelations. They also have such relations to nonindexicalpropositions. The A-proposition that S is now Fing at t7

entails the B -proposition tha t S is (timelessly) Fing at t?. A ll ofthis is rendered anomalous on the de re theory.

In defense of Perry's de re theory against the charge thatthere are conceptual absurdities in taking a de re proposition

to be an intentional accusative and the bearer of a truth-valu e, it could be said that his theory is no t supp lying us witha paraphrase type of analysis, such that the analysans is categorically substitutible for the analysandum in every context,but rather with a semantic or truth-condition type of analysis. It is not claiming that instead of saying "Jones believesthat here is 42nd Street and Broadway," we could have said"Jones believes that" followed by an act of pointing to 42ndStreet and Broadway as we say "the incomplete sense of' is 42nd Street and B roa dw ay '." It is giving us the truth-conditions for Jones's indexical belief. If this is all it is doing,the de re theory is a very unperspicuous philosophical theory , since it fails to accoun t for the object of an indexical beliefand the like. And since the de dicto theory attempts toexplain these things, it is, to that extent, a better theory.

The quasi-indicator theory. Castaneda has developed a theory

of "quasi-indicators" or quasi-indexicals in a series of articlesthat represent the finest treatment of indexicals by any philosopher, and he has put this theory to use in refuting Kretz-

82



The omniscience-imm utability argument

mann's temporal version of the omniscience-immutabilityargum ent - the on e directed against the om nitem pora l G od.20

But since his refutation can easily be ex ten de d so as to ap plyto a timeless version of the arg um en t, it will be consid ered inthis section; for, if it succeeds against the latter, it also succeeds against the former, since anything that can be knowntimelessly can be k no w n at a time (but possibly no t vice versa).Basically, this rebuttal is directed against premise /\e ofCoburn's argument, because it attempts to show that an A-proposition can be expressed by the use of a n on-A -sentenc e

in oratio obliqua containing a quasi-indicator, thereby refutingthe claim that an A-proposition is expressible only by thetokening of an A-sentence. This refutation is extremely subtle and requires that we first get control over the notion of aquasi-indicator.

Briefly, a quasi-indicator makes an anaphoric referencewithin an oratio obliqua construction attributing a prep ositional attitude to someone and indicates what sort of indexi

cal reference this person made. In the sentence "Jones believes that he himself is rich," "he himself, in the first place,makes an anaphoric reference within the scope of a referringexpression ("Jones") outside of the clause. But it is more thanyour ordinary bound variable or relative pronoun, for it alsoindicates that Jones's belief is expressed or thought by himthrough the use of a first-person indexical reference, such as"I." Had we simply said "Jones believes that he is rich," wewould not attribute any such first-personal indexical belief toJones, since the use of "he" in the oratio obliqua clause, ifindexical, would express the point of view of the speaker.And this wo uld result in our und erde term ining the prop osition Jones believes, since w e wou ld be leaving un de term ine dhow he referred to the subject of his belief.

Now for the controversial part of his theory. Through theuse of a quasi-indicator, we can express in oratio obliqua thevery same propo sition as is exp ressed b y som eon e elsethrough the use of an indexical sentence. If I say "YesterdayJones said that it is then t7," the occurrence of "then" withinthe oratio obliqua clause, in addition to making an anaphoric

83



The omniscknce-immuiab&ily argument

reference to yes terd ay in virtue of falling w ithin the scop e of"yesterday," also indicates that Jones made a present-tenseindexical reference to that time by tokening a sentence synonymous with "It is now t7" in my language. Through theuse of the quasi-indicator "then," I succeed in expressingin oratio obliqua the very same A-pro position tha t Jo nesexpressed in oratio recta through the tok ening of an A-sentence.

This doctrine is pressed into service in refuting Kretz-mann's contention that an omnitemporal God cannot know

at any time the very same A-prop ositions that he or so m eo neelse could ex press in oratio recta at other times by the tok enin gof A-sentences. First, Castaneda shows how God couldexpress all of these propositions at a single time; for if hecan't express them h e can't know them. N ext, he sh ow s ho whe can know them at a single time.

Kretzmann's somewhat confused claim that

First such a being (God) knows that it is now t, (and that it isnot now t2) and then it knows that it is now t2 (and that itis not now ttf1

is perspicuously rendered by the use of quasi-indicators as

6. At f, X know s (tenselessly) tha t it is (tense lessly) then tubut not t2, and at t2l later than f„ knows that it is then t2, butnot t, .2i

This, in turn, is recast so as to show how at a single time, tuX can know that someone knows at t2 an A-prop osition thatis expressed by X at t, via the use of a quasi-indictor.

7. Time t2 is later than t„ and at f, X kn ow s bo th (1) tha t itis then,, f„ but not t2, and (2) that somebody knows (orwould know) at t2 that it is (would be) then< 212/ but not t ,23

in which the quasi-indicators "then,," and "then,2" carry sub

scripts that indicate which date-name outside of the oratioobliqua clause binds them. It will be important later to realizethat a subscript does not turn the quasi-indicator "then" into

84



The omnisciime-imrnutability argument

a cons tant but only serves as an a id in loca t ing the cons tantt h a t b i n d s t h e s u b s c r i p t e d w o r d .

Pr op osi t ion 7 d o e s no t by it se lf sh o w tha t a t t, X k n o w st h e v e r y s a m e A - p r o p o s i t i o n t h a t h e o r s o m e o n e e l s e k n o w sa t f2 in oratio recta. A n a d d i t i o n a l p r e m i s e is n e e d e d , a n dCastaneda f inds i t in the pr inc ip le of the t rans i t iv i ty ofk n o w l e d g e :

T. If a sentence of the form 'X knows that a person Y knowsthat . . . ' form ulates a true statem ent then the pers on X

knows the statement formulated by the clause filling theblank.24

He goes on to po in t ou t tha t

this principle must be carefully understood: it establishes asort of transitivity of knowledge, but it does not say anyth ing

; abo ut deta chin g expr essions of the form 'Y knows tha t . . . '.23

All tha t remains to be done i s to p lug in to formula T there levant par ts of 7 . This resul t s in

7a . At t, X k n o w s . . . t ha t so m eb od y kn ow s at tz that it ist h e n ,2 t2

and has the consequence tha t a t £ , X k n o w s t h e v e r y s a m eA - p r o p o s i t i o n t h a t s o m e o n e k n o w s a t t2 - the on e tha t hew o u l d e x p r e s s a t t2 by token ing " I t i s now t2."

Before we evaluate th is beaut i fu l ly ingenious refu ta t ion , i twi l l be shown tha t i t can and ough t to be re fo rmula ted so asto ap pl y to a t im ele ss Dei ty . To see w h y th is ou gh t to bedone, le t us begin wi th the fo l lowing objec t ion:

Even if your Deity X at every time can know in oratio obliquathe very same A-propositions that are expressible in oratiorecta by the tokening of A-sentences at other times, it still isthe case that he must change the way in which he expresses

these A-propositions from one time to another; for instance,at f, he expresses in oratio recta that it is th en( j t , but does notdo so for the A-proposition that at t2 it is theh( 2 t.v but he

85



The omniscimce-iMMUtaMtfty argument

changes at f2 in that he no longer expresses the former A-proposition in oratio recta but must do so in oratio obliqua. Andthis change in the way God formulates his beliefs compromises his immutability.

An easy way ou t of this prob lem is sup plied by the realization that there is no need to place X (God) at a moment oftime in orde r for X to kno w an A-p ropo sition in oratio obliqua.To see this we o nly ne ed to rew rite 6 by sub stituting "an d Xtimelessly knows" for "and at t1 X knows" and make our

time-bound knower someone other than X.7 '. Time t2 is later th an tu and X timelessly kno w s both (1)

that at t, some one k no w s (or w ou ld kn ow ) that it is then, t tv

but not f2 and (2) that somebody knows (or would know) att2 that it is (would be) then ,212, but not t, .26

The reason why we can place our oratio obliqua knower X"ou tside " of time is that the sen tenc es that expre ss his beliefs

(e.g., "at t,

som eon e kn ow s that it is th en( l t ,"), because theycontain quasi-indicators, are freely repeatable B-sentences.A nd this is the vital nerv e in the refutation - that, pace 4 c anA-propo sition ca n be expressed b y a B-sentence that c ontainsquasi-indicators.

Unfortunately, although 7' helps to avoid the mutabilityproblem , it crea tes a problem of its ow n in that it clashes w ithreligious orthodoxy. According to theism God is essentiallyomniscient, that is,

8. God is omniscient in every possible world in which heexists.

It also is the case for theism that God is free to create finite,t ime-bound knowers or not to. Therefore,

9. There are possible worlds in which both God and timeexist but no time-bound knowers.

But the Castaneda-style solution using 7' saves God's omniscience in a tem por al wo rld by ma kin g him into an epistemicC ou nt Dracula w h o m us t live off the b lood of his tim e-boun d

86



The omriiscience^-iminutaMily argument

oratio recta knowers of A-propositions, which has the consequence that

10. There are no possible worlds in which both God isomniscient and time exists but no t ime-bound knowers do.

From 8 and 10 it follows that

11. It is no t the case that there are possible world s in whichboth God and time exist but no t ime-bound knowers do.

And this flatly contradicts theism's belief in 9.

Thus , the choice betw een em ploy ing 7 or 7' in a refutationposes a dilem m a, for we wind up either com promising G od'simmutability or departing from theistic beliefs. But it willturn ou t that ther e is an escape from this dilem m a, because a7'-style refutation really does not require that there be finite,time-bound knowers. But first it will be shown that both ofthese refutations suffer from an additional flaw: The principle of the tran sitivity of kno w led ge T no t only is false but a lsois m isus ed in these refutations.

This principle is subject to the same sort of counterexamples and restrictions as is the previo us principle that a per soncannot know that the proposition expressed by another istrue without knowing this very proposition. Just as I couldknow that the proposition expressed by another is true butnot k now it becau se I lack a suitable justification for thisprop osition, I can know that ano ther kn ow s this propositionand lack this needed justification for it. Both principles, fur

thermore, must be restricted to cases in which the proposition expressed or known by the other person is completelyspecified, and this means that both the subject and predicatesense components be specified or expressed. If propositionalunderdetermination is allowed within the blank space of T,the following sort of counterexample can be constructed: Inthe case in wh ich "X kn ow s that Y knows that concerning (orof) the F that it is G," the propo sition that Y know s is unde r-determined because we are not told how Y refers to the F; hecould have been ignorant that this individual is the F andinstead referred to it by a proper name. And, therefore, it

8 7



The omniscience-immtotabiUty argument

doe s not follow that X kn ow s the very sam e proposition thatY doe s.

The relevant part of 7' (and 7) says "X knows timelessly(know s at tt) tha t somebo dy, Y know s a t t2 that it is the n,2t2."Th e crucial qu est ion is w he th er this also is a case of p ropo si-tional un de rde term ina tion in the final "thaf '-claus e - the on etha t is inserted in the blan k spa ce of T. It could be argue d withsome plausibility that it is not, because the quasi-indicator"then,2" tells us h o w Y refers, nam ely , th rou gh the use of apresent-tense indexical word such as "now." Unfortunately,

its final that-clause - "that it is then,2 t2" - does not alonespecify a complete proposition since "then/ 2" occurs within itas a free variable, n o longe r being bo u n d b y the occurrence of"at t2" outsid e of th e that-clause. T hu s, the relevant p art of 7'(7) is not fit to be plugged into the blank space of T once T isrestricted, as it m u st b e, to com pletely specified prop osition s.

It might be thought that this problem of failing to specify acomplete proposition in the final that-clause is easily solved

by moving the "at t2" from outside to inside the final that-clause, resulting in "X knows timelessly (at t,) that somebod y, Y, kn ow s th at at t2 it is t hen ,212." But now it becomesclear that neither Y no r X kn ow an y A -propo sition at all. Theproposition that at t2 it is then ,212 is identical with th e tautology that t2 is t2, w hic h is a B-proposition since it is expressedby a freely repeatable sentence. Assuming, pace Perry, thatthere are de dicto A-propositions, that it is now t2, by Leibniz's law, is not identical with that t2 is t2 (or that at t2 it isthen,2 t2) be ca use o nly the latter h as th e prope rty of beingknowable a priori. God, of course, can timelessly know thatt2 is t2, but this gives no reas on to thin k th at he can timelesslyknow the A-proposition that i t is now t2. That he can knowthese tautologies in oratio recta shows that nothing is gainedby making God into an epistemic Dracula who must live offthe oratio recta blood of t ime-bound knowers.

The respon se to this might be that the p roposition th at at t2

it is then,212 cannot be identified with the proposition that t2

is t2/ since only the former identifies an A-proposition,namely, the one that is expressible by tokening at t2 the

88



The omni$eieH£e~MfHvttabtiity argument

A-sentence "It is now t2." This response has a spurious plausibility because it fails to distinguish between the specifyingand nonspecifying mode of identifying a proposition. While"that at t2 it is then,., t2" does identify an A-proposition, itdoes not do so in a specifying manner, as does the tokeningnow of "that it is now f2." As seen , T requ ires that a com pletespecification of a proposition be inserted in its blank space.While "that at t2 it is then,2 t2" does completely specify aproposition, it is the B-proposition that t2 is t2. What is trickyis that it also identifies in a nonspecifying manner an A-proposition. Thus, the phrase "the proposition that at t2 it isthen,212" identifies both an A- and B-proposition but does soin a specifying manner only for the latter.

The limitations on a perfect being's omniscience. This refutationaccepts Co burn 's arg um en t for 4 bu t escapes the co ntradictory conclusion of the omniscience-immutabili ty argumentby rejecting the account of God's omniscience given in premise 1 in favor of a w ea ke r on e that n ot only pe rm its bu t conceptually requires that God not know A-propositions. It willbe recalled from Chapter 4 that it proved necessary to givethe following restricted definition of God's omnipotence:

0 3 . For any proposition p, if it is logically possible thatGod bring it about that p, then God can bring it about that p.

The strategy is to restrict God's omniscience in an analogousmanner to what it is logically consistent for God to know:

Oy For any prop osition p, if it is true that p and it is logically possible that God know that p, then God know s that p.

What must be done is to demonstrate that from 05 , in conjunction with other necessarily true propositions, it can beded uce d that it is conceptually impossible that God kn ow anA-proposition, thereby showing that his failure to do so doesnot impugn his omniscience, assuming that 05 is true.

This de m on stratio n goes as follows: It beg ins with the conceptual truth that

89



The omni$cieme~kfmutabiliiy argument

12. God is an absolutely perfect being.

Then it argues that necessarily13. An absolutely perfect being is timeless,

by appeal to the alleged conceptual truth that

14. Any temporal being has the possibility of ceasing toexist.

Since having even the mere logical possibility of ceasing toexist is an imperfection, it follows that necessarily

15. God (i.e., an absolutely perfect being) is timeless.

And by appeal to

4. It is concep tually impossib le that a timeless being k no wan A-proposition,

it deduces the desired conclusion that

16. It is conceptually impossible that God know an A-proposition.And since God's omniscience, as defined by C"5, excuses himfrom having to know any proposition that it is logically orconceptually impossible that God know, his failure to knowA-facts does not constitute a counterexample to his omniscience.

This refutation of the omniscience-immutability argumentfaces two insuperable difficulties. First, it creates a very viru

lent version of the paradox of perfection, by comparison withwhich the Chap ter 1 parado x of perfection created by the0 3 and 04 definitions of God's omnipotence is quite mild.W hereas m any are conten t to live w ith the latter, few, including myself, will be able to stomach the former. While we arenot crushed by the tho ug ht that Go d, in virtue of his absolu teperfection, is barred from playing football, committing suicide, or performing a morally wrong act, we are by thethoug ht that it precludes his kno w ing any A -proposition. HisC yt y pe om niscience app ears far inferior to tha t po sses sed bysome possible nonperfect being who, in virtue of not having

90




to be timeless, can know everything that God knows plus allthe true A-propositions. (The conceptual possibility of such anonperfect omniscient being assumes that a being can havesome but not all of the divine perfections, and thus that thedoctrine of the divine simplicity is false, which has alreadybeen argued for in Chapter t.)

God's inability to know A-propositions in virtue of histimeless eternality, not only is a blight on his omniscience,but also is another factor that contributes to his being religiously unavailable. A religiously available God is one with

whom we can commune. But we can commune only withbeings who share our tensed perspectives, who know how itgoes with us right now, and not just how it goes timelesslywith us at t7, even if now is t7. In other words, they must bebeings of whom the song title "My Time Is Your Time" istrue. It matters very much to us when we seek solace, guidance, and en cou rag em ent from a perso n that they are awareof ou r pres en t situation an d, in turn, can be a living presenceto us when they offer such solace, guidance, and encouragement. Only a being possessed of knowledge of A-facts couldbe a suitable object of such intentions of ours. The God whois the ap pa re nt object of direct experiences of God's pres enceinteracts with us as a living presence, and thus possessesknowledge of A-facts concerning us. He carries on an A-typediscourse with us. Thus, if we are to be true to the phenomenology of religious experiences of God's presence, wemust conceive of the God who is the object of such experiences as a being in time who is possessed of knowledge ofA-facts concerning us.

A second major difficulty with the restricted-omnisciencew ay out of the om niscie nce -im m utab ility arg um en t is that itsclaim that 14 is a conceptual truth is quite dubious, sincethere is a long-standing tradition within Western theism,deriving from the Bible, that accepts as meaningful the concept of an omnitemporally eternal God. That the language

game involving the concept of an omnitemporal Deity hasbeen played by so many for so long counts against the claimthat 14 is a necessary conceptual truth. I know of no good

9*



The ommscieme^kMmuteMHtf argument

arg um en t for it bein g necessary that any being in time ha s th elogical possibility of ceasing to be. Nor, as will be argued inChapter 6, does an ontological argument have any morechance of succeeding for a timelessly eternal God than it hasfor an omnitemporally eternal one.

Not only does there appear to be no good argument forw hy a highest being , a being than wh ich n one g reater can b econceived, must be timeless, there is a good argument forw hy it m ust be tem poral. Such a being mus t be a pe rso n, bu tonly a temporal being can qualify as a person. Part of the

justification for this was given in Chapter 1, when it wasargu ed that a per son m ust have a m ind an d that a m ind m us thave temporal endurance, the reason being that consciousness is a temporally elongated process. A conscious being isa Sarrrian for-itself that is always running ahead of itself intothe future. The concept of a strictly immutable, completelyactualized personal God is an impossible one, an attempt, asSartre brilliantly portrayed in his Being and Nothingness, tohave a God that is both a for-itself and at the same time anin-itself, that is, a being whose existence is wholly containedwithin a single moment, which could weil be a timelesspresent that has no past or future, and yet transcends thismoment into i ts future. The penetrating phenomenologicalanalyses of the temporally "ekstatic" mode of existence ofpersons given by Heidegger and, before him, by James,Bergson, Dewey, and Husserl, bring out the necessity forsuch a self-conscious b eing to hav e retentio ns of its pa st th atinform its present experience in the formation of futureprojects.

The results of a conceptual analysis of a perso n ag ree w iththose of the previous phenomenological analyses. A perso n ,within forensic and moral contexts, must be an agent thatperform s inten tional actions so as to brin g abou t som e goal o rend. But to have a goal or end, the agent must have desiresan d v alues. But only a temporally incom plete being can ha v ea desire or intention or will something, since one cannotdesire, intend, or will what one already has. The rich man

92




does desire and intend to be rich, but what he desires is the

continuation of his richness into the future. And he does notyet have these future wealthy states.An agent, furtherm ore, is a morally respons ible bein g. Bu t

someone can be morally responsible only in a community ofother morally responsible agents with whom he interacts, aswill be argued at some length in Chapter 9. A timeless Godcanno t interact with other ag ent s, since he can no t be causallyaffected by what they do, nor, as just argued, can he affectthem in personal ways that matter. It is not an accident thatthose theists who conceived of God as timeless, althoughthey thought he could timelessly bring about effects in theworld, were careful to deny that anything in the world couldin turn affect him. (The petitionary prayer of working theistsm akes no sense on this view.) This doctrine of one -wa y interactionism makes God a nonagent, since he is rendered unable to be a member of a community of agents who interactwith each other and m utually hold each other morally re spo nsible. And without community there can be no communion,thereby rendering the timeless God religiously unavailable.The latter requires two-way interactionism between Deityand creatures. God, or should I say we, pay too great a pricefor his aseity. If this is absolute perfection, my vote is forsomething less exalted.

If we are right that a timeless God, in virtue of being aperso n, m ust be tempo ral, his inability to know A-facts is theleast of his worries. // the argument for the former is sound,there is no need for the om niscience-imm utabili ty arg um en t,since the former alone establish es the con ceptual ab sur dity ofthe concept of a timelessly eternal God. But this is a big if,since philosophical arg um en ts never are conclusive , an d th usit is wise to make use of as man y w ea po ns as are available tous in our attempt to undermine the concept of a timelesslyeternal God. Even if our multipronged attack on this conceptis not decisive, it can still be shown that the concept of an

om nitem pora l Deity is a m ore attractive and serviceable o ne .Not only does it render God religiously available but also

93



The om niscienceîmmutakiHfy argument

gives us an easy way out of the omniscience-immutabilityargument, as will now be shown, something that was notpossible for the timeless Deity.

The version of the omniscience-immutabili ty argumentthat is directed against the omnitemporal God begins withthe same initial set as does the timeless version:

1. God essentially knows all and believes only true propositions,

2. God essentially does not change from one time to another in respect to any nonrelational property; and

3. There are true A-propositions.

It then goes on to argue that it is necessary that

17. Any being who knows every true A-proposition willhave to change the way in which he expresses or is disposedto express his indexical beliefs from one time to another,

from w hich it follow s, in conjunction w ith 1 an d 3, th at

18. God (conceived of as omnitemporally eternal) changesin the w ay he expr esse s or is dispo sed to express his index icalbeliefs from one time to another.

And this contradicts 2. Thus, God both is and is notimmutable.

The atheologian's argum ent for 17 dra w s upo n th e p oin tsm ade in the above discussion of Castaned a. Assu m ing thereto be moments of time t, and t2, at tu but not t2, a being w hokn ow s every A-fact token s or is disposed to token a s en ten cesynonymous with our "It is now f," and at t2l but not tu

a sentence synonymous with "It is now t2." Such a being,therefore, changes from f, to t2 in either the man ner in w hichshe formulates her A-beliefs in oratio recta or its dispositionsso to formulate her A-beliefs.

The theist w h o conceives of God as being om nitem porallyeternal will righ tly reject the premise 2 de m an d for God to be

strictly immutable. This can be supported by showing, first,that the traditional arguments for strict immutability areunacceptable, and, second, that there are ways of restricting

94



The omniscknce-'imrnutabM ty argum ent

God's immutability that do not endanger his exalted status asthat than which none greater can be conceived. And this iswhat will now be attempted.

The re are tw o traditional a rg um en ts for strict imm utability:the Platonic argument that any change in an absolutely perfect being would represent a deterioration in it and therebydisqualify it as being absolutely perfect; and the Aristotelianargument, which is the one that weighed most heavily withthe great medieval theists, that an absolutely perfect beingmust enjoy complete actuality.

Of the two arguments, the Platonic is the most vulnerable,since it is implausible to assume that any change in God is forthe wo rse; it m ight ha ve a neu tral outcom e.27 The resp ons e tothe Aristotelian argument, which also applies to the Platonicone, is that any b eing th at has strict imm utability is a no np er-son for all of the reasons previously given: He cannot, forexam ple, be conscious, perform intentional actions, have anygoals or en ds , will any thing , interact with other persons , an d

so on. Aristotle's god of pure actuality might be an object ofintellectual love, as might be Euclidian geometry, and serveas an explanation of the final causes of worldly changes, buthe is conceptually unfit to be a relatum in a two-way loverelationship or one of communion, this being due to hisstatus as a nonperson.

What is to be rejected is the Greek assumption that underlies both the Platonic and Aristotelian arguments - that truebeing, a highest form of being, must be timeless. Such anassumption, no doubt fathered by a fear of death and thevicissitudes of time, as well as a disdain for the changeableworld that was the object of concern for the "lower" class ofpeo ple in their society, as John D ew ey so insightfully poin tedout, must be rejected by any theism that postulates a personal Go d. It is a totally p erve rse a ssu m ptio n, an d it is ama zing that the great medieval theists should have accepted it.Philosophy makes for strange bedfellows indeed.

We know that God, in virtue of being a person, cannotbe strictly immutable. The only question is how mutablehe is, and this must be answered in a way that does not

95



The omniscieim-*-im<mMtabilit$f argument

com prom ise his statu s as an absolutely perfect bein g. That anomnitemporal God must undergo some change is an immediate consequence of his enduring in time. If Don Ricklesshould be fortunate enough to meet God at the pearly gates,he could say to God what he has been saying for years toHenny Youngman - "You're old " While in one sense this isobviously true - God is not only old but infinitely old - itis false in the sense intended; for, unlike Youngman (andwho says names have descriptive sense), God has not become decrepit and feeble, has not aged in this pejorative

sens e. If aging or grow ing olde r in th e chronological se ns e isa purely relational change, involving having different external relations to different events or moments at differentmoments of time, God's aging presents no problem. But, if,as I suspect is the case, this change involves a change in hismonadic properties, the omnitemporal version of God'simmutability must be restricted to properties for which it isnot a logical consequence of his omnitemporality that hechange in respect to them.

Furthermore, God's immutability must not be so strict as topreclude his kn ow ing A-facts. This requires that he con tinually change in the manner in which he expresses or is disposed to express his A-beliefs in oratio recta discourse. Suchchanges do not compromise his absolute perfection by lessening in any way his omniscience, for his change of mindfrom one time to another does not result in his holdingincompatible beliefs - his believing at one time what he had

disbelieved at an earlier on e. Fu rth erm ore , he kn ow s at everytime everything that it is logically possible to know at thatt ime; since at every t ime he know s both every A-pro positiontruly expressible in oratio recta at that time an d for everyA-proposition truly expressible in oratio recta only at anothertime a differently indexed counterpart A-proposition thatentails it . As argued, there is ground for contending that heeven knows the latter and not just entailers of it.

What is required of an absolutely perfect being is not thathe not change in any of his nonrelational properties but thathe not change in respect to any one of his perfections, such

96





Chapter 4The deductive argument from evil

The existence of evil poses both a logical and empirical problem for theism, forming the basis of both a deductive argument that attempts to deduce a contradiction from theexistence of both God and evil and an inductive argumentthat contends that it is improbable that God is the creator ofa world tha t contains the am ou nt and k inds of evils found inthis one. Since the deductive argument tries to prove that

there is no po ssible wo rld that co ntains both G od a nd evil, allthe theist need do to rebut it is to show that it is logicallypossible that God and evil coexist, and this can best, maybeonly, be done by actually describing a possible world inwhich both exist. In such a world God will have a goodreason or morally exonerating excuse for permitting evil.Such a response consti tutes a defense. To rebut the inductiveargument, the theist must construct a theodicy, w hich is adefense plus an empirical argument for the actual existenceof the possible world articulated in the defense.

Often a defense is confounded with a theodicy, resultingin the bogus demand that the theist produce a theodicy as arebuttal of th e dedu ctive arg um ent. H . J. M cCloskey, as anexample, first espouses the deductive argument. "Evil is aproblem for the theist/ ' he says, "in that a contradiction isinvolved in the fact of evil on the one hand, and the belief inthe omnipotence and perfection of God on the other."1 H e

shows a surprisingly short memory span because later in hisessay he says of a certain defense that "it does not in itselfprovide a justification for the evil in the universe. It shows

98







The deMctmargummi

Now for the needed necessary additional premises:

However, the contradiction does not arise immediately; toshow it we need some additional premises, or perhaps somequasi-logical rules connecting the terms "good," "evil," and"omnipotence." These additional principles are that good isopposed to evil, in such a way that a good thing alwayseliminates evil as far as it can, and that there are no limits towhat an omnipotent thing can do. From these it follows thata good omnipotent thing eliminates evil completely, and thenthe propositions that a good omnipotent thing exists, and thatevil exists, are incompatible.

Thus , we are to add these two additional premises:

4. A wholly good (omn ibenevolent) being eliminates an dprevents every evil he can; and

5. There are no limits to w hat an om ni po ten t being can do

in which I have taken the liberty of adding "and prevents" in

4, since a being would not qualify as omnibenevolent if heallowed some preventable evil to occur, even if he thereaftereliminated it. (Think of all the volunteer firemen who torchbuildings so that they can have the fun of putting out thefires.) With these extra premises it is easy to deduce a contradiction in the following way:

6. It is no t the case that evil exists [from 1, 2, 4, and 5]; an d7. Evil exists and it is no t the case tha t evil exists [from 3

and 6j.Faced with this contradiction, the theist must give up at leastone of the propositions in the initial set, since the additionalpremises must be necessary. But are they?

Mackie is well aware that they must be. That is why hecalls them "quasi-logical rules." Why did he not call them"definitions"? That would fit nicely with his implied linguistic theory of necessity. The reason, no doubt, is that such

definitions are not to be found in any ordinary dictionary.Nevertheless, they are supposed to be true in virtue of normative rules or conventions that are implicit in our use of

1 0 1



The ieimiim w$ument from evil

language, such that speakers are willing to offer and acceptcorrection when there is a deviation from them.

The trouble with calling 4 and 5 quasi-logical rules is thatou r, and certainly th e theist 's, use of langu age doe s not squarewith them. They misdescribe the language games in whichthe concerned terms are used. Pace 4, we often feel justifiedin bringing ab ou t or no t prev en ting som e evil so that a greaterevil can be av oid ed (e.g., t he surg eon causing a patien t to feelpain in the course of am pu tat in g a limb or a paren t disciplining a child) or an outweighing good realized (e.g., suffering

the agony of w ritin g a pa pe r refuting the FWD). An d on e willfind very few the ists, if an y, w ho accept the prem ise 5 accou nt of om ni po ten ce . M ackie is well aw are of this, for w he nhe co nsiders vari ou s de fen ses he is willing to allow the theistto use "omnipotence" and "benevolence" in ways that deviate from these rules. While he lodges objections to thesedefenses, h e d oe s no t object to them on th e gro un ds th at theiruse of these terms is at variance with 4 and 5. (One wonderswhat perversity lead Mackie initially to formulate his deductive argument in a way that he knew was unacceptable.)There are numerous places where he shows a will ingness toreplace 4 and /or 5 w ith so m e m ore restricted propo sition.

One such place is his consideration of the defense thatholds that God could not create a world containing goodwithout permitting some evil, because there logically cannotbe one without the other, this being based upon some significant contrast principle. (Surprisingly, it turns out that if thePink Panther were the Deity, he couldn't make everythingpink .) This def ens e ob viou sly em plo ys a w eaker version of 5,namely,

5,. An omnipotent being can bring about any logicallypossible state of affairs.

Mackie is willing to give his theistic opponent this weakeraccount when he wri tes:

It may be replied that these limits are always presupposed,and the omnipotence has never meant the power to do what

1 0 2



The deductive argtmwnt

is logically impossible, and on the present view the existenceof good without evil would be a logical impossibility. This

interpretation of omnipotence may, indeed, be accepted as amodification of our original account which does not rejectanything that is essential to theism, and I shall in generalassume it in the subsequent discussion.5

Mackie ' s d i scuss ion o f the good- requ i res -ev i l de fense a l somakes c lear tha t i t re jec ts 4 . That God cannot logica l ly br inga b o u t g o o d w i t h o u t a l l o w i n g s o m e e vi l is s u p p o s e d t o m o ra l ly exonera te h im fo r pe rmi t t ing some preven tab le ev i l , pace4; for th is w as the pr ice tha t had to be pa id to h av e an y go oda t a l l , and the good supposed ly fa r ou twe ighs the ev i l .

Mackie has a response to th is , but i t i s not based upon th isde fense ' s depar tu re f rom 4 .

It will provide a solution for the problem of evil only if one isprepared to say'Evil exists, but only just enough evil to serveas the counte rpar t of go od '. I do ub t wh eth er a ny theist will

accept this.6

Here in Mack ie shows h imse l f wi l l ing to accep t th i s de fense ' srep lac em en t of h i s o r ig ina l 4 wi th so m et h i n g m or e qua l i f i ed ,such as

4 ,. A n o m n i b e n e v o l e n t b e i n g a t t e m p t s t o b r i n g a b o u t t h ebest overal l s i tuat ion he can; or

42. A n o m n i b e n e v o l e n t b e i n g t ri e s t o e l i m i n a t e a n d p r event every moral ly unjus t i f ied evi l ,

in w hi ch a m oral ly unjus t i f ied evil i s o n e tha t is no t ne ce ss ar yfo r e i the r the rea l i za t ion o f an ou twe igh ing good o r the p reve nt io n of a gre a ter evi l . To b lock th is de fen se M ackie i simpl ic it ly r ep lac ing h i s o r ig ina l p r em ise 3 w i th th e s t ro ng er

3 , . Th er e are evi l s in excess of w h at i s req ui r ed for the re tobe a s ignif icant contrast with evi l .

I t i s not c lear tha t the proponent of th is defense i s wi l l ing togran t 3 , .

103



The deductive argument from evil

Another example of Mackie's willingness to permit thereplacement of his 4 and 5, respectively, by 4, or 42 and 5, ishis discussion of the defense that holds that the universeis better off as a whole with some evil in it than if there werenone. Sometimes this is given an aesthetic twist, but moreoften it is cashed in in terms of the soul-building defense: Itis a better overall state of affairs for there to be a gradualovercoming of evil in the course of the spiritual and moraldevelopment of free creatures than a statically good world.His account of the latter defense makes clear that it accepts

ne ith er 4 nor 5. It rejects 5 because it ho lds that it is logicallyim poss ible for God to bring abou t this overall better situationwithout permitt ing some preventable evil , which plainlydo es n ot imp ut e totally unlim ited po w er to G od. As for 4, thefollowing shows that Mackie replaces it with 42:

But does it still hold that good and evil are opposed? Not,clearly, in the sense that we set out originally: good does nottend to eliminate evil in general. Instead, we have a modified,a more complex pattern. First order good (e.g. happiness)contrasts with first order evil (e.g. misery): these two are opposed in a fairly mechanical way; some second order goods(e.g. benevolence) try to maximize first order good and minimize first order evil; but God's goodness is not this, it is ratherthe will to maximize second order good. We might, therefore,call God's goodness an example of a third order good.7

Th e rele van t p oin t is that the existence of first-order evil isa logically necessary condition for the existence of second-or de r goo d con sisting in the good character traits of cou rage,fortitude, sympathy, and benevolence in the face of suffering. Not even an omnipotent God could bring about the latterw ith ou t th e former, and , mo reover, it is better overall to havea world that contains the two than one containing first-ordergood alone.

Again we find Mackie objecting to a defense that eschewshis original 4 and 5 in favor of their weaker cousins ongrounds other than this, for he counters this defense on the

104



7-ke&e$MCtive argument

ground that i t leads to a v ic ious inf in i te regress . We beginwi th the p rob lem of f ind ing a mora l ly exonera t ing excuse fo rGod ' s pe rmi t t ing f i r s t -o rder ev i l s , such as phys ica l and menta l suffer ing. This defense holds tha t such evi ls are logica l lyneces s a ry c o n d i t i o ns fo r t h e eme rgence o f s econd -o rde rgo od s of be ne vo len ce , courag e , an d the l ike an d tha t the segoods fa r ou tweigh the f i r s t -o rder ev i l s , the reby exonera ting God for a l lowing them. But there i s s t i l l the problem ofsec on d-o rde r ev il - ba dn es s of cha rac te r, su ch as malevolence , cowardice , c ruel ty. And th is type of evi l i s far more

ser ious than f i rs t -order evi l .

We should, therefore, state the problem of evil in terms ofsecond order evil, and against this form of the problem thepresent solution is useless. An attempt might be made to usethis solution again at a higher level, to explain the occurrenceof evil (2), i.e. second order evil. . . . But even if evil (2) couldbe explained in this way, it is fairly clear that there would be

third order evils contrasting with this third order good: andwe shou ld be well on the way to an infinite r egre ss, wh ere thesolution of a problem of evil, stated in terms of evil (n), indicated the existence of an evil (n + 1), and a further problem tobe solved.

Th is in f in it e r egr es s , su pp os ed ly, i s v ic iou s , bec aus e the ve rypro b le m tha t i s ad dr es se d by the acco un t on the f ir st l eve lrecurs on each l eve l in the regress . The bogey man nevers tops chas ing us , a lways be ing jus t one l eve l away f romca tch ing us .

T he re is a fa tal f law in th is v ic io us inf in i te re gre ss arg ument . I t fa i l s to sa t i s fy a necessary condi t ion for such ana rg um e n t t o suc ce e d , n a m e ly, t ha t eve ry p ropo s i t i on r epo r ting the exis tence of the problem on level n logically entail ap ropos i t ion repor t ing the ex i s tence o f the same type o f p roblem o n level n + 1. Mackie 's claim that the existence of evi l on

any level n " indica tes" the exis tence of evi l on level n + 1doesn ' t s a t i s f y t h i s e n t a i lme n t r equ i r em en t , bec ause he d o esno t sh o w t ha t if th er e is a go od o n an y level n th er e also is evi l

105




on level n + i . His claim that "there would be" is at best anempirical gen eraliza tion . As a resu lt, it is op en to th e theist tostop the regress by claiming that there is good on some levelwithout any accompanying evil on that level; for instance, itcould be said that the re is good (3) - inten tion s to pr od uc egoo dne ss of character - w ith ou t any evil (3) - inte ntio ns toproduce badness of character. It could be an empirical factthat benevolent beings are capable of forming higher-orderintentions than malevolent beings. After all, they're morehighly intentioned

Yet another place where Mackie operates with weaker versions of 4 and 5 is in his discussion of the FWD, in which hesees it as a special version of the preceding n-order-evils-are-justified-because-they-are-necessary-for-outweigh ing-n + 1-order-g ood s. Herein the free w ill of created p er so ns is a th ird-order good that far ou tw eig hs the secon d- and first-order evilsit occasions. Acco rding to M ackie's accou nt of the FWD, m oralresponsibility for these evils stop s with th ese created m idd lem en; it does not reach throu gh to Go d. A s w e shall see w h enwe discuss versions of the FWD published subsequently toMackie's essay, responsibility, but not blame, is allowed toreach through to God. Mackie presents us with a highlyimplausible, strawman version of the FWD.9 He makes thewild claim that according to the FWD, "second order evils,such as cruelty, are logically necessary accompaniments offreedom, just as pain is a logically necessary pre-condition ofsympathy."10 He demolishes this fantastic claim by pointing

out the logical possibility of all free persons always freelydoing what is right. The point here is not that he presents astrawman version of the FWD but that he is willing to allowa defense to operate with weaker versions of 4 and 5.

The crucial question is w hat h ap pe n s to M ackie 's ded uctiv eargument when his quasi-logical rules are replaced by something m ore acceptable, for insta nce , 5 by 5, an d 4 by 4, or 4JNo contradiction can now be deduced; for, from

1. God is omnipotent;2. God is omnibenevolent;

1 0 6



The inductive argutttent frotrt cotnplete enumeration

3. Evil exists;4,. An omnibenevolent being attempts to bring about the

best overall situation it can; and5,. An omnipotent being can bring about any logically

possible state of affairs

it follows only that

6,. The actual world is the best overall situation, that is,the best of all possible worlds.

And 6„ unlike

6. It is not the case that evil exists,

is not obviously inconsistent with 3. And when 4, is replacedwith

4,. An omnibenevolent being tries to eliminate and prevent every morally unjustified evil,

it follows only that

62. It is not the case that morally unjustified evil exists.And again there does not appear to be any logical clash with3. Mackie might try to salvage his argument by replacing 3with

3,. Morally unjustified evil exists,

but the theist is not going to grant this. The fate of Mackie'sargument is that either i t must use additional premises thatare unacceptably strong or it must require the theist to granttoo much in regard to what kinds of evils exist.

T H E I N D U C T I V E A R G U M E N T F R O MC O M P L E T E E N U M E R AT I O N

We should not be surprised that the deductive argument

does not work. Philosophy just isn't that easy. A battlebetween two long-lived philosophical traditions, such as theism a nd athe ism , is not go ing to en d like a Mike Tyson fight

107



The deductive argument from mtil

with just one haymaker. But the logical problem of evil can

survive the demise of the deductive argument. With a littleimagination one can screw out of the essays by Mackie andMcCloskey an inductive argument from complete enumeration, laced with a good deal of conceptual analysis, for theproposition that it is logically impossible that both God andevil exist. Attempts to show that some proposition is necessary or impossible through an inductive argument are notuncommon. The use of a computer to establish some proposition as a theorem appeals to various empirical factsconcerning the working of a computer. In philosophy, arguments for something being a necessary conceptual truthoften proceed piecemeal by an empirical inquiry into the waylanguage is used.

The inductive argument from complete enumeration,unlike the previous inductive argument that posed the empirical problem of evil, ma kes no men tion of the am ou nt andkinds of evils that exist. It begins with some conceptual anal

ysis. If it is not true that it is logically impossible that bothGod and evil exist, then there must be some possible world inwhich both God and evil exist. In any such world, God musthave a morally exonerating excuse for permitting evil. Therefore, it must be logically possible that God have such anexcuse. It is at this point that the inductive argument fromcomplete enumeration is brought to bear to show that all ofthe different ty pes of excuses that we can think of are conceptually inapplicable to God in virtue of his essentially beingomniman. Unlike men, God cannot plead any limitation inpower, opportunity, or knowledge. This inductive enumeration, of course, might be incomplete; but it is up to the theistto show that it overlooks a kind of excuse that is available toGod in the face of evil.11

The theist cannot sit back and say that this enumerativeinductive argument fails to establish conclusively that it islogically impossible that God have a morally exonerating

excu se for pe rm itting evil, since our po w er of imagination islimited an d there fore it is logically poss ible tha t there is som esuch excuse that we failed to take into consideration. Nor can

108



The inductive argument front complete enumeration

the theist claim that the onus rests with the atheologian to

prove conclusively that it is logically impossible for God tohave such an excuse, and not on the theist to argue that it ispossible.12 This response won't do for at least two reasons.First, while it is true that the atheologian is required to support his claim of impossibility, the theist in turn is required tosupport its possibility, since it is an essential tenet of theismthat such an excuse should be available to God; otherwise nodefense of God in the face of evil is possible. Each has theburden of supporting their respective modal claims. Neitherhas the option of sitting out this dance. Second, as TerencePenelhum has shown, the theist attributes a certain sort ofmoral character to God, namely, one that agrees with themo ral code that the theist acce pts, tho ug h she m ay fail to liveu p to it in practice, and this grea tly limits the k ind of justification that could be given for God's permitting evil; and, forthis reason, the induction by complete enumeration argument has considerable plausibility.13 It is not being claimed

that Mackie and McCloskey intended to structure their argument in this way, only that they do in fact try to show why allthe types of excuses we can think of could not apply to God,and that this could serve as the basis of an induction bycomplete enum erat ion argum ent .

Theists certainly have not sat out this dance. Their tradition is rich in attempts to construct a defense in which God'smorally exonerating excuse is articulated in great detail. We

have already come upon a smattering of them in our crit iqueof Mackie's essay. And we now turn to our main task - thecritical asse ssm en t of these defe nse s. We can forget ab ou t th eaesthetic defense that justifies evil by the contribution itmakes to the overall beauty and interest of the totality; for toimpute such a purpose to God does not square with themoral code of traditional theism, as represented in Judaismand Christianity. That God would have, for example, theoverriding Leibnizian aim of creating a world that achieves ama xim um of diversity w ith a m inim um of explan atory principles is a distinctively Greek ideal that might apply to theDemiurge of the Timaeus or his m ode rn-day cou nterpart , the

109



The deductive argum ent from evil

God of Sir James Jeans, or even the god of Oscar Wilde, werehe to hav e one , but it could no t app ly to the D eity of biblicallybased theism.

Of more pertinence is the soul-building theodicy of St.Irenaeus and John Hick. According to this theodicy, whichalso can double as a defense, God has created a moral proving ground in which imperfect men, in an environmentfrought with evils, have the opportunity to develop, accordingto their own free will, in their moral, spiritual, and intellectualperfection, ultimately to come to know and love God. That

God should have such a distinctively unutilitarian goal is areflection of the theist's own antiutilitarian moral intuitions.The theist's moral code places great stress on intrinsicallyvaluable states for a person to be in and gives equal importance to the manner in which they are brought about,namely, by the subject 's own free endeavorings. This is atype of causal theory of value.

The FWD shares the same causal theory of value and alsoaccords great importance to the existence of free will, somuch so that it uses it to exonerate God for permitting theevils that result from the misuse of free will by created persons. The two defenses complement each other, the soul-building one justifying certain natural evils, such as diseaseand famine, as necessary for the significant employment offree w ill by cre atu res, the FW D justifying th e evils that resu ltfrom this em ploym ent. Tho ug h they com plem ent each other,they face different sorts of challenge, such that one of them

m ight fail and the other succ eed. Th e en su in g discussio n w illfocus exclusively on the FWD, but it will turn out that thevery reason w hy it fails, nam ely, th at God ha s a freedom-canceling control over created beings, also undermines the soul-building defense, since it too requires that God create freepersons in the manner described by the FWD.

THE FREE WILL DEFENSE

This defense concerns itself only with so-called moral evil -evil that is in some way attributable to the free will of finite

n o



The free will defense

persons, not necessarily members of our biological species.Initially it has nothing to say about nonmoral evils, oftenmisleadingly called natural or physical evils, misleadinglybecause it implies that the self and its faculty of free will areunnatural or nonphysical, which begs some pretty big quest ions. The usual examples of such evils are floods, earthquakes, famines, and the like. If we take a broad enoughview of "attributable to," so that not only positive use of freewill but also failure to use it or use it effectively can count, itmight turn out that these examples of natural evils really aremoral evils.

But even if this is granted, the FWD is still powerless torebut the atheological argument based on natural evil. It willno t do to say, a s do es Alvin Plantinga, th at all the natu ral evilsappealed to in this argument might be species of moral evil,for instance, caused by very powerful free nonhuman agents,such as the devil. While this is true, it fails to realize that theatheological argument based on natural evil is an impure

atheological one, due to the proposition that there is naturalevil being taken to be only contingent by the theist. In denying that there is in fact any natural evil, it is not shown thatthe initial set of this argument does not entail a contradiction.And, if it does, so does the proposition that the conjunctionof the propositions in its initial set is possibly true. Thus, toneutralize the deductive argument based on natural evil ,Plantinga must show not just that every alleged natural evilreally is or cou ld be a moral evil but that it is logically impossiblethat there be a natural evil. An d that he has not don e. N or do Ithink it can be done. And if so, wc must recognize that theFWD can work as a defense of God only for moral evil.

Before we get embroiled in the different versions of theFWD, it will be helpful to have an overview of what theyshare in common. All of them employ weaker versions ofGod's omnipotence and benevolence than that given by 5and 4. They begin with the "intention premise":

8. It is G od 's intentio n to create the bes t overall situa tion orthe best world that he can.14

1 1 1




This is followed by the "norm ative p rem ise" that rank orderspossible worlds in terms of goodness:

9. A world containing free persons who freely performboth right and wrong actions, but for the most part go right,is better than any possible world devoid of free persons.

The next two premises, which are respectively the "incom-patibil ist pr em ise" and the "God-could-be-unlucky prem ise,"have the job of blocking the objection that any free personscreated by God will always freely go right, resulting in moral

good sans moral evil:10. God ca nn ot cause or determ ine in any way w ha t a cre

ated person freely does; and11. It is logically possib le that God is con ting entl y un able

to create free persons who always go right.

In the possible world in which God is unlucky in the mannerdescribed in 11, he is morally exonerated for creating Mr.Rogers-type people, you know, "the very same people whoare bad some time are good some time."

The intention premise, 8, faces the objection from Chapter1 tha t th ere is n o w orld that is the best of all possible w orlds.To avo id this difficulty, it could be reco ns trued as ho ldingthat it is God's intention to create a "very good world," thatis, a world in which there is moral good sans moral evil. Hisbenev olence w ou ld not be im pu gn ed by his creating a worldcontaining 1,000 free persons who are paragons of moral

virtue, when he could have created a world containing 1,001such beings, and so on. In the ensuing discussion we willun ders tand 8 in this wa tered-dow n ma nner.

Even atheologians, with one minor exception, are willingto gra nt 9.15 Co nsid erab le fire ha s been directed at 10, but them os t vulne rab le pre m ise by far is 11. Th e trick is to restrictone of God's omniproperties so as to account for how hecould be unlucky and not be able to realize his first choiceand thereby have to settle for second best - the world with afavorable balance of moral good over moral evil - withoutcompromising this omniproperty so that i t renders God less

1 1 2




than that tha n which no ne greater can be conceived. By vary

ing the story we tell about the circumstances under whichGod is unlucky, we get different versions of the FWD thatmake different types of morally exonerating excuse availableto God.

The story told in the Plantinga version tinkers with God'somnipotence. His being able to create moral do-gooders depends upon contingent facts beyond his control concerningwhat possible free persons would do if God actualized them.Herein G od 's excuse is that of a lack of op po rtu n ity , sim ilar toI can't play the piano now because no piano is present or Icouldn't save the patient because I lacked the necessary surgical instruments or medicine. The story in the Robert M.Adams version finds God's exoneration in an excusable lackof know ledge, since there is su pp ose d to be no way he cou ldhave known in advance the moral evils that would resultfrom his creation of free p ers on s. (How could I hav e kn ow nthat the apple I gave to the trick-or-treater contain ed de ad ly

poison, since I bo ug ht i t at my friendly neigh borh ood sup ermarket.) Herein it is God's omniscience that is being playedwith. Both versions will face the problem of whether theyhave watered down the concerned omniproperty too much.Our exposition of these two versions will begin with Plan-tinga's and then bring in Adams's version as a way of escaping a formidable o bjection, b ut it will tur n o ut th at the A da m sversion faces even more formidable objections, thereby leaving us with no clearly viable version of a FWD. An d the cau seof the failure of the FWD - the impossibility for the God oftraditional theism to create free persons - also underminesthe soul-building defense.

The Plantinga version

Plantinga's version of the FWD is a thing of beauty that, it issafe to say, will serve as one of the cornerstones in theism'sresp on se to evil no t just for m an y ye ars to com e bu t for m an ycenturies. Because of its importance, it will be subject to the

113




sort of careful, detailed critical exposition that would be anexercise in misplaced pettiness for a lesser work. His firstpublished formulation, titled "The Free Will Defence," appeared in Philosophy in America (hereafter PA).16 This articlewas reprinted verbatim as chapter 6 in his God and OtherMinds with the addition of a final section that briefly attempts to neutralize the inductive argument from evil andthe m ysterious deletion of a pa rag raph on p age 195 of PA tha texplains why God cannot control the free acts of persons. Anew version, employing the ontology of possible worlds,

appeared as chapter 9 of his The Nature of Necessity (hereafterN N ) . A more popu lar presenta tion of this version w as givenin his God, Freedom, and Evil (hereafter GFE). In the c ou rse ofmy exposition it will be shown that the PA and possible-worlds versions differ only in their terminology, not theirontology. (His argu m ent do es gain in formal rigor, ho w eve r,by the use of the possible-worlds terminology.) My exposition will jum p aro un d betwe en the se different pu blis he d p resentations of his FWD. In his "Self-Profile" in the 1985 AlvinPlantinga Profiles V olume, (hereafter SP),17 he attempts tom eet objections to the incompatibilist prem ise an d s ug ges ts aw ay in which the FWD can be formu lated so that it do es notrequire that God foreknows w ha t will result from his creationof free persons, thereby moving quite close to the Adamsversion , wh ose defense is built on ju st this limitation in G od 'sknowledge. For this reason we will hold off discussion of thenew version of the FWD in SP until we discuss the Adams

version.As indicated, the heart of the FWD is the story that is told

of the possible circumstances under which God is unable torealize his first choice and thereby is morally exonerated forcreating the morally mixed-bag world. For Plantinga thestory will involve some contingent factor beyond God's control that renders him unable, through a lack of opportunity,to realize the goody-goody w orld . Mackie and his cohorts are

going to argue that God's omniproperties logically precludethere being any such factor. Herein we find an instance ofwhat has become the standard objection to any FWD - the

114



The free w ill defense

God-can-do-more objection. No one has yet pressed the op

posite objection that God cannot do as much as Plantingarequ ires with ou t neg ating the freedom of the person s he creates. It will turn out that this is the telling objection.

The most prominent of the God-can-do-more objectionsare based on either causal or theological compatibilism, theformer arg ui ng tha t since an act can be both free and causallydetermined, God could have assured the realization of hisfirst choice by a suitable selection of deterministic causal lawsand the initial state of the universe, the latter that since an actcan be both free and determined supernaturally by God, hecould ha ve realized this simply by willing that it be so. Thesesupernaturally determined events need not be causally determined by anything within the universe. Plantinga developshis FWD in a highly dialectical manner in the course of responding to these two compatibilist-based objections, beginning with that of the causal compatibilist. We will follow thedialectic of his exposition.

The causal compatibilist objection

Th is objection ho lds th at since an action or choice can be bothfree and causally determined in the sense of admitting of aHempelian deductive-nomological explanation (i .e. , a proposition reporting the action or choice is deducible from the

conjunction of causal laws and p ropositions repo rting conditions prior to it, without being entailed by the latter alone),God can make certain that all of the free persons he createsalways freely go right by (A) ordaining that certain strictde term inis tic law s hold , an d (13) creating p er so ns in circumstances such that they would be determined by these lawsalways freely to go right. '8 It takes considerable intelligenceto find just the right combination of laws and initial conditions, but this should be no problem for an omniscient being.Once God has done (A) and (B), he can sit back like the Godof De ism and just watch the inevitable beneficient unfoldingof what their conjunction entails.

" 5




Plantinga do es no t consider the case in w hich e ven ts adm itof only a Hempelian inductive-statistical-type explanation,no doubt because such a case could not serve as the basis foran objection to the FWD. The reason is that if God were todetermine that probabilistic laws hold, regardless of howhigh the probabilities, and determine the initial state of theuniverse, it would not logically ensure that persons alwaysfreely go right. Were God to try to determine this morallydesirable situation in this fashion, he could be faulted by thecausal compatibilist objector for not availing himself of the

better option of laying down strict deterministic laws so as tologically ensure this outcome.Plantinga's initial response to the causal compatibilist

objection is to deny that it is consistent to say that some actis both free and causally determined:

When we say that Jones acts freely on a given occasion, whatwe say entails . . . that either his action on that occasion is notcausally determined, or else he has previously performed an

undetermined action which is a causal ancestor of the one inquestion. (PA 189)

He expands on this purely negative account of freedom onpages 165-6 of NN:

If a person S is free with respect to a given action, then he isfree to perform that action and free to refrain; no causal lawsand antecedent conditions determine either that he will perform the action, or that he will not. (see also GFE 29)

The dispute between Plantinga and the causal compatibilist is the m od ern -da y c ontinu ation of the m edieval dispu tebetween Molina and Scotus over whether freedom is to beunderstood in terms of the "freedom of indifference" or onlythe "freedom of spontaneity." Whereas the modern-daydefender of the freedom of spontaneity - the causal compatibilist - basically follows the Hobbes-Hume account of freedom as absence of external compulsion (with the added

116




r equ i r e me n t o f an abse nce o f i n t e rna l o r p sycho log i ca l c o mpu l s io n ) , P l a n t i nga , t h e L ibe r t a r i an , r equ i r e s t h a t t he agen thas bo th the power to do the ac t and the power to re f ra in ,w i th t h i s r equ i r i n g an absence o f c aus a l de t e rmina t i on bu tn o t a g e n t d e t e r m i n a t i o n .

P lan t inga , ho w ev er , do es no t t ry to re fu te the causa l co m -pat ib i l i s t .1 9 I n s t ea d he m ak es a p r e sen t of t he wo rd " f r eed o m " to F l ew, com pa t i b i l i st ob j ec to r, and r e fo rm u la t e s h i sFW D us ing a d i ffe ren t locu t ion :

The Free Will Defender can simply make Flew a present of theword 'freedom' and state his case using other locutions. Hemight hold, for example, not that God made men free andthat a world in which men freely do both good and evil ismore valuable than a world in which they unfreely do onlywh at is good; but rather that God ma de m en such that som eof their actions are unfettered (both free in Flew's sen se an dalso causally undetermined) and that a world in which menperform both good and evil unfettered actions is superior toone in which they perform only good, but fettered, actions.By sub stitutin g 'unfe tter ed ' for 'free' thro ug ho ut this acco unt,the Free Will Defender can elude Flew's objection altogether.(PA 189)

This s t r a tegy fo r e lud ing the causa l compat ib i l i s t ' s ob jec t ionis en do r se d , a l t h o u gh no t s t a t ed , i n N N a n d G FE . O n pa ge s170-1 o f NN he wr i t es :

An d a pe rson is free w ith respe ct to an action A at a time f onlyif no causal laws and antecedent conditions determine eitherthat he performs A at t or that he refrains from so doing. Thisis not a comment upon the ordinary use of the word 'free' ;that use may or may not coincide with the Free WillDefender's.

And when he cons ide rs F lew ' s ob jec t ion in foo tno te 14 onpa ge 32 of GF E, he refers th e rea de r bac k to h is ear l ie r d i sc u ss ion in PA in w hi ch he de ve lo ps the m ake-F lew-a-g i f t -o f - the -w o r d r e s p o n s e .

117




This strategy for meeting the causal compatibilist's objection is, in my opinion, a disaster.20 First, if Plantinga reallyintend s to make a presen t of the wo rd "freedo m " to the com -patibilists, and thereby let them decide its meaning, hisaccount turns out to be inconsistent because almost all ofthem have insisted that a necessary condition for an action'sbeing free is that it is causally d ete rm ined . They have claimedthat a causally undetermined action is a chance, fortuitousoccurrence that cannot be attributed to the doer of the actionin such a way that he is morally responsible for it.21 When it

is realized that the compatibilist requires that an action is freeonly if it is causally determined, as well as physically andpsychologically uncoerced, Plantinga's account is seen to beinconsistent; for it may then be stated as follows:

God made men such that some of their actions are unfettered(both free in Flew's sense of being both causally determinedand neither physically nor psychologically coerced and alsocausally undetermined).

The consistency problem is easily solved by redefining"un fettered" as "ne ither physically or psychologically coercednor causally deter m ine d," only no w to face the objection thatthere is nothing particularly valuable about such "unfettered" actions; and, as a consequence, the objector is nolonger willing to grant the normative premise when it isreformulated in terms of such actions. The reaso n for this, asstated, is that they equate a causally undetermined actionwith a random or chance occurrence for which the agentcannot be held responsible. Mackie, the causal compatibilist,writes:

If freedom is randomness, how can it be a characteristic ofwill? And, still more, how can it be the most important good?What value or merit would there be in free choices if these

were random actions which were not determined by thenature of the agent.22

118



The free will defeat

Thus, Plantinga's attempt to reformulate his FWD in terms of

"unfettered" actions, even when formulated consistently,fails to achieve the desideratum of avoiding an ultimate clashof intuitions between himself and the causal compatibilist critic.All he achieves by this maneuver is to shift the clash of intuitions from the incompatibilist premise to the normative one.

By the time Plantinga came to write his "Self-Profile," herealized that his former strategy of making a present of theword "freedom" does not work for the reasons just given,and he disowned it with the rather understated remark, "Inow think this way with the difficulty too short" (SP 45). Herealizes that th ere is no w ay he can avoid a clash of intu ition swith the compatibilists, but now he is willing to tough it outand assert his incompatibilist intuition in opposition totheirs. As he sees it, the nub of the dispute is whether hisLibertarian conception of a free act is viable in that the agentcan be held responsible for it. A free act, for the Libertarian,is not a m ere chanc e, ran do m , or accidental occurrence, since

it is "agent determined" by a nonempirical self in a way thatdefies scientific description and explanation.23 Plantingadoes not fill in his Libertarian theory, but from what little hesays it ap pe ar s tha t he is in the same c am p as A ristotle, Jam es(in his gre at acc ount of the will in The Principles of Psychology),Campbell, Chisholm, Taylor, and Sartre. In support of theviability of his theory, he gives God as an example - theult imate uncaused-causer whom we emulate every t ime weserve as the unmoved-mover of one of our own limbs.Plantinga thinks that it is unreasonable for his critic not togrant the viability of this conception of God.24

By cha llengin g the com patibilist objector, Plan tinga do esnot achieve his original desideratum of avoiding a clash ofintuitions, but at least he is confident that his Libertarianconcept of a free act is a viable one. He fails to see that hisclash with his compatibilist opponent is not confined solelyto the viability issue but also breaks out on the normativelevel; for it is conceivable that his opponent would (begrudg-ingly) grant him the viability of his concept of a free act but

119



The deductive argument from evi

then go on to deny that such acts have the sort of value thatis required for the normative premise to be acceptable. Andwere the dispute to shift to this normative level, there is nodoubt that Plantinga again would be willing to get into the"Tis-Tisn't" sort of shouting match with his opponent.

W hat are we to m ake of these clash es of intuitio ns over theviability and, assuming viability, the value of a Libertarianfree act? It would be naive in the extreme to think that thesedisp utes are to be resolved by ap pea l to "ordin ary lan gu ag e";for while "ordinary usag e" at Calvin College and N otre D am e

University might favor Plantinga's intuitions, what counts asordin ary usage at the establish m ents frequented by Flew andMackie might well support theirs. Just how shaky our intuitions are in this matter is app aren t to anyo ne w ho has taug htthe free will-determinism issue to a group of introductorystudents. I t takes no more than thirty seconds to turn theminto a wild-eyed lynch mob of incompatibilists and aboutsixty seconds to turn them back into compatibilists. In myopinion, nothing short of an extensive metaphysical theoryof the self conjoined with a normative theory about the forensic and ethical dimensions of personal responsibility will"settle" this disp ute, m eanin g th at this is no t the sort of issuethat gets settled. It certainly is too quick to dismiss Plantinga's FWD because one's intuitions do not support its normative premise. And since the FWD is cou nterin g the chargeof logical inconsistency against theism, maybe all that shouldbe requ ired of it is that it be valid an d no t em plo y a ny plainly

implausible premise. Judged by this weak standard, Plantinga earns at least a draw with his compatibilist critic. TheFWD is still alive.

I believe that there is available to Plantinga a workablestrategy for achieving his original desideratum of avoidingany clash of intuitions with the causal compatibilist: findsome freedom-canceling cond ition accep ted by the latter thatis presen t in the case in which Go d p rede term ines the behavior of persons in the nomic ma nn er m en tion ed. W hile causalcompatibilists do not recognize the mere fact that an action iscausally determ ined as freedom canceling, t hey d o accept cer-

1 2 0



The free will defence

tain ways in which an action is causally determined as freedom canceling. Even the crudest of them recognize physicalcompulsion as freedom canceling, and this is extended bym ore recen t comp atibilists to actions that result from psy chological compulsions. Most of them, with one minor exception, recognize ways in which one person can usurp anotherperson's free will by gaining a habitual ascendancy over theirwill, such that all or most of their behavior is caused by, orunder the causal control of, the former.25 They accept, forexample, cases of extensive brainwashing, posthypnotic suggestion, or intentional control over the inputs to a brain-in-a-vat as freedom canceling. There are recognized limits to howfar one person can go in causally controlling the behavior ofanother person without negating the latter 's freedom. NowGod is a person, but his control over created persons is evenmore extreme than in these man-man cases; not only doeshe sufficiently cause all of their behavior by bringing aboutcertain instantial conditions and have the counterfactual

power to produce alternative behavior, he also creates thecausal structure of the universe, whereas finite controllersmerely take ad va nta ge of a given causal struc ture. The cau sal-compatibilist-based objection, therefore, winds up beinginco nsisten t, for it im pu tes to God this agreed u po n freedom-canceling control over created persons and yet holds thatthey are free.26

The causal compatibilist might search for some disanalogy

between the m an -m an cases and the God -m an case thatprecludes applying to the latter the same freedom-cancelingprinciples that hold for the former. The actions of the finitecontrollers are events in the universe that event-cause insome nomic manner the behavior of the victims. But God'sact of willing is not an event in the universe, and it does notevent-cause in some nomic manner the instantial conditionsin the universe that in turn event-cause the behavior of thecreated persons, but instead agent-causes these conditions.Thus the causal chain linking God's initial act of will withthe subsequent behavior of created persons is a mixed bagof agent and event causes. It could then be contended that

1 2 1






of the g rea t the i s t s ha ve bee n theo log ica l com pa t ib i l i s t s . Su chthe i s t s m us t f ind a de f en se tha t is in d ep en d e n t of the FWD .

Plant inga 's way with this object ion is , f i rs t , to ferret out i tsc rucial p r esu pp os i t io n , t he n a rg ue aga ins t i t, an d f ina lly u t il i ze th i s a rgument fo r the purpose o f desc r ib ing a poss ib lewor ld in which God has a mora l ly exonera t ing excuse fo ra l lowing mora l ev i l . He beg ins by cons ide r ing Mack ie ' s fo ll o w i n g p r e s e n t a t i o n of t h e t h e o l o g i c a l - c o m p a t i b i l i s t - b a s e dobject ion:

First I shou ld query the ass um ptio n that second o rde r evilsare logically necessary accompaniments of freedom. I shouldask this: if God has made men such that in their free choicesthey sometimes prefer what is good and sometimes what isevil, why could he not have made men such that they alwaysfreely choose the good? If there is no logical impossibility in aman's freely choosing the good on one, or on several, occasions, there cannot be a logical impossibility in his freelychoosing the good on every occasion. God was not, then,

faced with a choice between making innocent automata andmaking beings who, in acting freely, would sometimes gowrong: there was open to him the obviously better possibilityof making beings who would act freely but always go right.Clearly, his failure to avail himself of this possibility is inconsistent with his being omnipotent and wholly good.27

It w ou ld seem tha t Mack ie i s ap pe a l in g to the Le ibn iz ian do ct r ine , ca l led "Leibniz ' s lapse" by Plant inga for reasons tha t

wi l l soon emerge , tha tLL. God can ac tua l i ze any poss ib le wor ld .

Th i s is a co ns eq ue nc e o f the s t a t ed log ica lly r e s t r i c t ed acc ou n to f o m n i p o t e n c e :

5,. A n o m n i p o t e n t b e i n g c a n b r i n g a b o u t a n y l o g i c a l l ypossible s tate of affairs .

N o p a r t y t o t h e d i s p u t e w a n t s t o d o u b t t h a t a m o n g t h elog ica l ly poss ib le wor lds a re ones tha t con ta in mora l goodsans moral evi l , so the only i ssue i s the acceptabi l i ty of LL.

123




Plantinga is quick to present counterexamples to it. If God isnot a necessary being, LL and 5, must be restricted to possible worlds or states of affairs in which God exists. But thismight not be enough; for, if there are necessary beings suchas numbers and properties that do not admit of creation ordestruction, they must undergo yet further restriction. And,also, let us not forget about the absurdity of God's having toplay causa sui by actualizing himself. To meet these counterexamples, they can be recast as:

LL,. God can actualize any possible world w such that it islogically consistent that God actualizes w; and

5J. An omnipotent being can bring about any state of affairs s such that it is logically consistent that an omnipotentbeing bring it about that s.

Mackie's objection can be reformulated so as to accommodate these more restricted accounts of what God, an omnipotent being, can do. It has been agreed that among the possible worlds are morally perfect ones. And since it is notlogically inconsistent that God actualize one of them, it follows that God can do so; and, given his benevolence, itfollows that any possible world containing free persons thatGod actualizes contains moral good sans moral evil. Butplainly this is not the case.

The crucial presupposition in Mackie's reformulated objection is

M. It is necessary that God can actualize a possible worldin which all free persons always freely go right.This is necessary because what it is logically consistent forGod to actualize cannot vary across possible worlds. Plantinga is going to argue for the negation of M:

- M . It is possible that God cannot actualize a possibleworld in which all free persons always freely go right.

This will be called the "depravity proposition" because itclaims that every possible free person could be depraved inthat were God to attempt to actualize any one of them, the

124



Tke free wili defeme

resulting concrete person would freely go wrong at leastonce . As a step along the way to proving - M , Plantinga w illprove that

CL. It is necessary that there is some possible free personthat God cannot actualize,

which will be referred to as the "counter-Leibniz proposition." O nce - M is established, P lantinga can com plete hisFWD by ad din g that in the possible world in which w hat - Mclaims to be possible is realized, God is morally exoneratedfor settling for his second choice - the creation of the Mr.Rogers- type persons.

Before w e can un de rsta nd P lantinga's arg um en ts for - Mand CL, we must see what he means by "possible person"and "possible world," as well as their instantiation and actualization respe ctively. Th e PA version w orks w ith po ssiblepe rso ns , the N N a nd GFE version w ith possible w orlds, but,it will be seen, the ontology is exactly the same, only the

terminology being different. A possible person is a higher-order Platonic entity, being a set of abstract properties thatmay or may not be instantiated, while a possible world is aset of abstract states of affairs (or propositions) that may ormay not obtain (or be true). More precisely, a possible personis a maximal compossible set of properties each of whichcould be possessed by a single person (PA 193).29 The set iscompossible in that it admits of the logical possibility of coin-stantiation by a single concrete individual, and it is maximalbecause for every property that could be possessed by aperson, either this property or i ts complement is included inthe set. A possible person is the same as a Leibnizian complete concept of a person. A possible world, however, is amaximal compossible set of states of affairs or properties. Iwill use "universe" to refer to a maximal spatiotemporalaggregate - the totality of what exists and happens in spaceand /or t ime - an d "co sm os" to refer to the maximal aggreg ateof actually existent entities. Whatever exists in the universealso exists in the cosmos but not necessarily vice versa, sinceabstract entit ies such as numbers exist in the cosmos alone.

125




For Plantinga th ere m us t be one an d only o ne cosm os, as wellas universe, an d a given cosmos m ust actualize one and onlyone possible world.

Another way in which Plantinga conceives of a possibleworld is as a maximal compossible set of abstract propositions, that is, all the propositions that would be true werethat world to be actualized. Such a set is called a "worldbook," and corresponding to every possible world is such aworld book. A given cosmos makes true one and only oneworld book. It is important to emphasize that possible per

sons, possible worlds, and world books are abstract possi-bilia. Each is a way things might be, and that is notsomething locatable in space and/or t ime. Moreover, thesepossibilia are necessa ry en tities and th u s exist in every po ssible world.

It should be apparent that the same Platonic ontologyinforms both the possible person and possible world formulations of the FWD. What is more, given Plantinga's definit ions, it follows that every possible person is identical withsom e possible w orld, th ou gh n ot conversely, since som e possible worlds are devoid of persons. This is a consequence ofthe fact that the only restriction Plantinga places on a property that is a m em be r of the set that is a possible per so n is thatit is possible for a person to have it. This allows us to includerelational prop ertie s and by usin g sufficiently rich spatial andtemporal relational properties, for instance, being born somany years after the last dinosaur died, being a person who

dies while there is a full eclipse of the sun, and so on, we cancome out with a complete description of an entire cosmos,including all of the other personal inhabitants of it.

Of special interest for our purpose is a possible significantly free person, that is, a possible person containing thepro pe rty of be ing free in respect to som e action ha vin g mo ralsignificance. For such a person there is at least one morallysignificant action A and set of circumstances C such that this

person includes the disjunctive property of either-freely-doing-/4-or-freely-refraining-from-doing-/4-in-C. Since a possible person is m axim al, it also inclu des th e pro pe rty of do ing

126



The free wiM defense

A or the property of not doing A. For every possible free

person containing the property of freely-doing-/ -in-C thereis a numerically distinct possible person that includes all ofthe sam e prop erties except for inclu ding freely-refraining-from-doing-/\-in-C instead. Let us call such a pair of possiblefree persons an "incompatible pair." Whenever you freelyperform an action, you instantiate one member of such a pairto the exclusion of the other. In what follows we shall consider only significantly free possible persons and, for short,will call them "possible persons."

With these terminological and ontological points out of theway , we can begin to app roach Plantinga's argu m en t for CL,the counter-Leibniz proposition. The argument attempts toshow that for any incompatible pair, God will be contingently unable to actualize one person in the pair. Let ourspecimen incompatible pair be P and P,, who include all ofthe same properties save for P's including freely-doing-A-in-C and P,'s instead includ ing freely-refraining-from-doing-/4-

in-C. Our revised Leibnizian of the LL, type will hold thateach is such that God can actualize or instantiate it, thoughhe cannot, of course, coinstantiate them. Plantinga, the Libertarian, disagrees, because he thinks it is logically inconsistent that God actualizes or instantiates either of them. Thereason is that God's actualizing or instantiating P, for example, consists in his causing there to exist a person having allof the properties included in P, and thereby God causes or

de term ine s this person freely to do A in C; but this is logicallyinconsistent according to the incompatibilist premise. For theLibertarian it must be the agent herself that is the (agent-)cause of a free act, not some condition external to the agent,though such conditions might limit the range of possibilitiesfrom which the agent can choose and "incline" it in a certaindirection, requiring greater effort of will for it to pursue anopposite course of action.

If God cannot actualize a possible person simply by super-naturally willing that it be actualized, how does he do it? It ishere that Plantinga has an incredibly ingenious and controversial story to tell. Again w e mu st begin w ith som e term ino log y.

127



The deductive argument ffem evil

What God does is to actualize what I will call a "diminishedpossible person" and then leave it up to the created personwhat it will freely do. Each possible person contains a diminished possible pers on that is its largest pro per subse t of p roperties that is such that for any action A, it neithe r include s orentails freely doing A no r inclu des or entails freely refrain ingfrom doing A, in which a property F includes or entails another property G just in case it is logically imposs ible th at F beinsta ntia ted and G not be. (W hat I call a "diminished possibleperson" corresponds with Plantinga's second definition of a

"possible person" on page 196 of PA. His dual use of "possible person" in PA creates difficulties for the reader, which,I hope, my terminology helps to avoid.) We will also refer tosuch a subset as a "freedom neutral" set of properties. Eachproperty included in a set of properties could be freedom-neutral and yet the set as a whole not be, for the set couldcontain either - freely - doing - A - or - freely - refraining - from -doing-A and doing A, Any incompatible pair will contain aspro pe r subsets the same diminished perso n or set of freedom-neutral properties. Thus, P's diminished person, DP, is numerically one and the same as DPt - the diminished personcorresponding to P,. God performs the same creative actw he n he endeavo rs to actualize P as he does wh en he endeavors to actualize Pu namely, he supernaturally wills that thediminished person DP be instantiated or actualized. Intuitively, we can think of this as God's creatively determiningevery feature of the universe up until the time at which the

created person, the instantiator of DP, freely does A or freelyrefrains from doing A. Since P and P, are each identicalwith some possible world, it follows that this is the samecreative act God performs in actualizing either of these twoworlds.

The question is what would result if God were to instantiate DP. Would the instantiator of this diminish ed pe rso n orset of freedom -neutral pro pe rties freely d o A or freely refrain?Plainly, the instantiator must do one or the other, since heha s t he disjunctive pro pe rty of either-freely-doing-A -or-freely-refraining-from-doing-A. Th us, it is either tru e th at

128



The free nM defense

F. If DP w ere instantiate d, the instantiator w ould freely do

A; orF'. If DP w ere instan tiated, the instantiator w ould freelyrefrain from doing A.

Let us call a subjunctive conditional whose antecedentreports the instantiation of a diminished possible person andconsequent the performance of a free action by the instantiator a "free will subjunctive conditional," for short, an "F-conditional."30 If F is true , the n w ere G od to instan tiate DP, itwould result in P's being actualized; whereas, if F' is true,w ere God to actualize DP, it w ou ld result in P,'s being actualized. S ince F an d F ' are logically in co m pa tible , it follows thatif F is true God is unable to actualize P„ and if F' is true Godis unable to actualize P. But necessarily, one of them is truean d therefore necessarily true tha t Go d can no t actualize P orcannot actualize P,, which proves CL.

This proof assumed that the law of conditional excludedmiddle holds for F-conditionals. Herein the necessarily truedisjunction is formed not from the disjunction of an F-conditional with its negation, as is the case when the weakerlaw of exc lude d m idd le is ap plie d, b ut from the disjunction ofan F-conditional with an F-conditional containing the sameantecedent and the denial of the former's consequent, as isthe pr ev iou s case w ith the disjun ction of F and F '. In SP,Plantinga gives a proof for CL that applies only the law of

excluded middle to F-conditionals. It begins with what Plantinga calls "Lewis's lemma," which, when translated into myterminology, says that God can actualize a possible person Pcontaining the property of freely doing A on ly if it is tru e tha tif God w ere to actualize its dim inish ed pers on DP, the instantiator would freely do A. It next is claimed by appeal to thelaw of exclu ded m iddl e th at it is eith er tru e or false tha t F. Ifit is false, the n, given Lew is's lem m a, G od ca nn ot actualize P;and, if it is true, then he cannot actualize P,. As we shall seewhen we consider Adams's version of the FWD, it is controversial whether an F-conditional could be true. Even if theyall were necessarily false, the law of excluded middle version

129




of the proof of CL would work; for if both F and F' are false,then God can actualize neither P nor P,.

From Plantinga's argument for CL, it is only a few shortsteps to the establishment of the deprav ity proposition , - M .At the outset let us confine ou rselves to possible pe rso ns thatinclude the property of being free with respect to only oneaction, such as our forementioned persons P and P,. Whatwe establish then can be generalized to more complex possible persons. Any incompatible pair of such simplified persons is a Dr. Jeckyl and Mr. Hyde pair, the former being the

one that contains the property of freely doing A (which isthe morally right thin g to do ), the latter the pro pe rty of freelyrefraining from doing A (which is the morally wrong thing todo). It has already been shown in the argument for CL thatGod might not be able to actualize P, the Dr. Jeckyl memberof the pair, since F could be false. But what could be true forthis particular Dr. Jeckyl and Mr. Hyde pair could be truefor all of them. Every incompatible pair of this sort could besuch that it is true that if God were to instantiate the diminished possible person common to both, the instantiatorwould freely do the morally wrong alternative. Under suchunfortunate circumstances, God can actualize only Mr.Hydes, and therefore will not attempt to instantiate any ofthese simple possible persons , assu m ing that his bra nd of benevolence req uires that th ere be a favorable b alance of moralgood over moral evil.

The result can be generalized so as to apply to more rich

possible persons that contain the property of being free inrespect to more than one action. It could still be the case forevery such person that it is true that if God were to actualizeits diminished person, the instantiator would freely gowrong with respect to at least one of these actions. And thissuffices to establish - M - th at it is possible that Go d ca nn otactualize a possible world in which all free persons alwaysfreely go right.

At this point Plantinga can complete his FWD by claimingthat in the possible world in which the truth-values of theF-conditionals preclude God from actualizing any Dr. Jeckyls

130




or, more generally, possible perso ns contain ing the prop erty

of always freely doing what is right, he is excused for creating the Mr. Rogers sort of persons, provided for the mostpa rt they freely go right, wh ich I am s u re is true for th eMr. Rogers types. In such a world God can plead that hedid the best he could but was screwed by those damned F-conditionals.31 This compares with Scotty saying to CaptainKirk, "Sorry about that, Skipper, but I gave it all the phaserpower I had." This completes our rough sketch of Plantinga'sFWD account of the possible world in which God is unluckyand thereby morally exonerated for allowing moral evil.

Those damned F-conditionals

The account given so far of Plantinga's story of creation failsto deal with some key issues concerning F-conditionals: the irmodal status; God's knowledge of them; and from whencethey derive their truth-values, assum ing tha t they have themat all. Different ways of handling these issues produce different versions of the FWD. It will now be shown that Plantinga's FWD is committed to the following theses about F-conditionals:

I. Every F-conditional has a con tinge nt tru th-v alu e, th at is,is contingently true or contingently false,

II. God kn ow s the truth-value of all F-conditionals prior tohis creative decision, and

III. God does no t determ ine the tru th-v alue s of F-conditionals.

Theses I and II together make up the doctrine of God's"middle knowledge," which doctrine is rejected by theAdams's version. Each of these theses will now be examinedin some detail.

Another way of formulating the doctrine of God's middleknowledge is that God foreknows for every diminished possible person what free actions would be performed were thatperson to be instantiated. This is an im po rtan t conside ration,

131




s ince in as se ss i ng one ' s respo ns ib i l i ty an d b la m e for so m ee v e n t , q u e s t i o n s of w h a t w a s k n o w n a n d w h e n it w a s k n o w nare r e l eva n t . We can imag ine a be l l ige ren t new sca s te r a sk ing"Jus t wha t d id God know? and When d id he know i t ?" ForP lan t ing a , God ca nn o t ge t aw ay wi th the Rona ld Re aga n-t y p e r e s p o n s e :

In the first place, there wasn't anything to be known (sinceF-conditionals lack a truth-value). In the second place, no onetold me what their truth-values were. And, moreover, if theyhad, I w ou ldn 't have understoo d w hat they were talkingabout. And, furthermore, I do n't remem ber wha t ha pp en ed ,but if I did r em em ber any thing it w ould be that I w as toldabout their truth-values too late - after I had m ade m y creativedecision.

(Tha t Reagan can ge t away wi th th i s type o f answer and Godca nno t sh ow s tha t Re aga n is m ore pow er fu l tha n G od , s incehe can do someth ing tha t God canno t . ) Many o f the th ingstha t P lan t inga sa ys in the PA an d NN -GF E fo rmula t ions p rec lude God ' s t ak ing th i s know-no th ing t ack :

Now God is said to be omniscient and hence knows, withrespect to any person he proposes to create, whether thatperson would or would not commit morally evil acts. (PA 192)

The athe ologian 's pro pe r retort, I think, is as follows. Sup pos ew e con clude that n ot even G od can cause it to be the case thatI freely refrain from A. Even so, he can cause me to be freewith respect to A, an d to be in some set S of circu ms tance sincluding appropriate laws and antecedent conditions. Hemay also know, furthermore, that if he creates me and causesme to be free in these circumstances, I will refrain from A.(NN 172)

Th ese tw o qu o ta t ion s a re no t dec i s ive , bec ause P lan t inga i sa t t r ibu t ing a be li ef in G od ' s m idd le kno w led ge to h i s the o logica l compat ib i l i s t objec tor. But he does not cha l lenge th isbelief. The fo l lowing quo ta t ions a re dec i s ive , because P lant inga speaks in the f i r s t person:

132




Our question is really whether there is something Curleywould have done had this state of affairs been actual. Wouldan omniscient being know what Curley would have done -would he know, that is, either that Curley would have takenthe bribe or that he would have rejected it? The answer, Ishould think, is obvious and affirmative. There is somethingCurley would have done, had that state of affairs obtained.(NN 180)

The fo l lowing makes c lea r tha t God possesses h i s midd le

knowledge p r io r to h i s c rea t ive dec i s ion :

Further, God knows in advance what Curley would do ifcreated and placed in these states of affairs. (NN 185-6)

S imi la r remarks a re made in GFE, Cur ley, the c rooked po l i t ic ian , be ing rep laced by Maur ice , the oa tmea l ea te r :

Now God no doubt knows what Maurice will do at t ime t, ifS' obtains. He knows which action Maurice would freely perform if S ' were to be actual. (GFE 42-3)

The fol lowing is of specia l impor tance , s ince i t c lear ly a t t r ibu tes a con t ingen t t ru th -va lue to an F-condi t iona l :

Now we merely supposed that DP is such that if it is instantiated, its instantiator will perform A. And this supposition, iftrue at all, is merely c on ting en tly tru e. (PA 195; I hav e a lteredPlantinga's names so as to match my above example of anF-conditional.)

Toge ther these quo ta t ions make c lea r tha t P lan t ingaa t t ri b u t e s m i d d l e k n o w l e d g e to G o d a n d t h e r e b y p r e v e n t sh im f rom tak ing the Reagan "excusab le" ignorance o r l ack o fknowledge way ou t ; f o r when God c r ea t e s f r ee pe r sons , he

a l r eady knows wha t t hey w i l l do and t hus doesn ' t wa t ch t heunfo ld ing o f the i r pe rsona l h i s to r ies in the way in which"hockey pa ren t s , " f o r example , wa t ch t he i r son p l ay i n agame , ho ld ing t he i r b r ea th , hop ing t ha t he wi l l com e t h r o u g h

133




and being disappointed w he n he does not - O h He just gaveup the puck in his ow n zone Oh He just took a du m bpenalty This sho w s that the Holiday Inn slogan "The bestsurprise is no surprise" isn't always true, judging by theamount of moral evil there is.

While it is clear that P lantinga accepts th ese s I an d II in t heversio n of the FWD given in the FA an d N N-GFE formulations, our question is whether this version requires that heaccept them. From what little he says about omniscience, itap pea rs that h e think s tha t II is entailed by I: If F-conditionals

have contingent truth-v alues, then God kno ws w hat they areat the time of his creative decision. This entailment holds ifwe accept, as he seems to, the tough account of omnisciencegiven by traditional theism - that God essentially knowsevery true proposition.32 For this reason, the only issue iswhether his FWD requires thesis I, and we can confine ourselves to this issue alone.

Given Plantinga's Libertarian account of F-conditionals, itis not obvious that thesis I is true; for, according to thisLibertarian theory, the consequent of an F-conditional is notdetermined by any antec eden t condition(s). Such conditions,at best, can render it probable that the consequent will berealized once the antecedent is actualized by inclining thefree person in the direction of what it reports. We shall findthat it is on this basis that Adams argues that it is conceptually or logically impossible that F-conditionals be true andthat they thereby could not have a contingent truth-value;

they are necessarily not true, which is the case if they areeither necessarily false or necessarily neither-true-nor-false,that is, truth-valueless.

Ad am s's arg um en t for the denial of thesis I w as pub lishe dafter Plantinga's three published formulations of his FWD."When Plantinga subsequently reconsidered his FWD in SP,he was again ready to take the ecumenical route to avoid aclash with h is critic, in this case A da m s, by reform ulatin g hisFWD in terms of m utua lly acceptable prop ositio ns. (We recallthe disastrous result of his earlier ecumenical effort when hemade a gift of the word "freedom" to his causal compatibilist

134



The free urill defense

objec to r and re fo rmula ted h i s FWD in t e rms o f the t echn ica lno t ion of "u nf e t t e re d . " ) His ho pe n o w i s to f ind a fo rmu lat ion tha t wi ll no t re qu ire e i th er the s is I or I I, the reb y av oid inga c l a s h w i t h Ada ms :

An interesting project would be to develop in detail a versionof the FWD that does not involve either middle knowledge orF-condition als (cou nterfactua ls of freed om ). (SP 50)

Th a t t he r e a r e F - c o n d i t i ona l s m ea ns " t h a t F - cond it i onal s h av ea de te rm in a t e t ru th va lu e" (SP 49). I a s su m e tha t by "de te rm ina t e t r u th va lu e" he m ea ns " con t i nge n t t r u th va l ue " andnot tha t f a ta l i s t i c no t ion o f now be ing unpreven tab ly t rue o runp reve n t a b l y f a l s e , w h ich , a cco r d ing t o t he A rab i c comm e n t a t o r s o n C h a p t e r 9 of de Interpretatione is p robab ly w ha tAr i s to t l e mean t when he sa id o f p ropos i t ions abou t fu tu recon t ingen t s tha t they a re no t ye t e i the r de te rmina te ly t rue o rde te rmina te ly fa l se . Tha t the re a re F-cond i t iona l s , the reby, i sjus t ano ther way o f express ing thes i s I .

W ha t fo l lows is s u p p o se d to be a "s im pl i f ied " ( ) vers ion oft he a r gu m en t fr om N N for CL an d - M , wh ich h a s a l r e adybe en s ke tc he d. It m ak e s expl ic i t w h a t is only impl ic i t in theNN fo rm u la t i on , na m e ly, Lewi s ' s l emma , wh ich s ays , i n myte rmino logy, tha t God can ac tua l i ze a poss ib le pe rson conta i n in g the pr op er t ie s of f ree ly d oi n g A, . . . A„ only if it ist ru e tha t if h e we re to ins tan t i a te i ts d im in i s he d pe rso n , theins tan t i a to r would f ree ly do A, . . . A„. Th i s a r gu m en t , l ike

the NN one , app l ies on ly the l aw of exc luded midd le toF-cond i t iona l s , no t the s t ronger l aw of cond i t iona l exc ludedm idd le , a n d t he reb y wo rk s ev en if eve ry F-cond i t iona l isfalse. Unde r t h i s c o nd i t i on , g i v en Lewi s ' s l emma , God ca nno t a c t u a l i z e any p o s s i b l e pe r son a n d t he r eby canno t a c t u a li ze one con ta in ing the p roper ty o f a lways f ree ly go ing r igh t .For th i s r eason he conc ludes :

Indeed, strictly speaking the present argument doesn'tdepend upon the assumption that any F-conditionals are true;it could be, for all the argument presupposes, that all suchF-conditionals are false. (SP 52)

*35




But that the argu m ents for CL and - M go thro ug h, andmust go through, if all F-conditionals are false hardly showsthat his FWD works even if there are no F-conditionals,which is what he thinks he accomplishes by formulating hisargument without appealing to the law of conditional exclud ed m idd le. It is on e thing for all F-con ditionals to be falseand quite another for them to lack a contingent truth-valuebecause they are necessarily not true.

Plantinga seems to forget that his FWD does not consistonly in the proofs of CL and - M bu t also mu st go on todescribe, utilizing - M , the possible wo rld in wh ich Go d ismorally exonera ted for perm itting m oral evil. In such a w orldGod succeeds in actualizing a possible person, from whichit follows in accordance with Lewis's lemma that some F-con dition al is tru e in that wo rld; an d , the refo re, it is false tha tevery F-conditional is necessarily not true , pace A da m s. T hu swe see that the FWD has both a negative and a positive component. The negative one concerns the conditions underwhich God is unable to create certain kinds of possible per

sons: The argu m ents for CL and - M perform this negative task.The positive com po nen t concerns the cond itions un de r w hichGo d is able to create certain free p ers on s: This is co nta ine d inthe description of the "unlucky" world in which God is excused for allowing moral evil. Plantinga's ecumenical effortfails, because the positive component requires thesis I.

It does not work even for the negative component. Thiswill be sho w n by ded uc ing an explicit contradiction from theconjunction of the premises of his argument for CL in NN(which is the same as that in SP) with the denial of thesis I -that F-conditionals necessarily are not true. This argumentattempts to prove that given any two possible worlds inwh ich God exists, W an d W", such th at there is a pe rso n w hofreely does action A in W bu t freely refrains from do in g A inW', either it is contingently impossible that God actualize Wor contingently impossible that God actualize W*.

For let W be a world where God exists, where Curley is freewith respect to the action of taking a $20,000 bribe, and where

136



The pee wi® defense

he accepts it; and as before, let T be the largest s tate of affairs

God stron gly actualizes in W.

G od's actualizing T(GT) includesneither Curley's accepting the bribe (A) nor his rejecting it{-A); so there is a wo rld W where God strongly actualizes Tand in which Curley rejects the bribe. Now(24) GT -> Ais either true or false. If (24) is true then by the previousarg um ent God could no t have actualized IV . O n th e oth erhand, if (24) is false, then God could not have actualized W(NN 182-3),

i n wh ich " -»" m ea ns " sub junc t ive cond i t i ona l ly im p l i e s " a n d"God ' s s t rong ly ac tua l i z ing T" m e a n s t h a t G o d ' s s u p e r n a t ural will is the sole free cause of T,34 in w hi ch T, ac co rd in g tom y te rm ino logy , is t he d im in i sh ed C ur l ey poss ib l e pe r so n o rCur l ey ' s f r eedo m -neu t r a l p rop e r t i e s .

My indirect proof of thes is I , re la t ive to the premises of th isa rgumen t , w i l l u t i l i z e t hese abb rev ia t ions : "L" a n d " M | "

m ea n , resp ect iv ely, " i t i s logical ly ne ce ss ar y th a t" an d " i t i slog ica l ly poss ib le tha t" and "Mc" m ea ns " it i s con t ing en t lypossible that ." I t begins wi th the denial of thes is I :

-•I. L | ev ery F-c ond i t ional is no t t ru e ,

t o wh ich a re t hen add ed the p re m ises of t he fo rmer a rg um en t . It i s ap pa re n t tha t the law of exc lud ed m idd le is a ppl ied to F-c on di t io nals , g iving us

12. Ev ery F-c on dit io na l is t ru e or false. 'O the r obv ious p remises a r e

13. W is a poss ib le wo r ld in w hic h G od ex is t s ; an d14. W* is a pos s ible wo rld in w hi ch G od ex is ts .

The nex t two p remises a r e based on Lewis ' s l emma:

15. L| [(it is true (G T -> A) z> (~ MC God actual izes W*)] ;a n d

16. L, [(It is false (G T -* A) 3 ( - Mf God ac tua l i ze s W)] .

P l an t inga a s s um es , fu r the rm ore , t ha t it i s l ogi ca lly, a l t h ou ghmaybe no t con t ingen t ly, pos s ib l e t ha t God ac tua l i ze any possible world in which he exis ts :

137




17. L| {w)(w is a possible world in which God exists => M\God actualizes w).

From - I it follows that

18. (p)(p is an F-conditional z> Upis not true).

It is the case that

19. (G T -» A) is an F-conditional.

And from 18 an d 19 it follows tha t

20. L\{GT -> /I) is not true.From 16 it follows that

21. L, [Mc God actualizes W 3 It is true (G T -> A)].

And by an uncontested theorem of modal logic it followsfrom 20 and 21 that

22. Li - Mc God actualizes w.

And this entails that

23. - M | God actualizes w.

Given that w is a possible world in which God exists, it follows from 17 that

24. M| God actualizes iv.

And by conjoining 23 with 24, we have our explicit contradiction.

It must be reiterated that this is not intended to be anindirect proof of thesis I, only a demonstration of the logicalcom m itmen t to it by the prem ises of Plantinga's argu m ent forCL. Its success as an indirect proof of thesis I hinges on theacceptability of Lewis's lemma, of which 15 and 16 are specialinstances; but we must postpone discussion of this until wehave completed our expositions of both the Plantinga andAdams versions of the FWD. Right now our only concern iswith how F-conditionals enter into Plantinga's FWD. It hasalready been shown that his FWD is committed to theses Iand II - the doctrine of God's middle knowledge - that F-

138



The free mill defense

conditionals have contingent truth-values that are known toCod. But whence do they derive their truth-values?

One possible answer, which Plantinga rules out by appealto his incompatibilist premise 10, is that God determinestheir truth-values. God's sovereignty must be restricted, inaccordance with the <j2 accou nt of om nip ote nc e, to w ha t itis logically consistent for God to bring about or determine.Plantinga claims that if God were both to determine an F-conditional's truth-value and actualize its antecedent, hewould in effect be determining what the created personfreely does, which is inconsistent with it being done freely(PA 195). Such du al dete rm ina tion is freedom canceling in thesame way that God's determining both what deterministiclaws hold and the initial state of the universe negates the freedom of the actions that inevitably result from their conjunction.

Even if Plantinga's incompatibilism is acceptable, it showsonly that God cannot both determine an F-conditional's truth-value and actualize or instantiate its antecedent, not that he

couldn't do either of these alone. Supposedly, given God'sfree will, there arc some possible worlds in which God doesnot elect to create free persons. Why couldn't he determinethe truth-values of F-conditionals in such worlds? Mightthere not be possible worlds in which God comes upon pre-existent persons who instantiate some diminished possibleperson and then determines the truth-values of the relevantF-conditionals concerning them? In both kinds of possibleworld, God does only half the job, determining the truth-values of the F-conditionals, rather than determining onlywhich F-conditionals, if any, get their antecedent instantiated, as is the case in the Plantinga story of creation. Plainly,

. ne ither Plan tinga no r the theistic traditio n in genera l iswilling to countenance such possible worlds, but no one,to my kn ow ledge, has devised an y explicit argu m ent againstthem. Three arguments will now be advanced for the conclusion that God does not determine the truth-values of F-conditionals in any possible world.

The first argument was given by William Wainwright ina letter to me. The following is a modified version of this

139



The deductive a rgument from evil

argument. It attempts to deduce a contradiction from theassumption that God makes any F-conditional true. Let ourrandomly chosen F-conditional be our old favorite:

F. If DP were instantiated, its instantiator would freely do A.

Assume for the purpose of indirect proof that

26. Go d ma kes it true tha t if DP w ere to be ins tantia ted, itsinstantiator would freely do A.

This entails the following two propositions:

27. If DP we re instantiated by G od, its instan tiator w ouldfreely do A; and28. If DP w ere instantiated by God, G od w ould m ake it

true that its instantiator freely does A.

Next comes the incompatibilist premise of the FWD:

10. God cannot cause or determine in any way what acreated person freely does.

An d from the conjun ction of 10 an d 28, it follows tha t29. If DP were instantiated by G od, i ts instantiator wo uld

not freely do A.30. Pro pos itions 27 an d 29 are logically inc on sisten t; a nd31. It is not true that God makes it true that if DP were

instantiated, its instantiator would freely do A.

This argument is vulnerable on two counts. The inferenceof 28 from 26 is du bi ou s; and 30 rests on th e contro versia l lawof conditional excluded middle. Still the argument is notwithout some interest and merit.

The next two arguments, which also appeal to some theological incompatibilist premise, are based upon the followingprinciple concern ing G od's sovereign ty, w hich I will call the"principle of sovereign type-invariance":

P. If in on e w orld God dete rm ine s wh eth er or not a certaintype of entity exists, happens, or is the case in that world orpart thereof then God determ ine s in every world in which

140




he exists whether or not this type of entity exists, and so on,in that world and every part thereof; and , if God dete rm inesin any world the truth-value of a proposition, then he determines the truth-value of every proposition of that type inevery world in which he exists.

This principle is consistent with 52. It allows there to betypes of states of affairs that fall outside the domain of God'ssovereignty, such as the existence of abstract entities and thetruth of necessary propositions. Although the principle ofsovereign type-invariance has never been explicitly stated, Ibelieve that it is implicitly assumed by the traditionalacc ou nts of G od's sovere ignty. The denial of P leads to an omalies. If we permitted God to exercise sovereignty over theexistence, say, of rocks in world iv but not in ano ther wo rldw' in wh ich he also exists or in one region of w but not another, i t would raise anomalous questions: "Why does God'ssovereignty extend just that far and no further?" "Why doeshe have the rock concession in w bu t n o t . a ; 7 " "Is the re

someone else who also must get a piece of the action?" OkayBig Al, you control the southside of Chicago and I got thenorthside.

This is the first argument based upon P:

32. W hethe r or not the re exist free pe rso ns , that is,whether or not any F-conditionals have their antecedentsinstantiated, is the type of thing that falls within God's sovereignty [premise];

33. In some possible world God determines whether anyF-conditionals get their antecedent instantiated [from 32];

34. In every world in which God exists, he determineswhether any F-conditionals get their antecedent instantiated[from P and 33];

35. In some world God determines that some F-condi-tional's antecedent is instantiated [premise];

36. It is logically inconsistent for God to determine both

that some F-conditional's antecedent is instantiated and itstruth-value [one version of theological incompatibilism];

141




37. In some possible world God does not determine thetruth-value of F-conditionals [from 35, 36, and P]; and

38. In no possible world does God determine the truth-value of an F-conditional [from P and 37].

I believe that this argument does the trick. Its one controversial point is the theological incompatibilist premise 36. Itwill be argued for w h en I raise m y Go d-cann ot-do-as-m uchobjection to Plantinga's FWD.

The second arg um en t based u po n P, which also appeals to36, is based on th e natu re of God's decision w he the r or not tocreate free persons. It first attempts to prove that God musthave middle knowledge when he makes this decision, andthereby that F-conditionals have truth-values prior to his decision. Next it gives an indirect proof for it being impossiblethat God be the determiner of these prior truth-values. Theargument, formally mounted, goes this way:

34. In every world in which God exists, he determineswhether any F-conditionals get their antecedent instantiated

[from P and 33];39. It w ou ld be reckless to decide w he the r or not to instan

tiate an F-conditional 's antecedent without first knowing itstruth-value [premise];

40. An omnibenevolent being would never make a reckless decision [premise];

41. God is essentially omnibenevolent [premise];42. God would never make a reckless decision [from 40

and 41];43. In every possible wo rld in wh ich God exists, he k no w s

the truth-values of the F-conditionals prior to his decidingwhether or not to instantiate any of their antecedents35 [from39 and 42];

44. In every po ssible w orld in w hich God exists, F-conditionals have truth-va lues prior to God's deciding wh etheror not to instantiate any of their antecedents [from 43];

45. In some possible world God determines the truth-val

ue s of the F-conditionals prior to his decision wh eth er or n ot

142




to instantiate any F-conditional 's antecedent [assumption forindirect proof];

46. God's determining the truth-values of the F-condition-als is in effect a decision not to actualize any of their antecedents [from 36];

47. God dec ides not to instantiate the an tece den t of any F-conditional prior to his decision whether or not to instantiateany F-conditional's antecedent [from 45 and 46];

48. There is no possible world in which God determinesthe truth-values of the F-conditionals prior to his decisionwhether or not to instantiate any F-conditional 's antecedent[from 45 to 48 by indirect proof]; and

49. There is no possible world in which God determinesthe truth-values of the F-condiu'onals [from 44 and 48].

Th is third ar gu m en t for thesis III is far less com pelling tha nthe second one. The premise 40 claim that an omnibenevo-lent being would never make a reckless decision - a decisionw ho se outco m e is un kn ow n - will be seen to be the basis foran objection to any version of a FWD sans middle knowledge; and it will be found to be without plausibility. In spiteof misgivings about the first and third arguments, I believethat together these three arguments make out a pretty goodcase for thesis III.

If God does not determine the truth-values of the F-conditionals, who or what does? There is an answer to this

that is implicit in the Platonic ontology employed in Plan-tinga's FWD. Since possible persons, including diminishedpossible persons, are sets of abstract properties, they exist inevery possible world, as do all abstracta. Abstract entitieshave both essential and accidental properties. The number 2has the property of being even in every possible world buthas the property of being Igor's favorite object in only some.Our old friend, diminished possible person DP, bei ng a set of

properties, has the property of containing the same properties in every possible world, such as the property of beingfree with respect to A. However, i t also has some accidental

143



The deductive argum ent pom evU

prop erties, am on g wh ich is the following: being-such-that-if-i t -were- instant ia ted-i ts- instant ia tor-would-freely-do-A. Insome worlds it has it and in others not. In virtue of this, theF-conditional, that if DP were instantiated, its instantiatorwould freely do A, is true in som e wo rlds but n ot o the rs. It isall right to call this funny property of DP a "dispositionalproperty" provided we are clear that it is not a disposition ofDP to freely perform A if instantiated (abstract entities, withthe possible exception of God, cannot perform actions) but adisposition to have its instantiator freely do A.

But what, it will be asked, determines whether a diminished person has one of these funny dispositions? As theyused to say in the Bronx, "Don't ask " Here's where the regress of explanations hits the brick wall of brute, unexplain-able contingency. There are no further elephants or tortoisesupon whose back this contingency rests. This brings to aclose our account of how those "damned F-conditionals"enter into Plantinga's FWD, and we are now in a position toconsider objections to this defense.

Objections to Plantinga's FWD

The omnipotence objection. In Plantinga's story of creation, theF-conditionals are God's kryptonite, limiting his power in asimilar way to that in which fate limits the powers of theGreek go ds. In both cases ther e is a force or po w er ab ove an dbeyond the control of individuals that limits their powers todo what they want. The idea that God must be lucky, that hemust be dealt a favorable poker hand of F-conditional facts, ifhe is to be able to realize his first choice - the creation of auniverse containing moral good sans moral evil - strikessome as blasp hem ou s, as a radical distortion of the orthod oxconcept of God's omnipotence.

Has Plantinga gone too far in watering d ow n G od's om nipotence? Th ere is som e cause for alarm . It isn 't tha t the limita

tion s placed up on God clearly violate the 52 (an d O3 an d O4)account of omn ipoten ce, for there is a wa y in wh ich the th eo -

144






concept of omnipotence some room within which to maneuver. Finally, he says of worlds in which God is screwed bythe contingent truth-values of the F-conditionals that he isless lucky in those worlds but not less perfect than in thoseworlds in which their truth-values allow him to realize hisfirst choice. In the latter-type world, "he would perhaps bebetter off bu t it is not clear that he w ou ld be better. "v Theproblem with this is that it understands God's perfectionsolely in terms of his benevolence. If meaning well is all thatis required of a perfect being, we wind up with a lot of bun

gling, incompetent "perfect" beings.Maybe the best response to the omnipotence objection isthat while Plantinga's account of omnipotence is not everytheist's cup of tea, it might be the one that will prove mostdigestible and healthy for theism in its effort to construct anade qu ate defe nse. As po inte d o ut in Chap ter 1, there is atremendous conceptual gap between the notion of "God almighty" in the Old Testament and its highly theoretic reconstruction by the great traditional theists. For this reason, thecontemporary theologian must not be frozen by fear ofoffending against "orthodoxy," that is, the theories of thesetraditional theists, any more than present-day physicistssho uld fear disagreeing w ith p ast phy sicists. In every area ofhuman endeavor that allows for some form of measurementor evaluation of performance, we find progress. Why shouldphilosophy be any different?

God-can-do-more objections. So far we have no t challengedPlantinga's various theological incompatibilist claims, mostnotably premise

10. God can not cause or determ ine in any way wh at a created person freely does; and

36. It is logically inconsistent for God to determine boththat some F-conditional's antecedent is instantiated and itstruth-value.

We shall critically evaluate three objections by the theological compatibilist to such incompatibilism: the Gabby Hayes

146



The free wffl defense

objection; the God's-gtaee objection; and the just-in-the-ntGk-

of-time objection.The Gabby Hayes objection. O ne strategy available to th e

theological compatibilist , which has been vigorously pushedin recent times by Nelson Pike, is to object that Plantingagives a misleading description of what God must do in orderto assure that created persons always freely do right. PacePlantinga, he does not create free persons and then causethem freely to do various actions, which would be inconsistent. Instead he creates people that he foreknows will freelydo these actions. According to Pike he is like the employerwho hires those whom he knows will do a good job on theirown rather than one who hires them and then makes themdo a good job. This presents us with the "enabling" view ofGod's creation of free persons and assures us that it is necessarily within God's power to instantiate the do-gooder-typepossible persons, given that there are possible persons containing the property of always freely going right.

Pike supports this enabling view of creation by an analogybetween God's instantiating a possible person containing theproperty of always freely doing right and the opening ofthe door of a cage that releases pe rson s wh om one forekno wswill always freely go right.3" This analogy relegates God to aminor Gabby Hayes-type role - a roustabout in a rodeo whomerely opens the chute of a stall (as he delivers one of his

imm ortal lines, "O ka y, m en , let's cut her loose ") en ablin gJohn Wayne to gain entrance onto the floor of the arena sothat he can flawlessly ride his bucking bronco, as Gabbyknew he would.3 9

Unfortunately, this analogy limps on all four hoofs. Tomake God out to be a Gabby Hayes-type enabler hides fromus the extent of his control over the behavior of created persons. When Gabby cuts John Wayne loose so that he can gain

entrance into the arena, his action is only a necessary causeof Wayne's subsequent brill iant performance. Gabby didnot create or determine the additional causal factors thatalso were necessary for this performance, especially the ones

147





the free wM liefctm

the funny dispositional property of being-such-that-if-it-were-instant iated-i ts- instant iator-would-always-freely-do-right. It is a contingent matter, necessarily beyond God'scontrol according to thesis HI, whether any of them havesuch a diminished possible person. If by some unfortunateluck none of them do, then God is unable to "open the door"of the goody-goody cage. In possible worlds in which theF-conditionals have unfavorable truth-values, God is no better off tha n a DT -imm obilized G abby H ayes, suffering fromDelirium Theologicum in those worlds.

The God's-grace objection. A da m s points ou t that certaineminent past theists, most notably Molina and Suarez, who,like Plantinga, also imputed middle knowledge to God, nevertheless went on to claim that in the case of every createdperson, God has the power through the knowing use of hisgrace to get this person freely to do what is right. He quotesSuarez:

It is alien to the comm on doctrine . . . and to the divine perfection and omnipotence, and is therefore of itself incredibleenough, to say that God cannot predetermine an honorablefree act, in particular and with all its circumstances, by Hisabsolute and effective will, the freedom of the created will stillbeing preserved.41

Adams endorses this :

There is an infinite variety of natural and supernatural waysin which God can work on us inwardly, assisting our reasoning, affecting our feelings and perhaps our beliefs and desires,without causally determining our responses.42

Since A da m s rejects causal com patibilism, I assu m e that th equalification "without causally determining our responses"means that although God's supernatural grace determinesour responses, it does not do so by bringing about conditionsthat are sufficient event-causes of them. Plantinga could

149



The deductive argument frwnevil

grant that God's behind-the-scenes machinations could render it more likely that a creature will freely do right, but itmust stop short of determining such a free act, pace bothSuarez and Adams.

Why are Suarez and Adams so sure that God can use hisgrace to get these persons freely to go right? Is if that theybelieve that there is no inconsistency in God predeterminingthis? This is theological compatibilism, an d reaso ns hav e justbeen given for rejecting its enabling version of God's creationof free pe rson s. Or are they ass um ing th at th e truth-value s of

the F-conditionals, al thou gh not God de term ine d, are alwayssuch that there is som ethin g God can d o tha t will bring abo utthat these people always freely do right? For example, whenGod sees W. C. Fields about to freely drink the fatal glass ofbeer, he remembers that it is true that if W. C. Fields at thatmoment were to have an image of his sweet wife and children, he would freely refrain from taking it. But this just begsthe question again st Plantinga's claim th at W. C. Fields, alongw ith everybo dy else, could be the instan tiatio n of a Mr. H yde

or a depraved diminished possible person.Another way of playing Adams's grace objection is as a

variant on the forementioned omnipotence objection, hereinthe point being that Plantinga's denial that it is logically necessary that God, by the use of his foreknowing grace, canmake everyone always freely do right is "alien to the common doctrine" concerning omnipotence, to use Suarez'sexpression. And to this the answer must be the one already

given. Treat the accounts of omnipotence given by Suarez,A quina s, Plantinga, and so on as so m an y theories, and th ensubject them to critical analysis to determine which of themoffers the best systematic account of God's nature and relation to the universe, especially its free creatures. We must notprejudge this issue by assuming that the final word was saidabout this hundreds of years ago, any more than we shouldass um e that the best physicists or hig h h ur dl er s existed in theM iddle Ages. It is repo rted that St. T ho m as 's extreme obesity

hu rt his t imes in the high hurdle s, th ou g h n ot as mu ch in thelow hurdles.

150




The just-in-the-nick-of-time objection. The theist gra nts th eexistence of both morally wrong free choices (and actions)and the suffering of innocent persons that results from them.A de fen se, acco rding ly, m us t be given for both types of evils.The answer of the FWD is that there are possible worlds inwhich God is contingently unable to create free persons without there being both types of evils. The objection countersthat God can use his middle knowledge, which he is conceded to have by the FWD, to prevent both types of evils bystepping in just in the nick of time.

The first type of evil is preventable by God's intercedingjust before individuals freely make or perform some morallywrong choice or action that God foresees will occur if he doesnot interfere. When God foresees that they will freely doright, he leaves them alone. By adopting this policy of selective interference, God prevents all moral evil but does notnegate the free will of created persons. Thus, he can have hiscake and eat it too, pace the FWD. The second type of evil ispreventable by God's interceding just after the morallywrong free choice or action occurs so as to causally quarantine the culprits from the surr ou ndin g world, rendering themcausally impotent like the derelicts I used to pass on my wayto school w h o m I kne w were forming the most evil inten tionsbut fortunately were too wasted to carry them out - if onlythey could have gotten their fingers on that red button, or me

for that matter. Again, God succeeds in preventing the typeof evil in question without negating the free will of createdpersons .4 3

The Free Will Defender has a ready response to both typesof preventive strategies. The objection to the first strategyis that it is logically inconsistent for God to both create freepersons and have the intention of preventing them fromchoosing or performing the morally wrong alternative that

their freedom makes available to them. Given that God hasthis intention and that his will is necessarily effective, theycannot avoid doing what is morally right. But a choice or action is free o nly if avo idable. Therefore, the se created pe rs on s

151




aren't free after all.44 The principle that freedom entailsavoidability has not been without doubters, both past and

present. There is Locke's famous counterexample of the manlocked in a roo m w h o gladly an d willingly rem ain s because ofthe delightful company. The moral is that even though hisremaining in the room is unavoidable, he still remains thereof his own free will.45 I think that Locke and his sup po rtersare confounding his remaining in the room gladly with hisremaining there freely. It is a fact, a very d istres sin g o ne , th atsome people gladly accept being in bondage; just look at theprevalence of "bo nd age " An othe r wrinkle in the Lockeexample is that the man might be glad to remain in partbecause he is ignorant of the fact that he cannot leave. Werehe to believe that he freely remains, he would be mistaken.That he might be the sort of pe rso n w ho d oes not m ind actingunder false beliefs, provided he is conscious in a pleasantmanner, is irrelevant to whether he really is free.

Richard Swinburne, in his treatment of evil in The Existence of God, gives a rather convincing objection to the causal-

quarantining strategy. What is of value according to themoral intuitions of the theist is not having just free will butsignificant free will in virtue of which created persons havecontrol not only o ver their ow n life bu t over oth er p arts of theuniverse as well, including the well-being and future of theirspecies. Fu rther m ore , if God we re always to interv ene just inthe nick of time, created persons no longer would attach anyimportance to their possession of free will. They could indulge themselves, knowing that no harm will result: "I have

a great idea, dear, let's kill the children." And if it should besuggested that God could take a page out of the book of theEvil Demon and systematically deceive these persons so thatthey are not aware that "superman" always appears just inthe nick of time, the reply is that this is inconsistent with hisnot being a deceiver.

The God-cannot-do-as-much objection. This objection stands insharp contrast to the preceding God-can-do-more objections,for it claims that Go d can not consistently d o as m u ch as he is

152



The free will defame

required to do by the FWD in his creation of free persons. It

will be argued at some length that God, in virtue of havingmidd le kn ow ledg e, h as a freedom-canceling control over created persons. And because these created middlemen aren' tfree, the buck of moral blame for seeming moral evils cannotstop with them but must reach through to God, which destroys the FW D's attem pt to show ho w G od can escape blam e,although not responsibility, for these evils. I will begin bymaking the distinction between blame and responsibility.

In general, a person is responsible for an occurrence thatshe was fully able to prevent, that is, had the power, opportunity, and requisite knowledge to prevent. God, for example,is respo nsible for moral evil, since he could hav e pr ev en ted itby electing not to create any free persons. An especially pertinent case is that in which a person delegates some of herpower to another but retains the power to revoke the delegated power.46 In a dual-control student-driver car, theinstructor can throw a switch that gives the student controlover the car but still retain the power to regain control overthe car by flipping the switch the othe r wa y. If the car sh ou ldbe involved in some foreseeable untoward incident while thestudent is in control, the instructor, along with the student,is responsible, but it could be that only the student is blamewo rthy. W hether the instructor shares blame will de pe nd onw heth er sh e has a good reason for not having reta ken controlof the car, for instance, the resulting harm was minor and thestudent can best learn by being left free to make mistakes.

The relation of God to created free persons is similar. Bycreating free persons God delegates some of his power tothem, but he still retains the power, called "overpower" byPike, to rescind their power, either in part or wholly. Thefirst version of the just-in-the-nick-of-time objection wasbased on God having overpower plus middle knowledge.Because G od can w ith dra w his gift of free will - flip the bigswitch - he is responsible along with created free persons forthe moral evil they cause. But, like the driving instructor, hemight have a good excuse that frees him from sharing theblame with those to whom he has delegated some of his

153



The deductive mguwmt fr<0tn evil

pow er. The FWD supplies such an excuse. God can be respo nsible but not blameworthy for the evils caused by createdbeings only if they are free. But, I will now argue, they arenot according to the Plantinga story of creation. He neversucceeded in flipping the switch that gave them the power tofreely control their own lives.

The first stage of my argument establishes that God causesthe actions of created free persons according to the FWD. I'llbegin with a fallacious argument for this that can then befixed up to work. According to the FWD, when God instanti

ates a diminished possible person, say DP, there is a trueF-conditional known to God to the effect that if DP wereinstantiated, its instantiator would freely do A. This showsthat DP's being instantiated is sufficient for the instantiator'sfreely doing A. But God's willing that DP be instantiated iscausally sufficient for DP's being instantiated. Since causationis transitive, it follows that God's willing that DP be instantiated is causally sufficient for the instantiator's freely doing A.

The fallacy in this argument for God being the cause of acreaturely free action jum ps right out at us . D P's being instantiated is only subjunctive-conditionally sufficient for the instantiator's freely doing A, not causally sufficient for it, at leastaccording to the Libertarian acc ount given by the FWD . Sinceone of the links in the sequence going from God to A isnoncausal, the principle of the transitivity of causation cannot be app lied. By interposing the indeterm inistic proce ssreported by the F-conditional between God's initial act of will

and the eventual doing of A, the FWD supposedly cuts thelink of causal sufficiency between them. God's act of will isonly causally necessary at best for A, since if he were not tohave willed as he did, A would not have occurred.

Th e first thing tha t m us t be do ne in fixing u p this fallaciousargument is to show that under certain circumstances, a sufficient cause can reach through the interposition of a relationof subjunctive-conditional sufficiency of an indeterministicsort. T he n it will be show n th at this very circum stance obta insin the FWD. Consider this stochastic machine: When its button is pressed, a stochastic process, such as the decay of a

154



the frei Witt defense

radioactive element or the spinning of a wheel of fortune, is

triggered, the outcome of which determines whether a poisonous gas will be released into a crowded stadium that willresult in the deaths of fifty thousand innocent people. Itmight be necessary to add for you sports fans that theywould die before seeing the end of the game, this qualification being necessary so that you are convinced that this is atruly evil outcome. When the button is pressed, either thisou tco m e will en su e or it w on 't. Therefore, eith er it is true thatif the button were to be pressed, this horrendous outcomew ou ld e ns ue or it is true tha t if the bu tton we re press ed, thisoutcome would not ensue. Let us assume, furthermore, thatwe mortals can no t discover by any discursive m eth ods whichof these subjunctive conditionals is true, any more than wecan for similarly matched F-conditionals.

Imagine the case in which I chance on the scene and inadvertently press the button, resulting in the horrendous outcom e. Given tha t I did not have "mid dle kno w ledg e" of wh atwould result from pressing the button and did not intend tobring about or even risk bringing about this outcome, I amblameless for the resulting evils. Furthermore, I do not evencause these evils. Were we to infer by appeal to the transitivity of causation that 1 cause them, we would be guilty of theabove fallacy.

Let us chang e th e circumstances so that I now have m iddlekn ow led ge via som e ESP faculty and pre ss the bu tton so as tobring about the deaths. In this case my action is a sufficientcause of the deaths, and is so in spite of the interposition ofa stochastic process. Furthermore, I am blameworthy for thedeaths, unless I have got a mighty good excuse, for instance,they were British soccer fans.

While there is no doubt that this is what people on thestreet would say, it might be objected that their concept ofcausation is confused; for the only difference between the

tw o cases is m y psychological state, wha t I kno w and inten d,an d h ow can this dete rm ine wh eth er or not I cause thedeaths? If what was at issue was the physicist 's concept ofcausation, this would be a powerful objection. But this is not

155




the concept of causation in question. Rather, it is the forensicone that concerns moral and legal responsibility and blame,which is the very concept that figures in the FWD, since it isconcerned with the assignment of responsibility and blameto God and man.

Plantinga could grant that I do cause the deaths in thesecond case but claim that there is a crucial disanalogybetween my relation to the deaths and God's relation tomoral evil on his account that shows God not to be the causeof such evil. It is in the nature of the intervening stochastic

pro cess. In my case it involves a no nin ten tion al physical process, for instance, the spinning of a wheel of fortune or thelike, w he rea s in the God case it invo lves the free actions andchoices of created persons. This makes all the difference inregard to w he th er th e initial age nt is the caus e of and blameworthy for the resulting evils. A sufficient cause can reachthrough an intervening stochastic process only if the initialagent is morally blameworthy for the outcome. In my casethere weren't any later free interveners who could be blamedfor the deaths, but in God's case there are later free interveners, the created free persons, who can take the "fall" for therebeing moral evil. A nd , as a result, I am bla m ew orth y for andthereby the cause of the evil outcome, but God, in virtue ofhav ing later free mid dlem en to take the rap, is neithe r blameworthy for nor the cause of moral evil.

But it is not clear why causation and blame cannot reachth ro ug h even a "free choice" stochastic pro cess. Wh y co uld n't

both God and the free middlemen be causes of and blameworthy for moral evil? Against this possibility appeal mightbe made to the Hart-Honore thesis from their Causation andthe Law that the occurrence of a vo lun tary or free act betw eenthe initial and outcome events "negatives" the former as acause of the latter, and thereby relieves the initial perpetratorof any blame for the outcome.

This is one of the very few philoso phical thes es I kn ow ofthat is plainly false. (In some rare cases philosophy is easy.)The clearest cou nterin stan ce is that of the contract m urd er. Inspite of the subsequent free action of the hired killer, the

156



f te pet wrfMfltfemi

contractor does not escape sharing both causal responsibility

and blame for the resulting death.But Plantinga need not resort to the implausible Hart-Honore thesis to render God blameless for moral evil. Whileboth God and created free persons were fully able to haveprevented moral evil, only the latter are to blame for it, sinceonly God could have a morally exonerating excuse for notdoing so. God, not they, is the creator of the universe, andthus he alone could say that his allowing such evil was theprice that had to be paid for there existing any free persons atall, which is something like my "They were British soccerfans " excuse for intentiona lly pu sh in g th e bu tto n. In bo thcases, the evil in question was necessary for the realization ofan outweighing good.

Notice that the respon se that has been m ad e on Plantinga'sbehalf does not claim that God does not cause moral evil,only that he is not blameworthy for it since he has an excusethat cannot be available to created free middlemen. Thisexcuse collapses if these middlemen are not free, since thenthe buck of blame could not stop with them. And this is justwhat I will now argue.

Since God creates free persons with middle knowledge ofwhat will ensue, he sufficiently causes the free choices andactions of these persons. This alone does not negate thefreedom with which these acts are done, for one person cancause another to act without thereby rendering the act

unfree. As a rule, the more that the external event only triggers a deep-seated character trait or natural disposition of theagent, the less difficulty there is in treating it as not abrogating the free will of the affected agent.47 When I induce aperson of amorous nature to call Alice for a date by tellinghim that she is des irous of going ou t with h im , I cau se him toact but do not usurp his free will since prominent among thecauses of his action are his own deep-seated character traits,

which w ere not imp osed on him by m e. I di dn 't hav e to"work on him" - drug, hypnotize, brainwash him - to callAlice. Unfortunately, God's way of causing created personsto act is not of this innocent sort. It is freedom canceling.

*57




My argument for this is anthropomorphic in that it appliesthe same freedom-canceling principles that apply to man-mancases to the God-man case. Whether it is permissible to reaso n in this anth ropo m orph ic m ann er will be considered subsequently. Obviously, any analogy between man and Godwill be an imperfect one, since there are such striking dis-analogies between the two. For this reason I do not see myarg um en t as in any way conclusive. At best, it m igh t take th esmirk off the face of a Free Will Defender and replace it witha worried grin. I will try to derive these freedom-canceling

principles by examining paradigm cases in which on e hu m anor finite perso n h as a freedom-canceling control over an oth er.

The case of the sinister cyberneticist. Imagine a Stepford Wives -type situation in which a cyberneticist operates on his wife'sbrain or replaces i t with a preprogrammed computer analogu e, so that h e can inculcate in his wife th e des ired psy chological makeup comprising various desires, wants, dispositions, and so on. As a consequence, she is always am oro us,

anxiou s to cook, clean, and so on. To an uninform ed observerher actions will appear free and voluntary, since theyemanate from and are explainable by her own psychologicalmakeup. But her cyberneticist husband has imposed thismakeup on her. Her lack of freedom of the will is not due tothe fact that this makeup has been determined by factorsexternal to herself (no man being either an island or a causasui), but rather to the manner in which it has been determined, namely, through the machinations of another personfor the purpose of controlling her responses to stimuli. Thecases of the Insidious Hypnotist- and Manchurian Candidate-type Barbaric Brainw asher wh o have gained a hab itual ascendancy over the will of another by inculcating in them acertain psychological makeup are similar.

Our intuitions about these cases suggest the followingfreedom-canceling sufficient condition for man-man cases:

C,. If M / s actions and choices result from psych olog icalconditions that are intentionally determined by ano ther m anM,, then these actions and choices are not free.

158



T ie free wM defense

U nde r these circumstances, M2 has a freedom-canceling control over M„ not in virtue of determining M,'s actions andchoices, but rather causing M, not to have a mind or will ofhis ow n. It isn't so mu ch M ,'s actions and choices that are notfree but M, himself; and in virtue of M,'s lack of global freedom, his specific actions and choices are not free.

The case of the evil puppeteer. Stromboli has poor Pinnochiow ired u p in such a wa y that he controls his every m ovem ent.An observer who fails to notice the wires might falsely

believe that Pinnochio's behavior was fully free and voluntary. Stromboli controls Pinnochio, not via having imposedon him an inner network of dispositions, motivations, intentions, and the like, but by exerting a compulsive force overhim that renders such inner factors irrelevant. There neednot be actual wires connecting the controller with the "puppet." It could be a wireless radio hookup such as existsbetween a controller and a remote-control toy airplane or

between the Horrible Dr. Input and a brain in a vat that inturn has a radio-control hookup with a shell body in themanner described in Daniel Dennett 's "Where Am I?"

By a coincidence that rivals that of the preestablished harmony case, it could be the case that every time the externalcontroller causes the puppet to perform some movement, thepuppet endeavors on its own to perform this movement.This is a case of causal overdetermination in which there ismore than one sufficient cause of a given occurrence. Whilethe pup pe t 's action is unav oidab le in that it wo uld hav e m adethis movement even if it had not endeavored to, there arethose, like Locke, who would still call it free. Reasons havealready been given for rejecting their claim.

What is it about these cases that makes us say the controller, be it the Evil Stromboli or the Horrible Dr. Input, hasa freedom-canceling control? It is that most of the "victim's"behavior is caused by and subject to the whim of the controller. This suggests that

C2. M2 has a freedom-canceling control over M, if M2causes most of M,'s behavior.

159




Is God's relation to created p er so ns in the FWD such that itsatisfies C, and/or C2? If it satisfies either, no less both, theFWD is in trouble, as would be the soul-building defense aswell. I submit that it satisfies both, and thus it is time for thenervous smile to replace the smirk.

It is clear that it satisfies C„ since according to the FWD,God intentionally causes a created free person to have all ofher freedom-neutral properties, which include her psychological makeup. The Free Will Defender will make the Libertarian claim that these inner traits only "incline," but do not

causally determine, the person to perform various actions oract in a certain regular manner, but this does not make theGod-man case significantly disanalogous to the type-iman-man cases; for even if we imagine that our intentionalpsychological-trait inducers could render it only probableaccording to various statistical laws that their victims wouldbehave in certain characteristic ways, they still would exercise a global freedom -canceling con trol in w hich the pe rso n isrendered nonfree due to her not having a mind of her own.

The God-man relation in the FWD also satisfies C2; for,w he n God instantiates dim inish ed possible person s or sets offreedom-neutral properties, he does have middle knowledgeof what choices and actions will result, and thereby sufficiently causes them. And he does so quite independently ofw heth er or not he is blameless for the unto w ard ones am ongthem.

Because God sufficiently causes the actions of created per

sons, Plantinga's attempt to distinguish between "determinate" (i .e. , freedom-neutral) and "indeterminate" propertiesat pag e 195 of PA in term s of w h at Go d can an d can no t causea person to have, as well as his distinction between God's"strong ly" and "w eak ly" actua lizing a state of affairs at pa ge173 of NN in terms of w ha t G od do es and do es not cause tobe actual, do not work; for whatever God weakly actualizeshe sufficiently causes to be actual. A t pag e 49 of SP, he drawsthe distinction between strong and weak actualization in away that does not rest on the cause-not cause distinction:

160






This slide from some to all is suspect. P eople have differentintuition s abo ut th e all case. Michael Slote sugg ested that M ,would feel dependent upon M2's ubiquitous timely advice, somuch so as to feel bereft of the sort of independence that isnecessary for being free. Just think of the cold sweat thatwould engulf Mt if he were called on to act before M2 calledhi m . He could say to M2 w hat m any a m othe r has said to heroffspring, "W hat's the m atter, you never call." M , w ou ldha ve the sam e sort of radical doub ts about his ow n agency ashe would if M2 were correctly to predict his every action, in

spite of his best efforts to falsify them.Be this as it may, G od's wa y of causing our actions in Plan-tinga's FWD is less benign than is M2's, for God's instantiating a diminished person alone sufficiently causes all of theinstantiator's actions. (The truth of the relevant F-conditionalis not am on g th e causes of these acts, since a prop osition canno t cause an ythi ng .) To be on the safe side, C2 migh t be narrowed so that it is required that M2 sufficiently cause most ofM,'s behavior.

A noth er objection to C2, even in its narrow version, is thatfor M2 to have freedom-canceling control over M„ it is notenough that M2 sufficiently cause most of M / s behav ior: M2also m ust hav e counterfactual control over M, in virtu e of w hichM2 can cause M, to beh ave in way s other th an tho se in whichM, in fact behaves. Whereas Stromboli and Dr. Input havethis additional counterfactual control over their victims, Goddoes not have it over created persons; for while God causes

the instantiator of DP to d o A, he does not have the power tocause this instantiator to do other tha n A, given th at it is truethat if DP were instantiated, its instantiator would freely doA. God could have prevented the instantiator from doing Aby not instantiating DP, bu t this is no t causin g t he instantiator to do other than A - nonexistent persons do not act.

Granted that there is this disanalogy between God and ourfinite controllers in that only the latter have this sort of counterfactual control, what follows? Not that God does not havefreedom-canceling control over created persons in virtueof satisfying C2 (as well as C,), but that there is a stronger

162



Tkefree will defe«M

sufficient c on dition for hav ing freedom -canceling control thathe does not satisfy, namely,

C3. M2 has a freedom-canceling control over M, if M2

causes mo st of M /s b ehav ior and also has the counterfactualpower to cause M, to act differently from the way in whichM t in fact acts.

But that there is this additional sufficient condition for oneman to have a freedom-canceling control over another doesnot show that C2 and C, are not also each sufficient conditions; for, in general, to satisfy one sufficient condition forbeing X do es not req uire satisfying eve ry sufficient conditionfor being X.

The objector might retort that having counterfactual control is necessary for having freedom-canceling control: The"if" in C3, accordingly, is to be replaced by "only if." This isnot particularly plausible for two reasons. First, if C, isturned into a necessary condition, it follows that C, is unac

ceptable and that, therefore, the Sinister Cyberneticist andthe like do not have freedom-canceling control, which is notwhat we want to say. Second, God, although lacking counterfactual control, has an additional power over created persons that Stromboli and Dr. Input do not have - God bothcreates and determines the psychological makeup of his"victims." This add itional po w er of Go d sh ould at least counterbalance his lack of counterfactual power and thereby makehim at least as good a candidate as our finite controllers forhaving freedom-canceling control.

Furthermore, i t should be obvious by now that the FWD'sgambit of having something other than God determine thetruth-values of the F-conditionals does not succeed in showing that God does not cause the free acts of created persons.Stromboli and Dr. Input were not excused from being thecause of their victim's behavior because they did only halfthe job - determined the causally relevant instantial conditions but not which causal laws hold. Analogously, God isnot excused from being the cause of the free acts of createdpe rson s beca use he did only half the job - d eterm ined which

163



The deductiw yvgument from evil

diminished possible persons get instantiated but not thetruth-values of the relevant F-conditionals. If this does notconvince you, try these counterfactual thought experiments.Our finite controllers do only half the job by determiningwhich causal laws hold after they come upon their victim insome instantial state, and God does only half the job bydetermining the truth-values of the F-conditionals after hecomes upon concrete instantiations of various diminishedpossible persons. Certainly, we want to say of both God andthe finite controllers in these thought experiments that they

cause their victim's behavior and have a freedom-cancelingcontrol in virtue of C2 alon e.So far, it appears that God's relation to created persons

satisfies both C, and C2 (but not C3) and that he thereby hasa freedom-canceling control over them. But there still remainsome disanalogies between the God-man and man-mancases that ha ve not been explored. O ne of them conce rns thefact that th e finite controllers in our t yp e 1 an d 2 cases w erea sinister bunch who meant no good for their victims,whereas God is benevolent and intends the best for his created beings. This makes no difference in regard to havingfreedom-canceling control but only in how the movie istitled. One is titled "The Horrible (Sinister, Insidious, Barbaric) Dr. Input (Cyberneticist, Hypnotist, Brainwasher)"while the other is titled "The Incredible (Fabulous, etc.)Supern atural Pred eterm iner." O ne is a horror movie, an d theother is not; but neither involves free persons.

An othe r tack is to argu e that G od's relation to man is so dis-analog ous to m an's relation to ma n as to ren de r the freedom-canceling principles, such as C, and C2, that hold for thelatter inapp licable to the former. It is no t just that Go d is quitedifferent from men - in the lingo of the streets, "He is somethin g else " - bu t that he is different in just those r esp ect sthat m ake thes e principles inapplicable to his relation to created persons. He is literally out of it, not a part of the universe. No insult intended, but he is as unnatural as you canget; in fact, he is sup ern atu ral. H e doe s not cramp o ur elbow

164




room in the way in which finite men do. Unlike these universe mates who block our path and physically compel andcoerce us, God is not pushing, elbowing, or kneeing anyonein the subway, or putting a gun to anyone's head ("Yourmoney or your salvation "). In these respects he is cruciallyunlike our bevy of sinister finite controllers. These peopleride herd on their fellowman. God does not do so (shades ofGabby Hayes and John Wayne again). This is not an episte-mological point concerning our being unaware of God'scausal efficacy in bringing about things in the world, but an

ontological one having to do with the radical difference in theway his causal efficacy works from that in which a finitecontroller's works.

It is just such an tianthrop om orph ic conside rations that areat the foundation of theological compatibilism. And I amvery sympathetic to them. Were I to be a theologian in mynext reincarnation as a result of my sinful life in this one, thisis the line I w ou ld ru n. U nfortunately, Plantinga can not avail

himself of this strategy for averting the objection that Godassumes a freedom-canceling control over created persons inhis FWD. The reason is that his FWD mu st take the a nth rop omorphic route in its rejection of theological compatibilism,for it claims that God cannot determine the free acts of persons without negating their freedom. And the only basis forthis claim is that if one man were to do this to another, itwould be freedom canceling. In other words, God cannot getaway with determining the free actions of men, because thiswould violate C, and/or C2 - the very principles tha t o pera tein man-man cases.

We canno t allow P lantinga to be a good-time anth rop om orphism to reason anthropomorphically when warding off theobjection of the theological compatibilist, and then refuse todo so for the purpose of rebutting the charge that God hasassumed a freedom-canceling control over created persons.Thus , Plantinga is caught on the horns of a dilemma. If he

reasons anthropomorphically, his FWD collapses because itimputes to God a freedom-canceling control over created

165



The deductive mgument pom evil

persons. And if he does not reason anthropomorphically,again his FWD collapses, this time because it has n o reply tothe objection of the theological compatibilist. But either hereasons anthropomorphically or he does not. Therefore, hisFWD collapses.

But it would be quite premature for me to write "Q.E.D."after my above argument for the God-cannot-do-as-muchobjection that Plantinga's FWD is logically inconsistent; forPlantinga, or one of his followers, could mount a counterargument for why God does not cause the free actions of

created persons in his story of creation. One such counterargument, which I owe to the fertile minds of JonathanDancy and Donald Turner, goes as follows: They begin bytaking seriously Planting a's claim th at w he n God creates freepersons, he leaves it up to them what they shall freely do.This has the consequence that it is they who determine thetruth-values of F-conditionals through their own free actions.A conse que nce of this view is that F-conditionals tha t do no tget their antec ed en t instantiated are truth-valu eless. Since anF-conditional has its antecedent instantiated in some but notall possible worlds, an F-conditional is not necessarily eithertrue or false: In the actual world there are F-conditionals thatare neither true nor false. This is not such a bizarre result asto totally discredit their argument, since F-conditionals aresuch queer birds. My argument for God being the sufficientcause of all the actions of created persons was based onPlanting a's thesis II that God has m iddle know ledge w he n hecreated them. But, Dancy and Turner urge, the actions ofthese persons is the cause of their middle knowledge - theirknowing the F-conditionals in question. God's decision toinstantiate an F-conditional's antecedent is determined orexplained by his middle knowledge of what will result fromhis doing so. Therefore, God's decision to instantiate an F-cond itional's an tec ed en t is caused or explained by the in stan-tiator's action, rather than vice versa as I would have it; and,thereby, God is not the cause of this action, since we are not

going to allow a circle of causes or explanations.

166



the pee will

The following is an explicit mounting of their argument:

50. An F-conditional is made true or false by the instantia-tor's action [premise];

51. God's decision to instantiate the antecedent of an F-conditional is determined or explained by his middle knowledge [premise];

52. God's middle knowledge is determined or explainedby the action of the instantiator [from 50];

53. God's decision to instantiate the antecedent of an F-conditional is determined or explained by the instantiator'saction [from 51 and 52]; and

54. God's decision to instantiate the antecedent of an F-conditional does not determine or explain the instantiator'saction, that is, is not the cause of this action [from 53 in virtueof causation being asymmetric].

The Dancy-Turner argument is clearly in the spirit of

Plantinga's FWD, especially in its imputation of middle knowledge to God at the time or stage at w hich h e deliberates ab ou tinstantiating the antecedents of different F-conditionals. Atthat time or stage, he is fully apprised of the truth-values ofevery F-conditional that has a truth-value, which, accordingto prem ise 50 of the Dan cy-Turner a rgu m en t, m ean s everyF-conditional whose antecedent he will instantiate, sinceonly they have a truth-value.

This can be shown to lead to a contradiction. My reductioargument will avail itself of an omnitemporal God that engages in deliberation prior to his creative decision, but everything I say can be recast in terms of a timelessly eternal Godand the order of explanation or determin ation in his n on tem -poral deliberation and decision (see note 35 on this point). Myargument is as follows:

55. It is impossible that an agent deliberate w hile kn ow ingw hat decision she will make or know ing of an e vent for w hich

her decision is a causally necessary condition [premise];

167




56. God 's instan tiation of an F-conditional's antece de nt is acausally necessary condition for the existence of its instantiator, as well as its subsequent action [premise];

57. It is impossible that while God deliberates about instantiating an F-conditional's anteceden t, he know either tha tits ins tan tiato r exists or that it performs som e action [from 55and 56];

58. Go d has m iddle know ledge while he deliberates aboutinstantiating an F-conditional's antecedent [Plantinga's thesisII an d g ran ted by prem ise 51 of the Dancy-Turner arg um ent];

59. God ha s m idd le know ledg e of an F-conditional only ifhe instan tiates its anteced ent [from prem ise 50 of the D an cy -Turner argument] ;

60. It is possib le that the re is som e F-conditional such tha twhile God deliberates about instantiating its antecedent, healready knows whether or not he decides to instantiate it[from 58 and 59]; and

61. It is impossible that there is some F-conditional suchthat wh ile God deliberates about instantiating its an teced ent,he already knows whether or not he decides to instantiate it[from 57].

I conc lude tha t my God-cannot-do-as-m uch objection h asconsiderable force against the Plantinga version of the FWD,since it im pu tes m iddle kno w ledge to Go d. W ithout it, God'sinstantiating the antec ede nt of a true F-conditional wou ld nolonger count as a case of his causing the instantiator to do

what the consequent reports. Created persons, then, couldserve as the sole suitable scapegoats for moral evil. This naturally gives rise to the question whether a viable version ofa FWD can be constructed that denies middle knowledgeto God.

Versions of the FWD sans middle knowledge

Since God's middle knowledge is comprised of the conjunction of

168




I. Every F-conditional has a contingent truth-value; andII. Go d kn ow s the truth-va lue of all F-conditionals prior to

his creative decision,there will be two ways of constructing a FWD sans middleknowledge. The first version, which is ably championed byA da m s, den ies thesis I an d the reb y II, since not even anom niscient being can know w ha t isn' t true . This rend ers Godblameless for permitting moral evil, since he could not haveknown in advance the moral evils that would result from hiscreation of free persons. The second version, which is hintedat by Plantinga on page 52 of SP, accepts thesis I but deniesII, again rendering God blameless in virtue of an excusablelack of kn ow ledg e. He rein there was som ethin g to be kn ow n,unlike the first version, but there was no way in which Godcould have divinely known it . God winds up as a hockeyparent after all.

Because both versions have God instan tiate the antec ede ntof F-conditionals without foreknowledge of what the created

persons will freely do, they face the objection that God is acting in a recklessly immoral way by shooting craps at our expense.48 The reckless objection has already been ad um bra tedby these two premises in the third argument for thesis III:

39. It w ou ld be reckless to dec ide wh et h er or not to instantiate an F-conditional's antecedent without first knowing its truth-value; and

40. An omnibenevolent being would never make areckless decision.

An omnibenevolent being would not create free personswithout middle knowledge because it would be immoral totake such a risk.

No red-blooded theist would accept the wimpy moral intuition underlying the reckless objection and would give God'screation of free persons in both versions as a counterex

ample. The objection also faces an ad hominem-type rebuttalin that n o existent pe rso n, except for a few g ripers, are apt tomake it; for, if God hadn't elected to roll the dice, they

169




w ou ldn 't even exist , and, su pp osed ly, the y are glad that theydo.49

Although both versions make the same excusable lack ofkno wledg e excu se available to Go d, the y differ significantlyin their epistemological and metaphysical underpinnings,and thus require separate consideration. The outcome of ourdiscussion will be that both versions face formidable objections: The first version, because of its denial of thesis I, renders it ano m alou s how God is able to create any free pe rso ns ;and the second version w aters dow n G od's om niscience in an

unacceptably anthropomorphic manner.

First version. Adams develops this version in the course ofdefending the attack on middle knowledge by certain late-sixteenth-century Dom inicans against their Jesuit op po ne nt s,Molina and Su arez. A dam s's denial of thesis I is built u p o nthe Libertarian accoun t of F-conditionals, according to w hichthe act reported by the consequent is neither causally deter

mined nor determined by anything other than the instantia-tor of the antecedent. Given this account, he does not seehow an F-conditional could possibly be true. Thus, it appearsto be logically or conceptua lly im possible for it to be t ru e an dtherefore it necessarily lacks a truth-value. Unlike the "neuter" or "indeterminate" propositions of Lukasiewicz, i t doesnot become true or false with the passage of time. Adamssays that he do ub ts if they "ev er we re, o r ever will be , tru e. 5°This m eans that even if the anteced ent shou ld be instantiate dand the instantiator subseq uently perform the action re porte dby its consequent, the F-conditional does not become true.And a fortiori this sequence of events does not show that theF-conditional w a s true all alon g, for ther e is no pre se nt tru thto cast a backw ard sh ado w . A dam s gives us a choice be tw eenF-conditionals being necessarily false and being necessarilyneither-true-nor-false. The common denominator of theseoptions is that F-conditionals necessarily are not true.

Ad am s's ar gu m en t for the denial of thes is I is no t m ade fullyexplicit. One who denies thesis I is typically a warranted-assertibility theorist who holds a proposition to be true only

170




if it is in principle epistemically supportable. But this isn'tA dam s's line, for he says that a prop osition rep ortin g a futurecontingent (e.g., the sea fight tomorrow, which we finallyknow didn't happen since Khrushchev decided not to challenge Ken ned y's blockade of Cuba), altho ug h in principle no twarrantedly assertible, "can be true by correspondence to theactual occurrence of the event they predict."51 This suggeststhat for Ad am s a necessary cond ition for a pro pos ition beingtrue is that it have an external correspondent. The "external"qualification precludes the correspondent of the proposition

that p being the fact that p; for, g iven that a fact is a tr ueproposition, this would make that p the correspondent ofthat p.

In the previous discussion of III, it was suggested thatfor Plantinga the external correspondent of the true F-conditional, that if DP were instan tiated, its in stantiato rwould freely do A , is the abstract diminished possible personDP having the funny dispositional property of being-such-that-if-i t-were-instantiated-its-instantiator-would-freely-do-A. According to Adams's commentary, Suarez also held thisview. W hile not objecting to this accoun t's Platon ism, A da m sobjects that he does not

have any conception, primitive or otherwise, of the sort ofhabitudo or property that Suarez ascribes to possible agentswith respect to their acts under possible considerations.52

This autobiographical fact hardly constitutes a refutation ofthe Suarez-Plantinga position. The rejection of a positionby saying "I do not quite understand this" has mercifullygone the way of rejection by appeal to desert landscapes.No t being able to und ersta nd is not alway s an achiev em ent.I suspect that the reason Adams does not understand whatit is for a diminished possible person to have one of the"dispositional" properties is that he does not see how an

F-conditional can be true, and his reason for this will emergein our discussion of whether thesis I is true, to which wenow turn.

171



The deductive argum ent pom evil

In our previous defense of thesis I, which was relativizedto the premises of Plantinga's FWD, it was shown that forPlantinga, I is a consequence of the law of excluded middle:

Our question is really whether there is something Curleywould have done (had he been offered the bribe). . . . Theanswer . . . is obvious and affirmative. There is somethingCurley would have done, had that state of affairs obtained

, (NN 181).

If DP were ins tan tiate d, given that its instantiator has the d isjunc tive prop erty of either-freely-doing-/4-or-freely-refraining-from-doirtg-A, its instantiator either freely does A or freelyrefrains. Thus, it is either true that

F. If DP w ere in stantia ted, its instantiator w ould freely doA; or true that

F '. If DP were instantiated, its instantiator would freelyrefrain from doing A.

So far 1 ca nn ot find an yth ing to object to. Acco rding toLewis's lemma, if neither F nor F' is true, it follows that Godis una ble to insta ntia te either P or P„ in which P and P ' forman incom patible pa ir having D P in com m on an d differing onlyin that P contains freely-doing-A while P' contains freely-refra ining-from -do ing -/i; for if it is no t true tha t if God w ereto instantiate DP, its instantiator would freely do A, God isun ab le to instan tiate P, an d likewise m utatis m utan dis for P,.And this can be generalized, so that if no F-conditional ist rue, God cannot instantiate any possible person, whichw ou ld be the ship w rec k of the FWD, since any possible w orldin which God is morally exonerated for allowing moral evil ison e in which he succ eed s in creating free pers ons and thu s isable to do so.

Adams, surprisingly for a Free Will Defender, says something that seems to imply that God is unable to instantiate

any possible person: "In other words, I deny that God couldhave made free creatures who would always have freely do neright."53 This seems to say that God couldn't have instanti-

172



The freewill defense

ated any possible person containing the property of always

freely doing right by instantiating its diminished possibleperson, because the relevant F-conditionals are not true. Iass um e that this is w hy Ad am s put would in italics. But thisreason for God's being unable to instantiate any goody-goody possible person can be generalized to all possible persons , since no F-conditional is true. Thus, God is unable toinstantiate ev en a M r. Rog ers-type possible perso n. A da m s isno t rejecting Lew is's lemm a -q u it e to the con trary. It is exactlybecause he accepts it and denies thesis I that he reaches theanomalous conclusion that God cannot instantiate any possible person.

Maybe my interpretation of Adams is unduly uncharitable,for certainly as a Free Will Defender he does not want towind up with this anomalous conclusion. Another way ofun de rsta nd in g him is as den ying bo th thesis I and Lewis 'slem m a, rathe r than de ny ing thesis I and a ccepting Lewis'slemma, which resulted in anomaly. On this new interpretation, that God instantiates DP and this is followed by theinstantiator freely doing A entails that God was able toinsta ntia te P bu t not tha t F wa s true. But this new position isnot without the appearance of anomaly, for i t seems obviousthat F w ould be ren de red true by this sequenc e of even ts.Adams better have a very good argument for why our intuitions deceive u s in this m atter; how eve r, I can 't find anysuch argument in his essay. I get the feeling that he is makingimpossible demands on an F-conditional: First, neither its

antecedent nor anything else can nomically necessitate i tsconsequent; and, second, for it to be true, there must be thisvery relation of nomic necessitation between them. That theantecedent probabilize the consequent in virtue of variousdispositions and inclinations of the instantiator of DP isn'tenough.

At this point, Adams, no doubt, will throw the ball backinto Plantinga's court and ask him to supply an explanation of how F could possibly be true. Plantinga's previouslystated claim that there is something Curley would havedone if offered the bribe suggests a "minimalist" account of F-

173




conditionals. While F-conditionals usually involve a probabi-lizing relation between their antecedent and consequent dueto a set of freedom -neutral pro per ties often co ntaining d ispositional properties, they need not. For an F-conditional to betrue, it suffices (though isn't necessary) that the instantiationof its anteced ent both pre ce de and b e necessary for the ev entreported by its con seq uen t. T his minimalist view could aptlybe called the "stage-setting" account. F is true if the instantiation of DP is in fact follow ed by the instantia tor 's freelydoing A, since the former is necessary for the ensual of the

latter. While the law of cond itional excluded m idd le hold s forstage-setting subjunctive conditionals, it does not apply to allsubjunctives, such as "If I were now to snap my fingers,Caesar would have freely crossed the Rubicon," in whichthere isn't the right sort of con ditio na l-en su al relation, in thiscase because there is neither the right sort of temporal orderbetween what is reported by antecedent and consequent northe required stage-setting relation. Because a stage-settingF-conditional need not involve anything more than theinstantiation of its antecedent both preceding and being necessary for the event reported by its consequent, it will resista possible worlds analysis in terms of the closeness relationbetween different worlds;54 for a wo rld in which th e ins tanti-ator of DP freely does A and one in which it freely refrainsfrom doing A need no t involv e any othe r difference betw eenthem, such as a difference in their deterministic or statisticallaws.

I think that this stage-setting analysis of F-conditionals interms of their truth conditions captures the ordinary wayin which they are used a nd e xplains wh y we are perplexe d bythe claim that DP's instantiation could be followed by itsinstantiator freely doing A without it being true that if DPwere instantiated, its instantiator would freely do A. This isjust as anomalous as saying that God's instantiation of DPcould be followed by the instantiator freely doing A withoutGod having been able to instantiate P.

Would it help if this first ve rsion of a FWD sans m iddle knowledge were to be changed so as to allow an F-conditional, ini-

174



The free will

t ially truth-valueless as Adams contends, to become true orfalse when its instantiator performs her free action? Dancy andTurner, after being confronted with my reductio counterargument to their above argument, developed a variant that isbased on this Lukasiewicz manner of viewing F-conditionals.According to their variant on the Adams FWD, an F-conditional lacks a truth-value at the time or stage of God'sdeliberation, thereby denying thesis H, and acquires a truth-value only after God has instantiated its antecedent and theinstantiator does the deed reported by the consequent.

(Those that go uninstantiated remain forever truth-valueless- poor devils.) It is still the case that the instantiator makesthe F-conditional true; and, moreover, since God lacks middle knowledge, i t destroys my God-cannot-do-as-muchobjection. God is in very much the same situation as theperson who touched the button of the stochastic machine inigno ranc e of w hat the outcom e wou ld be, the only differencebeing that the former, unlike God, did not know that his

action created a certain risk of there being an evil outcome,thereby rendering God, unlike this person, responsible forwhat ensues. But, supposedly, since it was worth taking therisk, God is rendered blameless for any evil outcome of his"p res sin g the but to n. " This arg um en t, */ successful, establishes that God is not the cause of the actions of createdpersons, thereby neutralizing my God-cannot-do-as-muchobjection.

This Dancy-Turner variant on the Adams FWD is subjectto the same basic objections. By denying that F-conditionalshave any truth-value, and thereby are not true, at the time orstage at which God deliberates and makes his creative decision, i t winds up denying Lewis's lemma, just as the AdamsFWD did. A nd this rend ers i t anom alous ho w God is able toinstantiate any possible person, say P, wh o includes diminished possible person DP and freely doing A; for if it is nottrue that if DP were to be instantiated its instantiator wouldfreely do A, it seems impossible for God to instantiate P, andsimilarly for any other possible person you choose. Furthermore, if God instantiates DP an d the instan tiator freely do es

175




A, the concerned F-conditional thereby "becomes true," asDancy and Turner maintain. But does this not cast a shadowbackw ard, ren de ring it true all alon g that if DP w ere instantiated, its instantiator would freely do A, pace their claim? Ibelieve that the only response that can be made to theseobjections must invoke a warranted assertibility theory oftruth; but this ru ns in to the "Tru e" and "Fact" cigarette problem (see footnote 51).

Second version. This version acce pts both thesis I an d Lew is's

lemma, understood in terms of minimalist stage-setting F-conditionals, and thereby escapes both types of anomalies.Its special gimmick is to deny II, thereby enabling God toreply to the newscaster's question by saying that althoughthere was a truth of the matter at the time he deliberatedabout creating free persons, there was no way he could haveknown then that these moral evils would result from hisdecision to instantiate certain diminished possible persons.This second version, because it accepts theses I and III, alsomakes available to God, in addition to the excuse of ignorance, the I-w as-screwed -by-those-dam ned-F-co nditionalsexcuse of Plan tinga 's FWD. The latter excuse, of course, couldbe offered by God only after the fact, not at the time of hiscreative decision, as it can be in the Plantinga version.

It is plain that in accepting thesis I and rejecting II, thesecond version rejects the traditional definition of God'somniscience according to which God knows every true

proposition and is replacing it with the weaker definitionthat

0 5 . For any proposition p, if it is true that p and it is logically possible that God know that p, then God knows that p.

It wa s show n in Ch apter 3 that 05 creates a very virulentversion of the paradox of perfection in that it winds upimputing to God a lesser degree of omniscience than thatposse ssed by so m e possible non perfect being. But there is noreason for the second version to use such a watered-downaccount of God's omniscience. It can, instead, employ

176




0 6 . For any proposition p, if it is true that p and it islogically possible for any on e to kno w tha t p, then God know sthat p

in w hich G od's om niscienc e is not relativized to w ha t it is logically possible for God, a perfect being, to know but to whatit is logically possible for anyone to know, thereby eludingthe paradox of perfection.

The pro blem is wh y is it logically im po ssible for an y on e toknow an F-conditional in advanc e? The an sw er is that kno w ledge requires a justificatory explanation and none could begiven in advance for an F-conditional, since neither its antecedent nor anything else determines i ts consequent. Thiscertainly holds true for finite persons, since their knowledgeis discursive and thus in need of a justificatory explanation.But God is supposed to know things in his own inimitablesupernatural way that is denied to us finite beings, just as heis supposed to be able to do or bring about things that we

cannot, such as create things ex nihilo. Moreover, we cannoteven conceive of how God knows and does these things,though we can, pace Berkeley, conceive of there being thingsthat we cannot conceive of. Maybe God's justification forbelieving an F-conditional that has not yet had its antecedentinstantiated is that he knows that he is God and therebyomniscient and thus whatever he believes is true. To restrictGod's omnipowers to what we can conceive of is a radicalanthropomorphizing of God. For this reason a traditionaltheist cannot adopt this second version.

But this is no t the end of the ma tter, since the "trad ition al"account of God's omniscience is a theoretical reconstructionof the biblical notion of an all-knowing personal Deity that inspite of his essential omniscience, is through and throughtemporalistic. Because God is t ime bound, he does not knoweverything about the future, such as what future decisions hehimself will make55 and what responses will be made by cre

ated persons to his overtures. He periodically turns intoCharles Bronson because of his disappointment at these respo nse s. Co ntem po rary process theologians give a theoretical

177



The deductive argummi pQtn evil

reconstruction of God's omniscience that is closer to thetemporalistic God of the Bible. Such a view dovetails nicelywith the overall thru st of Ch ap ter s 2 an d 3 - tha t a strictlyimm utable timeless God is a no np er so n t ha t is no t religiouslyavailable.

If the price that must be paid to have an a de qu ate defenseof God in the face of moral evil is to go wi th th is b iblical-stylenotion of God's omniscience, then it is worth it. One mightworry whether such a sharp divergence from the traditionalaccount does not change the referent of the w ord "G od " from

what it had when used by the great medieval theists. If theaccount in Ch apter 1 of ho w "G o d" refers is basically co rrect,the answer is no; for contemporary process theologians arenot rejecting any one of God's hard-core properties, such asbeing that than which none greater can be conceived and being eminently worshipable. They are only giving a differentanalysis from that given by the medievals of what constitutesthe lower-level, soft-core determiners of these hard-coreemergent properties. Furthermore, and of greatest importance, they are members of the same ongoing religious community, the role of the God-idea in their religious languagegame being the same as it was for the medievals.

[78



Chapter 5

The argument from world-relat ive actual i ty

This atheological argument differs sharply from the onesconsidered in the preceding three chapters in two importantrespects. First, it has never been given before. Second, it failsto undermine in any way the traditional conception of Godand therefore is not an occasion for theists to go back to thedrawing board and redesign this conception.

Why, then, consider it? The reason is that it brings intoclear relief an overlooked feature of this conception concerning God's "absolute actuality." God, as creator, has a standpoint that is "outside" of or independent of possible worlds- the infinitely many different ways that things could be.God's existence or actuality is absolute, not being in any wayrelative to one of these worlds: He exists or is actual simpliciter. In his role as creator, God contemplates them and makes

a unique creative decision as to which one alone shall beactual simpliciter. His choice is unique because, of all thecreative choices he might have made, this choice alone isactual simpliciter, not just actual relative to some world. As aresult of the absolute nature of God's creative choice, oneworld has a special sort of ontological honor bestowed uponit consisting in it alone having absolute actuality, that is,being the actual world simpliciter. Thus, the absoluteness of

G od's persp ectiv e as creator is prop agate d do w n the line viahis unique creative choice to a particular world. And what ismore, the universe - the largest spatiotemporal aggregate ofobjects and e ve nt s - that is created by his un iqu e actualization

179

X



The argument from world-relative actuality

of one possible world itself has an absolute, unqualified existence. It exists simpliciter.

This theistic view of actuality and existence as absolute ornon-world relative is challenged by David Lewis's extremeversion of modal realism according to which "every way thata world could possibly be is a way that some world is."1

Logical space is a plenitude of isolated physical worlds, eachbeing the actualization of some way in which a world couldbe, that bear neither spatiotemporal nor causal relations toeach other.2 This m eans that no wo rld is absolutely actua l" or

"actual simpliciter. "3

The formal m o de version of this thesi s isthat the predicate "is the actual world (or is actual)" really isa disgu ised two -place pred icate of th e form " is actual at

." There is an ontological parity b etw een w orld s b ecausefor every world w, it is the case that w is actual at w and onlyat w.4 No world qualifies, therefore, for the coveted title ofbeing the actual world. There is no such title to be won sincethere is no such thing as being actual simpliciter, "is actual"being a two-place predicate. This is the thesis of ontologicalparity . Not even G od qualifies as be ing actual simpliciter. Hisactuality, along with the actuality of his creative choice, alsois wo rld relative. H e exists in m an y different w orlds; and , foreach world w in which he exists, he chooses in w that w beactualized. No one of these creative choices is special orunique, which, if true, completely undermines traditionaltheism.

The ontological parity thesis helps to clarify and support

Lewis's shocking claim that worlds other than the one thatwe take to be alone actual do not differ in "kind" or their"manner of existing" but only in their contents from thisworld.5 Since our w orld is a max imal agg regate of spatiotem -porally related objects and events, so are these other worlds.This results in the plenitude of isolated concrete worlds.

Lewis has another account of what a possible world is interms of a "way things could have been."6 The ontologicalparity thesis not only does not support but undermines thisaccount. From the following four propositions,

180



The m'ptmmt pom world-relative actuality

1. Other worlds do not differ in kind from the world wetake to be actual;

2. Our world is a maximal spatiotemporal aggregate;3. A spatiotem poral aggregate has a spatiote m poral loca

tion determined by the spatiotemporal regions occupied byits parts, that is, it occupies the same spatiotemporal regionsas do the objects and events that form its parts; and

4. A way thing s could have been (or are) ha s no spa tiote mporal location;it follows that

5. A world is not a way things could have been.7

Lewis has an argument for the thesis that actuality is worldrelative that is based upon the premise that "actual" is anindexical term.8 This argum ent, al thoug h not intend ed assuch, constitutes the basis of an atheological argument. It isin no way bas ed up on the claim that a poss ible w orld is a waythings could have been. Peter van Inwagen makes the baf

fling claim that "Lewis's only argument for the existence ofmore than one possible world depends upon his identifyingworlds with ways things could have been."9 To be sure, thisdefinition does occur as the key premise in Lewis's preliminary Quinean-style argument for our prima facie ontologicalcommitment to possible worlds in virtue of our quantifyingover them w he n w e say that there are m any way s things couldhave been.10 The indexical argument, however, is quite distinct from this argument, although both support the conclusion that there is a plurality of worlds. Van Inwagen errs,therefore, when he speaks of "Lewis's only argument for theexistence of more than one possible world."

The indexical arg um en t receives its first and fullest pr es entation in "Anselm and Actuality."11 In his later Counterfactualsand On the Plurality of Worlds, Lewis gives a brief summary ofthe argu m ent, along with some minor refine m ents , a nd refershis reader back to the fuller presentation in "Anselm andActuality."12 Accordingly, we will focus on this presentationand bring in the later refinements when relevant.

181



The argum ent from world-relative actuality

The key premise of the indexical argument for world-relative actuality is that "actual" and cognate expressions areindexical expressions that are analogous to "now," "here,"and "I":

I suggest that "actual" and its cognates should be analyzed asindexical terms: terms whose reference varies, depending onrelevant features of the context of utterance. The relevantfeature of context, for the term "actual," is the world at w hicha given utterance occurs. According to the indexical analysis Ipropose, "actual" (in its primary sense) refers at any world w tothe world w. "Actual" is analogous to "present," an indexicalterm w hose reference varies depending on a different featureof context: "present" refers at any time t to the time t."Actual" is analogous also to "here," " I," "you," "th is," and"aforementioned" - indexical terms depending for their reference respectively on the place, the speaker, the intended

••'«• audience, the speaker's acts of pointing, and the aforementioned discourse.13

Lewis's use of "I suggest" and "I propose" permits twointerpretations. They can be understo od as hav ing assertionalillocutionary force that is somewhat weaker than that of "Iassert (state)." Herein Lewis would be purporting to describethe actual w ay in wh ich "actua l" is Used, it being claimed tha t"actua l" functions in the way that indexical terms do in ordinary discourse. Or they can be understood as having theillocutionary force of "I hereby stipulate that." Herein there

would be no claim that "actual" is used indexically, only thatthe author stipulates that he will so use it. If Lewis is onlystipulating th at we use "actual" in the way in wh ich indexicalterms are used, he would be unable to support his ontologi-cal parity thesis by this move, since this thesis is about ourordinary concept of actuality - not some made-up one thatren der s his ontological parity thesis true by definition. T hu s,we will interpret Lewis as purporting to describe our ordinary use of "actual." There is, of course, plenty of middleground between purely descriptive and stipulative definitions, such a s revisionary definitions tha t are justified by var-

182



The argument from wmrld-rtWkie actuality

ious arguments. If Lewis's account were revisionary in thissen se, he wo uld ha ve to justify his revision of ord inary usag e(e.g. , by appeal to his cost-benefi t argument adumbrated infootnote 8), and thus the indexical argument would not standon its own two feet, which supposedly it does.

That "actual" refers at any world w to w is further refinedso as to deal with its cognates and their different semanticrelations:

To speak more precisely: at any world w, the name "the actualworld" denotes or names w; the predicate "is actual" designatesor is true ofw and whatever exists in u>; the operator "actually"is true of propositions true at w, and so on for cognate term s ofother categories.14

Thus, it isn't only "is actual" that is a disguised two-placepredicate but also "is true" and "exists": the latter requireexp ans ion respec tively into "is true in w orld " an d

"exists in w orld ." O rdina rily, w h e n it is said that p istrue, i t means that p is true in the actual world. Thus, if "isactu al" is indexical, so is "is tru e" and "ex ists." Wh en w ha t issaid to be true or existent is necessary, it will not be important to expand the grammatical one-place predicate, but i t iswhen the item is contingent. It will be futile, therefore, toargue for the monadic nature of actuality by appeal to amonadic concept of truth or existence. An example of suchfutility is Robert M. Adams's argument that there is only onew orld that is actua l sim pliciter bec ause the re is only on e w orldthat has a book containing propositions all of which are truesimpliciter, in which a book for a given world is the set of allpropositions true in that world, that is, that would be truewere that world to be actualized.15

Alongside of the world-relative use of "exists," Lewis recognizes an absolute, non-world relative use, though this isnot clearly m ark ed ou t in ordin ary la ng ua ge . There is for himan existential quantifier that is totally unrestricted in itsdomain, ranging over all the possible worlds and their contents. Thus, Lewis can say without contradiction that there are

183

<



The argument from umtd-relative actuality

(unrestricted) things that do not (actually) exist (restricted).16

This is analogous to there being a temporal quantifier that isunrestricted among times so that it can be said without contradiction that there are (timelessly) times that are not nowpresent.

If "the actual world" and "is actual" are indexical, thereshould be some explicitly indexical phrase that can replacethem without change of sense. Based on what Lewis says,there appear to be two different indexical expressions thatcan do this job, one employing "I" and the other the demon

strative "this." According to the first-person parsing, "theactual world" and "is actual" mean, respectively, "the worldin which I exist" and "is in the world in which I exist." Thefollowing quotations speak in favor of this construal:

If we take an a priori point of view and ignore our ownlocation among the worlds the big difference between theactual world and other worlds should vanish.17

I said that when I use it, 'actual' applies to my world and myworldmates; that is, to the world I am part of and to otherparts of that world.18

W hen discussing the above distinction betw een th e restrictedand unrestricted quantifiers, he says that

they may be restricted to our own world and things in it.Taking them as thus restricted, we can truly say that there

exist[s] nothing but our own world and its inhabitants.19

Furthermore, Lewis's reply to Adams's indifference argum ent rests on such a first-person analysis. Ad am s's o bjectionto the ontological parity thesis is that it is unable to accountfor our strong want or preference that the actual world bewith out evil, to which Lewis resp on ded that this involves t heirreducibly indexical want that the world in which I exist bewithout evil.20

The demonstrative rende ring of "the actual w orld " and "isactual" takes the form "this world" (or "the maximal spatio-

184



The argument frimtioortfl'Mtiiist&e actuality

t empora l agg rega t e con t a in ing t h i s " ) and " i s spa t i o t empo-ra l ly re la ted to th i s" (o r " i s pa r t o f the maximal spa t io tempo-r a l agg rega t e con t a in ing t h i s " ) . Th i s r ende r ing i s suppo r t edby t he r epea t ed u se o f " t h i s wor ld" and " th i s -wor ld ly"t h r o u g h o u t On the Plurality of Worlds a s pa r ap h ra se s o f " t h eactu al w o r l d ." La ter it wi ll be sh o w n th at ne i th er the f i rs t -p e r s o n n o r d e m o n s t r a t i v e p a r a p h r a s e s h o l d salva veritate.

O n e i th e r of th es e two index ica l pa rap hr as es i t is r eq u i re dtha t ind iv idua l s do no t ex i s t in more than one wor ld ; o therw ise " th e w or ld in w hic h I ex i s t " o r " th e ma xima l s pa t io tem -po ra l ag g re ga t e co n t a in ing t h i s " w ou ld fail t o hav e a un iqu ereferent . Given that Lewis holds to the indexical analys is ofactual i ty, i t i s no accident tha t he a lso holds to a theory ofwor ld -bound ind iv idua l s , o f wh ich coun te rpa r t t heo ry i s oneexample . Thus , any ob jec t ion to the theory tha t no ind iv idua l( save some abs t rac ta ) ex i s t s in more than one wor ld a l so i s anobject ion to the indexical theory of ac tual i ty. In my opinion,th i s i s m os t un fo r tu na te fo r Lew is , s ince it se em s ob v iou s to

m e , a s w ell as to a h o st of his cr i t ics , tha t w h e n I say of s o m eperson tha t he cou ld have done o ther than he in fac t d id , I amsp eak ing abo u t ho w tha t ve ry pe r so n ac ts in som e c oun te r-fac tua l s i tua t ion a n d no t ab ou t so m e co un te rp ar t to h i s inan o t he r w or ld . S ince I am no t p r ep a re d t o d i s cus s co un te rpa r t th eo ry in th is bo ok , I wi ll no t pres s th is ob ject ion.

There wi l l be much to say l a te r about whe ther "ac tua l" i sindex ica l , bu t r igh t no w o ur con cern is in f igur ing ou t h o w

th e on tolo gica l pa r i ty thes is i s s u p p o s ed to fo llow from thepr em ise th a t "a c tu a l" is index ica l .2 1 Supposed ly, t he r e i ssome a rgumen t t ha t shows t ha t no t ime , p l ace , o r pe r son i sresp ec t ive ly n o w , h er e , o r I abso lu te ly o r s impl ic it e r in v i r tu eo f " n o w, " " h e r e , " a n d "I" be ing indexical tha t works in anan a lo go us w ay fo r " ac tu a l . " Lew i s g ive s u s som e h in t h o w tof i l l in th i s do- i t -yourse l f a rgument when he wr i t es :

If we take a timeless point of view and ignore our own location in time, the big difference between the present time andother times vanishes. That is not because we regard all timesas equally present, but rather because if we ignore our own

185




location in time we cannot use temporally indexical terms like"present" at all. And similarly, I claim, if we take an a prioripoint of view and ignore our own location among the worldsthe big difference between the actual world and other w orldsshould vanish."

Before we try to develop this into an argument, it will benecessary to give a brief account of indexicality, similar tothat which was given in Chapter 3.

As Lewis puts it, what is peculiar to indexical terms is thattheir "reference varies, depending on relevant features of thecontext of utterance."2 3 An indexical term could be said tohave as its meaning a rule that takes us from a context oftokening to a value or referent. The relevant contextual feature differs from one indexical term to another. A token of"now" denotes the t ime of the tokening, "here" (when usedwithout an act of ostension) the place of the tokening, "I" theperson doing the tokening, and so on. "The actual world" issupposed to behave analogously to these indexical terms.

"The relevant feature of context, for the term 'actual,' is theworld at which a given utterance occurs."24 A token of "theactual world" denotes the world in which the token occurs.

On the basis of this analogy, the following argument couldbe constructed, in which an analogy is drawn between "theactual world" and "now" or "the present":25

THE INDEXICAL ARGUMENT

6. The token ing at any time t of "N ow (or the prese nt) is t"expresses a true proposition [based on the tokening rule for"now"];

7. No time is now simpliciter [from 6];8. "The actual world" is an indexical term that is anal

ogous to "now" [Lewis 's premise];9. The tok en ing at any world w of "The actual world is w"

exp resses a tr ue pro pos ition [from 8 by changin g the re levantcontextual feature of utterance from a time to a world]; and

10. No world is the actual world simpliciter [from 9].

186



The indexical argument

It will be shown that this argument fails on both materialand formal grounds. Two of i ts premises, namely, 8 and 9,are false; an d it co nta ins two seriou s non se qu itu rs - the inferences from 6 to 7 and from 9 to 10. "The actual world" is notindexical, and, even if it were, nothing would follow aboutactuality being world relative.

To beg in, it will be sh ow n w hy "the actual wo rld " is notindexical. Four significant disanalogies between it and"now" (or the indexical terms of your choice - it doesn'tmatter) will be unearthed, namely, that unlike "now," "the

actual wo rld" (i) do es n ot alw ays take the w idest scop e, (ii) isnot a rigid designator, (iii) is not replaceable salva veritate byany recognized indexical expression, and (iv) does not contribute to different prop ositions being expressed w he n a sentence containing it is tokened at different relevant contexts.

(i) In ou r discu ssion of Hector Ca stan ed a's a cco unt ofquasi-indicators in Chapter 3, it was shown that an indexicalword always has the widest scope, even if it occurs in anoratio obliqua clause, because it expresses the speaker 's ownindexical reference. As a rule, w he n I asse rt the sen ten ce"Jones believes that Mary is baking pies now," the oratioobliqua occurrence of "no w " expresses only m y ow n indexicalreference, and th ereby I un de rd ete rm ine the con tent ofJones's prop ositional belief, since I do no t specify the m a nn erin which Jones referred to the time at which Mary is bakingpies - whether by an indexical expression, proper name, ordefinite des criptio n. I could just as well ha ve m ad e theexplicitly de re assertion "In regard to now, Jones believesthat Mary is baking pies at this time," in which "this time" isa bou nd v ariable. By contrast, wh en I assert the se nten ce"Jones believes that the actual world is a place of evil," I donot un de rde term ine the proposition believed by Jone s, sincethe em be dd ed o ccurrence of "the actual w orld " exp resses n otmy but Jones's reference. The explicit de re formulation "Inregard to the actual world, Jones believes that it is a place of

evil" is not synonymous in this case. That an oratio obliquaoccurrence of an indexical term, unlike "the actual world,"takes the wide scope should become even more apparent by

187




considering th e case in which I assert th at Jones believes thatthis world is not the actual world. While I am attributing toJones a radically false belief concerning what is actual, I amnot attributing to him the even m ore radical confusion of believing w ha t is expressed by the pragm atically self-falsifyingsentence "This world is not the actual world," in which apragmatically self-falsifying sentence is one every tokeningof which expresses a contingently false proposition. He maybe a bit crazy, but he is not that crazy.

(ii) Whereas indexical terms are rigid designators accord

ing to the account given in Chapter 3, namely, "d" is a rigiddesignator just in case the proposition expressed by a tokening of "d migh t n ot have been d" is false, "the actual world"is not. Any tokening of

11. Now (I, here) might not have been now (I, here)

expresses a false propo sition bu t not so for every tok en ing of

12. The actual world might not have been the actual w orld.

A tokening of 12 expresses a true prop osition w he n th e n arrow scope is given to its modal operator - there is a worldthat is the actual world and it is possible that it not be theactual world - rather than the wide scope - it is possible thatthere is a world that both is the actual world and is not theactual world. Numerous counterfactual stories can be told inwhich the world that in fact is the actual world fails to be. Ifany feature of the actual world were to have been different,

for instance, someone were to have chosen other than heactually did, some other world would have been the actualworld. But no counterfactual story can be told in which now- this very time - is a different time than it in fact is. Certainordinary ways of talking might suggest that this can be done.When the impatient bridegroom says to his bride during thewe dding cerem ony, "If it w ere now 9 P.M. rather than 3 P.M.,we wou ld be mak ing love," he is not ima gining a counterfactual situation in which now is not identical with the time it is

in fact identical with, namely, 3 P.M., but is saying only thatw herea s they a re (regretfully) not m ak ing love n ow at 3 P.M.,

188



The indexMl argument

they will be a t 9 P.M. Similarly, w hen I say to you that if I wereyou I wo uld buy treasury bo nd s, I am not imagining a situation in which I am identical with som eon e else, but am sayingonly that an yo ne in your situation prudentially ou ght to buytreasury bonds.

Lewis has a ready response to the charge that "actual" isnot indexical because it is not always used rigidly. His strategy is not to den y th e latter but instea d to save the analogy bydenying that every indexical term is always used rigidly. Hedist inguishes between a primary sense of "actual" in which itrefers to the world of utterance, even in a context whereanother world is under consideration, and a secondary sensein which it shifts its reference in such a context.26 A primaryuse is rigid, referring to the world in which it is used oruttered even when another world is envisaged, but a secondary use is nonrigid because the reference shifts to theenvisioned world. But something similar happens to somebut not all indexicals, thereby salvaging the analogy:

Compare 'now,' which is normally rigidified, with 'present,'which may or may not be. So you say 'Yesterday it was colderthan it is now,' and even in the scope of the time-shiftingadverb, 'now' still refers to the time of utterance. Likewiseyou say 'Yesterday it was colder than it is at present,' and thereference of 'present' is unshifted. But i( you say 'Every pastevent was once present,' then the time-shifting tensed verbshifts the reference of 'p resent.' 1 suggest that 'actual' andits cognates are like 'present': sometimes rigidified, sometimes not.27

In this passage Lewis fails to show that "at present" hasboth a rigid and nonrigid use, for his example of its nonrigiduse is "Every past event was once present." Herein "at present" does not even occur, only the nondesignator "present,"which in this use means the same as "happening." "Everypast event was once at present" is either informed or contradictory. One might think that a more favorable example forLewis's purpose is "he," which has both an indexical use in

189



The argument from wwW<-*f Wfee actuality

which i t m ean s "this ma n" a nd an ana phoric use as a relativepronoun. But this is an uninteresting case of an equivocationin which the latter has no connection with the former. Theoccurrence of "present" in "Mary is baking pies at present"and "Mary's baking pies once was present" are a similar sortof un intere sting equ ivocation. By sho w ing th at "the actualworld" can shift its reference in some counterfactual assertions, Lewis is unwittingly undermining his indexical theoryof actuality, since it makes it appear as if "the actual world"is a nonrigid definite description, similar to "the first person

to write out a jazz orc hes tratio n," w hich shifts in its referencewhen used in the counterfactual sentence "If the first personto write out a jazz orchestration w ere to hav e been W. C.Handy instead of Jelly Roll Morton, then Jelly Roll Mortonwould still claim to have been the first." (He had a big ego.)

(iii) Another reason for believing that "the actual world" isnot indexical is that it is not replaceable salva veritate by anexplicit indexical phrase and, in particular, not by Lewis'stw o can dida tes - "the w orld in w hich I exist" and "thisworld" or their variants.28 We will begin w ith the first-personanalysis.

Whereas it is true that

13. It is possible that I do not exist in the actual world,

it is false that

14. It is possible that I d o not exist in the world in which Iexist.

This cou nterex am ple to the first-person a nalysis does not begthe question against counterpart theory, since it does notrequire that I inhabit more than one world.

Th ings go n o better for the analy sis in term s of "this w orld "or "the maximal spatiotemporal aggregate containing this."Because "this w or ld" is rigid a nd "th e actual wo rld" is not, it

is true that12. The actual w orld migh t no t hav e be en th e actual world

190



The inimical afgmm0M

but false that

15. This world might not have been this world."The maximal spatiotemporal aggregate containing this" alsoflunks the substitutible salva veritate test, for w hile it cou ld betrue that

16. Jones disbelieves that the actual world is the m ax im umspatiotemporal aggregate containing this,

it almost certainly is false, unless he is in pretty bad shape,

that17. Jones disbelieves that the actual world is the actual

world.

W hile it is relatively easy to criticize the firs t-pe rson anddemonstrative analyses, it is more difficult to explain howvery able philosophers could have accepted them. My speculative explanation is that their acce ptan ce is base d on a failure

to distinguish between the meaning of a designator and theway in which we fix its reference, to make use of Kripke'sdistinction. It probably is the case that many of us are prepared to fix the reference of "the actual world" by employingthe indexical phrase "this world" or "the world in which Iexist." Were our madman Jones to ask me which world is theactual one, I m ight reply "This wo rld ," as I po int to so m eobject aro un d m e. That 1 fix the reference of "th e a ctua lworld" by using "this world" does not show that they havethe same sense, and thereby will be intersubstitutible salvaveritate.

(iv) Because "the actual world," unlike "now," is a non-rigid definite description, we should expect there to be thisfurther disanalogy: A sentence con taining " the actual w orld "expresses the same proposition when tokened at differentworlds whereas one containing "now" expresses differentpropositions when tokened at different t imes. It is apparent

that tokenings of "tx is now" at different times involve a shiftin the time referred to by "now" and express different propositions; but, disanalogously, due to the fact that "the actual

191




world" is a nonrigid definite description, tokenings of "w1 isthe actual world" at different worlds, although involving ashift in the world referred to by "the actual world," expressthe same proposition. This disanalogy, in turn, gives rise tothis further disanalogy: While it is true that

6. The tokening at any time t of "Now is t" expresses atrue propositions,

it is false that

9. The tokening at any world W of "The actual world is w"

expresses a true proposition.That there is this disanalogy, no doubt, is not obvious.

Some might quibble that 6 is true only if it is explicitly relativized to the English langu age, since an oth er lang uag e m ightcontain a sign d esig n similar to "N ow is t" that has a differentm eaning . Fu rtherm ore, it is true that successive tok ening s of"f, is now" express different propositions only if the successive tokenings are in English.29 I gladly accept these am end

ments. But even when similar amendments are made in 9, itis false. It is a universal proposition, but it doesn't hold forevery case. Imagine that the actual world is not w but w,. Inthat case it would not be true that the proposition expressedby the tokening at w of "w is the actual w orl d" is true. We canprotect 9 from this kind of counterexample by restricting itstruth-ascription to the world of the tokening. This gives us

9'. The tokening at any wo rld W of "The actual world is w"

expresses a true proposition at w. 3

°But this gives the game away, for there was no need to rela-tivize the truth-ascription of 6 analogously to the time of thetokening, as is done by

6'. The toke ning at any time t of "N ow is t" expresses a trueproposition at t.

It is interesting to note that w he n "this w orld " is s ub stituted

for "the actual world" in 9, it results in the true proposition that

192





The argument from world-relnkiw actuality

different contexts, which is a familiar story from Chapter 3.The proposition expressed by tokening today "Today is whenevent E occurs" is at least logically equivalent to, if not identical with, the proposition expressed by tokening tomorrowthe differently indexed counterpart to this sentence, "Yesterday is w he n event E occurred." But 1 cannot imagine a nythinganalogous for worlds, since I have no idea of what would bea differently "indexed" counterpart to a sentence containing"the actual world."

The indexical argument, in addition to containing the radi

cally false premises 8 and 9, also commits two serious nonsequiturs. It is totally mysterious how

7. No time is now simpliciter

is supposed to follow from

6. The tokening at an y time t of "N ow is t" expresses a trueproposition.

And likewise for how

10. No world is the actual world simpliciter

follows from

9. The tokening at any world w of "The actual world is w"expresses a true proposition.

We kno w in the case of 7 an d 6 that the en tailm ent do esn 'thold, beca use 6 is true an d 7 false an d enta ilm ent is tru th

preserving. But is 7 really false? Lewis does not think so, forhe writes:

If I am right, the ontological arguer who says that his world is* special because his world alone is the actual world is as foolish

as a man w ho boasts that he has the special fortune to be aliveat a unique moment in history: the present.32

Pace Lew is, the boaster obviously is right. W ould Lew is alsoconten d that wh en I say that I am Richard Gale (thou gh thisis no thin g to boast abo ut) this is no t tru e if it m ea ns that I am

194



The indexkal atgummt

Richard Gale simpliciter or absolutely? Certainly my state

ment is true without any relativization to persons - that I amRichard G ale at Richard G ale. Similarly, w h en I asse rt at th etime I am w riting this senten ce th at time f, is now , w ha t I sayis true and not in need of any relativization to a time - that t,is now at f1# Another way of putting this is that the proposition I express by tokening at f, "t, is now" is true simpliciterand, moreover, true without relativization to any time, as wesaw in the case of 6. Ag ain, we find that Lew is's accou nt d oe snot square with actual usage, and he cannot take the revi-sionary route since this winds up in the wonderful world ofhab erda she ry. Similar points hold concerning the connec tionbetween 10 and 9, with the difference that 9, unlike 6, is false.The predicate "is the actual world' ' ' is no more a disguisedtwo-place predicate than is "is now (the present)."

The problem, again, is to give some therapeutic account ofhow a bright person could be misled into accepting such afalse thesis. The explanation, again, might be in terms of the

failure to distinguish between the way in which the referenceof a token of a referring term is fixed and the sense this tokenhas . The following describes the tokening rule for fixing thereference of a token of "the present" in English:

18. The tokening of "the present" at any time t refers to t.

From 18 it m igh t wro ngly be inferred that

19. The tokening of "the present" at any time t m eans thesam e as "the time that is pre sen t at f."

A nd from 19 it do es follow that

20. The predicate "is the present" is a two-place predicate.

A similar argument can be given for "is the actual world"being a two-place predicate. From the following descriptionof the rule for fixing the referent of a token of "the actual

world ,"21. The tokening of "the actual world" at any world w

refers to w,

*95



The argum ent from worl4~Hl«tme acttmlisty

it is wrongly inferred that

22. The tokening of "the actual world" at any world wmeans the same as "the world that is actual at w."

This would establish the ontological parity thesis, namely,

23. The predica te "is the actual w orld " is a two-place p redicate.

U nfo rtun ately for this thesis , the inferences of 19 from 18an d 22 from 21 fail because 18 a nd 21 are each true and 19 an d

22 each false. Theses 18 and 21 each truly describes the referring rule in English for the use of "the present" and "theactual world" respectively. That 19 and 22 are each false canbe establishe d by showin g that the token s they claim to havethe same sense are not substi tutive salva veritate.

It is true that at any time f that

24. It is knowable a priori that t is the time that is presentat t,

but it is not true that at any time t that

25. It is knowable a priori that t is the present.

Similarly, it is true that in any world w

26. It is kn ow ab le a priori that w is the world that is actual

in w,

but it is not true that in any world w

27. It is knowable a priori that w is the actual world.Given these differences in sense between tokenings of theexpressions in question in the very same relevant contexts(time or world), it follows that the expressions will not besubsti tutable salva veritate in pro positional-attitude contexts. Icould believe that w is the world that is actual in w but notbelieve that w is the actual world, or know that time t is the

time that is present at t but not know that t is the present.This completes my critique of the indexical argument. It ismisleading to speak of it as the indexical arg um en t, since

196



Conclusion

there are other arguments that begin with Lewis's premises

that "actual" is indexical and terminate in his conclusion thatbeing actual is world relative. Maybe a more interesting andviable indexical argument can be constructed on Lewis'sbehalf; but I m us t leave this task to m y su pe rior s, since I ha vedone the best I can.

C O N C L U S I O N

One and only one theory of actuality - the indexical theory -has been considered, and it has been found wanting, therebysaving theism from an atheological argument based uponthis theory. The results of my discussion, however, are notwholly negative . It has been sho w n that "the actual w orld" isa nonrigid definite description and "is actual" is a monadicpredicate. While these theses do not constitute a full-blowntheory of actuality, any such theory must square with them.

It is not my purpose in this book to construct such a theory,since it would take me too far afield. Nor am I up to the task.But I w an t to cau tion those w h o atte m pt it. Because "isactu al" is m on ad ic, it m ust no t be inferred that being actual isa simple property that one and only one world possesses.This is not so m uc h w ro ng as un hel pfu l, for there are p rop erties and there are "properties."3 3 Ord inary, paradigmaticproperties, such as being red, are the sort of thing that an

individual can possess in different modal manners, forinstan ce, actually as oppo sed to m erely po ssibly being red . Ifbeing actual is supposed to jump through the same conceptual hoops as do these ordinary properties, a vicious infiniteregress is generated by the simple property theory of actuality. It analyzes "X is actual" into "X has the property ofactuality," but the latter means "X actually (as opposed toonly possibly) has th e prop erty of actua lity," w hich, in accordwith this theory, must be analyzed into "X has the propertyof actually having the property of actuality," and so on adinfinitum.34 The regress is vicious because it is one of meanings or analyses rather than one of only entailments. It would

197



The argument from w& rii-rehtwe actuality

app ear to be fundamentally wr ong to try to reduc e a mo dalityto a prop erty. The moda l m ann er in which an individual ha sa property is not itself a property. The actual world does notdiffer from other worlds in that it alone possesses the property of actuality but in that it alone actually has existence.M aybe it is existence tha t is the sou gh t after simp le pr op erty .

This completes my discussion of atheological arguments. Itrust that you have found your excursion through them tohave positive value in helping find a more adequate anddefensible conception of God. It is now time to turn to the

theological argu m en ts for it being rational to believe that G odexists. We shall begin with the epistemological or truth-directed arguments and then go on in the final chapter toconsider some pragmatic arguments for faith.

198



Theological arguments



Chapter 6.Ontological arguments

An ontological argument attempts to deduce the existence ofGod from an analysis of the conception of God, therebyshowing that it is necessary that God exists. From the merelogical possibility that this conception of God is instantiated,it is supposed to follow that it is necessarily instantiated. Ithas been dogmatically pronounced by a long line of philosophers from Hume up through the logical positivists that an

ontological argument cannot work, since existence can neverbe a logical conse quen ce of an entity's esse nce . This do gm a iscalled into question by what appear to be perfectly legitimateontological arguments for certain types of abstract entities,that is, entities that could not logically be located in eitherspace and/or t ime, such as num bers , prop erties, an d propo sitions. God, as conceived of by the great medieval theists, issuch an abstract entity, though differing from these abstractain having a life (an illimitable one at that, which is supposedto be h ad all at once ) an d also hav ing a causally efficaciouswill that can timelessly bring about effects in the universe.1

Given God's status as an abstract entity, the question naturally arises whether the same style of deductive reasoningfrom prem ises kno wa ble a priori to an existential c onclu sion,such as figures so prominently in mathematics in which existential questions are internal to the system, being decidableby deductions from axioms and definitions, might not also

apply to the existence of God. It certainly would be a blatantbegging of the question to deny that it can. It is no accidentthat the ontological argument is the darling of the mathemat-

2 0 1



Ontological arguments

ically inclined theistic philosophers in the Platonic tradition,those who think that the mathematical method of reasoningis the pathway to true knowledge.

This chap ter will consider ma ny different type s of ontological arguments. The first is based on a weak version of theprinciple of sufficient reason according to which everythinghas a possible explanation and argues that a contradictionfollows from the assumption that God does not exist consisting in there both being and not being a possible explanationfor this neg ative state of affairs. T he secon d versio n, of which

there are several variants, is based on some premise that issupposed to formulate a necessary truth about abstract entities in general, from which, in conjunction with other necessary premises, God's existence is supposed to follow. Thethird is the famous argument in chapter 2 of St. Anselm'sProslogion, and attempts to deduce a contradiction from theassumption that a being than which none greater can beconceived does not actually exist, namely, that the beingthan which none greater can be conceived could be greater(or could be conceived to be greater). The fourth, and finalversion to be consid ered , is based on a tou gh er set of requirements than those supposedly laid down by Anselm for qualifying as a being than which none greater can be conceivedand employs necessary existence, rather than plain old existence, as one of the great-making properties. It will turn outthat none of them works. The moral that will emerge fromour discussion is that the philosopher who gives an ontolog

ical argument is playing with fire. To make it work she mustup the ante, soup up her conception of God so much that itopens the door for an ontological disproof based on deductions from this very conception, which is just what will beseen to befall the fourth version in particular. More generally, it will be shown to be a fundamental mistake either toargue for the necessary existence of God or to conceive ofGod as having necessary existence. God should not be the

sort of entity th at necessarily exists; for, if he is so conceived ,it follows that he does not and cannot exist.

2 0 2






for the fact that x exists or the fact that x does not exist [weakversion of the principle of sufficient reason];

3. God does not exist [assumption for indirect proof];4. It is pos sible that th ere is an exp lana tion for the fact tha t

God does not exist [from 2 and 3];5. It is not possib le tha t there is an exp lana tion for the fact

that God does not exist [from 1];6. It is and it is not po ssible tha t the re is an ex plan ation for

the fact that God does not exist [from 4 and 5]; and7. It is false tha t God do es no t exist [from 3 th ru 6 by in di

rect proof].And, since a contradiction was deduced from the propositionthat God does not exist, it shows that this proposition is notonly false but necessarily false. It is necessary that God exists

While there is little to qu estion with eith er prem ise 1 or 2,the inference of 5 from 1 is most du bi ou s. It implicitlyassumes that the only possible explanation for the fact thatsome individual does not exist is a causal one in terms ofsomething that (causally) prevents the existence of this individual. But this is plainly not the case. One possible explanation for why some individual does not exist is an argumentthat shows a logical absurdity, such as a contradiction, in theclaim that it exists. For instance, we can give a noncausalexplanation for why the largest integer does not exist consisting in a reductio argument against its existence. Anothertype of a noncausal explanation of a negative proposition is

in terms of some positive reality that logically precludes thedenied state of affairs. For instance, the reason that I am notnow in San Francisco is that I am now in Pittsburgh. Mybeing in Pittsburgh now does not causally prevent but logically prec ludes my being in San Francisco no w . Similarly, th ereason the rug is not blue is that it is red, and so on.3

It is exactly such an explanation for a negative propositionin terms of a positive reality that logically excludes t he de ni edstate of affairs that is possible for the nonexistence of God.For instance, the reaso n tha t God d oe s not exist is that the reexists some positive state of affairs that logically precludes

2 4



The weak sufficient wm&a persfew

his existence, such as a morally unjustified evil, that is, an evil

that is such that it is logically necessary that God would notallow it to occur, such as an evil that serves no good pu rp os e.Such evils will serve as the linchpin on which will hinge myargument later in this chapter against it being necessarilytrue that God exists.

What is more, the proponent of the above argument isinconsistent in inferring 5 from 1. Just as it is not logicallypossible that anything causally prevents God's existence, it isno t logically possible that an yth ing cau ses his existence e ither;yet according to this argument, there is an explanation for hisexistence consisting in this very ontological argu m en t. Here inthe ontological arguer does recognize an explanation for anindiv idua l's existence or nonex istence that is not causal. If allexplanations of existence and nonexistence must be causal,there cannot be a successful ontological proof or disproof.An d this und ercu ts the very purpose of the above a rgu m ent,since it is intended to be an ontological proof of, and thereby

explanation for, God's existence.

Ontological arguments from God's abstractness

This is a neglected, but nevertheless interesting type of apriori argument for it being necessary that God exists thatis based upon the conception of God as an abstract entity.

What is distinctive about this kind of argument is that item ploys some prem ise that is such that the argum ent wo rksonly if it represents a necessary truth about abstract entities.Four different specimens of it will be critically evaluated - St.Anselm's argument in Part I of his Reply to Gaunilo, alongwith N orm an M alcolm's two variations on it in his "A nselm 'sOntological A rg um en t," plus a simplified v ersion of my ow nthat clearly brings out the essence of these arguments. It willbe shown that all four arguments face the objection that theiralleged necessary truth about abstract entities is questionbegging. Furthermore, each argument will be shown to beuncompelling by construction of a parallel argument that

205



OHtdogmlatguments

contains its alleged necessary truth about abstract entitiesand other premises that are no less plausible than those ofthe original but whose conclusion is logically incompatiblewith that of the original. And, in a later section of this chapter, it will be shown that there are independent reasons forthinking it false that it is necessary that God exists and thusfor holding suspect any argument that has this proposition asits conclusion.

Anselm writes:

I insist, however, that simply if it can be thought it is necessary that it exists. For 'that-than-which-a-greater-cannot-be-thought' cannot be thought save as being without a beginning. But whatever can be thought as existing and doesnot actually exist can be thought as having a beginning of itsexistence. Consequently, 'that-than-which-a-greater-cannot-be-thought' cannot be thought as existing and yet not actuallyexist. If, therefore, it can be thought as existing, it exists ofnecessity.4

This is a subtle and beautiful argument; but it cannot be fullyappreciated, no less shown to be valid, until it is translatedinto the logistical form of contemporary modal logic.

The argument attempts to prove that if "that-than-which-a-greater-cannot-be-thought" (to be abbreviated as "God")can be tho ug ht , t he n God exists of nece ssity. It takes the formof a cond itional proof in wh ich we begin w ith the ass um ptionof the antecedent proposition, which will be understood as

8. It is possible that God exists,

based on the assumption that for Anselm, whatever "can bethought" is logically possible.

Two premises are then added. Both are intended as necessary tru ths , as th ey m ust b e if Anselm is to be allowed to usethem in his attempt to deduce from 8 that God "exists ofnecessity," which is to be rendered as

C. It is necessary that God exists.

2 0 6



The weak sufficient reason version

The first says that God "cannot be though* save as beingwithout a beginning," which wil l be represented as

9. It is not possible that God begins.

That G od c ann ot ha ve a begin nin g in t ime is a logical consequence of Anselm's conception of God as an abstract entity.Following the Boethian-Augustinian conception of God'seternality as involving timelessness rather than everlasting-ness or omnitemporality, Anselm rightly deduces that i t isnot possible that God begins to exist; for, were God to do so,he would begin in t ime and thus be in t ime. The secondadd ed p rem ise, also assum ed to be necessary, is sup po sed tohold, or at any rate must hold if the argument is to succeed,for every abstract entity. It says that "whatever can betho ug ht as existing and doe s not actually exist can be t ho ug htas having a beg innin g." In accordance with the prev ious wa yof un der stan din g "can be thou gh t ," this will be para ph rasedas

10. It is necessary that, for any x, if it is possible that xexists and x does not exist, then it is possible that x begins.

From 8-10 we are supposed to deduce C, at which point wecan close the scope of our initial assumption 8 and deducethat if 8 then C. A nd since Anselm 's op po ne nt is sup po sed togrant him the truth of 8, C can be deduced by m od us po nen s .

To demonstrate that C is entailed by 8-10 it is helpful totrans late 8-10 into a C. I. Lewis-style m od al logical form an dmake use of the axioms of both S4 and S5. The following isthe Lewis-style translation and argument.

Anselm's argument from God's abstractness

8. MGod exists5 [assumption for conditional proof];9. - M G o d begins [necessary truth];10. L(x)[Mx exists an d -x exists) ID Mx begins] [necessary

truth];

2 7




11. L[(MGod exists and -God exists) =) MGod begins][from 10 by univer sal in stan tiation ];

12. L - G o d b egin s [from 9 by definition of L];13. LL - G o d beg ins [from 12 by axiom of S4];14. L [ - M God begins z> - (M Go d exists an d - Go d exists)]

[from 11 by tran spo sition ];15. L[L - G o d begins => (MG od exists and - G o d exists)]

[from 14 by definition of L];16. L - (M G o d exists and - G o d exists) [from 13 an d 15 by

axiom of modal logic L(p r> q) n (Lp Z) Lq)];

17. L( - M G o d exists or God exists) [from 16 by DeM organaxiom];18. L(M Go d exists r> God exists) [from 17 by defini tion of

3 ] ;19. LMGod exists [from 8 by axiom of S5];20. LGod exists [from 18 an d 19 by axiom of m od al logic in

step 16]; and21. MGod exists ID LGod exists [from 8-20 by conditional

proof].

Notice that if the material implication in 10 is not itselfnecessary, it can only be deduced that if it is possible thatGod exists, then G od exists. A nd this isn't a bad day's wo rk.The need to appeal to the axioms of S4 and S5 will not bequestioned no w . W hat they collectively say is that a pro pos ition's modality is invariant among possible worlds.

Norman Malcolm's two variations on this argument aredirected against an opponent who says that although nothing could prevent God from existing, it nevertheless is thecase that G od's existence is contin gent in that it is bo th pos sible that God exists and possible that he does not. In the firstargument, Malcolm replies that "from the supposition that itcould happen that God did not exist it would follow that, ifHe existed, He would have mere duration an d not eternity,"6

which, I assume, is meant to be a substitution instance of a

necessary universal truth. This consequence, however, isabsurd, since we conceive of God's eternality as excluding"as senseless all senten ces that imply that H e has du ra tio n."7

2 8





Ontologiml urgwments

He cannot have come into existence (for the reasons given),nor can He cease to exist, for nothing could cause Him tocease to exist nor could it just happ en that He ceased to exist.So if God exists His existence is necessary. Thus God's existence is either impossible or necessary. It can be the formeronly if the concept of such a being is self-contradictory or insome way logically absurd. Assuming that this is not so, itfollows that He necessarily exists.8

This very complex arg um en t can be sho w n to be valid w ithout the use of either S4 or S5. The following is an explicitformulation of the argument.

Malcolm's second argument

34. - M G o d is limited [necessary, truth];35. L[God came into existence r> (God is caused to do so

or God happened to do so)] [necessary truth];36. L(God is cau sed to do so r> God is limited) [necessary

truth];37. L( - G o d is limited o - G o d is caused to do so) [from

36 by transposition];38. L -G o d is limited [from 34 by definition of L];39. - M G o d is cause d to com e into existence [from 37 and

38 by axiom L(p z> q) Z2 (Lp z> Lq)};40. - M G o d h ap pe ne d to come into existence [by same

kind of argument as given for 39];41. L[

-(God is caused to come into existence or Godhappened to come into existence) z> -God came into existence] [from 35 by transposition];

42. - M G o d is caused to come into exis tence and - M G o dha p pe n ed to com e into existence [from 39 and 40 by conjunction];

43. - M (G o d is caused to come into existence or God ha ppened to come into existence) [from 42 by axiom of modallogic -M(p or q) as ( -M p an d -Mq)\;

44. L -(God is caused to come into existence or God happ en ed to com e int o existence) [from 43 by definition of L);

2 1 0



The weak sufficient teason version

45. L -G o d cam e into existence [from 41 an d 44 by axiom

in step 39];46. God does not exist [assumption for conditional proof];47. L(x)[(x does not exist and Mx exists) ID Mx comes into

existence] [necessary truth];48. L[(God does not exist and MGod exists) r> MGod

com es into existence] [from 47 by u niversal insta ntiati on ];49. - M G o d came into existence [from 45 by definition of L];50. -(God does not exist and MGod exists) [from 48, and

49 by m od us tollens];51. God exists or - M G o d exists [from 50 by D eM organ

axiom];52: - M G o d exists [from 46 and 51 by Disjunctive Syllo

gism];53. G od do es not exist 3 - M G o d exists [from 46-52 by

conditional proof];54. God exists [assumption for conditional proof];55. L(God exists z> —M God beg ins or cea ses to exist) [nec

essary truth];56. - M G o d beg ins or ceases to exist [from 54 an d 55 bymodus ponens] ;

57. L(x)[(x exists and Mx does not exist) 3 Mx begins orceases to exist] [necessary truth];

58. L[(God exists and MGod does not exist) r> MGodbeg ins or ceases to exist] [from 57 by u niv ersal ins tant iatio n];

59. -(God exists and MGod does not exist) [from 56 and58 by modus tollens];

60. - G o d exists or —M God do es not exist [from 59 byDeMorgan axiom];

61. - M G o d doe s not exist [from 54 an d 60 by disjunctivesyllogism];

62. G od exists r> - M G o d d oe s not exist [from 54 -61 byconditional proof];

63. (God does not exist n - M G o d exists) an d (God exists3 - M G o d do es not exist) [from 53 and 62 by conjunction];

64. God exists or God does not exist [tautology];65. - M G o d exists or -M G o d d oes not exis t [from 63 and64 by constructive dilem m a];

2 1 1




66. MGod exists [granted by Malcolm's opponent];67. - M G o d does not exist [from 65 an d 66 by d isjunctive

syllogism]; and68. LG od exists [from 67 by defin ition of L].

While this argument is far more complex than its two predecessors, it is of a piece with them. Each of these argumentcontains a premise that is either identical with or a closecousin of one of its premises. This argument's premise

47. L(x)[(x does not exist and Mx exists) 3 Mx comes intoexistence]

appears to be identical with Anselm's premise

10. L(x)[(Mx exists and -x do es not exist) r> Mx begins],

assuming that "come into existence" means the same as "begin." Also, its premise

57. L(x)[(x exists and Mx does not exist) 3 Mx begins orceases to exist]

is a close cousin of Malcolm's first argument's premise,

24. L(x)[M -x do es not exist n (x exists z> x has duration )]

agreeing in their antecedents when exportation is applied to24 and differing only slightly in their consequents. What iscommon to all four propositions, which will become evidentas we proceed, is that they are true only if no abstract entitycan have a contingent existence.

Before showing why this requirement is question-begging,it will be helpful to construct a simplified version of thesethree arguments that makes this requirement explicit.

The simplified argument

69. L{x)[x is an abstract entity 3 (L x exists or Lx does notexist)] [necessary truth];

70. L[God is an abstract entity r> (LGod exists or LGoddoes not exist)] [from 69 by un ivers al in stan tiati on ];

2 1 2



Objections to arguments from God's abstractness

71. God is an abstract entity [necessary truth];72. LG od ex ists or LGod d oe s not exist [from 70 an d 71 by

modus ponens] ;73. -LGod does not exist [granted by opponent of argu

ment] ; and74. LG od ex ists [from 72 an d 73 by disjunctive syllogism].

O B J E C T I O N S T O A R G U M E N T SF R O M G O D ' S A B S T R A C T N E S S

The first objection to be lodged against all four arguments istha t each con tain s as a prem ise an alleged necessary truth tha tis question-begging. The conception of a question-beggingargument is a dialectical one, since it concerns what propositions the op po ne nt of the argum ent is will ing to concede priorto the prese ntation of the arg um en t. N ot every valid ded uctive argument will be question begging. The valid argument

that Socrates is mortal because all men are mortal and Socrates is a man will beg the question only if the person towhom it is directed initially rejects either of its premises. Itwill not be fair for the opponent of the argument, after it hasbeen given and discovered to be valid, to charge that itbegged the question, for at that time he was willing to grantits premises.

What is the dialectical context of our four arguments? Thisgets down to what are the prior beliefs of the op pon en t ofthese arguments. The opponent is characterized by Malcolmas having the p rior belief that Go d is a contingen t being - that"it might just happen that He did not exist."9 When thisbelief is combined with the opponent's prior acceptance ofthe Anselm-Malcolm conception of God as an abstract entity,it bec om es clear tha t the op po ne nt w ill not grant prem ises 69,10, 24, and 57; for he could offer God himself as a counterexample to each of them.

He can offer this abstract, but contingent God as a counterexa m ple to th e prem ise 69 claim that necessarily everyabstract entity is either a nece ssary or im possible being. A nd

213





Objections to arguments from God 's abstradness

ises of the original arg um en ts du biou s, in particular prem ises69, 10, 24, and 57.

It might be questioned why it is logically impossible forboth God and the GCAD to exist. It is not because therew ou ld exist two om nipo tent beings, but rather because therewould exist two sovereign beings. A sovereign being completely and uniquely determines every feature of any worldin which it exists, which is such that it is logically consistentthat this being completely and uniquely determines it. Thus,if two providential beings existed in a given world, each

w ou ld com pletely and uniquely determ ine everything in thatworld that it is consistent for either one of them alone completely and uniquely to determine, thereby resulting in thecontradiction that some things in this world both are and arenot completely and uniquely determined by one of the twosovereign beings.

The only way to block this refutation is to argue that it isnot possible for a being to have one of the divine perfectionsw ith ou t ha ving all of them . In Chap ter 1 it w as argue d thatthere is no good argument for this and, more strongly, thatthe doctrine of the divine simplicity is a most implausibledoctrine.

It is important to note that even if it could be proven thata being could not have one of the divine perfections withoutha vin g all of the m , it w ould not save An selm's arg um en t an dMalcolm's second argument from refutation by a parallelargument that uses the concept of an omnitemporally eternalGod in place of that of a timelessly eternal God. The theistictraditio n is pretty m uch split do w n th e m iddle with respect tothese two different ways of understanding God's eternality.All of the great medieval theists, as well as Berkeley andSchleiermacher, have defended the timelessly eternal view,while the omnitemporally eternal view is that of the Bible,along with the Nicene Creed and most Protestant theologians, not to mention present-day process theologians. It has

already been argue d at some length in Ch apters 2 and 3 thatthe religiously available God of the everyday working theistcannot be timelessly eternal. Malcolm could not consistently

215




attack the intelligibility or viability of this omnitemporal conception of God, for he defends the viability of his own time

less conception of God on the grounds that this conceptionenters into language games that are actually played. Butexactly the same can be said, only more so, on behalf of theomnitemporal conception of God.I0 W hat Malcolm m us t do isto argue that it is necessary that any temporal being has thepossibility of ceasing to exist; but, as argued in Chapter 3, itis unlikely that any such argument can succeed. There seemsto be no thin g con tradictory or ab su rd ab out the conception ofan omnitemporally eternal God, that is, a God that existsthroughout a beginningless and endless t ime and logicallycould not begin or cease to exist. Plainly, the onus is onMalcolm to show that this conception of God, which hasentered into the language games of working theists throughout the ages, is not viable. This he has not done.

Anselm's argument and Malcolm's second argument workequally well for the greatest conceivable omnitemporallyeternal being.11 But then w e again w ind u p with incom patible

conclusions, this time between a sovereign timeless God anda sovereign omnitemporal God. One does not escape thispro blem by app ea l to the doc trine of the Trinity, in which oneof its members is the timelessly eternal and another theomnitemporally eternal God. The reason is that, if taken in aliteral, nonmystical manner, it violates Leibniz's law, sincethese Gods are supposed to be identical and yet differ insome of their properties, one being temporal and the othernot. It was suggested in Chapter 2 that this doctrine must be

interpreted as a mystical doctrine, but then it cannot beap pe ale d to in order to save a piece of rational theolog y, suchas the arguments of Anselm and Malcolm.

Again, the moral to be drawn is that the premises, in particular 10 an d 57, of the original arg um en ts are suspect. A ndsince 10 and 57 are of a piece with premise 24 in Malcolm'sfirst argument in that the same sort of metaphysical intuitions that lead one to accept or reject one of the former dolikewise for the latter, it also renders this argument uncom-pelling . Th ere is an oth er way to discredit our four ar gu m en ts,

2 1 6



Anselm's i>ersi@H

based on an arg um en t for i t being im possible that Go d necessarily exists, that will be explored later.

ANSELM'S VERSION

Our concern now is with the version of the ontological argum ent that Saint An selm pre sen ts in C ha pte r 2 of P roslogion.Many sil ly arguments have been attributed to Anselm bypeople who obviously have never read the text. Sometimes itis said that he engaged in the sleight of hand maneuver ofdefining the word "God" in such a way that i t included,among other divine perfections, the property of having existence. From this it follows only that no nonexistent individual could be G od. W here as a non existent being , for instanc e,Merlan, could be a magician, no such merely fictional ormythical being could qualify as God. But this is perfectlyconsistent with the proposition that there exists no individ

ual that is God. To conceive of God as being essentially existent leads to the same dead end, for it entails only thetautology that God is existent in every possible world inw hich G od exists, w hich, aga in, is quite consistent w ith the reexisting no individual that is God. It does not help to add theadditiona l prem ise that God , so defined or conceived, is apossible being or could exist, since it would follow only thatthe individual that is God in some possible world is existentin that world, not that this world is the actual world or thatthis individual actually exists.

Fortun ately, this is not St. Anselm's argu m en t, for he do esnot arg ue from, bu t to, God having existence. H e beg ins w iththe conception of Go d as a being than w hich no ne greater canbe conceived, leaving it open whether any individual thatqualifies as such m us t be existent. To be su re, th ere is anadditional premise having to do with existence being a great-making property, but there is no attempt to include beingexisten t in the conc eptio n or definition of God or a be ing th anwhich none greater can be conceived. There is no doubt thatin Chapter 2 he is employing the concept of plain old exis-

2 1 7



Ontologiml arguments

fence, not necessary existence. Whether he had other versions of an ontological argument in Chapter 3 an d his Reply toGaunilo that employed such modalized existence is a controversial issue that will be left alone. (I believe that the text isnot decisive and that if he had such a version, it was notclearly worked out in his own mind.)

Let us begin by quoting Anselm's formulation of the argument in Chapter 2.

Now we believe that You are something than which nothing

greater can be thought. . . . Even the Fool . . . is forced toagree that something-than-which-nothing-greater-can-be-thought exists in the mind, since he understands this when hehears it, and whatever is understood is in the mind. Andsurely that-than-which-a-greater-cannot-be-thought cannotexist in the mind alone. For if it exists solely in the mind even,it can be thought to exist in reality also, which is greater. Ifthen that-than-which-a-greater-cannot-be-thought exists inthe mind alone, this same that-than-which-a-greater-ca«nof-be-thought is that-than-which-a-greater-can-be-thought. Butthis is obviously impossible. Therefore there is absolutelyno doubt that something-than-which-a-greater-cannot-be-thought exists both in the mind and in reality.12

This is a turgid piece of reasoning that cries out for clarification. My strategy will be to give at the outset my own interpretat ion of it in the form of an explicit indirect proof, bearingin mind that it is only one am ong several w ays of draw ing an

argument out of the text, and then go on to clarify its premises and indicate their textual basis.

Anselm's Chapter 2 argument

75. God is that than which a greater cannot be thought[true by definition];

76. It is possible that that than which a greater cannot bethought exists [granted by fool opponent];

218



Anseim's virion

JJ. That than w hich a greater cann ot be tho ug ht does notexist [assumption for indirect proof];

78. For any individual x, if it is possible that x exists an d xdoes not exist, then x could be greater [necessary truth];

79. If it is po ssible th at that than which a greater can no t bethought exists and that than which a greater cannot bethought does not exist, then that than which a greater cannot be thought could be greater [from 78 by universal instantiation];

80. That than which a greater cannot be thought could be

greater than it is [from 76, yy, and 79 by mo du s pon ens);81. It is no t the case that that than which a greate r ca nnot

be thought does not exist [from 77-80 by indirect proof].

And since a contradiction, namely, proposition 80, has beendeduced from the assumption that God, conceived of as abeing than which a greater cannot be thought, does not exist,in which use was made of only necessarily true propositions,

it follows that it is necessary that God exists.Premise 75, the "definitional premise," should not occasion mu ch difficulty, but often s tud en ts wh o read Anse lm forthe first time object that it does not accurately describe theway in which they conceive of or define God. If it was alexical definition, its failure to agree with ordinary usagewould be a serious flaw. But for the purposes of Anselm'sargument, it can function as a purely stipulative definition,the point being that when God is so conceived or defined, itcan be demonstrated that he exists.

Another difficulty they have with the definitional premiseis that they do n ot un de rsta nd wh at is m eant by a being th anw hich a grea ter can not be tho ug ht. To mak e sense of thisconception, it must be placed in a religious context. God,understood as that than which a greater cannot be thought,plays a special role in the religious language game as anem inen tly w orsh ipa ble an d obeyable being. It will be recalled

from the d iscussio n in Ch apter 1 that this constituted on e ofGod's hard-c ore p rope rties, but a higher-level, emerge nt o ne.

219




There will be lower-level determiners of it, such as beingessentially omnipotent, omnibenevolent, omniscient, andthe like. This supplies some positive content to the conception, but there is also a negative component to it, since thisbeing is conceived of having each of these lower-level perfections in an unlimited manner. Herein Anselm goes a longway with the negative approach to the divine; but, unlike aPascal, who thinks that our inability qua finite creatures toform a fully positive conception of God's true nature oressence bars us from being able to argue either for or againstthe existence of God, Anselm sees it as no barrier, his earlierargument being proof that it isn't.

Premise 76, the "possibility premise," is far more controversial. The op po ne nt of the ar g u m en t - the Fool of th e Scriptures who denied that God exists - supposedly understoodthat whose existence he was denying. Anselm uses "understand," "understandable," and "in the mind" in a somewhattechnical way so that from the fact that any individual satisfies either of them, it follows that it has the logical possibilityof existing.13 If the Fool were to deny that the being thanwhich a greater cannot be thought is not understandable inthe sense of being logically possible, the onus would be onhim to show that this concept is in some way absurd or contradictory. Supposedly, the Fool is not able to do this, andthereby is committed to accepting 76.

Nor will it do for the fool to charge that 76 begs the question, since if it is not granted, Anselm cannot deduce his

desired conclusion 81; for this is an a fter-th e-arg um ent refusalto grant a prem ise, w hereas w ha t is in question in determ ining w hethe r an argum ent begs the qu estion is the before-the-argument bel iefs of the argument 's opponent . And suppos edly th e Fool is willing to gr an t 76 before h e kn ow s w ha tis coming.

Step JJ is the assumption for indirect proof, and it is ashortened version of Anselm's claim that "it exists solely inthe mind." What is at issue is not whether God has intentional existence as an object of thought but whether he hasfirst-class, mind-independent existence. From the assump-

2 2



Anselm's version

tion that he does not, a contradiction - 80 - is su pp os ed tofollow.

To sho w this it m us t be assum ed that existence is a grea t-making property. Premise 80, the "great-making premise,"presents one way of formulating this doctrine based onAn selm 's claim that "if it exists solely in the m ind e ve n, it canbe thought to exist in reality also, which is greater." Following the principle of minimal ordinance - that an argumentshould not have stronger premises than are absolutelyrequ ired for it to be valid - 80 is the w ea ke st in terp reta tion of

the text that will do the trick. There are stronger versions ofthe existence-is-a-great-making-property doctrine, forinstance, that any individual that exists is greater than anyindividual that does not. But why use them and run the riskof refutation if they are not essential to the argument's validity?

In 80 a com parison is dr aw n not betw een a n one xistententity and some other existen t entity in respe ct to their rela

tive degrees of greatness but between the greatness possessed by a single entity in two different circu m stanc es, inone of which it exists and the other in which it does not. Ifthere are no other relevant differences between these circumstances, the individual is greater in the circumstance inwhich it is existent. Some, such as C. D. Broad, find such acomparison to be suspect, because they believe that we cannot talk about the greatness of a nonexistent being. But it isnot obvious that this is so, for w e do talk ab ou t the gre atn essof purely fictional and mythical beings.

A no ther gr ou nd for attacking 80 is ba sed o n the den ial th atexistence is a great-making prop erty in anyth ing . In su pp or tof this, Norman Malcolm points out that if President Bushwere to ask his advisors to make up a list of the propertiesthat it is desirable for a secretary of defense to have, theymight put on their list such properties as being chaste, beinga teetotaler, having no previous business relationships with

defense companies, but they would not list having existence.14 The response to this is that we do not do so becausethe context makes clear that we are restricting ourselves to

2 2 1



Ontologkal arguments

existent beings. It is strange to find Malcolm going on todefend the doctrine that necessary existence, that is, the

impossibility of failing to exist, is a great-making property,for it would seem that in general the only reason why beingnecessarily or essentially F is great-m aking or desirable is thatbeing plain old F is great-making or desirable. We would notprefer that an individual be essentially benevolent ratherthan just benevolent unless we deemed benevolence itselfdesirable. But if what was in question was malevolence, wewould prefer the malevolent to the essentially malevolentbeing. Malcolm, therefore, seems to be inconsistent in upholding necessary existence as a great-making property butdenying that existence is.

The crucial step in the deduction is 80, the "reductio step."It is based on Anselm's claim "For if it exists solely in themind even, it can be thought to exist in reality also, which isgrea ter." My interpretation is based up on the earlier ass um ption that for Anselm "what can be thought" is what is possible or could exist. Thu s, if that than w hich a grea ter can no t bethou gh t can be th ou gh t to be greater than it is, it follows tha tthat than which a greater can be thought could be greatertha n it is. This is taken to be contradictory by A nse lm , so thathe can close the scope of the assumption made in JJ anddeduce its falsity.

But is the redu ctio step really contradictory? At first glanceit appears to be so. How could that than which a greatercannot be thought be greater than it is? A good case can be

made out that it is not.15

The first step in the refutation is torealize that an individual has existence or status in more thanone possible world and its greatness can vary across theseworlds. The Brando character in Chapter 1 w ho said th at h ecould have been a con tend er is implicitly saying t ha t he existsin other possible wo rlds than the actual one an d th at in som eof these he is a contender and thereby greater in that worldthan he is in the actual one. All of us believe that we couldhave done better than we in fact did and that thereby we aregreater in some possible world than we are in the actualworld. But we should take solace in the thought that there

2 2 2



Anselm's version

are possible w orld s in which w e do a lot w orse . Ah, the gloryand the misery of possible-worlds fantasizing.

Given that an individual exists in more than one world,there is a crucial ambiguity in

80. That than which a greater cannot be thought could begreater than it is.

The phrase "than it is" is incomplete, failing to specify aworld. Is it the actual world that is in question or the worldin which it realizes its greatest perfection, that is, the one in

which it is essentially (omnipotent, omniscient, omnibenevo-lent, and so on), which shall be called maximal excellence? Ifthis being were not to exist in the actual world, for instance,to have a merely mythical or fictional status in this world,then it would be true that it could be greater than it is in theactual world, though not in the world in which it realizesmaximal excellence. Superman, for example, in virtue of failing to exist in the actual world, could be greater than he is inthe actual world. Maybe he could even be greater than he isin the world depicted by Marvel Comics, for he could be freeof his vulnerability to kryptonite.

This way of giving a noncontradictory rendering of 80becomes quite compelling once a possible-worlds renderingis given of the possibility premise

76. It is possible that that than which a greater cannot bethought exists.

Given that what is possible is what is realized in at least onepossible world and that an individual's greatness must berelativized to a world since its greatness varies across worlds,76 is to be translated into

76t. Th ere is a possible world 10 and an individual x, suchthat x exists in 10 and the greatness of x in 10 is not exceededby that of any individual in any possible world.

While it is not possible that this individual x be greater thanit is in w, w he re in it realizes m axima l excellence, it is possiblethat x could be greater than it is in the actual world, for it

223





The necessary existence version

According to the basic intuition underlying the necessary

existence version, the determination of the greatness of abeing in some wo rld de pe nd s not only up o n ho w it goes w iththis being in that world but also upon how it goes with thisbeing in other possible worlds - the logical space that surrounds this being and determines its possibilities. This intuition is shared by the Anselm version, since it required thatthe being wh o instantiates a greatest greatnes s hav e in ev erypossible world in which it exists all of the great-makingomniproperties. I t did not go far enough, however, and seethat necessa ry existence - the impossib ility of failing to exist -was also required of a being than which a greater cannot bethou ght. Th us, the new version replaces j62 with the strongerpossibility premise that

763. There is a possible world w in which the property ofhaving unsurpassable greatness, that is, maximal excellenceplus necessary existence, is instantiated.

With this premise it can be deduced that the being, call itFred, who instantiates having unsurpassable greatness in wdoes exist and has maximal excellence in the actual world.Since Fred has essentially all the properties that comprisemaximal excellence, it has these properties in every possibleworld in which it exists. And, given that the actual world isa possible world and that Fred exists in every possible worldin virtue of having necessary existence, it follows that Fredexists and has maximal excellence in the actual world.*6

We would be imposing on the Fool opponent of the onto-logical argu m ent if we were to dem and that he grant th e n ewpossibility premise 76, on the grounds that he accepted theformer po ssibility prem ises 76-763. He m igh t be a fool but heis not a complete schmuck. He can point out that this newpossibil ity premise involving unsu rpas sab le greatn ess is considerably stronger than the previous ones involving onlymaximal excellence. Furthermore, he could charge 763 with

begging the question. That i t is blatantly question-beggingbecomes manifest wh en th e earlier arg um en t is reformulatedin logistical terms employing S5 as follows:

225



Ontdogical arguments

82. God is an unsurpassably great being [by definition];83. An unsurpassably great being is maximally excellent,

that is, essentially omnipotent, omniscient, and so on, andnecessarily existent [by definition];763. There is a possible world w in which the property of

un su rpa ssa ble greatnes s is instantiated [granted by fool (??)];84. M(3x)(x is max imally excellent an d necessa rily exis

ten t) [ from 763 by the principle that x exists in some possibleworld = It is possible that x ex ists];

85. ML( 3x)(x is maximally excellent) [from 84 by definition of "necessarily existent" - X is nece ssarily e xiste nt = It isnecessary that X exists]; and

86. L(3 x)(x is maximally excellent) [from 85 by th eo rem ofS3 thatMLp z> Lp].

And since it is necessary that there exists a maximally excellent being , there actually exists a maximally excellent being,w h o , by definition, is God. Such a being, in other words,exists in the actual world.

Before considering the crucial question of whether the Foolwill grant 76,, som ething ou ght to be said abo ut the S5 theorem, that MLp 3 Lp, which is appe aled to in the ded uctio n of86 from 85. It is not clear what this theorem means, no lessthat it is true. This theorem requires that the accessibilityrelation betw een w orlds is reflexive, transitive, and sym me trical. On e w orld is accessible to an oth er if a bein g (or po ssiblebeing) in the latter can know the former's world book, that is,all of the propositions true in that world. A proposition ispossible if it is true in some world, and necessary if true inevery world. Thus, if it is possible that it is necessary that p,it follows that that p is necessarily true in some world w, butit qualifies as necessarily true in this world because in thisworld a being could discover that it is included in every otherworld book. There are problems with the S5 doctrine thatevery world is accessible to every other one, especially inregard to indexical propositions and the epistemologicalgro un ds on w hich a person can determine the truth-values of

226



The ne&e$sary extetinee version

propositions in other worlds, but we will let them pass since

they don't directly concern the issue under consideration.The explicit S5 formulation of Plantinga's argument bringsout just what is involved in the Fool's acceptance of the possibility premise j61. It would be unreasonable to ask him toassent to a proposition until he fully understands what itmeans. Once the Fool understands what unsurpassablegreatness involves - in particular that it involves necessaryexistence - and that for an entity to have necessary existence

is for it to be necessary that it exists, and, moreover, acceptthe S5 theorem that what is possibly necessary is necessary,he wo uld hav e to be quite a schm uck to grant 763. In g ranting763, he is in effect granting that it is possible that it is necessary that God, conceived of as an unsurpassably great being,exists. A nd that is som ethin g the op po ne nt of the ontologicalargument would not want to do once he knows what isentailed by the combination of these modal operators within

S5. The principle of informed consent applies to disputationand not just within the area of biomedical ethics.W hile it see m s clear that 763 beg s the qu estio n, there re

mains the larger question of whether it is true. If people weresimply to deny it, they would be pitting their modal intuitionagainst Plantinga's. This would result in a philosophical stalemate that would satisfy Plantinga's larger purpose of showing that it is not irrational or epistemically impermissible tobelieve that God exists; for his argument is valid and haspremises that are just as probable or likely to be true as anyor all propositions that are incompatible with them.

A mo re p rom isin g stra tegy for rebu tting 763, as well asPlantinga's claim that even if his argument does not succeedas a piece of ration al the olo gy , at least it sh ow s that religiousbelief is epistemically permissible, is to find some propertythat (i) intuitively seems more likely to admit of the possibility of instantiation than does having unsurpassable greatnessand (ii) is strongly incompatible with it in that if either propertyis instan tiated in any po ssible w orld , the other is instantiatedin none. It is not difficult to produce properties that satisfy

227




(ii), for instance, Plantinga's own example of the property ofhaving near m aximality, which is possessed by a being just incase "it does not exist in every possible world but has adegree of greatness not exceeded by that of any being in anyworld."17 But such properties are intuitively on a par withhaving unsurpassable greatness, neither possessing greaterintuitive possibility of instantiation than the other; andthereby they do not satisfy condition (i).

Just slightly, but only slightly, less artificial is that of beinga maximally great omnimalevolent being - our GCAD who

essentially has all of the properties that make up maximalexcellence, save for malevolence in place of benevolence.18

That this property is strongly incompatible with unsurpassable greatness is due to the logical impossibility of thereexisting in the same w orld tw o beings that are both sovereignfor the reason already given. The ontological arguer mightchallenge the possibility of this property being instantiatedby appeal to Plantinga's ontological argument. While thiswould be question-begging, it still is not clear which of thetwo strongly incompatible properties (if not both) is to berelegated to the junk heap of the logically impossible.

But we need n ot reso rt to such artificial ex am ples , since wehave ready at hand properties that do satisfy both (i) and (ii).O ne exam ple is the pro per ty of being a morally unjustifiedevil, that is, an evil that is such that the God of traditionaltheism could not be morally excused for permitting it. Suchan evil, wh ich is familiar to us from C ha pte r 4, is one that can

neither be attributed to the use of (or failure to use) free willby finite persons nor be shown to be necessary for the realization of some outweighing good or the prevention of agreater evil. An example of such an evil is a supernova thatdestroys a planet containing sentient beings without thisresulting from any finite person's use of or failure to use freewill nor this being necessary for the prevention of an evengreater evil or the realization of an outweighing good; forinstance, we never would have known how much Jonesmeant to us if we hadn't lost him in that supernova. There isno possible world in wh ich bo th God, an unsu rpa ssa bly great

228




bein g, and morally unjustified evil exist, since it is a co nc eptual truth that God is both willing and able to prevent such anevil. But if the prope rty of havin g un sur pa ssa ble grea tnes s isinstantiated in any world, it is instantiated in every one.Therefore, the possibility prem ise 763 is logically inc om pa tible with

87. The re is a possible wo rld in which the pro pe rty of being a morally unjustified evil is instantiated.

A theist who does not impute a necessary existence to God

can accept 87 an d then go on to construct a defen se or th eo dicy for different types of moral and natural evil, dependingon whether it is the deductive or inductive argument fromevil that is being rebutted. But the problem of evil for thetheist who either accepts an ontological argument or imputesnecessary existence to God takes a far more acute form. It isthe modal problem of evil, concerning the very possibility ofthere being a morally unjustified evil, the theistic response towhich must take the form of an argument for the impossibility of there being an evil for which God does not have amorally exonerating excuse. It is not e no ug h to give a FWD ,since this still leaves op en the possibility of a morally un jus tified natural evil, such as the example of the su pe rn ov a. W ha tthese theists must establish is that it is more likely that it ispossible that having unsurpassable greatness is instantiatedtha n that being a morally unjustified evil is. At a m ini m um ,they can challenge the intuitions underlying 87 with theirown counterintuitions favoring 763.

There is an even more plausible property that satisfies (i)and (ii), namely, being a world in which every free personalway s freely doe s w ha t is morally w ro ng or, in other w o rd s,being a world that contains moral evil sans moral good. Thiswill prove to be an especially troublesome objection forPlantinga, since it will be show n that the very intuitions tha testablish tha t this pro pe rty satisfies (i) and (ii) are tho se tha t

enter into his own FWD. But this will have more than a meread hominem interest, since these intuitions are both widelyshared among professional theologians and working theists

229




and plausible on their own. First, it will be shown why it isreasonable to think that this property has the possibility ofbeing instan tiate d, then , w hy it is strongly incompatible w ithhaving unsurpassable greatness.

It is both a basic tenet of any FWD and granted by everyright-thinking person that

88. There is a possible world containing free persons.

The FWD is committed to this by its account of how it ispossible for God to create free persons, namely, by actualiz

ing their diminished possible persons. Whatever is possibleis the case in some possible world. Now a free person cannotonly make a morally wrong choice on any occasion but onevery on e as w ell. Th us,

89. There is a possible world in which some free personalways freely does what is morally wrong.

What holds true for any free pe rso n in this respect also h old sfor all free p er so ns. Therefore,

90. There is a possible world in which every free personalways freely does what is morally wrong.

While neither 89 nor 90 is entailed by Plantinga's FWD,they rest on the same sort of intuitions as do some of its keypremises. It would be most odd, although not contradictory,for someone to accept the FWD's depravity premise.

- M . It is possible that God cannot actualize a possibleworld in which all free persons always freely go right.

but deny 89 or 90. The reason is that the depravity premiserests on the possibility that the F-conditionals would havecontingent truth-values such that if any of their antecedentsw ere to be in sta nt iate d, it w ould result in the occu rrence of atleast some moral evil. According to the FWD, and rightly inmy opinion, their contingent truth-values could be even

more unfavorable, so that it would result in moral evil sansmoral good, that is, every free person always freely goingwrong. This still does not entail 90. What must be added is

23




that in some possible world in wh ich the F-conditionals ha vesuch unfavorable truth-values, some of them have their antecedents instantiated, be it by God or mere chance. (It willtu rn out that it is only by the latter, since it w ou ld be inc onsistent for Plantinga's God to do so.) This certainly seems possible. More needs to be said in defense of 90, but first it will beshown why it is logically incompatible with 763.

Plantinga's above ontological argument shows that from

763. There is a possible world w in which the property ofhaving unsurpassable greatness, that is, maximal excellenceplus necessary existence, is instantiated,it can be deduced that

86. L( x)(x is maximally excellent).

And 86 in turn entails that

91. The re is an individual, wh o, by definition, is Go d, an dboth exists and has maximal excellence in every possible

world.Since 763 entails 91, if 91 is incom patible with 90, so is 763. Thefollowing is a demonstration, based upon the premises of theFWD, of the incompatibility of 91, an d the reb y 763, w ith 90.

A logical consequence of gi is that

92. God both exists and ha s omnibe nevo lence, om niscience, and sovereignty in every possible world,

since, by definition, a maximally excellent be ing is essentially(omnibenevolent, omniscient, and sovereign). Since Godexercises his sovereignty in every possible world, he completely and solely determines every feature of every worldthat it is consistent for him completely and solely to determine. According to the FWD, given that it is consistent forGod to create free persons (though not also to determinewhat they freely do - that is why the F-conditionals do nothave their truth-values determined by God), it follows that

93. In every possible world in which there exist free persons, God creates them.

231






98. There is no possible world in which every free personalways freely does what is morally wrong.

And from this it can be concluded that the property of beinga world in which every free person always freely does whatis morally w ron g is strongly incom patible with having un surpassable greatness, that is, that 90 is incompatible with 763.Fu rth erm or e, since 90 rests on far stron ger intuitions thandoes 763, it is 763, and thereby Plantinga's ontological argument, that is to be rejected.

Theists who believe that God has necessary existence andthereby accept 763 are not going to accept this demonstrationof the falsity of 763. The sort of rebuttals that are available tothem will depend upon what tenets they hold in additionto 763. First w e shall cons ider the po ssible r espo ns es of th ePlantinga-type theist - the one wh o accepts 763 as well as th epremises and intuitions that underlie Plantinga's FWD. Suchtheists have very few options for rebutting the above demonstration of 90. For instan ce, they ca n't explain how G od cou ldpermit the possibility envisioned by 90 to come to pass interm s of his lacking midd le know ledg e, thereby morally exonerating him for creating free persons all of whom alwaysfreely go wrong on the basis of an excusable lack of knowledge. An Adams-type theist can say this but not a Plantinga-type, since the former, unlike the latter, denies that God hasmiddle knowledge.

One possible move is for the Plantinga-type theist to denythat God is essentially benevolent. This will require a redefinition of "maximal excellence" so that to qualify as such, abeing must be only plain old benevolent, though it still mustessentially have all of the other omniproperties. In any possible wo rld in which all free pe rso ns a lwa ys do wh at is mo rallywrong, God will exist and have middle knowledge. He justwon't be morally at his best. In fact, he will qualify as a realcad in such wo rlds. This do w ng rad ing of wha t is required of

a maximally excellent being, however, clashes with the firstpremise of the necessary existence version of the ontologicalargument:

233



Ontelogical arguments

82. God is an unsurpassably great being.

A being that has this sort of watered-down maximal excellence intuitively does not seem to qualify as an unsurpassably great being or a being than which a greater cannot bethought. It presents us with an insuperable case of the paradox of perfection.

The Plantinga-type theists might just outright reject 90 byappeal to their intuitions that it is more likely that havingunsurpassable greatness admits of the possibility of instantiation than does being a world in which all free persons freelydo what is morally wrong. The problem with this is that, asargued above, the very premises and intuitions that enterinto their FWD su pp or t the possibility of the latter, tha t is, 90.An other pro blem for them is that eve n if they could establishon some gro un d that 90 is false, they cou ld still be c ha llengedby a we aker an d thereby m ore intuitively powerful ver sion of90, say

90'. There is a possible wo rld in w hich the majority of actsfreely performed by persons are morally wrong.This w eak en ed v ersion of 90 is incom patible with the no rm ative premise of their FWD

96. A world containing free persons who freely performboth right and wrong actions, but for the most part go right, isbetter than any possible world devoid of free persons,

since it requ ires a favorable ba lanc e of moral goo d ove r m ora levil, something not realized in the 90'-type possible world.

Plainly, theis ts w ho accept 763 bu t do n ot bu y in on all th epremises of Plantinga's FWD, especially the doctrine of middle kno w ledg e, ha ve a better chan ce of refuting 90. Th ey canappeal to one of the versions of a FWD sans middle knowledge from Chapter 4 to show how 90 is consistent with 763.It has already been argued in that chapter that such versionsdo not work. But even if one of them did, they would still

face two problems. Their FWD at best meets the challengepose d by 90 bu t no t that bas ed u p on the seem ing possibilityof there b eing a morally unjustified evil. Furtherm ore, by den y-

234



The necessary existence Version

ing middle knowledge to God, they significantly limit hisomniscience, and this seems hard to square with their identification of God with an un surp assab ly great being in pre m ise82 of their ontological argument.

The most reasonable move for this type of theist to make isto challenge my claim that intuitively it seems that being amorally unjustified evil or a world in which all free personsalways freely do wrong more likely admits of the possibilityof instantiation than does having unsurpassable greatness.Both Phil Quinn and Peter van Inwagen wrote to me that

their modal intuitions are the reverse of mine. My responseto them is tha t our prima facie modal intu ition s spea k in favorof my intuitions, and thus the onus is on them to show thatthey are not to be trusted. This might be don e by sho w ing , inregard to 90 for exam ple, that up on a deep er an alysis, it tu rn sout to be impossible that all free persons would always freelygo wrong. It might be argued that a free person must berational and that on at least some occasions, rational beingswill realize that it is in their own best interest to do themorally right thing, since thereby they keep afloat the practice or institution of following moral rules, which furtherstheir own prudential interests. This Kantian-style argumentaga inst 90 is weak, since w e can conceive of a pos sible w orld ,such as one in which the Evil Demon is in charge, in whichon every occasion that a free choice is to be made, the agenthas good reason to believe that the pru de ntia l th ing to chooseis the morally wrong alternative.

I conveyed these considerations in a letter to Phil Quinn,and he wrote back that he rejected my placing the onus onhim. His justification was that immediately upon openingand reading my letter in the office of the philosophy department at Notre Dame, he asked the first seven people he metwhether they thought it was possible that being a morallyunjustified evil or a world in which everyone always freelygoes wrong be instantiated, and to a person they denied its

possibility. I, in turn, con duc ted my ow n m od al intuition pollat the office of the philosophy department at the Universityof Pittsburg h after I op en ed his letter, an d, no t surprisin gly,

235




everyone I con sulte d tho ug ht it possible that thes e pro pertie sbe instantiated. This clash of modal intuitions between the

persons at these two departments raises an intriguing possibility. Might it not be a good idea to complement the longstanding rivalry between the football teams representingthese two great institutions with a Modal Intuition Bowlinvolving these two rival departments?

If the participants in this bowl are restricted to the members of the two philosophy departments, it might seem thatthe result will be a tie (unless someone is throwing thegame), the reason being that the participants are sufficientlytutored in philosophy to know what will result from theirassenting to or denying the possibility of these propertiesbeing instantiated. But there are ways of finding a winner.O ne way w e mig ht try to resolve the deadlock is to w iden thescope of our poll and ask some "ordinary" people who havenot been perverted by the study of philosophy to give theirmodal intuitions. I actually left the philosophy departmentoffice and pu t the question to such pe op le - a janitor, m ainte

nance man, barmaid, and the like - and each had modalintuitions that matched mine.Fu rther m ore, I question the consistency of the No tre

Dame-types who refuse to countenance the possibility ofthese properties being instantiated because they know theproblem such an admission would pose for their brand oftheism. If their modal intuitions really did favor 763 over thepossibility of these properties being instantiated, why do som any of them feel it necessary to constru ct a defense, as wellas a theodicy, for what seem to be morally un justified evils?Premise 763 alone constitutes a perfectly adequate defense ofGod against an y type of evil. It wo uld a pp ear that the reasonthey feel a need to construct a defense other than that basedon an acceptance of 763 is that the y are no t all tha t su re of 763.Thereby they violate one of the basic rules of the ModalIntuition Bowl - be consistent in you r mo da l intuition s Furthe rm ore , the ir m od al intuition in favor of 763 over its rivals

seems to be mo tivated by their desire to us e it for the pu rp os eof giving the stated ontological argument in which it occurs

236





Chapter 7

Cosmological arguments

One of the fondest memories of most of us is the time wenailed our parents with the regress of why-questions, thecrowning point of which was when we triumphantly asked,"Yeah, so who made God?" At this point we were told to goplay marbles in traffic. Little did our parents realize that thecosmological argument could have served as Jim Dandy tothe rescue; for it demonstrates that our initial explanatory

demand must ultimately terminate with a self-explainingexplainer, in which a self-explaining being is one whose existence is entailed by its nature or essence, that is, one forwhom there is a successful ontological argument.

But if there is a successful ontological argument, as thecosmological argument supposedly proves, why screwaround with the cosmological argument? Why not go rightfor the jugular and give this ontological argument, thereby

having an end to the matter of whether God exists? Thiswas one of Kant's objections to the cosmological argument.This charge of redundancy misses the mark, since therecould be a successful ontological argument even if we wereunable to give it. Thus, we could know on the basis of thecosmological argument that there is a successful ontologicalargument and thereby that God exists, but not be smartenough to give it.

The most telling objection that can be lodged against thecosmological argum en t is tha t it is imp ossible for such a bein gto exist, thereby showing that this argument's conclusion isnecessarily false. Any ar gu m en t for an impossible c onclu sion,

238



Costnological arguments

to say the least, leaves something to be desired, whether inthe way of validity or so un dn es s. This will be the grou nd onwhich every version of the cosmological argument will beattacked. The points made in the Chapter 6 criticisms ofontological arguments will form the basis of my polemic.

Before getting do w n to a detailed analy sis of different versions of the cosmological argument, it will be helpful todescribe their generic nature. A cosmological argument ismade up of three components: (a) a contingent existentialfact, (b) an explanatory argument, and (c) a version of the

principle of sufficient reason (hereafter PSR). It begins bydemanding an explanation for the existential fact in (a) orsome fact that it entails in conjunction with certain otherprem ises or assum ptions . This is followed by an ar gu m en t in(b) that purports to demonstrate that this fact or the factgenerated by it can be explained only in terms of the causalefficacy of a self-explaining God, or som eth ing very like G od ,which will be called a "theistic explanation." And, finally,there is in (c) a version of PSR, suitably tailored so as torequire that there be an explanation for the sort of fact ofwhich the initial fact or the one entailed by it is an instance.Th e conjunction of (a)-(c) entails that ther e is a theistic explanation of the fact in question and thereby that there exists aself-explaining God.

Different versions of the cosmological argument resultfrom different explanatory demands. St. Thomas, in his FirstWay, begin s with the fact that on e object is moved by an oth er ,

in the Second Way that one object's existence is simultaneously dependent upon another. Sometimes the existentialfact is quite humble, such as that there exists somethingrather than nothing or that there exists at least one beingwhose existence is dependent upon the causal efficacy ofanother. Often it is a more rich fact, for instance, that thereexists this universe, comprised of all the specific objects andevents it contains. Herein what requires explanation is amaximal spatiotemporal object.

This existential fact component, as well as the PSR one, isnot knowable a priori. It is desirable that the existential fact

239




be either und isp uta ble or, even better, of a sort one of w ho sespecies or determiners is bound to obtain, thereby assuringthat n o m atter w ha t th e actual world is like, there will be su cha fact to be explained; for instance, no matter what possiblew orld gets actualized, the re will be some one world book tha tis true simpliciter and for which an explanation can be demanded. Few versions realize the latter, though some comeclose. For instance, if the explanatory demand concerns whythere exists a certain concrete aggregate of individuals thatcomprise the universe, it can be repeated no matter what

individu als th e univ erse contain s, provid ed there is at least aspatio tem por al agg regate of some kind. A cosmological argument that begins with the existential fact that there existssomething also comes very close to realizing this desideratum, if not fully realizing it.

It is also desirable that the existential fact be of a sort thatcan no t in principle either be scientifically explained or requireno explanation relative to some possible scientific theory; forinstance, that there exists something. Flistory amply demonstrates th at it is un w ise to mak e a theistic explanation a competitor to a scientific explanation. This desideratum is oftenviolated by the cosmological arguments given by scientists,especially cosmologists, who, if they are steady-state theorists, ask what causes hydrogen atoms continually to comeinto existence or, if big-b ang th eorists, w hat p recipitated theprimordial cosmic explosion. They might even hark back toDescartes and ask why material objects persist in being or

why matter is conserved.1

They indict science with beingnecessarily inco m plete b ecau se it cann ot in principle explainthe cosmological fact in question.

What they fail to realize is that their explanatory demandconcerns a kind of fact that is either in principle scientificallyexplainable or no t in ne ed of any explanation relative to som epossible scientific theory. Certainly, it is conceivable thatsomeday science would be able to explain what broughtabout the initial cosmic explosion that presently serves as aunique point singularity. Furthermore, whether i t is incum-

24



Cosmot&gkal arguments

bent on us to explain why certain objects persist in being or,contrariwise, why hydrogen atoms come into being willdepend on the scientific theory we accept, in particular whatsort of cons erva tion laws it inco rpo rates . R elative to A ristotelian science, the fact of motion required explanation but notso relative to Newtonian science with its first law of motion.And if what requires an explanation is why some law ortheory holds, it is possible that it would be explainable bysom e m ore inclusive law or theor y, su ch as was the case w iththe explanation of the Galilean law of free fall in terms of

Newton's law of gravity. The strategy of the cosmologicalarguer should not be to ask why some specific law or theoryholds but rather why there is any lawlike regularity at all. Allof the versions of the cosmological argument that will beconsidered in this chapter have explanatory demands thatnecessarily fall outside the purview of science.

I t is the explanatory argument component that dis t inguishes a cosmological argument from inductive theological

arguments, such as the teleological argument or the argument from the widespread lawlike regularity and simplicityof the wo rld . Th e latter are inferenc es to the best exp lana tion ,the claim being that the theistic explanation of these existential facts is superior to its competitors or that these facts makeit likely or probable that there exists a being very much likeGod whose causal efficacy explains these facts. The explanatory argument of a cosmological argument, in contradistinction to these inductive arguments, tr ies to demonstrate thatthe only possible explanation of these facts is the theistic onein terms of God's will. Sometimes it is not clear whether anargument based upon facts of natural design, order, beauty,and the like is a cosmological or inductive argument due to itnot being made manifest whether the theistic explanation ofthes e facts is su pp os ed to be the only pos sible one or only themost probable one.

There are numerous versions of the PSR that vary in

strength, from the strongest version requiring that every trueproposition have an explanation to a relatively weak one

241





Concrete-aggregate versions

C O N C R E T E - A G G R E G AT E V E R S I O N S

Scien t i f i c exp lana t ions a re by na tu re p iecemea l , exp la in ingo n e part o f the un iverse in t e rms o f some o ther part, w i t happea l be ing made to some sor t o f nomic connec t ion o f ade te rmin i s t i c o r s t a t i s t i ca l so r t tha t connec t s them toge ther.But wha t , a sks the p roponen t o f one ve rs ion o f the concre te -aggrega te ve rs ion o f the cosmolog ica l a rgument , exp la ins theexis tence of the universe as a whole? We f ind Leibniz , in "Onthe Ul t imate Or ig in o f Th ings , " wr i t ing :

And even if you imagine the world eternal, nevertheless sinceyou posit nothing but a succession of states and as you find asufficient rea son for them in no ne of them w hatso eve r, an d a sany number of them whatever does not aid you in giving areason for them, it is evident that the reason must be soughtelsewhere.

And , in more recen t t imes , R ichard Tay lor has made the samedemand tha t the re be an exp lana t ion o f the ex i s tence o f theu n i v e r s e - a s - a - w h o l e :

But it is at least very odd and arbitrary to deny of this existingworld the need for any sufficient reason, whether independe nt of itself or not , while pre sup pos ing that there is a reas onfor every other thing that ever exists.2

Both au thors assume some vers ion o f the PSR requ i r ingtha t every ind iv id ua l ha s an exp lana t io n o f i ts ex i s t ence an dthen go on to t rea t the universe as a whole as an individual ,and thereby subjec t to th is vers ion of the PSR. Leibniz g ivesa b ri ef ske t ch of t he exp l a n a to ry a rg um en t t ha t de m on s t r a t e sthe imposs ib i l i ty of expla in ing the exis tence of the universeas a whole in a non the i s t i c way, fo r ins tance , by appea l toso m e ind iv idu a l wi th in the un iv erse . To do so in the l a t t e r

manne r, s u p p o se d ly, wou ld v io l a t e t he p r i nc ip l e t ha t noind iv idua l can be a se l f -cause , s ince the in t raun iverse ind iv idu a l th a t i s inv oke d to exp la in the ex i s tence of the un ive rse

243




as a whole would have to explain, among other things, i tsown existence.

The concrete-aggregate version finds its fullest and mostforceful presentation in Part II of Samuel Clarke's 1705Demonstration of the Being and Attributes of God, The argumentis presented in two stages. The first attempts to prove thatthere is a being who necessarily exists and is the causalexplainer of the world of dependent beings, the second thatthis being is the God of traditional theism. The argument ofthe first stage, which will be our primary concern, employs a

fairly strong version of the PSR that holds that whateverexists (and not just whatever comes into existence), even if itbe omnitemporally eternal, has an explanation of its existen ce. Th e expla natio n for a bein g's existence will be eith er anexternalistic one in terms of the causal efficacy of some otherbein g, in wh ich case it is a "d ep en de nt bei ng ," or an intern al-istic one in terms of its own nature, it then being an "independent being." Given this distinction between the two waysin which a being's existence can be explained, Clarke's version of the PSR can be expressed as:

PSR,. Every existing thing has a reason for its existenceeither in the necessity of its own nature or in the causal efficacy of some other being.

Armed with PSR,, Clarke gives the following argument forthe existence of a necessary being:

1. Every being is either a d ep en de nt being or an in de pe ndent being [PSR,];2. Either there exists an in de pe nd en t being or every being

is d ep en d en t [from 1 by logical equiv alence];3. It is false that every being is dependent [premise];4. Th ere exists an in de pe nd en t being [from 2 and 3 by dis

junctive syllogism]; and5. Th ere exists a necessary b eing [from 4 by defin ition of

an independent being as on e for whic h the re is a succes sfulontological argument] .

244





Cosmologkri mgrnmnU

two) or open with local closed causal loops, such as portrayed by Robert Heinlein in his stories "All You Zombies"and "By His Bootstraps," in which a person literally is hisown mother and father due to a sex-change operation followed by a backward time journey to enable this person tocopulate with his earlier self. Clarke, being a good Newtonian, did not countenance the possibility of such closedcausal loops, no less a topologically closed time, but theirpossibility does not und erm ine his arg um en t; for if time w ereto form one big closed causal loop comprised of a finite n u m

ber of dependent beings, he could still demand an explanation, in the name of PSRj, of the entire causal loop, just as hehad demanded an explanation of the entire infinite linearsuccession of dependent beings.4

There is an existential assumption underlying Clarke'sargument that has not yet been made explicit, and, withoutwhich, the subsidiary argument fails. Before presenting themain argument, Clarke assumes that at every moment of

time in an infinite past, there h as existed som e being or o ther .Actually, he can get by with the weaker existential assumption that there exists at least one being, be it dependent orind ep en de nt. Unless he makes an ass um ptio n that is at leastas strong as the latter, it is not true that

3b. If the whole of existing things consists of an infinitesuccession of dep en de nt being s, the infinite succession itselfmust have an explanation;

for if there were to exist nothing, it would be true that thewhole of existing things consists of dependent beings, since(x)Fx = —(3x)—Fx, but there would not be an infinite succession of dependent beings to be explained.5

Thus, i t turns out that when Clarke's argument is completely fleshed out, the initial existential fact that comprisesco m po ne nt (a), that there exists at least one bein g, is no t thefact about which the explanatory de m an d is m ad e and wh ich

is shown in component (b) to admit of no explanation otherthan a theistic one. Rather, this initial existential fact, in con-

246






14. There is no explanation for th e existenc e of this infinitesuccession of de pe nd en t beings [from 11 to 13 by modus tol-

lens and DeMorgan theorem];15. There both is and is not an explanation for the exis

tence of this infinite succession of dependent beings [from 10and 14 by conjunction];

16. It is false that every being is a dependent being [from7 to 15 by indirect proof];

4. There exists an ind ep en de nt be ing [from 2 an d 16 bydisjunctive syllogism]; and

5. There exists a necessary being [from 4 by definition ofan independent being].

The problematic premises in this argument are 9 and 13,which correspond, respectively, to 3b and 3e in the subsidiary argument, and we shall now consider some objections to them. Since PSR, quantifies over individuals, theinfinite succession of dependent beings is not covered byPSR, unless it qualifies as an indiv idu al. For different rea sons

it has been doubted that it so qualifies. These scepticaldoubts would also apply to the maximal spatiotemporalaggregate, and thus undermine every concrete-aggregateversion of the cosmological argument. The fact that we havenames for these entities - "the history of the world" for theinfinite succession of dependent beings and "the universe"for the maximal spatiotemporal aggregate - is not decisive inallaying such doubts. But what are the grounds for doubt?

In opposition to Clarke's PSR,-based demand that there bean explanation for the infinite succession of temporal beingssince it qualifies ̂ as an ind ividu al in its ow n righ t, Row ewrites:

But such a view of the infinite collection is implausible, if notplainly incorrect. Many collections of physical things cannotpossibly be themselves concrete entities. Think, for example,of the collection whose members are the largest prehistoricbeast, Socrates, and the Empire State Building. By any stretchof the imagination can we view this collection as itself a concrete thing? Clearly we cannot. . . . At best our know ledge of

248



C&ilerete-aggregate ii&sions

the things (both past and present) comprising the universeand our knowledge of their interrelations would have to bemuch greater before we would be entitled to view the sum ofconcrete th ing s, past and prese nt, as itself som ething concrete.(p. 135; see also p. 145)

R o w e ' s a r g u m e n t a g a i n s t p r e m i s e 9 of m y c o m b i n e d f o r m u l at ion and 3b o f the subs id ia ry a rgument needs some f i l l ingou t , e spec ia l ly in regard to wha t a "concre te en t i ty" i s . Furthe rm o re , i t i s no t c l e a r wha t s o r t s o f know ledge we wou ld

have to possess be fore we would be en t i t l ed to t ake thewho le o f de p enden t be i ngs a s a s i ng l e conc r e t e en t i t y.

Rowe g ives us no reason why the l a rges t p reh i s to r ic beas t ,Socra tes , and the Empi re S ta te Bui ld ing (which sounds l ikethe be g i nn in g of a Jo hn ny C arso n "W hat do . . . ha ve in common?") a re no t pa r t s o f a s ing le concre te en t i ty, and we mus ts u p p l y s o m e a c c o u n t o n h i s behalf. I t w ou ld see m th a t aconcre te en t i ty mus t have a na tu re o r e ssence - the wha t - i t - i s

- tha t spec i f i e s the m a n ne r in w hich i t s spa t i a l pa r t s a re conca tena ted and func t ion in re la t ion to each o the r. Th ink o f thesong "Dry Bones" as a case in po in t . For th i s r eason a concre te en t i ty c an be iden t i f ied as be ing of a cer ta in sor t or t yp ein v i r tue o f i t s s i ze and shape and , mos t impor tan t , by theway i t s spat ia l par ts are in ter re la ted so as to achieve cer ta ind i s t in c t iv e func t io ns . Of ten su ch ob jec ts m ov e as a w ho le . Aba r ke r a t an auc t i on w ho w an t e d m e t o m ove do w n f ron t s a id"M ov e the fee t an d the bo dy fo l lows ." F ina lly, an d m os timpor tan t , an ob jec t ' s na tu re wi l l p resen t us wi th a way o fco nc ei vin g of it th a t i s e i th er prag m at ica l ly use ful in h el p i ngus to get around effec t ive ly in the wor ld or in te l lec tual ly sa ti s fy ing by employ ing sc ien t i f i c no t ions tha t he lp us to unders t a nd w h y t he obj ect beha ves a s it doe s . Th e m an ne r i n wh ichthe pe r iod ic t ab le de f ines the na tu re o f the d i ffe ren t e lementsis a go od ex a m pl e of th e la t ter. Let us ca ll an objec t tha t h a sa na tu re tha t sa t i s f i e s the p reced ing cond i t ions a " sor ta l

ob j ec t " a n d i t s n a tu r e a " so r t a l na tu r e . " Obv i o u s ly, t he c o ncre t e ag gr eg a t e co m po se d of Ro we ' s th ree d i s jo in ted ob jec t sis not a sor ta l objec t or, as he would say, a "concre te ent i ty."

249




The underlying assumption in Rowe's objection is that ifthere is a causal explanation of X then X is a sortal object. The

reason why it is illegitimate to ask for the causal explanationof the infinite su ccession of de p en d en t be ings or the m aximalspatio tem po ral a gg rega te is that it do es no t qualify as a sortalobject. But Rowe's assumption is false, facing numerousclear-cut counterexamples. A heap of twenty rocks is not asortal object, n ot m ov ing as a w ho le or hav ing a sortal n atu re,yet it makes sense to ask for the cause of its existence, whichcould be the person who assembled the rocks into a heap.

The distinction between sortal and nonsortal objects hasonly a rough event analogue, since there seem s to be no temporal analogue to the essentially spatial notion of an Aristotelian natural kind that has just been articulated. And do notconfuse th e issue by bringin g up the concep t of unnatural acts.A sortal event would be one whose concept prescribes themanner in which its temporal parts are concatenated, forinstan ce, a baseball ga m e. The cause of a sortal event, such asa war, can be something above and beyond the several

causes of the battles that make up the war, such as the kingwhose decree brought about the war. Similar considerationshold for a nonsortal event, such as the "event" comprised ofthe succession of Jon es run nin g in place, Smith falling do w n,and Brown singing "Dixie." There could be a common causeof this conjunctive event consisting in some person hiringthese three people to perform their respective acts, just asNero is the cause of a strange succession of events by hisissuing the order "Let the games begin." Much more will besaid shortly about common causes of concrete aggregates,both sortal and nonsortal, but for now we can conclude thatRowe's objection to 9 and 3b fails because its assumption isnot true in general.

Instead of denying that the infinite succession of dependent beings or the maximum spatiotemporal aggregate qualifies as an "individual," it might be said that an adequateformulation of the PSR should not count them as individualsand thereby require that they have an explanation. The reason they should not be subject to the PSR is that they are

2 5 0




unique in virtue of involving the totality. Bertrand Russell, inhis BBC debate with Father Copleston, seemed to make thispoint wh en he said in opposition to Co pleston's dem and thatthere be an explanation for the universe as a whole that"you're looking for something which can't be got" and "youhave to grasp this sorry scheme of things entire to do whatyou wan t, an d that we can't do."6 Russell is m akin g a stro ng erclaim than that we have no right to assume that there is anexplanation for the existence of the maximum spatiotemporalagg rega te; for this w ou ld just be to call into que stion the PSR

- a perfectly legitimate thing to do. Rather, he is making thestronger claim that there (conceptually) could not be such anexplanation. But why not?

There are two reasons that he might give. When askedwhy it is illegitimate to explain the existence of the universeas a whole, he responded that such a request rested on thefallacy of composition in which it is assumed that a wholemust have the same properties as i ts parts:

I can illustrate what seems to me your fallacy. Every man whoexists has a m other, and it seems to me your argum ent is thattherefore the human race must have a mother, but obviouslythe human race hasn't a mother - that's a different logicalsphere.7

In saying th at the hu m an race is in "a different logical sph er e"than are individual men, Russell seems to be taking the human race to be an abstract set. To be sure, abstract entities donot have ca us es , bu t this exam ple is crucially disan alog ous tothe case of the universe versus parts of the universe in whichboth members are concrete entit ies, the universe occupyingthe same space-time region as does i ts parts. Thus, theredoes not seem to be any obvious category mistake in askingfor the cause of the universe. As Rowe has pointed out, thecosmological arguer who requires a causal explanation for

the existence of the universe does not do so on the basis ofeach of i ts parts having such an exp lanation, w hich w ould besubject to the charge of a fallacy of composition, but rather on

251




the basis of the PSR and the universe being an individual inits own right.

But there is another way in which Russell might supporthis claim that there cannot be an explanation for the totality.It would be based on his scientism - his deep-seated beliefthat all knowledge is scientific knowledge. Knowledgerequires an explanation, and, in the case of scientific knowledge, such an explanation will be a piecemeal, intraworldaffair in which one state or event within the universe is nom-ically hooked up with another. Since all knowledge is scientific and all scientific explanations are intraworld, therecann ot be any expla nation for the existence of the to tality, theworld itself, at least none that could qualify as knowledgeconstituting. I have heard m any p hilo sop he rs of science echoa similar sentiment. But that's all it is - a sentiment. Thecosmological arguer will rightly charge Russell with beggingthe question by assuming at the outset that all knowledge isscientific because only scientific explanations qualify as realexplanations.

Now for wh at is generally concede d to be the most pow erful objection - tha t prem ises 13 an d 3e w ron gly as su m e t hatthere can be an explanation of a whole or aggregate aboveand beyond the several explanations for the parts or members that make it up. This objection is raised by Hume in theninth paragraph of Part 9 of his Dialogues Concerning NaturalReligion:

Did I show you the particular causes of each individual in acollection of twenty particles of matter, I should think it veryunreasonable should you afterwards ask me what was thecause of the whole twenty. This is sufficiently explained inexplaining the parts.

This implies that since each one of the infinitely m any de pe ndent beings in Clarke's succession has an explanation, theinfinite succession is thereby explained, pace premises 13

and 3c This objection has been forcefully reiterated in recent

252




t imes by Paul Ed w ard s, w h o imagines a gro up of five Eskimosstanding together in New York City:

Let us assume that we have now explained in the case of eachof the five Eskimos why he or she is in New York. Somebodythen asks: "All right, but what about the group as a whole;why is it in New York?" This would plainly be an absurdquestion. There is no group over and above the five mem bers,and if we have explained why each of the five members is inNew York we have ipso facto explained w hy the group is there.It is just as absurd to ask for the cause of the series as a wholeas distinct from asking for the causes of individual members.8

The principle underlying the Hume-Edwards objection is:

H E. If the existence of every member of a collection(group, succession, and the like) is explained, the existenceof the collection, and the like, is thereby explained.

There are two things to note about this principle. First, it

ap plies to con crete collections or agg regates, since a gro up offive Eskimos and a collection (heap) of twenty particles ofmatter have spatiotemporal location. Second, it is intendedas a necessary conceptual truth. This is especially clear in thecase of Edwards, who says that i t would "plainly be absurd"to ask why a group exists after the existence of each of itsm em be rs ha s be en explaine d. The moda l status of HE is lessclear in Hume's case, for he says only that it would be "veryunreasonable" to raise this question about the collection. Heusua lly is un de rstoo d as hold ing HE to be necessary (and m ycolleague Annette Baier, a real Hume expert, assures me thatthis is what he intended, which is enough for me). And thisis how he will be understood.

Th e HE principle is just plain false. N um er ou s c ounterexamples have already been given to it, involving both sortaland nonsortal objects. The existence of each and every rockin a heap of rocks could be causally explained, each one

owing its existence to some prisoner swinging a sledgehammer, and yet there could be a causal explanation for the

253




existence of the hea p, for instan ce, in terms of the assem blingactivity of some other prisoner, above and beyond the several explanations for the rocks in the heap. Of course, theremight not be, but that there might be suffices to refute HE,understood as being a necessary conceptual truth. A sortalobject, such as an automobile, could have an explanation forits existence in terms of the activities in a Detroit assemblyplant that is above and beyond the several explanations foreach part of the automobile, for instance, the carburetor wasm ade by De ko-R em y in Chicago, th e starter mo tor by U nitedMotors in Kansas City, and so on.

The exact cause of HE's failure is that explanation is notclosed with respect to conjunctive intro du ction , th at is, is no t"agg lom erativ e," to use Michael Slote's helpful term inolo gy.9

What this means is that the following is an invalid argumentform:

There is an explanation for X (or the fact that p);The re is an ex planatio n for Y (or the fact that q); th erefore,Th ere is an ex pla na tion for X and Y (or the fact that p

and q).

It could be noncoincidental that A is in the Sludge Falls Bankat T - there is an ex plan ation for w h y h e is there at T, forinstance, his boss, B, sent him there to make a deposit. Andit also could be non coin cide ntal tha t C is in this bank at T forinstance, his boss, D, sent him to make a withdrawal. Andyet it could be coincidental that A an d C are both in the ban kat T. Th ere might be no exp lan atio n of this conjunctive fact.And, again, there could be an explanation of this fact aboveand beyond the several explanations of its conjuncts, forinstance, B and D have the same bo ss, £ , w ho o rdered themto order their subordinates, A and C, respectively, to be inthe bank at T so he cou ld wi n a w age r.

No t only could the re be an exp lana tion of a concrete agg regate or conjunctive fact above and beyond the several explanati on s of its parts or con junc ts, in som e cases it is reaso nableto assume that there is such an explanation. As WesleySalmon has convincingly argued, when "confronted with

254




what appears to be an improbable coincidence, we seek acom mo n cause of this imp robab ility."10 Ed w ards 's exam ple ofthe group of five Eskimos standing together in New York,unfortunately for its intended purpose, is just such a case. Itis reasonable to seek an explanation for the entire group'sbeing there even after we have for each of the Eskimos anexplanation for why he or she is there: for instance, Eskimo1 won a trip to New York in a lottery; Eskimo 2 is on assignment from the Sludge Falls Courier, and so on. There couldbe a Mister Big operating behind the scenes to bring about

the several causes of these Eskim os be ing there ; he rigged thelottery, bribed the editor of the paper to send the reporterthere, and the like.

It would be a serious mistake, however, for the cosmolog-ical arguer to appeal to the common-cause principle in support of her demand that there be an explanation for thetotality above and beyond the several explanations of itsparts, to defend, a la Mark Lane and Cyril Wecht, the conspiracy theory of the universe - "Don't try to tell me thatthere isn't some capo de tutti de capi, a celestial CarlosM arcello, beh ind all thes e electron s, stars , an d ga laxies ." Thecommon-cause principle can be applied only to an improbablecoincidence, but the universe or totality, being a unique,one-shot affair, could not be said to be such since there couldbe no prior evidence concerning the frequency with whichsuch totalities came about. But the cosmological arguer neednot make the strong claim that it is probable that there is acommon cause of the universe as a whole, only that it ispossible. And this, in conjunction with the PSR, entails thatthere is such an explanation.

Another mistake that underlies the HE-based objection to3e an d 13, in add ition to the as su m ptio n that explan ation isagglomerative, is the belief that it is extensional, that is, thatthe principle of substitutivity salva veritate of core ferentialexpressions and coreporting sentences holds within the

blan k spaces of " expla ins ___ ." Plain ly, E dw ard s isassuming this when he says that "there is no group over andabove the five mem bers, an d if we ha ve explained w hy each

255




of the five members is in New York we have ipso factoexplained w hy the group is there ." Supp osedly, be cause thegroup is nothing "over and above the five members," that is,is identical with these members, any explanation of the latteripso facto is an explanation of the former.

Unfortunately for Ed wa rds, "explains" is a no nex tens iona lcontext. That the president of the Sludge Falls Bank waspublicly convicted of embezzlement, in conjunction with certain generalizations of folk psychology, explains why therewas a run on the Sludge Falls Bank, but not why the event

tha t led to the downfall of Sludge Falls occu rred at that tim e,even though the run on the Sludge Falls Bank is the eventthat led to the downfall of Sludge Falls. And that the onlyman in Sludge Falls with 712 hairs on his head was publiclyconvicted of embezzlement, along with the same generalizations of folk psychology, does not explain why there was therun on the bank, even though the president of the SludgeFalls Bank is the only man in Sludge Falls who has 712 hairson his head.

To make sense of the nonextensionality of explanation, aswell as for the fact that its explanans entails (in some cases)its explanandum, we must view explanation as a relation between abstract propositions, that is, we explain one fact, thatis a true proposition, by reference to other facts. This point isoften unperspicuously formulated in terms of the explanandum and explanans containing events or objects "under adescription."

In order to square with the nonextensionality of explanation, the concrete-aggregate versions of the cosmologicalargument must be recast so that what requires explanationis not some concrete aggregate, such as the infinite succession of dependent beings or the universe as a whole, butinstead the true proposition that reports the existence of thes econcreta.

But exactly what is the proposition that requires explana

tion in a concre te-aggrega te version of the cosmological arg ument? There are three different ways of construing theex pla na nd um , yielding three different v arian ts of each of th e

256




different versions of the cosmological argument: (1) that Aand B an d so on are con caten ated in the way they are; (2) thatthere exists A an d B and so on together; and (3) that thereexists A a nd B and so on tog ether an d that they are concatenated in the way they are." Whether or not the infinite succession of dependent beings or maximal spatiotemporalaggregate admits of the possibility of explanation by somedependent being or member of this spatiotemporal aggregaterespectively might depend upon which of these three interpretat ions we adopt .

If we adopt the "assemblying" interpretation (1), we areassimilating the demand to explain the infinite succession orthe unive rse to the earlier de m an d to explain w ha t caused thevarious pa rts of an autom obile to become concatenated in theway they are, which was explained in terms of w ha t wen t onin a Detroit assembly plant. Herein there is no demand toexplain the very existence of the different parts, only theirbeing con catena ted as they are. Supp osed ly, the explainer ofthe manner in which the inf ini tely many dependent beingsget successively concatenated cannot itself be a member ofthis succession, since then it would have to operate on itself.

But it is not obvious that this is impossible. I can be theassem bler of an orchestra of which I am a m em ber and evenconduct from the piano. There might, however, be a specialproblem with a member of the infinite succession of dependent beings that terminates in the present serving as theassembler of it, since this would require that this being

causally operate on, for instance, manipulate or control,de p en de n t objects that preced e it in the temporal succession.But causation cannot go backward from a later to an earliert ime.12 This response presupposes that t ime is open, that is ,that its generating relation is later than, and thus fails toad dr es s the possibility of a closed causal loop. Th e cosm ological arguer, in my opinion, should grant this limitation to herargument and add as an a posteriori premise that there isgood empirical evidence that the time of the actual world isopen. Herein she makes i t manifest that her argument doesno t hold in every possible world , w hich is hardly new s, since

257




it was obvious all along that the PSR does not hold in everypossible world. That the argument works in the actual worldshould be enough to satisfy the interests of theism.

Interp retation (2) of the explanatory d em an d esca pes theneed to make these messy qualifications, for it requires anexplanation for the fact that A and B and so on all existtoge ther. H erein it is ap pa ren t that the causal explainer of theexistence of each and every dependent being cannot itself bea de pe nd en t being, for then i t w ould hav e to be the proximatecause of its own existence; and this is impossible even if it ispossible that an object be a remote cause of its own existencein a closed time or Heinlein-type closed causal loop.

Interpretation (3) is a conjunction of (1) and (2), andthereby inherits both of their strengths, making a strongerexplanatory demand than does either alone, but escapes theproblem that confronted the assemblying interpretation. Anadequate explanation of the totality must explain both theexistence of each and every part of it as well as their comingto be con catena ted as they are. Even if an internalistic account

could be given for the concatenation, it could not be given forthe existence of the pa rts. T hu s, ev en if the cosmological failson interpretation (1), it could still work on interpretation (3),as well as (2). Plainly, the cosmological argument, whateverversion it takes, should go with interpretation (2) or, evenbetter, (3).

T H E G E N E R A L FA C T V E R S I O N

This version is ably defended by William Rowe in his important book, The Cosmological Argument. It has the same mainargument as Clarke's and also attempts an indirect proof ofits premise

3. It is false that every being is dependent,ba sed u po n the ass um ptio n tha t 3 is false. But, w he rea s

Clarke generated from this assumption, in conjunction with6. There exists at least one being,

258



The general fact version

the de m an d to explain the existence of the infinite suc ces sionof dependent beings, understood as a concrete aggregate,Rowe derives from them the quite different demand toexplain why there exists at least one dependent being. And,furthermore, both argue that on the assumption that everybeing is dependent, their explanatory demand cannot be metwithout some sort of vicious circularity, be it in the order ofcausation or explanation. Given that their explanatorydemands are so different, it could be that one of their viciouscircularity arguments works while the other does not.1 3 I

have given a som ew hat idealized recon struction of R ow e'sactual arg um en t; and , before I explicitly formu late it in th esame combined manner as I did Clarke's , I will show how itnaturally evolves out of a critical analysis of the text.

The point at which Rowe departs from Clarke's version isin his interpretation of the infinite succession of dependentbeings for which the subsidiary argument demands an explanation. Whereas Clarke took it to be a concrete ag gregate ,having the same spatiotemporal location as i ts parts, Roweconceives of it as an abstract set . '4 Rowe is well aware that itis conc eptually im possible for there to be a causal ex pla na tionfor the existence of an abstract set. To avoid this absurdity heinterprets the question "W hy does set X exist?" as "W hy do esset X hav e the mem bers it does rathe r than no ne at all?"Notice that it is possible to explain the existence of everym em ber of X witho ut thereby exp laining w hy X has at leastone member rather than none at all. This would refute an HE

principle that applies to abstract sets, which, not surprisingly, is how Rowe understands this principle.Th ere is an ambigu ity in Rowe's acc oun t of w ha t is involv ed

in explaining the existence of a set. When we are asked toexplain why some individual a is a member of set X are werequired to explain why a exists or instead w hy a ha s theset-defin ing prop erty? If w e are to explain , for ex am ple , w h ya is a m em be r of the set of con ting ent b ein gs, are we req uir edto explain why a exists or why a is a contingent being, forwhich the explanation might be that a is a horse and horsesare essentially contingent beings? Rowe does not explicitly

259




resolve this ambiguity, but his discussion for the most partfavors the former interpretation, and this is how he will be

un de rsto od . T hu s, to explain the existence of set X is toexplain w hy th ere exist the various m em bers of X rathe r tha nnone at all.

Whereas, for the most part, Rowe construes the demand toexplain the existence of X in this manner (see pp. 138, 139,144, 145, 154, 159), there are places where he demands moreof an explanation of a set. At two places he requires that anexplanation of the existence of the set of dependent beings,A, tell us why "A has the members it has rather than someother members or none at all." The addition of "rather thansom e other m em be rs" will raise the hackles of those p hilosoph er s wh o recoil from n egative ev ents in horror because theythink we lack adequate criteria of identity for unrealized possibilities. That sinister character, Smokey the Bear, the veryon e w ho ate a Boy Scout alive just to get his hat, w as fond ofsnortin g, "O nly you can prev ent forest fires " I gue ss on eaccomplishes this feat by not dropping matches in a dry

forest and the like. But ho w m any times did you not do the sesorts of things yesterday? How many forest fires failed tocome into being as a result? Likewise for unrealized possibledependent beings. How many times did you not copulateyesterday? And how many possible dependent beings didyou deny existence to as a consequence (and remember toinclude possible twin s, an d so on)? In Chap ter 2 of my Negation and Non-Being, I argue that we do have adequate exten-sional criteria of identity for unrealized possible individuals,but it is too long a story to be told here. And, moreover, it isnot necessary to do so, since Rowe drops this extra explanatory demand when he presents his argument.

When Rowe formulates his argument against the possibility of giving an internalistic explanation for the existence ofset A, he switches to yet another set of explanatory demandsthat require us to explain why A has now and always hadmembers rather than never having had any members at al l .

H is arg um en t is pre sen ted first in terms of the exactly p arallel case of attempting to explain the existence of the infinite

260




set M of men in terms of the causal efficacy of the membersof M. H e con stru es "W hy do es M exist?" as "W hy is it that Mhas now and always had members rather than never havinghad any members at all?" (155). His argument against thepossibility of explaining the latter solely in terms of the causalefficacy of men is this:

Surely we have not learned the answer to this question whenwe have learned that there always have been members of Mand that each member's existence is explained by the causal

efficacy of some other member. (155)Rowe does not tell us the name of the malady that afflictssuch an internalistic explanation of M, but I take it that it issupposed to be viciously circular. Neither does he tell usexactly why the proffered explanation of M is viciously circular, probably because he thought i t too obvious to requirefurther explanation and justification. It will turn out that farfrom being obvious, it is highly dubious that a noncircularexplanation cannot be given; but before we turn to this task,it will be help ful to pin po int the exact pa rt of Row e's explanatory demands that supposedly cannot be given a noncircularinternalistic explanation.

According to his least demanding account of what it is toexplain the existence of the set of dependent beings, A, it isrequired that we explain each of the following two facts:

C. A has the members i t does; andD. A has any members at all, that is, that there exists at

least one member of A.

According to the more demanding account that figures in hisvicious circularity argument, it is required that we explaineach of these three facts:

E. There now exists a member of A;F. There always have existed members of A; and

D. There exists at least one dependent being.It will be shown that Rowe's argument against the possibility of giving a noncircular explanation of A's existence really

261




concerns or ought to concern fact D alone, and thus it doesnot matter whether he operates with the weak or strong setof explanatory demands.

Noncircular internalistic explanations can be given for C,E, an d F. Such an exp lana tion for C con sis ts in the infinitelymany explanations for each and every dependent being interms of the causal efficacy of some other dependent being.W hen Rowe says that the question "W h y is i t that A has nowand always had members rather than never having had anymem bers at all?" is not adequ ately a ns w er ed by the explanation "that there always have been members of A and thateach mem ber 's existence is exp lained by th e causal efficacy ofsome other member," he is not clearly locating the exactsource oi \h e dii i icutty, \h a\ bein g \o explain D. He -wrongtymakes it appear that noncircular internalistic explanationscannot be given for E and F.

That this can be done is shown by the following noncircular internalistic explanations for F and E. For F,

17. If at any time t there exis ts a de p en de n t being d, thenat some earlier time t' there exis ts another dependent beingd' who causes the existence of d;

18. There no w exists a de pe nd en t being d, that is , a member of A; therefore,

F. There always have existed merfibers of A.

For E,

19. If at any time t there exists a dependent being d, thenat some later time t' there exis ts another dependent being d'whose existence is caused by d;

20. There exists a dependent being d at some time t that isearlier than now; therefore,

E. There now exists a member of A.

The propositions that occur in the explanans of these twoexplanations are such that no one of them alone entails theexplanandum, thereby avoiding any charge of vicious circularity.

262




It would seem that it is D alone that is recalcitrant to inter-nalistic explan ation. This holds w he the r or not A has a finiteor infinite number of members. Imagine that A has a finitenu m be r of m em be rs d u e to there being a circle of cause s in them an ne r p reviou sly env ision ed. Th ere will still be a D-typ e factab ou t the set of dep en de n t bein gs to the effect that it contain sat least one existent member. And it is this fact that supposedly cannot be given a noncircular internalistic explanation.

Given that it is only explanatory demand D that cannot bem et noncircularly if all bein gs are de pe n de n t, w e can give the

following formulation of Rowe's "true" argument, in whichthe main and subsidiary arguments are combined in the wayin which we already did for Clarke.

The combined general fact version

6. There exists at least one being [existential-fact premise];i . Every individual is either a de pe nd en t or inde pe nd en t

being [PSR,];2. Either every being is dependent or there exists an inde

pendent being [from i by logical equivalence];7. Every being is a de pe nd en t being [assum ption for indi

rect proof];D. Th ere exists at least on e de p en de nt being [from 6 and 7];21. There is an explanation for the fact that there exists at

least one dependent being [from 1 and D by universal instantiation];

22. The explanation for the fact that there exists at leaston e de p e n d en t bein g is in term s of the causal efficacy of eithera dep en de nt or an ind ep en de nt being [from 1 and 21];

23. The explanation for the fact that there exists at leastone dependent being is not in terms of the causal efficacy ofan independent being [from 7];

24. The explanation for the fact that there exists at least

on e d e p e n d e n t being cann ot be in term s of the causal efficacyof a dependent being [premise];

263






the controversial premise in this combined argument is

24. The explanation for the fact that there exists at leastone d ep en de nt being canno t be in term s of the causal efficacyof a dependent being.

We have already considered Clarke's vicious circularityobjection to the possibility of explaining the existence of theinfinite succession of de pe nd en t bein gs in term s of the ca usalefficacy of a dependent being and found it to have somemerit, but it is not obvious that there is any vicious circularity

in explaining the fact that there exists at least one d e p en d en tbeing in this way. Rowe's sketchily presented vicious circularity argument against this possibility needs further probing. Two objections against this argument will be considered:First, it rests on an unacceptable set of explanatory requirements; and, second, even in terms of these requirements, anoncircular explanation can be given for fact D.

While Rowe never explicitly states the requirements for anexplanation, he seems to require of an explanation that itsexplan ans (i) entails the expla nan du m and (ii) has no pr op osition that alone does. Requirement (i) is too strong, becauseit precludes inductive-statistical explanations; but this willnot concern us. It is (ii) that is at the core of Rowe's viciouscircularity argument, but it also is too strong, facing counterexamples in mathematical and personal explanations.Each step in a mathematical proof of a theorem alone entailsit, since the other steps in the deduction are all necessary;

but the explanation is not vitiated on this count. A perfectlygood personal explanation for why my arm went up is that Iraised it, or for wh y the re exists at least on e de pe n de nt bein gthat God willed that there be. Herein we have an explananscomposed of a single proposition that alone entails the explanandum. These are decisive counterexamples to the Hu-mean thesis that in an adequate causal explanation, theproposition reporting the cause must not entail the proposition rep ortin g the effect. R equ irem ent (ii) w ou ld ap pe arto rest on a hasty generalization from the case of Hempeliandeductive-nomological type explanations in which it is required

265




that no statement of a particular fact alone entail the expla-nandum, since it would render the covering law(s) otiose in

the deduct ion.1 6

\It might be retorted that my explanation of my arm's rising

is elliptical, for I endeavored (tried, and so on) to raise myarm; and there are causal laws that say roughly that whenever someone endeavors in this manner to raise their arm insuch circum stances, their arm goes up. Herein the statem entof the cause does not entail the statement of the effect. Thisfull-blown explanation imputes to me the belief that myintentio nal actio ns are causally deter m ined ; but I do n ot hav ethis belief. Furthermore, this sort of covering-law explanation could have no application to the explanation of D interms of God's will, since the connection between God's willand its effects is certified by the conceptual truth that he isomnipotent rather than by some contingent causal law connecting God's endeavorings with their worldly effects.

Even if requirements (i) and (ii) are accepted, counterexamples can be given to Rowe's claim that it is impossible togive a non circular formal explan ation of D in term s of de pe ndent beings. The following is an example of just such anexplanation:

27. Fred exists;28. Fred is a dependent being; therefore,D. There exists at least one dependent being.

It m igh t be cou ntered that 27 and 28 each alon e entails D,

since each is expressed by a sentence containing a propername of a dependent being. But by this reasoning, it alsowould follow that each alone entails that there exists at leaston e fireman , if Fred shou ld ha pp en to be a fireman Being adependent being (as well as being a fireman) need not beincluded in the sense of the prop er nam e "Fred." A nd this isho w the nam e "F red" is to be understo od in my above explanation. Some might find a special problem with 28, sincethey think that it has existential import and thereby entailsthat Fred exists, with the result that it alone entails D. Butther e is no rea son wh y 28 m us t be interp reted in this w ay.

266




Certainly we can predicate a property of an individual without committing ourselves to its existence, as in "Santa Clausis a man." This is how 28 is to be read. Maybe it would beclearer if it were replaced by

28' . For any x, if x is identical with F red, th en x is a dep endent being.

Even if my explanation is not formally circular, it might befelt to fail as an explanation on the ground that it does notgive any cause for D. This can be rectified as follows:

29. Fred endeavored to create a dependent being;30. If Fred were to endeavor to create a dependent being,

he would succeed; therefore,D. There exists at least one dependent being,

in which the pro per na m e "Fred" again den otes a dep end entbeing, thereby explaining D in terms of the causal efficacy ofa dependent being.

It might be objected that 29 violates (ii) because it aloneentails D. If Fred e nd ea vo rs to do X, he b rings ab ou t or agen t-determines his act of endeavoring, from which it follows thatthere exists at least on e de pe nd en t bein g, his act of en dea voring. This objection is not decisive because of the murkynature of agent determination.

Even if my two "explanations" of D, and for sure the firstone, do not fail on formal grounds since they satisfy (i) and(ii), they still might fail on pragmatic grounds. The pragmatic

aspect of explanation has to do with the basic fact that anexplanation is an attempt to answer some person's why-question and thereby must be geared to their knowledge,beliefs, skills, interests, and purposes. It is this pragmaticaspect of explanation that accounts for why an explanatorycontext is no ne xte nsio na l; as well as the fact tha t explanationis not closed under deduction, that is, an explanation of pdoes not explain every proposition that p entails, the reasonbeing that the person to whom the explanation is directedmight be ignorant of the entailment, just as he might beignorant of the fact that X = Y and thus be satisfied by an

267




explanation of w hy X is F but not find it adeq uate as an explanation of why Y is F. Putnam's account of why the physicist's description of the microstructure of a round peg and asquare hole on a board is not an explanation of the fact thatthe peg do esn 't go in the hole rests on suc h p ragm atic cons iderations concerning the knowledge and skill of the person towhom the explanation is directed. '7 Most people are notknowledgeable and quick enoug h to derive the m acrodescrip-tion of the peg and hole from the microdescription, eventhough it is entailed by the latter.

Granted that there are these pragmatic requirements for asuccessful explanation, in exactly what way are my twoexp lanations of D pragm atically deficient?18 Let us say that anexplanation is pragmatically circular if the p er so n to w ho m it isaddressed is in an epistemic situation that precludes hisbeing able to know some proposition in the explanans without know ing the explan and um . I imagine that all ordinaryhuman beings are in epistemological circumstances thatmake it impossible for us to know one or more of the propositions in the analy san s of my tw o explan ations of D w itho utknowing D.

In respect to the second explan ation, it is very du bio us thatwe could know that

30. If Fred were to endeavor to create a dependent being,he would succeed

without knowing that some dependent being exists and

thereby knowing D. If 30 resembles an F-conditional in having a con sequ ent tha t is not deter m ined by its ante ced ent, wecannot know it since we are not possessed of some mysterious faculty of "middle knowledge." We can know a subjunctive like 30 only if the re is som e nom ic conn ectio n b etw ee n itsantecedent and conclusion and we have observed past casesof Fred's successfully creating dependent beings, therebyknowing of certain individual beings, namely, the ones created by Fred in the past, that they are de pe nd en t being s. A ndsince we know certain singular propositions that predicatebeing a dependent being of some individual, we know D as

268



The general pet version

well, since D is entailed by such a singular proposition invirtue of the principle of existential generalization. We mustassume that the person to whom the second, as well as thefirst, explanation is directed reasons in accordance with thisprinciple; otherwise the explanation could not work for himsince it m ust ap peal to this principle in de du cing its e xp lana n-dum from its explanans.

The first explanation also seems to suffer from pragmaticcircularity. It is hard to imagine how beings in our epistemiccircumstances could know

28. Fred is a dependent being

witho ut know ing D. The problem gets do w n to w he the r w ecould know 28 withou t know ing that Fred exists. O ne w aywe could get to know 28 is by observing Fred's coming intoexistence through the causal efficacy of some other being,but this involves our knowing that Fred exists. Another waywe could learn that 28 is true is by finding out that Fred is afictional or mythical being who is described in some story asa dependent being; but, again, this involves knowing thatsome being is dependent, since we know that the story itselfis such a dependent being, since being created by itsauthor(s). Althoug h I am sure that my two explanations arepragm atically deficient, I am no t confident that I hav e correctly brought out the manner in which they are, and I mustleave it to my philosophical sup eriors to do a bette r job th anI have been able to do.

So far we have tried with dubious success to underminethe combined general fact version of the cosmological argument by finding an internalistic explanation for D in terms ofa dependent being. But a more effective way of criticizing theargument is to show that

22. The explanation for the fact that there exists at leastone dependent being is in terms of the causal efficacy ofei ther a dependent or an independent being

does not, as claimed by this argument, follow from the conjunction of

269




1. Every individual is either a dependent or independentbeing; and

21. There is an explanation for the fact that there exists atleast one dependent being.

The strategy is to find some possible explanation for the factD that does not invoke any dep end ent being or causally efficacious ind ep en de nt bei ng . Let us call such an explana tion of Dan "atheistic explanation." We are permitting an atheisticexplanation to make use of independent beings provided

that they are not capable of being causal explainers, ex am ple s-of which are numbers, properties, and so on. Plainly, 22 isincompatible with the possibility of such an atheistic explanation or D.

Rowe shows that he is aware of this strategy for attacking22, for he considers the possibility of giving some kind ofatheistic explanation for why D is necessary. Such an explanation would also explain why D is true since we philosophers - the people to whom the explanation is directed - areaware that a necessary proposition is true. After correctlyfaulting Clarke for inferring from the proposition that everydependent being has the possibility of not existing that itis possible that no dependent being exist, Rowe presentsan arg um ent of his ow n for the non necess ity of D based up o nGod's omnipotence that avoids Clarke's de re-de dictoconfusion.

Th e prem ises of Row e's argu m ent are that it is possible that

God exists an d tha t it is necessa ry th at if Go d exists, G od canbrin g it abo ut that n o de pe nd en t being s exist (166). W hile h isargument is valid, its second premise is question-begging,provided that God's omnipotence is restricted, as it ought tobe, to that which it is logically consistent for God to bringabou t. If it shou ld tu rn o ut to be necessary th at there exists atleast one dependent being, then not even an omnipotentbeing could bring it about that none exist. Two argumentswill now be presented for D being necessary. The premisesfrom which the necessity of D is deduced must each be necessary, and there are grounds for doubting that they are.

27




For instance, each argument contains the PSR as a premise,and the PSR hardly seems necessary. But even if each premise is only possible, prov ided that the prem ises are com possi-ble and entail D, it suffices to establish that, pace Rowe's 22,it is possible to have an atheistic explanation for D. Thus, anatheistic explanation need not worry about whether the PSRis necessary, only that it is true in some possible world, andcertainly we'll grant at least this much.

The necessity of time argument

31. It is neces sary that time exists [prem ise];32. It is necessary that if time exists there exist empirical

beings [premise];33. It is necessary that there exists at least one empirical

being [from 31 and 32];34. It is necessary that every being is either an indepen

dent or dependent being [PSR,];35. It is necessary that no empirical being is an independent being [premise];

36. It is necessary that every empirical being is a dependent being [from 34 and 35]; and

37. It is necessary that there exists at least one dependentbeing [from 33 and 36].

Serious challeng es can be m ade to both 31 an d 32, as well

as 34, but the latter issue will not concern us for the previously stated reason. Two sorts of objections might be raisedagainst 31. W hile all bu t so m e mystics an d mystically inclinedmetaphysicians would grant that t ime is real, many woulddeny that it is necessary. Arguments for the latter can befound in Kant's Transcendental Aesthetic. I have developed avery different argument for its necessity in my Negation andNon-Being based upon the need for a spatioremporal receptacle of a topologically and metrically amorphous nature toserve as the ground for the possibility of forms being instantiated and, moreover, being multiply instantiated, space and

271




time herein serving as the ultimate ground of individuationof qualitatively similar empirical particulars. T he basic u nd er

lying metaphysical intuition is that possibility is grounded inactuality. It is conceptually necessary that the Platonic formsexist. They specify possibilities for existence, and it is necessary that there are these possibilities. But such possibilitiesare possibilities for the instantiation of a form, and this requires a realm in which such instantiations could take place.And what could this be but space-time. There are many whoare not convinced by these arguments, but the only important issue is w he th er it is at least possible tha t time exists, a ndabout this there should be no argument, unless it can bepresented very quick ly, tha t is, in n o time at all.

It might be objected that 31 refers to an independent being - time - since there is an ontological argument of sorts,such as my above Timaeus -inspired ar gu m en t for its existence. But an atheistic explanation of D is allowed to invokean independent being, provided it is not capable of being acausal agent. And remember, neither time can wither norcustom stale If 31 is recast, as som e think it mu st be , so th atit is only contingently true that time exists, the explanationinvokes a being that is neither an inde pe nd en t no r a d ep endent being, since time itself cannot cause anything. While itmight be reaso nab le to define an atheistic exp lana tion of D sothat it is perm itted to use individuals that are neither de pe ndent nor independent beings, there would be a clash with34-PSR,. This requires that the necessity of time employ amore restricted version of PSR„ but I do not think that it is

very difficult to find a version that will do the trick.The most seriou s difficulty con cern s 32's att em pt to gr ou nd

an ontological claim on empirical considerations. Since it isno t possible for us to identify tim es except by reference to th eempirical particulars that exist in time, time itself could notexist without such particulars. Even though 32 has beenaccepted by such luminaries as Plato, Aristotle, and St.A ugu stine, to men tion just a few, I find it du bio us. A nd if 32isn't acceptable, neither are its weaker cousins that claimonly that it is contingently true or possibly true, for the only

272



1%e general fact version

reason we could have for accepting these weaker claims is

that it is a necessary conceptual truth that time could not existwithout empirical particulars.

The argument from negation

38. It is necessary that every being is either an independent or dependent being [PSR,];

39. It is necessa ry that every con tingen t being is a de pe ndent being [from 38];40. It is nece ssary that for every contin ge nt being x, if it is

a fact that x exists, there is an explanation of this fact, and ifx does not exist, there is an explanation of this negative fact[a version of the PSR];

41. It is necessary that an explanation for the fact that acontingent being does not exist be in terms of some existent

contingent being(s) whose properties logically preclude itsexistence [premise];42. It is nece ssary tha t som e con tingent beings exist [from

41]; and43. It is necessary that at least one dependent being exist

[from 39 and 42].

W hat this argu m ent am oun ts to is an ontological ar gu m en tfor the existence of a world of contingent or dependent be

ing s. It is borro w ed from Bergson's accoun t of noth ing ne ss inCreative Evolution and was advanced by him for the expressedpu rp os e of precluding the need to invoke necessary, t imelessabstracta - Platonic forms - in explaining why there is something rather than nothing. '9 Again, w e nee d not w orry abou tthe modal status of the different versions of the PSR thatappear in 38 and 40, only that they are at least possible. Furthermore, there is no question about this being an atheistic

explanation of D, since it invokes no independent beings atall, not even causally inefficacious ones.Prem ise 41 is the vulnerable spot in the arg um ent, an d

objections to it cannot be sidestepped by downgrading it to a

273






into existence. It might be retorted by the theist that thepositive fact that explains the failure of hippogriffs to existcould be that God willed that they not exist. How rationallysatisfying w e find such an explanation will de pe nd u p o n ourass essm en t of the overall ade qua cy of theistic ex pla na tion s ofworldly facts. And about this philosophers disagree.

The ontological disproof argument

This argument attempts to prove only that i t is possible thatthere is an atheistic exp lanation for D, not that there nec essa rily is; but this suffices to refute the combined general factversion's attempt to deduce from the assum ption that eve rybeing is dependent that there is no explanation for D. It goesas follows:

i ' . Th ere is an explan ation for every fact [PSRJ;D. There exists at least one dependent being [assumption

for conditional proof];21. There is an explanation for the fact that there exists

at least one d ep en de nt be ing [from 1 an d D by univ ersa linstantiation];

44. The explanation for the fact that there exists at leastone de pe nd en t being is either an atheistic on e or in ter m s ofthe causal efficacy of an independent being [premise];

45. It is logically impossible that there exists an independent being capable of causally explaining the fact that thereexists at least one dependent being [premise];

46. The explanation for the fact that there exists at leastone d e p en d en t being cann ot be in term s of the causal efficacyof an independent being [from 45];

47. The explanation for the fact that there exists at leastone dependent being is an atheistic one [from 44 and 46 bydisjunctive syllogism];

48. If there exists at least one dependent being, then the

explanation for the fact that there exists at least one depende nt bein g is an atheistic one [from D, 21, an d 4 4- 47 byconditional proof];

275







Cosmalogical arguments

from it by dropping out all of its conjuncts that are eithernecessary or entail tha t God exists. Th ere are two reasons fordirecting the explanatory demand at this depleted conjunction. First, it avoids the horrendous problem of what constitutes an explanation of a necessary proposition, for instance,some fact about linguistic use or an essential property of aPlatonic entity and the like. Second, it escapes the possibilityof giving an internalistic explanation in terms of the conjuncts themselves. If God were to exist in W, B would containsuch G od-en tailing prop ositio ns as that G od has willed that all

of the propositions in B' be true together, thereby explainingthe conjunction formed from the propositions in B'. And thisinternalistic explanation of the B '-de rived conjunction, inconjunction with whatever internalistic explanation might beoffered for the necessary propositions in B (e.g., the Platonicor linguistic account), might be thought together to constitut e an intern alistic explan ation of th e big conjunction formedfrom B, thereby pulling the rug from under the explanatoryargument that attempts to show that it is impossible for thereto be an internalistic explanation of this big conjunction.

The question is whether the explanatory argument succeeds when directed at showing the impossibility of givingan internalistic explanation for the B '-derived conjunction.Let us imagine that this conjunction contains the contingentpropositions p an d q and r, some of which are atomic andoth ers lawlike. It w ou ld ap pe ar tha t it is impossible to explainthe conjunction (p and q a nd r) in term s of its conjuncts. Evenif each and ever y m em be r of this conjun ction had a n internalistic explanation, for instance, p is explained in terms of q, qin terms of r, and r in te rm s of p, this would not constitute anexplanation for the fact that they are all true together, due tothe nonagglomerative nature of explanation. Even thoughthe several internalistic explanations of p, q, and r do notcon stitute an expla natio n of their conjunc tion, might it not bethe case that some member of this conjunction explains the

entire conjunction? I don't see how this is possible. Supposeit w ere co njunct p that performed this explanatory task; con-

278



The world fadfc version

junct p would have to explain, among other things, each andevery conjunct and thereby would have to explain itself. Butthis it cannot do, since it is only a contingent proposition.Herein we face what seems to be an irresolvably vicious circularity of explanation. My imagination is quite limited, andI might have overlooked some foxy way of giving an internal-istic explanation of B'; however, our prima facie intuitionsare against this being possible, and thereby I place the onuson tho se w h o think it can be do ne . I await my next letter fromPhil Q ui nn tell ing me that my intuitions are not sha red by th e

maintenance people at Notre Dame on this issue.All that remains is the PSR component of the world bookver sion , an d it is obv ious w ha t form it will take, nam ely, tha tof PSR2 requiring that every fact, including truth-fu nction alones, have an explanation. Thus, in the name of PSR2, theremust be an explanation for a B '-derived conjunction. A ndsince the explanatory argument has already established thatit cannot be an internalistic one, it must be in terms of thecausal efficacy of som e ind ep en de nt being w ho is very m uchlike God.

Some general objections

Our critical evaluations of the different versions of the cos-mological arg um en t have mostly been confined to an assessment of their relative strengths and weakness, with thew orld bo ok version em erg ing as the mos t attractive of the lot;but there are several objections that apply to all of them thathave not yet been considered. And we now turn to this finaltask so as to reach an absolute evaluation of them.

Denial of the PSR. This is the obvious wa y of cu tting off at thepass any version of the cosmological argument: refuse togrant, even as a contingent truth, the version of the PSR thatit requires - PSR2 in the case of the world book and general

existential fact versions and PSR, in that of the concrete-aggregate version.

279



Cosmological argummts

The cosmological arguer's response to this objectionshould be to recast her argument so that its conclusion nolonger is that t he re is in fact a theistic-type ex planation for theexistential fact in qu es tion in t erm s of the causal efficacy of anindependent being but that this is the only possible explanation of this fact. If there is to be an explanation for thisexistential fact, it must be the theistic one. If successful, thisargument would have a very important and startling result -tha t only God ca n satisfy ou r rational longing to have a reallyintellectually satisfying e xpla nation of the univ ers e. All along

this has been the real goal of the cosmological argument.Furthermore, if this conditional version of the cosmologicalargument works, it can be urged that we ought to accept thetheistic explanation by the principle that anything is betterthan nothing or, even better, by the currently fashionableprinciple of inference to the best explanation. The critic willrespond that these principles do not hold when the onlyavailable or best explanation is a bad one.

Regressiveness of a theistic explanation. It will be recalled thatthe aim of the cosmological argument was to shut up precocious brats by putting an end to their regress of why-questions. The point of this objection is that the cosmologicalargument fails to do this because it must employ in its explanation of the existential fact in question some contingentproposition, such as that God or the independent beingwilled that this be a fact; and, since this proposition is not a

self-explainer, it is a legitimate subject of a why-question."Why did God will that way?" The Spinozisitic theist canrespond that this proposition is necessary, because God wasnot free to make any creative choice but this one. But fewtheists find this w ay o ut of the regre ss of explanation s attractive, since, in addition to denying free will to God, it entailsthat there is only one possible world, given that it is necessary that God exists.

Fortunately, there is a less drastic way for the theist toblock the regress of why-questions - argue along Libertarian lines that a free act is not a legitimate subject of an

28



The world book version

explanatory demand. According to the intuit ions of the Libertarian, a free choice canno t ha ve a sufficient eve nt-ca useand thereby cannot be causally explained, though the agentcan make her choice intelligible by giving reasons for it.Thus, the proper response to the child 's request to explainw hat caused Go d's creative choice is, "Sh ut up It w as do nefreely." Whereas the fact that God exists supposedly can beexplained by deducing his existence from his essence, God'screative decision admits of no explanation whatever andthereby is a suitable place at which to stop the regress of

explanations. The sort of theistic explanation that a cosmo-logical argument offers for its initial existential fact could becalled, in contradistinction to a scientific explanation, a "finalexplanation," in that each proposition in i ts explanans iseither self-explained via some ontological argument orrequires no explanation on conceptual grounds.

The ontological disproof. Each of the above three cosmologicalarguments had a different existential fact for which itdemanded an explanation, but in each case the conclusionwas drawn that there exists a necessary being that is thecause of the fact in question - that th e univ ers e or the infinitesuccession of dependent beings exists, that D is true, andthat the B '-derived conjunction is true. Several arg um en tswill no w be given for their con clusio ns be ing im possib le. T heunderlying strategy in each of these arguments is to showthat the necessary being referred to in their conclusion is an

unsurpassably great being - a being that is both maximallyexcellent, that is, essentially (omnipotent, omniscient, omni-benevolent, providential , etc.) , and necessarily existent. Appeal will then be m ade to the Ch ap ter 6 arg um en t for it beinglogically impossible that there exist an unsurpassably greatbeing based on there being properties, such as being a morally unjustified evil or a world in which all free personsalways freely go wrong, that are both strongly incompatiblewith having unsurpassable greatness (i .e. , if either has thepossibility of instantiation the other does not) and intuitivelym ore likely can did ates for the possibility of instantiatio n th an

281



Cosmokgical arguments

is having unsurpassable greatness. My arguments will bedirected against the concrete-aggregate version, but theycan be deployed with equal effectiveness against the otherversions.

Th e first arg um en t is an ad ho m ine m on e, since it do es n otde m on strate that it is absolutely impossible that th ere exist anecessary being who determines the existence of the universe, and the like, only that it is impossible relative to thepropositions that the cosmological arguer is committed toaccepting, nam ely, that this necessa ry being is the G od of tra

ditional theism. It was Clarke himself who went on to arguein the second stage of his cosmological arg um en t in the Demonstration that this necessary be ing h as all of th e essen tialom nipro perties of the God of traditional W estern the ism , thatis, maximal excellence.

The argument begins with the conclusion of the concrete-aggregate argument as an assumption for indirect proof:

51. There exists a necessary being, N, w ho d eterm ines thatthe universe or the infinite succession of dependent beingsexists [assumption for indirect proof];

52. N is a maximally excellent being [g ranted , ev en arg ue dfor, by cosmological arguer];

53. N has necessary existence [from 51];54. N is an unsurpassably great being [from 52 and 53 by

definition of unsurpassable greatness];55. It is conceptually impossible that there exists an un su r

passably great being [proven by argument in Chapter 6];56. N doe s not exist [from 54 an d 55 by u niv ersa l in stan tiation];

57. N exists [from 51];58. N both do es and do es no t exist [from 56 an d 57 by co n

junction]; and59. It is not the case that th ere exists a necessa ry b eing , N ,

that determines that the universe and the like exists [from 51to 58 by indirect proof].

There is another path that leads from

282



The rmrtd book version

51. The re exists a necessary being, N, wh o determines thatthe universe, and the like, exists

to

54. N is an unsurpassably great being

that appeals to an assumption that is more plausible in itsown right than is

52. N is a maximally excellent being.

It holds that N is a universe determiner not just in the actual

world but in any world in which N exists:60. In every world in which N exists, N determines that

the universe exists [premise].

Basically, this is ap ply ing the Ch ap ter 4 principle of so vereigntype-invariance to N." This seem s a reason able extensio n ofthis principle, for any necessarily existent being that is auniverse determiner in one world should perform this function in every world. It would be extremely anomalous toallow it to be a universe determiner in one world but not inanother in which i t exists, prompting unanswerable questions as to what caused it to lose its job, how it could lose itsjob given that i t has such w on dero us p roperties as being necessarily existent, a nd so on. The next step in the derivation of54 from 51 is the quite plausible assumption, obviouslyacceptable to the cosmological arguer, as well as Anselm'sFool opponent, that it is at least possible that there exists a

maximally excellent being capable of completely and solelyde term inin g tha t the u niverse exists. Since a bein g is capableof doing something only if it does it in some possible world,this entails

61. There is a possible world, W, in which there exists amaximally excellent being, £, who completely and solelydetermines that the universe exists [granted by both Fool andcosmological arguer].

Since N is necessarily existent, it follows that

283



Cosm&bgical arguments

62. N exists in W [from 51].

But from 62 in conjunction with 60 it follows that63. N determines in W that the universe exists.

From 61 and 63 it follows that

64. N is numerically one and the same being as E in W,

the reason being that, since £ is the complete and sole universe determ iner in W, any be ing w ho is a unive rse determiner in W is identical w ith E. Since N is identical w ith E,a maximally excellent being, in virtue of Leibniz's Law, itfollows that

65. N is maximally excellent in W. '

Because maximal excellence is an essential property and Nhas necessary existence, it follows that

66. In the actual world, N is both maximally excellent andnecessarily existent,

and thus

54. N is (in the actual wo rld) an un sur pa ssa bly great being[by definition of unsurpassable greatness].

And from here the arg um ent p roceed s to ded uce a contradiction by appeal to the argument in Chapter 6.

My two arguments against the possibility of 51 constituteontological disp roo fs of the existence of the very so rt of beingw ho se existence is asser ted in the con clusion of every v ersionof the cosmological argument, thereby showing that thesearguments are radically defective. These ontological disproofs, however, do not pinpoint the defective spot in thesearguments. Hopefully, some guidance for accomplishing this .task has been supplied by the critical analyses of these arguments that have been given in this chapter.

284



Chapter 8Religious-experience arguments

Throughout history, in almost every society, persons haveclaimed to experience God or some ultimate reality in thesame sort of immediate way that they experience physicalobjects through their senses. Their experiences have beenquite diverse, running the gamut from extrovertive andintrovertive mystical experiences of some underlying unityinto which the exp erient is totally or partially inc orp orate d tonuminous experiences and direct experiences of God's presence as a very pow erful, loving, no nh um an person w h osh ow s concern for th e we ll-being of the exp erie nt. Differentdescriptions are given of the apparent object of their experiences - God, one's personal purusha, Brahman, the eternalone, the undifferentiated unity , and the like. In spite of th esedifferent characterizations, the subjects agree in taking theobject of their experience to be the ultimate reality, the really

real, and, most important, are convinced that man's greatestpossible good consists in achieving a direct, expe riential co ntact or union w ith this reality. For the se re as on s, I shall call allthese experiences "religious experiences," even though theyare not taken to be such by everyone who has them.

I shall duck the thorny issue of whether there is a commonphenomenological content to all or most of these experiences, at least the unitive ones, that supposedly gets interpreted differently by mystics in diverse cultures, as well asthe larger problem of whether the phenomenological-interpretative and observation-theoretical language distinctions are viable. While I am in basic agreement with the

285





Religious-experience arguments

epistemic justification for believing that there is such anobjective reality, they will serve as the basis for a good argument for the existence of God or this highest reality:

A. Religious experiences occur;B. Religious experiences constitute evidence for the exis

tence of their apparent object, God and the like; therefore,C. There is evidence that God and the like exist.

To have this sort of evidential status, these experiencesmust not only be veridical but cognitive as well. A type of

experience could be veridical without being cognitive. Itcould qualify as veridical in virtue of the apparent object ofthe experience cau sing the experience in the "right w ay ," bu tno t coun t as cogn itive since it does n ot constitute evide nce orwarrant for believing that this object exists and is as itappears to be in the experience. Even if we were to have agood argument for the veridicality of religious experiences,for instance, based on a successful ontological or cosmologi-cal arg um en t an d the analytic truth tha t God wo uld no t allowus to be duped on a large scale by experiences that seemed tobe veridical perceptions of him, it still would not show theseexperiences to be cognitive, since they alone wou ld n ot constitute a basis for our gaining knowledge of some objectivereality. A cognitive faculty should stand on its own two epis-temological feet. It will be argued that religious experiences,although possibly veridical, could not be cognitive. Even if itwere possible that their apparent object exist and be the right

sort of cause of the experience, we could never know on thebasis of thes e ex per ienc es either that this object exists or thatthe ex perien ce is cau sed in the "right wa y" by it. I shall go onto argue that a religious experience also could not qualify asa veridical perception of an objective reality, even if its apparent object were to exist and be the cause of the experience.

If a type of exp erien ce £ is cognitive, the n the o ccur renc e ofan E-type experience counts as evidence for the existence

of its intentional accusative, even for those who have not had thisexperience or even any E-type experiences. It was for this reasonthat prop osition s B an d C in the former arg um en t did not

287




relativize their claims about the evidential status of religiousexperiences to the people who have them. That a cognitiveexpe rience's evide ntial statu s is not pe rson-re lative is a fundam ental conceptual requ irem ent for an experience hav ing cognitive statu s, a po int that is often m issed by those w ho defendthe cognitivity of religious experiences. William James, in hisassessment of the cognitivity of mystical experiences in theconcluding section of The Varieties of Religious Experience, violated this epistemic counterpart to the moral principle ofuniversalizabili ty when he wrote, "Mystical states, when

well dev elope d, usu ally are, and have the right to be, abso lutelyauthoritative over the individuals to whom they come" (myitalics). But, he adds, "no authority emanates from themwhich should make it a duty for those who stand outside ofthem to accept their revelations uncritically." This makes itlook as if the occurrence of mystical experiences constitutesevidence for their veridicality for those who have them butnot for those who do not, which clearly violates the requirement that a cognitive experience's status as evidential isobserver-neutral. No doubt those who have mystical experiences will be caused to believe in what they seem to reveal,whereas no such causal compulsion will operate on the non-mystic, but this is a psychological issue and of no epistemo-logical interest. Further, mystics will be more certain of theoccurrence of mystical experiences than nonmystics, but thism ak es no difference to the evide ntial sta tus of mystical experiences, only to the certainty of their occurrence.

W hat argu m en t can be given for the cognitivity, as op pos edto the veridicality, of religious experiences of the sort underconsideration? Every argu m en t tha t I kn ow of is base d o n anana logy w ith sen se experien ce an d is a special instan ce of thisgeneral argument, to be called the "analogical argument forcognitivity":

i. Religious experiences are analogous to sense experi

ences;2. Sense experiences are cognitive; therefore,3. Religious experiences are cognitive.

288



Religiousêxperience arguments

Why should religious experiences have to ride the coattailsof sense experience in their quest for epistemic respectability? It is d u e to tw o factors . First, sense exp erien ces are a paradigm of cognitivity or knowledge-yielding experiences,unlike religious experiences, whose credentials have beenperennially challenged. Second, we are clearer about the conceptual lay of the land in regard to the former than the latter.As William Alston, on e of th e leadin g defend ers of the cognitivity of religious exp erie nc es, h as said, "It seem s reaso nab leto take principles which are well established with respect to

sense-percept ion, an area whe re we pret ty mu ch kn ow w ha tto say under a given set of conditions, and extend them to thediscussion of purported direct experiences of God, an areawhere i t is not at ail clear what one should say/" Thus, themo st likely way to esta blis h the cognitivity of religious exp eriences is to sh ow th at th ey are sufficiently an alo go us tosense experiences, tho se pa rad igm s of cognitive exp eriences.William Wainwright, another leading proponent of the analogical argument for cognitivity, is quite right when he saysthat "the analogy (or lack of it) between mystical experienceand sen se experience ap p ea rs , then , to be critically im po rtan tboth to those who ascribe cognitive value to mystical experience and to those who refuse to do so."2

While few will venture to challenge premise 2, for thisleads to com plete scepticism, a serious prob lem arises in respectto 1, the analogical premise. How close must the analogy beto support the inference of 3 from 1 a nd 2? Th ere is no agr eed

upon answer to this quest ion among the proponents of thedifferent versions of the analogical argument, since thesearguments fill out the analogical premise in different ways,some giving a very thin, high-level analogy that holds onlythat both kinds of experiences form the basis of ongoinglanguage games, and others a more thoroughgoing analogythat claims similarity in the sort of checks and tests or defeating conditions that apply to both. The problem of how closethe analogy must be is very acute, since everything is likeeverything else in some respect; for instance, any two entitiesshare the property of being self-identical. This, I take it, was

289




the point a burly Canadian good-old-boy truck driver, Bob,m ad e wh en he res po nd ed to my question about wha t it is l ike

driving his monstrous sixteen wheeler by saying that "it wasju st aboot like an yth ing else you d o, just aboot like d rink inga glass of beer; you just sit back and do it without making abig fuss." I was te m pte d to ask Bob if he w as con tem platingwriting a book about Zen and the art of truck driving, but Iknow he would have replied that "Zen is just aboot likeanything else," which, by the way, would be an excellentone-sentence summary of the Zen vision.

The analogical premise must specify not any old analogybetween religious and sense experiences, for instance, thatboth are among Bob's favorite type of experiences, but onethat is relevant to their cogn itivity. But, eve n w ith this restriction, there is no decision procedure for determining howclose the analogy must be for it to be reasonable to infer 3from 1 and 2. The best that can be done is to consider thedifferent versions of the analogical argu m en t and, for each ofthem, explore in detail and depth all of the relevant analogiesand disanalogies, with the hope that once this is accomplished it will become rea sonably clear w ha t is the right jud gment to pass. I hope to show that the various analogies arevery thin and that the disanalogies are very extensive anddeep , thereby destr oy ing t he different versions of the analogical arg um ent. I kn ow tha t this will leave Bob un m ov ed ; b ut,hopefully, not all of my readers are imbibing a Molson, asthey say "This book is just aboot like every other book."

The results of my discuss ion will have wid esp read interest,since they can be applied to several other nonsensory faculties that are claimed to be cognitive on th e basis of an a nalogy with sense experience, most notably Plato's faculty ofintellectual intuition and Moore's faculty of moral intuition,both of which are described in perceptual terms. The fate ofall thes e nonse nso ry faculties of "percep tion" are inte rtw inedsuch that if my polemic against the cognitivity of religiousexperience succeeds, it works equally well mutatis mutandisagain st these othe r no ns en so ry "perc eptua l" faculties. "C atchone of them and you catch all of them "

290



Religious-experieme arguments

We shall begin with the weakest version of the argumentfrom analogy and then progress to ever stronger versions.The wea kest version is based o n a very thin analogy be tw ee nreligious and sense experience. In its thinnest form it holdsonly that both s erv e as the basis on w hic h objective ex istentialclaims are made within different ongoing language games,such a game being a normatively rule-governed hu m an practice of usin g lan gu ag e. N o further similarity is claimed, b u t inthe stronger versions the analogy gets filled in further so thatin both games there are checks and tests or defeating condi

tions for these claims, and, in the strongest versions, thatthese tests are similar.The weakest version will be called "language-game fide-

ism," s ince it derives from Wittgenstein 's language-gameapproach to religion in his Lectures and Conversations on Aesthetics, Psychology and Religious Belief. There are several different versions of language-game fideism, all of which can findsom e basis in W ittgenstein's w ritings. Th ere are the c ognitiveand noncognitive versions, the former viewing the religiouslanguage game as involving truth claims for which epistemicjustification can be given, the latter as involving nothingmore than the expression of attitudes toward life, basic values , convictions, and the like. The latter version admits of afurther distinction between descriptive and revisionaryaccoun ts . Philosophers l ike H are , Phill ips , H ud son , an d Malcolm and Winch in some of their pronouncements , haveclaimed that the religious language game that we actuallyplay is noncognitive, whereas others, l ike Braithwaite andSchmidt, accept that the actual game is cognitive but propose, so as to avoid a clash with science, that we revise it sotha t its sen ten ce s no longe r play a fact-stating role bu t in stea dplay an evocative an d evincing role. We shall not both er w itheither of the two versions of noncognitive language-gamefideism, though it should be noted that both are atrocious.The descriptive version radically misdescribes the actual

enterprise of religion, and the revisionary account, if actuallyimplemented, would result in the death of religion. Thereis no doubt that religious believers, including Kierkegaard,

2 9 1




think they are saying things that are literally true about Godand his relation to creation; their activities of attempting tocommune with God would make no sense without suchbeliefs, for we do not think of ourselves as entering into apersonal relationship, especially one of love, with a nonexistent person. Even the star of Harvey thought the big whiterabbit to be real.

In the weakest version of cognitive language-game fide-ism, the analogy between sense and religious experiencesconsists in nothing more than both types of experiences serv

ing within their respective language games as an epistemicwarrant or justification for making an existential claim to theeffect that the apparent object of the experience really exists.W illiam A lston d escrib es a versio n of suc h f ide ism , of whic hthe following is a modified account.3

The first thing that must be done is to give some criterionfor individuating language games, since this version of fide-ism depends upon religious and sense experiences occurringwithin different language games. A language game is individuated by the part of its ontology - what it is about or refersto - that has its categoreal nature defined by the rules of thegame. Each language game has epistemological autonomy inthat it has its own principles of rationality and justification,especially in regard to the unique part of its ontology. Language games can overlap in their ontologies, categories, andepistemology . W hethe r a type of experience cou nts as cognitive is relative to a language game. An of-O-type experience- an experience that has O as its apparent object or intentional accusative - counts as cognitive relative to languagega m e L jus t in case: (a) Th e vast m ajority of th e L -playersupon having an of-O-type experience assert or believe thatthey are having an experience of an objectively existent O orthat there exists an objective O and agree that it is correct todo so ; an d (b) L con tains a nonin du ctive ly derive d rule ofpresumptive inference that if one has an of-O-type experience, then one probably has an experience of an objective O

or the re exists an objective 0 .4

The a priori pre sum ptiv e inference rule of (b) is grounded in the agreement in "language-

2 9 2




entry" assertions described in (a), an of-O-type experienceserving as both the occasion and the justification for asserting that one experiences an objective O or that O objectivelyexists.

This version of f ideism w ind s u p by p ointin g o ut th at religious experiences count as cognitive relative to the languagegames played within various religious traditions. In theseextant gam es, the having of an of-God (undifferentiatedunity, etc.) -type experience serves as both the occasion andthe a priori-based justification for the language-entry asser

tion that one experiences an objectively existent God (undifferentiated unity, etc.) or that God objectively exists. Thisversion ends with the Greatest Story Ever Told, that "thelanguage game is played."

Alston applies this fideist account not only to direct, non-sensory experiences of God's presence but also to sensoryexperiences of some worldly item as God-caused - what waspreviously called an "inferential perceiving-as." Herein Ithink he misdescribes the Christian language game. He saysthat "one learns to see good and ill fortune as blessings ofGod or as trials sent by God to test us or to provide theoccasion for moral and spiritual growth."5 Thus, one learnsto respond to an evil by making the language-entry assertionor believing that it is a test sent by God. Since this assertionentails that God exists, it follows that evil, by virtue of the apriori pre sum ptiv e inference rule that applies to it w ithin theChristian language game, counts as evidence for the exis

tence of GodWhat could be a more radical distortion of this game, at

least as played by Phil Quinn's "intellectually sophisticatedadults in our culture,"6 in which evil tries rather than confirms one's faith. Alston's "seeing-as" language-entry assertions involve a theoretically based inference to God as thecause of some worldly item. They are not the recognitionalsort of seeing-as, which alone are our concern, and thus we

shall not pursue the problem further.Alston is unable to em brace this extreme version of f ideismbecause it grates against his dee p realist intu ition s. H e objects

293



ReUgiw§rmipe»riitm arguments

that an assertion is true by virtue of reality's being the wayit is asserted to be. But, pace Alston, the correspondencetheory of truth leaves fideism untouched, since it can beincorporated into every language game. The religious experience language game can hold that the assertion that thereexists an Eternal One (God, etc.) is true by virtue of thereexisting an Eternal One (God, etc.). The correspondencetheory is the willing tool of a language game's assertibilityprinciples.

The obvious and decisive objection to extreme language-

game fideism is that it is based on too thin an analogy, onethat would impress only Bob. It does not enable us to inferthe cognitivity of religious experience from the admitted cog-nitivity of sens e experience ; for, w her ea s the existential claimsbased upon sense experience have only a prima facie warrant, since subject to a battery of interconnected tests andche cks, ther e ar e no defeating con ditio ns for existential claimsbased upon religious experience. Whatever the experient

takes to be the case on the basis of his religious experiencereally is the case. Because there are no tests for the veridical-ity of his experienc e, there is no bas is for dr aw ing the d istinction between his actually perceiving God and it just seemingto him as if he is perceiving God. By not providing for anychance of being wrong, the analogy fails to make religiousexperience relevantly similar to sense experiences, for whichthe veridical-unveridical distinction holds.

An even more devastating objection to extreme language-game fideism is that its religious experience language gameis not actually played. It is only a figment of my imagination that I have introduced as a foil by which to bring intoclear relief less severe versions of fideism that are to follow. The religious experience language game we actuallyplay allows for the possibility of error, because it containstests for the existential claims based on religious e xperien ces.Such experiences give only a prima facie epistemic justifica

tion for the former due to the existence of these defeatingconditions.

294



Langmge^game fMeism ttHth any old tests

LANGUAGE-GAME FIDEISM WITH ANY OLD TESTS

The crucial question is how close the analogy must be betw ee n th e sort of tests an d c hecks that en ter into the religiousand sense experience language games for the analogicalarg um en t to work. Let us begin with an analogical a rgu m en tbased upon the thinnest analogy between them, namely,that each game contains some test for the veridicality of anexistential claim based upon its respective experience.

This imputes considerable epistemological autonomy tothese gam es, each hav ing its ow n distinctive criteria of rationality and epistemic justification, and thereby achieves thefideistic end of protecting religion from science.7 This comesclose to embracing a multivocalist theory of meaning according to which "cognitive," "knowledge," "veridical," "rationality," "justification," and "objective existence" have a different meaning in a religious context or language game than

they have in a scientific or ordinary fact-stating one. Anexample of such a multivocalist-based defense of the cogni-tivity of religious experiences is Walter Stace's claim that suchexperiences, although not qualifying as objective by ordinarysensory-based standards, are objective in the sense of being"trans-subjective." Herein he seems to hold that there is inthe mystical context a different and equally legitimate senseof "objective," along with th e other related term s, to the o rdinary, sensory-based one.

This version of f ideism, a ltho ug h sup erior to the extre m eform, is still unacceptable. In the first place, it misdescribesthe religious language game. The players are not willing tosay that their religious experiences provide them with knowledge of an objective reality bu t in a different sen se of " kn ow ledge" and "objective reality" from their ordinary, sensory-base d o ne; for this w ou ld be to aw ard the boo by prize to theirgam e. Adm ittedly, a propo sition that predicates objective

reality or existence of a table ha s categoreally different enta ilments than does one that predicates objective reality of God

295




or the undifferentiated unity; but these differences are duenot to any difference in the meaning of "objective reality" inthe t w o prop osition s but a difference in the categoreal na tu reof the subjects of predication. Similarly, "and" is univocal in"Tables and chairs exist" and "Minds and sensations exist/ 'in spite of the radical differences in the categoreal kind ofpropositions they entail.

Furthermore, the religious-experience language game doesnot have the sort of self-containment and epistemologicalautonomy that is imputed to it by this version of fideism.

This game is a part of the larger religious language game,w hich em ploy s man y of the same principles of ded uctiv e a ndinductive logic as does the sensory-based physical object language in science and everyday life. The claim that God existsis not justified solely by appeal to religious experience butalso by inductive-type arguments, as well as ontological andcosmological ones employing principles of deductive logic.For the "intellectually sophisticated adults in our culture"who participate in the religious language game, the proposition that God exists is not basic in Plantinga's sense, sincethey are willing to give various sorts of epistemic justifications for it.

In general, the use of "language game" to characterize thedifferent w ays in which w e use languag e is m isleading , sinceit makes it look as if different uses of language enjoy the sortof self-sufficiency an d insularity that different ga m es do . Butthis is a my th. Th ere is only one big languag e gam e of wh ich

different so-called language games are interconnected parts.Th e religious-experience lang uage gam e, as well as the largerreligious language game, is part and parcel of the physical-world one, sharing many of its basic concepts concerningcognitivity, knowledge, existence, and warranted assertibil-ity. Pace Wittgenstein, the theist and atheist really do believeincompatible propositions when one asserts and the otherdenies that there is a Last Judgment.

The most devastating objection, however, to this versionof fideism is that it not only could, but actually does, justifyfar too much. By not requiring that the checks and tests for

296



Language-gmne fideism with any old tests

the veridicality of religiou s exp erience be significantly similarin cognitively relevant respects to those that enter into thesense-experience language game, i t requires us to count ascognitive those religious experiences that occur within acultist, religious-experience language game, in which theonly test or check is ba sed u po n w ha t the cult leader says . Forinstance, if a member of the PTL wants to find out if she hashad a veridical religious experience, she simply telephonesJim Bakker, whatever he says settling the matter. While thisis an exaggeration of the PTL game, it is not of some of the

other cultist games, which, sad to say, actually are playedand played with a vengeance. That this version of fideism iscommitted to saying that these cultist religious experiencescount as cognitive shows that there is something radicallywrong with this account of cognitivty. Anyone who doesn'tagree should be horsewhipped.

It might be urged against my cultist counterexamples tolanguage-game fideism with any old tests that they can be

ruled out on the grounds that there are good independentreasons for rejecting their outputs, for instance, that theyclash with the outputs of logic, science, or our deepest andmost widely shared moral convictions. But we can availourselves of merely possible counterexamples to this versionof fideism, such as a possible PTL language game in whichBakker and his followers are paragons of virtue and do notsay an yth ing that clashes with science an d the like. The pointis that even in such counterfactual circumstances, we wouldnot want to count their epistemic practice, based upon theedicts issued by Bakker being decisive, as cognitive.

Another difficulty is that there are a multiplicity of different religious-experience language games, each with i ts owntests for veridicality - its ow n h oly scriptu res, holy me n, an dparadigmatic religious experiences. A religious experiencecould count as veridical relative to the tests provided by oneof thes e gam es or religious traditions bu t not ano ther , wh ich,

in fact, is the case. Whether this failure among players ofreligious-experience language games to agree about theproper methods for determining veridicality is destructive of

297





Languag&game fideim with Analogous tests

be established by appeal to experiences other than religiousones. If we are willing to accept a circular justification in onecase so as to escape scepticism, we should be willing to do soin the oth er. It w ou ld n ot only violate the princ iple of paritybut also be an unwarranted display of chauvinism for us todemand that the reliability of the religious-experience doxas-tic practice be established by app eal to sen se ex pe rien ce. T hiswould be as unfair and chauvinistic as requiring that thesense-experience doxastic practice be justified by appeal toreligious experience.

Tha t th e tw o doxastic practices of belief form ation are to beaccorded parity is justified by showing that they are analogous in cognitively relevant respects, especially in regard toboth containing their own checks and tests. Some give theanalogy up front, while others present it in the manner ofrearguard action, first challenging someone to produce a dis-analogy that w ould justify our not exten din g the a priori ruleof presumptive inference to religious experiences and thenneutralizing any preferred disanalogy. For Richard Swinburne this rule takes the form of his principle of credulity: "Ifit seems (epistemically) to S that x is present," that is, Sbelieves on the basis of her experience th at x is pre sen t, "thatis good reason for S to believe that it is so, in the absence ofspecial considerations" that defeat the claim.11 Gary Guttingargues (convincingly) that this principle is too strong andbuilds his analogical argument on the weaker rule that if itseems to a subject that x is pres ent, then the re is "sign ificant

but not sufficient evidence" that x is present, hereby departing from Swinburne, who takes it to be sufficient in theabsence of canceling conditions. This is one of those minorin-house disputes that need not detain us. For WilliamAlston, the presumptive inference rule, as applied to religious exp erience s, takes this form: "The experience . . . ofGod provides prima facie epistemic justification for beliefsabout wh at God is doing or how God is 's i tuate d' vis-a-vis oneat the m om en t . ""

Before getting down to an in-depth comparison of the testsinvolved in the sense- and religious-experience language

299




games, it should be pointed out that we are being imposedon at the very ou tse t in being asked to coun tenance the latteras a separate language game or doxastic practice. Alstonholds that a language game or "a doxastic practice will typically operate on certain basic assumptions that cannot bejustified within the practice but rather serve as presuppositions of any exercise thereof. 13 Sense experience certainly isconstitutive of the language game or doxastic practice inwhich claims are made about the existence and properties ofphysical objects. Such claims m ust ultimately rest up on sen se

experienc e; an d it is a presu ppo sition, not justifiable w itho utcircularity within this practice, that such experiences generally are veridical. Doubt this presupposition and you cannoteng age in the practice. Disanalogously, claims abou t the existence and nature of God need not, and typically are not,based upon religious experience, and thus religious experiences are not constitutive of the religious language game.There are many believing participants in the religious langua ge gam e w ho m ake claims abou t the existence and na tur eof God who have not themselves had religious experiencesan d, m oreo ver , d o not use the religious experiences of oth ersas a means of anchoring their reference of the name "God,"but instead do so via the use of descriptions of God's hard-an d soft-core pro pe rties , to use the distinctions from C hap ter1. Religious experiences are not constitutive of any doxasticpractice, bu t instead are an inessential part of the en com pas sing religious language game.

Alston disagrees and offers the following causal theory ofthe reference of "God," in which the referent of "God" initially is pinned down via religious experiences with subsequent members of the religious community picking up thereferent from their predecessors, ultimately terminating inthese seminal "baptismal" bestowals of the name:

In the Judaeo-Christian community we take ourselves to be

worshipping, and otherwise referring to, "the God of Abraham , Isaac, and Israel", i.e. the being who appeared to suchworthies of our tradition, revealed Himself to them, made

3



Language-game fideism with analogous tests

convenants with them , and so on. If it should turn out that itwas actually Satan, rather than the creator of the heavens and

the earth, with whom they were in effective contact, wouldwe not have to admit that our religion, including the referential practices involved, is built on sand, or worse (muck,slime), and that we are a Satan-worshipping community, forall our bandying about of descriptions that fit the only trueGod?14

I have consulted a wide variety of religious believers, and all

of them answer Alston's rhetorical question in the negative.I do no t do ub t that there are O rtho do x Jews and ChristianFundamentalists whose religious language game fits Alston'scausal theory of reference, but it certainly does not properlydescribe the religious language game of Phil Quinn's "intel-lectualy sophisticated adults in our culture," many of whomare ignorant of which Biblical dignitaries were alleged to havehad religious experiences, and thus cannot be picking up thereferent of "God" from them and who, moreover, would notaccept the characterization of themselves as devil worshippers or LSD worshipers, if it should turn out to be the casethat the experiences of these dignitaries were caused by thedevil or LSD. I feel sorry for Alston's type of believers; theymust tremble every time they open their copy of the NationalEnquirer for fear that th ey will com e u p o n a headline thatreads "A nthro polo gists Discover that An cient Israelites WereAcid-Heads." Another fundamental inadequacy in Alston 's

causal theory of reference for "God" is that it cannot accountfor how two persons who use "God" can be coreferrers evenif God does not exist . As argued in Chapter 1, our use of"God" allows for such a case of intentional identity.

In spite of these misgivings about there being a separatereligious-experience doxastic practice or language game, wewill grant that there is for the sake of argument and will nowget on with our main task of exploring the analogies and

disanalogies between the tests for the two practices. We willbegin w ith a brief ove rview of the var iou s tests and checks forthe veridicality of sense experience, hereafter "sensory tests"

301



Religious-experieme argum ents

for short, mainly as seen by the analogical arguers, and then

see how our four analogical arguers attempt to show that fora sufficient number of them, there is a sufficiently similarmystical analogue.

SENSORY TESTS

i . Logical consistency. Are the sensory -based claims of th esubject logically consistent?

2. Empirical consistency. Are they consistent w ith w ell-established empirical facts?

3. Existence of apparent object. Are there good arg um en tseithe r for or aga inst the existence of the ap pa ren t object of th esense experience?

4. Reliability of subject. Do es the subject ha ve a good trackrecord in his past perceptual claims?

5. Agreement. Do the sense experiences had by othersun de r normal or stan dar d con ditions agree with the subject's?

6. Continuity between contents. Does the con tent of the exp erience stand in the right sort of lawlike relations to those ofthe vast majority of the subject's preceding and succeedingexperiences?

7. Prediction. Can future sen se experienc es of the subjectand others be predicted on the basis of the assumption thatthe experience is veridical?

8. Proper position. W as the subject pro per ly located in sp ac eand time so as to have been able to perceive the apparentobject of his sense experiences?

9. Physiological state of subject. Was the subject's se ns or yfaculty in good working order?

10. Psychological state of subject. Did th e subject hav e an ypsychological im pe dim en ts to his being able to perceive a dequately, for instance, phobias, prejudices, or the like?

11. Causal requirement. Did the ap pa re nt object cause th eexperiences in the right way?

The analogical arguer takes these tests to form a "cluster"in tha t no one of the m is either nec ess ary or sufficient for

3 2



Sensory tests

veridicality, but satisfaction of a sufficient number is such.Such vagueness, double vagueness in fact, should not sur

prise us if w e are dealing w ith a cluster co nce pt.Before p roce edin g further, I w ou ld like to dr op tests 1-4from further c onsid eration . Test 4 obviously is parasitic up onthe others, since the subject's past perceptual claims musthave their veridicality determined by them. Tests 1-3 havevery little positive probative force, both individually and collectively, serving a primarily eliminative function. The logicalconsistency test is not really a test but a sine qua non forbe ing a fit subject for testing. Th at the ap pa ren t acc usative of

a sense experience is logically and empirically consistent, oreven that it is existent, does almost nothing toward increasing th e probability of the experience 's veridicality, since m os tunveridical sense experiences, for instance, dreams, have anintentional accusative that satisfies these conditions. Furthermore, in regard to the religious-experience analogue to 3,the existence of the apparent object would have to be established either by an argum ent that is not based up on religiousexperiences, in which case it would be relevant only to the

veridicality of the exp erience b ut not its cognitivity, as previously indicated, or by appeal to other religious experiences,in which case appeal is made to one of the other tests, suchas agree m ent, thereby rende ring this test parasitic up on oth erones.

In eliminating tests 1-3 from further consideration, Iam really doing the defender of the cognitivity of religiousexperiences, at least of the mystical sort, a favor, since thereis serious doubt that they satisfy them. Their descriptions

are often either contradictory or in conflict with our best-estab lished empirical beliefs, such as that there exists a multiplicity of distinct objects and e ve nts in space and time . Wain-wright has argued skillfully in defense of the logical andempirical consistency of mystical claims. Basically, what hedoe s is to give wa tered-d ow n reconstructions of thes e claimsso that they are neither contradictory nor in opposition toempirical claims, but instead are making claims about a reality that is numerically distinct from any empirical reality, God

3 3




or the undifferentiated unity being distinct from the physicalwo rld. W ainw right's thesis do es not square with th e claims ofnature mysticism or cosmic consciousness, since these experiences seem to have the same apparent object as do senseexperiences, namely, nature. Furthermore, if God or theundifferentiated un ity is num erically distinct from the phy sical world, it belie s the claim of the monistic my stic that thereis only the One.

Wainwright claims that, in respect to mystical experiences,a sufficient number of analogues to the sensory tests can be

found within various religious communities and offers thetests em ploy ed by the C atholic mystical tradition as an example: (i) the subject's moral improvement as a result of theexperience, (2) the morally beneficial effects the experiencehas on others, (3) the depth of the mystic's description,(4) the concurrence of the mystical report with the holy scriptures and revealed truths of Catholicism, (5) resemblance ofthe experience with paradigmatic mystical experienceswithin the Catholic tradition, and (6) pronouncements ofauthorities and holy men on the experience's veridicality.These tests are tailored to theistic mystical experiences andneed to be modified slightly for monistic experiences. Tests(1) and (2) - the moral tests - will have to be altered to takeno te of the fact th at it is not a loving pe rsonal bein g th at is theapparent object but an impersonal undifferentiated unity.Accordingly, they will require good spiritual rather thanmoral consequences for the subject and others.15 Tests (i)-(3)

are instance s of the p rediction test an d (5) the agreem en t test,with the added provision that some member of the agreement set is paradigmatic. Supposedly, his privileged status,as well as the privileged epistemic status accorded to certainscriptures and men in (4) and (6), respectively, are to be justified by appeal to either other mystical experiences or arguments not based on religious experience. If it is the latter,tests (4) and (6), although relevant to the veridicality of theexperience being tested, contribute nothing in support of itsstatu s as cogn itive, that i s, as a w arra nt for a kn ow ledg e claim.

3 4



Sensory tests

William James, in his discussion of mysticism in Varieties ofReligious Experience, em ployed som ethin g like tests (i)-( 3).As a physician and psychologist he was well aware of thedarker side of mystical experiences, that many were diabolical rather than benevolent in their effects on the subject andothers. He claimed that the veridicality of mystical experiences must be determ ined not on the basis of their "m edical"cau se bu t on th e go od ne ss of their effects. "To pa ss a spiritualjudgment upon these s tates ," he wri tes , "we must not content ourselves with superficial medical talk, but inquire into

their fruits for life."Gutt ing also works with a t r immed down version of ( l )-(6). He is concerned with a nonmystical theistic experienceinvolving a nonsensory perception of a very powerful, loving, no nh um an perso n w ho is concerne d w ith our well-being.All his tests are predictability on es base d on the further experiences the subject an d oth ers can expect if the of-God expe rience is veridical: "(1) Those who have had such experiencesonce wo uld b e likely to hav e them again ; (2) oth er indiv idua lswill be found to have had similar experiences; (3) those having such experiences will find themselves aided in theirendeavors to lead morally better lives. '"6 Gutting's (2) and(3) are virtually identical with Wainwright's (1) and (2). Histest (2) also is an agreement test of sorts.

Alston, in defending the "objectivity" or cognitivity of theChristian ve rsion of the religious-experience langua ge ga m e,first points out that (a) the language game, along with theform of life that it involves, has proven itself viable over timeand (b) no assertion warranted by the rules of a sophisticatedversion of this gam e contradicts an yth ing that is w arran tedlyassertible within other language games we play. He thengoes on to argue that (c) religious experiences, as a group,cohere together in a way analogous to that in which senseexperiences, as a group, cohere together. In each case theexperiences are connected together in a way that is to be ex

pected, given the categoreal nature of the apparent objectof the experiences. "The behavior of God, as revealed in the

305



Religious+experience arguments

Christian language-game, is in line, roughly speaking, withwhat one could reasonably expect from the categoreal features of God, as depicted within this language-game."1 7 H eappeals to something like the previous tests of Wainwright,Gutting, and James based upon the moral and spiri tualgrowth of the subject, as well as that of agreement. "TheChristian life, or, in the present parlance, seriously playingthe Christian language-game, is self-justifying and self-fulfilling."18 And, in a later paper, he says, "Given standardtheistic conceptio ns an d beliefs, it is reasonab le to expect th at

prolonged communion, perceptual and otherwise, with Godwould lead to an increasing sanctity of character and a moreChristlike mode of life, ceteris paribus."1 9 Herein there is anappeal to a religious experience analogue to the sensoryagreement and prediction tests, the claim being that the religious-experience language game is internally consistentbecause things h ap p en by an d large in a m an ne r that is explicable in terms of the categoreal nature of the entities withinthe game's unique ontology.

Swinburne, w h o presen ts a rearguard-action version of th eanalogical argument, does not speak of tests but instead ofdefeating conditions - conditions that lower the probabilityof the experience be ing veridical an d th ereby its being subjectto his earlier mentioned principle of credulity. A defeatingcondition is a flunk ed test. He begin s by listing the defeatin gconditions for a sense experience: The subject or conditionsun de r which the appare nt perception w ere ma de are of a sort

that have proven to be unreliable; other ob servers in like conditions disagree with the subject 's pe rcep tua l claim; the exp erience w as not cau sed in the right w ay by its ap pa ren t object.It is then a rgue d th at religious exp erien ces, specifically directperceptions of Go d's presence, d o not flunk eno ug h of th esetests so as not to be subject to the principle of credulity, andthereby qualify as cognitive. As we shall see below, hisdefeating con ditions inco rporate all of the senso ry tests , sa vefor 9 and 10.20

With this overview completed, i t can now be consideredho w good an analogical case ha s be en m ad e ou t by the se four

306



Sensory tests

authors. It will be argued that their analogies betweenreligious- and sense-experience tests are weak and that theybe com e even weaker, appro aching the vanishing point, w he nwe replace the above cluster view of the sensory tests with amore adequate organic view.

Th at th e agreem ent test applies in analogous w ays to bothsense and mystical experiences was asserted by C. D. Broad:

When there is a nucleus of agreement between the experiences of men in different places, times, and traditions, and

when they all tend to put much the same kind of interpretation on the cognitive content of these experiences, it is reasonable to ascribe this agreement to their all being in contact w itha certain objective aspect of reality unless there be some positive reason to think otherwise. . . . I think it would be inconsistent to treat the experiences of religious mystics ondifferent principles. So far as they agree they should be provisionally accepted as veridical unless there be some positiveground for thinking that they are not.21

Subsequent writers have shown, however, that the application of the agreement test to mystical and religious experiences is not as straightforward as Broad suggested. Theagreement or disagreement of other observers is relevantonly if their observations are made under so-called normal orsta nd ard cond itions. We hav e a pretty good idea of what theyare for sense experience. Their mystical or religious ana

lo gu es vary from one mystical or religious tradition to an ot he rbut usually involve asceticism, meditation, breathing exercises, and the like.

So far the analogy looks fairly tight, but, as Wainwrightp o in ts ou t, it falters because m ystics (i) coun t as confirm atorymystical experiences that occur without following the prescribed "my stical wa y" and (ii) do no t cou nt failure to have amystical experience under the prescribed conditions as dis-confirmatory. The evidential asymmetry in (ii) is especiallydisturbing, since i t makes the mystical use of the agreementtest lo ok like a "h ea ds I w in , tails you lose" sort of con g am e.

3 7




The same holds mutatis mutandis for direct perceptions ofGod's presence.

There is an even more serious problem for the religiousagreement test. The application of this test to sense experience requires, as Swinburne points out, that we "know whatsense-organs and training you need, and how attentive youneed to be to perceive the object in question."22 The troubleis that we can specify none of these things with respect toreligious experiences. Whereas we have tests for when one'ssensory faculties are not in proper working condition and

when the context of perception is not normal or standard,there are no analogous tests for when one's faculty of religious intuition or the circumstances are not normal. Andsince the notion of a normal observer and normal conditionshas no application to religious experience, neither will tests8-10, since their requirements of a physiologically and psychologically normal observer properly stationed in space andtime are parts of the normal or proper conditions requirement. And since these tests, along with the agreement test,lack religious analogues, this greatly weakens the analogicalcase for cognitivity.

It might be objected that we do have some idea of thepsychological and moral conditions that facilitate having amy stical or religious experience, su ch as the previously m entioned features of the mystical way - asceticism and so on.This may be true, but it is irrelevant to the determination ofthe veridicality of a religious experience. Such causal knowl

ed ge as we ha ve pe rtain s to the occurrence of a psychologicaloccu rrenc e or state - the having of an of-God-type experience- and not to its veridicality.

In summary, it is highly dubious that there is a religious-experience analogue to the sensory agreement test, and forreasons that also call into question whether there are religious-experience analogues to the proper-conditions tests8-10. But the analogical arguer has not yet run out of tests.There still remain tests 6, 7, and 11, and the question ofw he th er they ha ve religious-experience analogues that m ightsuffice for their analogical argument.

3 8



Sensory tests

The continuity-betw een-co ntents an d prediction tests overlap. The latter, unlike the former, is not confined to the subject's experienc es n or doe s it req uire that th ere be strict no m icconnections between the content of the experience to betested and those of the predicted experiences. Since the apparent object of a religious experience, be it God or the undifferentiated unity, is not one object among others in a nomi-cally determined system, the continuity-between-contentstest cannot apply to it. But maybe probabilistic predictionscan be m ad e abo ut h ow and w he n this object will reveal itself

and what sort of impact it will make upon the subject andothers.Wainwright, Alston, and Gutting make the morally good

consequences of the theistic religious experience central intheir deployment of the prediction test to such experiences.We saw that James appealed to the same test in order todistinguish between veridical mystical experiences and theirunveridical diabolical cou nter par ts. In some of his u ng ua rde d

remarks, James endorsed a happiness criterion of truth ingeneral - true propositions being those that bear beneficialconsequences for their believers. One might raise against thegood-moral-consequences criterion for the veridicality of of-God experiences the standard objection to this criterion oftruth - h ap pin ess h as no thing to d o w ith truth or veridicality.Whatever force this objection might have against James'spragmatism, it has no application to the theistic religiousexperience's good-moral-consequences criterion, for in thecase of the latter there is a conceptual or categoreal linkbetween the nature of the apparent object of such an experience and good moral consequences. Since God is essentiallygood, it is probable that those who have a veridical experience of him will benefit morally and spiritually. The morallybeneficial consequences of the experience are not confirmatory qua good consequences (as a crude pragmatism mighthold) but qua being a fulfillment of a categoreally based

prediction.It would be askin g too mu ch to requ ire that precise pred ic

tions be made of the occurrence of of-God experiences. God's

309




na ture p recludes this. What is disturbing is that it is no t e ve nclear tha t any probabilistic estim ates at all can be m ad e. Sw in

bu rne hold s that "if there is a God , ther e is a greater prob ability that men will have such experiences."23 But on the basicassumption of theism, it is not clear that the opposite is notthe case. Theistic mystical experiences are often so overpowering that they psychologically necessitate the subject tobelieve in God. Such a belief, however, goes counter to thetheist's view that God wants man freely (in the Libertariansense) to elect to believe in him. This point would have even

greater force against Gutting's claim that "those who havehad" veridical of-God experiences "once would be likely tohave them again."24 This really w ou ld be goin g too far tow ardusurping a person's free will.

It could be objected that I have mislocated the point atwhich freedom is supposed to enter in for the theist. Thetheists who stress the importance of our freedom to choosefaith in God would not suppose that one is free to eitherperceive or not perceive som ethin g (once one is in a positionto do so), God include d. Furth erm ore , significant freedomenters the picture in terms of our efforts to have experientialcontact with God by following the prescribed mystical way.Thus, we exercise our free will in the manner in which wepu rsu e this way, not in our believing or no t believing in w ha tour religious experience appa ren tly revea ls. But this re sp on sefrom the theist fails to give us significant freedom to believeor not believe in the veridicality of one of our religious expe

riences. A nd , as a result, G od, by choosing to cause a p er so nto have an of-God experience, renders them unable to doother th an believe tha t he exists. This gives a very differentversion of our freedom to have faith in God from the onecommonly advanced by theists.

It is not even clear that, as Sw inbu rne claims, "ther e are nogrounds for supposing that if there is a God, the atheistwould have experience of him," thereby hoping to block

application of the principle of credulity to atheistic experiences of it seeming (epistemic) that God is not there.25 Theatheist is far more in need of God's grace than is the person

3 1 0



Sensory tests

who is already launched along the path of faith. Thus, theatheist would seem m ore likely to ha ve an of-God exp erien ceif God exists than would the believer, especially since theatheist's disbelief would help to offset the potentially freedom -canceling im pact of the experienc e. I am n ot, of cou rse ,sincere in these probabilist ic pronouncements, but merelywish to show how little basis there is for making any probabilistic estimates at all in this area. This radical unpredictability stands in marked contrast with the familiar sort ofprobabilist ic predictions that are made within quantum and

thermodynamic theory, as well as of the free actions of persons, even the Libertarian admitting that free actions arepredictable for the most part. Again, we find a cognitivelyrelevant disanaiogy betwe en the tests for sen se an d religiousexperience, this time pertaining to the prediction test ratherthan the agreement test .

Because the religious-experience prediction test is so liberal, it should not surprise us that all sorts of bizarre experiences that obviously are not cognitive pass it. In thisconnection let us recall Gutting's tripartite test for the objectivity or cognitivity of an of-God experience,

(1) those who have had such experiences once would belikely to have them again;

(2) other individuals will be found to have had similarexperiences;

(3) those having such experiences will find themselves

aided in their endeavors to lead morally better lives.Consider the experience of an orgasm. Most of us considerthis a par adig m of a subjective exp erienc e, an experien ce thathas a cognate accusative, "or gas m ing" being pe rsp icuo us for"experiencing an orgasm." But this experience satisfies allthree of Gutting's predictive requirements:

(1) Those who have had orgasms once would be likely tohave them again;

(2) other individuals will be found to have had similarexperiences;

311




(3) those who have orgasm s find them selve s aided in theirendeavors to lead sexually better lives.

in which "sexually better lives" replaces "mora lly better live s"in (3) so as to m atch th e na tur e of the accusa tive. O bv ious ly,Gutting's criteria for an objective or cognitive experience letin too much, since we don't want to say that people whohave an orgasm that satisfies Gutting's criteria experience ormerge with the Big O (and I don't mean Oscar Robertson),some kind of objective reality that has as its main function toproduce orgasms in people, sometimes employing ordinarypersons as its causal intermediaries, thank God

Alston's attempt in "The Christian Language-Game" tospell out the conditions u nd er wh ich a lang uag e gam e is "o bjective" or "in touch with reality" also lets in too much. Histripartite criterion requires: (A) The language game is playedsuccessfully in that the form of life it involves has provenviable, (B) no assertion warranted by the rules of the gamecontradicts anything that is warrantedly assertible within

other langu age gam es we play, and (C) the gam e is internallyconsistent because things happen by and large in a mannerthat is explicable in terms of the categoreal nature of theentities within the game's ontology. Condition C, as pointedout, incorporates the prediction test.

We all take the experience of pain to be a paradigm of asubjective experience, one in which the accusative does notexist independently of its being experienced and is not common to different subjects of experience. We accept the cognate accusative (adverbial) analysis of "X feels a pain" into "Xpains" ("X feels painfully"). But the language game in whichwe report our pains satisfies Alston's conditions (A)-(C) fora language gam e being objective: (A) The gam e is play ed, itsform of life in which we aid and commiserate with people indistress hav ing pro ven to be very valuable, (B) the lang uag e-entry assertions warranted within this game, for instance, "Ifeel a pain," do not contradict those warranted by any other

game we play, and (C) the game is internally consistent, forinstance, peop le wh o utter avowals of pain usu ally take step s

3 1 2



Sensory tests

to eliminate their pain and try to avoid pain, and we can predict under what conditions persons will feel pain. But wedon't want to say that the pain language game is "objective";for w hile there no doub t are pain s - o ccurrences of pain ing -there are no objectively existent pains, that is, pains or pain-ings that exist independently of their being experienced andare common objects of the experiences of different persons.Pains are nothing but certain psychological states or processes. Thus, the fact that the religious language game satisfies conditions (A)-(C) does not establish that of-God

experiences are objective. While there are of-God-type psychological states or experiences, they may take a cognaterather than an objective accusative. And unless of-God experiences are objective, they do not qualify as cognitive, since acognitive experience is one on the basis of which we gainknowledge of some objective reality. Much more needs to besaid about the objectivity of religious experiences. It will turnout that the really decisive disanalogy between religious andsense experiences concerns the concept of a veridical objective experience.

One might try to block my counterexamples to the Guttingand Alston criteria for the objectivity or cognitivity of a kindof experience by restricting these criteria to kinds of experiences that are either putatively perceptual or such that theapparent object of the experience would be objective if real.Th e latter restriction wou ld be no he lp, since pain s and oth ertypes of subjective experiences have apparent objects that

are real (there really are pains) but not objective in the senseof being mind independent and a common object of theexperiences of different persons. Such restrictions, furthermore, are viciously regressive, since people, especially philosophers, often disagree about whether a certain type ofexperience is putatively perceptual or such that its apparentobject is objective if real at all. Alston and Wainwright, forexample, think that religious experiences are putatively perceptual, but I view them as being only psychological, that is,as taking cognate rather than objective accusatives. Mooretook se nse exp eriences to hav e objective accusatives, but C. J.

313




Ducasse and Wilfrid Sellars thought their accusatives to beonly cogna te. W ha t is ne ed ed is a criterion for d istingu ishingbetween cognate and objective accusatives so that these dispu tes can be settled. T he criteria of Gu tting an d A lston a re ofno help in this regard. Surface grammar and ordinary languag e are ina deq uate , for there are som e verbs that obviouslytake a cognate accusative but for which there is no corresponding verb or adverb for its grammatical accusative. Forinstance, whereas we can convert "Jones played a trill" into"Jones trilled" we cannot convert "Jones played a sonata"

into "Jones sonataed," since ordinary language lacks such averb. But obv ious ly the latter, like the form er, tak es a cogna teaccusative. It is the aim of this chapter to supply just such acriterion of objectivity and thus for a verb taking an objectiverather than a cognate accusative.

Ha ving found the religious-expe rience prediction test to beinadequate, there remains only the caused-in-the-right-waytest to consider. While everyone would agree that veridicalreligious experiences must be caused in some way by theirintentional accusative, we find the defenders of their cogni-tivity in sha rp disa gree m ent ov er w he the r it can have a natural cause and, if so, whether it can be any kind of naturalcau se. At on e extrem e we find thos e like Stace with his principle of causal indifference, who hold that it makes no difference how a mystical experience is caused. Swinburne alsoplaces no restrictions on how an of-God experience is caus ed:

But if there is a God, he is omnipresent and all causal processes only operate because he sustains them. Hence anycausal processes at all which bring about my experience willhave God am ong their causes, and any experience of him willbe of him as present at a place where he is.26

Wainwright agrees. "If theism is true," the causal requirement condition "is met by theistic experiences, since God is

a cause of everything that exists."27

God, however, is essentially benevolent, and thereby would not cause an of-Godexperience in a devious or ignoble way. Swinburne con-

3M



Sensory tests

founds the true proposition that it is impossible for God tocause an of-God experience in the wrong way with the falseproposition that it is impossible for an of-God experience tobe caused in the wrong way, that is, a way that would preclude it having God as its cause. Many have felt that anof-God experience that was caused by ingestion of LSD orunconscious sexual desires could not be veridical.28

At the other extreme we find someone like Wayne Proud-foot who claims that by definition a mystical experience, likea miracle, can not h ave a natural ca use.29 This use of "m ystical

experience" is not only eccentric but also perverse, since itblurs the veridical-unv eridical distinction. Ce rtainly the ve rysam e phen om enolo gical-intentional state that is realized in averidical mystical experience could be realized in an unverid-ical on e. Alston o nce claime d that a veridical religious e xpe rience could not have a sufficient natural cause.30 Instead ofm akin g it true by definition, as doe s Prou dfoot, he pre sen tedthis argument: God must be at least a necessary cause of a

veridical religious experience, but if there were to be a sufficient natural cause of it, that is, a natural cause comprised ofevery necessary cause of it, God, not being a natural entity,would not be among them. This argument fails because, asseen in Chapter 4, causation can reach through a proximatesufficient cause; for instance, my hiring of an assassin whothen comm its the m urd er and my hand m oving the stick thatmoves the stone are necessary causes respectively of themurder and the moving of the stone. If I were not to havedo ne w hat I did, the effect w ou ld not hav e occurred. Thisshows that an event can be causally necessary for a givenevent even though there intervenes between them a causallysufficient event. It would seem, then, that a veridical religious experience can have a sufficient natural cause, provided it is consistent with the nature of the experience'sapparent object to allow such a cause.

The problem is how it could be determined that a relig

ious experience is caused in the right way by its apparentobject. There is a significant disanalogy between how this isdetermined for sense and religious experiences, which is not

3*5




objectionably chauvinistic by making sense experiences themeasure of all things, that renders it dubious that there is areligious experience analogue to the caused-in-the-right-waytest. Whereas w e can determ ine on the basis of sense experience that a given sense experience is caused in the righ t w ayby its apparent object, we cannot determine on the basis ofreligious experience alone that a given religious experience iscaused in the right way by its apparent object. Wainwrightwas quite right when he said,

One cannot determine whether the [caused-in-the-right-way]condition is met independently of metaphysical and theological considerations. The practical value of an appeal to this consideration is therefore limited.31

To det erm ine th at a religious experien ce is cau sed in the righ tway by its apparent object, we require an argument that isnot based on religious experience for the existence of thisobject and a theological theory con cern ing how it is related toexperiences of it.32

But why can't we have an argument based upon religiousexperiences for the existence of the ap pa rent object of a givenreligious experience and its bearing the right sort of causalrelation to the experience? There can be such an argumentonly if religious experiences count as cognitive. But they cancount as cognitive only if they are subject to similar tests tothose that sense experiences are. And it has already been

sho w n that this is not the case. Sensory te sts 6 an d 8-10 ha dno religious-experience analogues at all, while tests 5 and 7were found to have at best quite weak analogues. The mystical analogues to the agreement and prediction tests arefurther undermined by the diversity of different doxasticpractices for basing existential claims on religious experiences, since each has its ow n revealed tru th s, holy scrip tures ,paradigmatic religious experiences, holy men, authorities,and conception of what constitutes desirable moral and spiritual growth. Circumstances that count as confirmatory ofa given religious experience's veridicality relative to one of

316



The analogkal-argtter's response

these practices are not relative to another. Even if I havesomewhat overstated the diversity between the religious-experience doxastic practices of the great extant religions,there still is no analogous diversity of doxastic practices forbasin g claims ab ou t physical objects on se nse experience thatdiffer among themselves as to what counts as confirmatoryand disconfirmatory of a given sen se experience being veridical.33 This is a cognitively invidious disanalogy that shoulddestroy the requ irem ent to extend the presum ptive inferencerule from the sense experience to the religious-experience

doxastic practice in the name of the principle of parity. Theresho uld be parity in their trea tm en t only if the y are sufficientlyanalogous. Equals should be treated equally but not unequals.

For these reasons, the case for the cognitivity of religiousexperiences bas ed on langua ge-gam e f ideism w ith analogoustests is quite unconvincing. None of the objections that haveso far been made has escaped its defenders. We shall nowconsider their rebuttals to my objections. After showing thatthey have little force, the "big disanalogy" - the really decisive one - between the two doxastic practices or languagegames will be given.

T H E A N A L O G I C A L - A R G U E R ' S R E S P O N SE

The major objection to countenancing a religious-experienceanalogue to the sensory agreement test was that the former,unlike the latter, involved evidential asymmetry, as well asan inability to specify w ha t cons titutes a norm al observer an dstandard circumstances. Swinburne, Wainwright, and Alstonhave a response to it.

Swinburne does not think that the problems of evidentialasymmetry and the inapplicability of the normal-conditionsrequ irem ent can be rend ered completely harmless. They posea problem all right but only a small one:

If we do not know what experience would count against someperceptual claim (because we do not know which observers

317




could have been expected to have had an experience apparently of s if s had been there), that somewhat lessens theevidential force of an apparent perception - but only somewhat. This is because in tha t case we cannot have the confirming evidence of failure to find evidence which counts againstthe claim.34

The basis of Swinburne's "somewhat lessens the evidentialforce of" is ob scure . But this slight concessio n on Sw inb urn e'spart hides the seriousness of the challenge posed for the ana

logical argument by evidential asymmetry. The in-principleimp ossibility of disconf irm ing the verid icality of a religiousexperience by appeal to the testimony of other observers,rather than just lessening the probability of its veridicality,calls into ques tion the very applicability of the agre em ent testto it. If religious experience s cann ot flunk the ag reem ent test,the n th e test is no t applicable to them . A nd this is a far m oreserious matter than a mere lessening of the probability thatsuch an experience is veridical. Swinburne seems to haveforgotten that he is su pp os ed to be justifying the extension ofthe principle of credulity to religious experiences on theground that they are sufficiently analogous to sense experiences, especially in regard to their respective defeating conditions. For this purpose, it is not sufficient to show thatreligious experiences cannot flunk some test that sense experiences can, for this is to adm it that th is test doe s not app ly toreligious experiences, thereby undermining his analogical

argument. The same considerations hold for the admittedimpossibility of religious experiences running afoul of theproper-conditions tests 8-10, since such tests are not evenapplicable to them, thereby further undermining the analogical premise of this argument.

W ainwright and A lston adm it these disanalogies but try toneutralize them by claiming that they are due to a differencein the categoreal nature of the apparent object of sense andreligious experience. God is an infinite being who freelybestows his grace upon people according to no discerniblelaw or rule w he n h e causes them to have a direct p erception

318



The analogical-arguer' s response

of him, unlike physical objects that operate according todeterministic laws and therefore admit of fairly rigorous predictions of their behavior. This explains why the agreementand prediction tests will not app ly to religious expe rience s inthe same way that they do to sense experience. The mysticaltests, says Wainwright, although they must be "similar" tothe sensory tests so as to sustain the analogical argu m en t, areallowed to differ in their "n a tu re " bec aus e of the difference inthe natures of the apparent objects of mystical and senseexperiences. In one breath he says that mystical tests are

"similar to the tests which we employ in ordinary perceptualcases to determine whether an apparent perception of anobject is a gen uine p erce ption of it" and in the oth er tha t "th enature of the tests is not much alike."35

There are two difficulties with this response. First, toexplain why the tests for sense and religious experiences arenot analogous is not to explain away the disanalogies; andthese disanalogies are devastating to the analogical argument. This fundamental oversight also infects Alston's wayof responding not only to this disanalogy but to others aswell. For example, he responds to the claim that physical objects and God are disanalogous because only the former canbe picked out by ph eno m ena l qualit ies with a good n ew s- b adnews joke, though he does not realize it as such. He beginsby saying that the existence of this disanalogy is "the badnews," but then goes on to say, after he has explained whythere is this disanalogy in terms of the categoreal difference

between the nature of physical objects and God, that the"good news is that it explains why we would be in a positionof almost complete ignorance here even if there are basicphenomenal qual i t ies that make up the phenomenal character of divine appearances."3 6 This is like the doctor saying,"I 've got good news and bad news; the bad news is that youwill die of an untreatable cancer within six months and thegood news is that we know the reason why you have cancer." That there are these disanalogies destroys the analogical arg um ent, a nd in giving a con ceptu al expla nation of th em ,Alston is unwittingly doing the job of the critic of this

319




argument. To paraphrase Butler, everything is what it is, nomatter why it is what it is.

There is an oth er w ay in which Alston attem pts to neutralize the disanalogy based upon evidential asymmetry by appeal to the categoreal nature of the apparent objects:

But a more specific objection is that we have set things up sothat facts about the occurrence of religious experience areallowed to count only as positive evidence. About this I willsay the following, (i) This is true. Since we cannot hope toanticipate the patterns of the divine operation, God's nonrev-elation at any point cannot count as evidence against Hisexistence, but whenever He does reveal Himself that is a markin his favor. (2) But / didn 't set it up this way. This " set-up" isdeeply imbedded in the Christian . . . conception of G od.Now if the features of that concept responsible for this asymmetry were developed in order to protect the scheme fromdisconfirmation, this would be highly suspicious. But, andthis is the crucial point, the complete blend of positive andnegative theology that we have in this language-game developed, I believe, from quite other roots. Hence the evidential asymmetry is a consequence of integral features ofthe language-game that were basically a response to otherdemands.37

Why should the motivation behind a doctrine that falls preyto evidential asymmetry be relevant to excusing it from thecharge of fraudulence? The con game of "heads I win, tails

yo u los e" is objectionable as a disho nes t gam e even if its rule swere formulated not for the purpose of separating suckersfrom their money but instead from some lofty Spinozisiticnecessitarian principle that denies the reality of chance. Butthe fact that the author's intention in creating a game of"chance" that involves no chance is of this sort makes it noless of a con game.

Another difficulty with the Wainwright response is thatthere is a tension, if not a downright inconsistency, in theclaim that the tests are similar, but different in nature. If thetests do not have to be similar in nature, Wainwright's ana-

32



The analogical-arguer's response

logical argument degenerates into a version of the previouslydiscredited language game fideism with any old tests. Hereinw e see Wainwright fluctuating betw een a univocalist accou ntof cognitivity (the tests are similar) and a multivocalist one(the tests differ in nature).'8 The tests must be similar to support his analogical argument, but they must also be dissimilar in nature due to the categoreal difference between theapparent objects of mystical and sense experiences.39

There is a similar tension, if not inconsistency, in Alston'swo rk. It is be tw een his de m an ds , on the one ha nd , for parity

of treatment of the religious- and sense-experience doxasticpractices and, on the other, that we not be chauvinistic, thatwe not uphold the epistemological principles of one practiceas a standard by which to judge the adequacy of others.Alston's demand for parity of treatment of the two practicesis based on the claim that the religious-experience doxasticpractice is sufficiently analogous to the sense-experiencedoxastic practice so as to be subject to all of the cognitiverights and privileges thereunto appertaining to the latter. Insaying that he "wants to explore and defend the idea that theexp erience . . . of G od plays an epistem ic role with respect tobeliefs about God importantly analogous to that played bysense perception with respect to beliefs about the physicalworld,"40 Alston is espousing the analogical argument andthereby accepting the version of chauvinism that attributes aspecial sort of paradigmatic status to the sense-experiencedoxastic practice, such that a doxastic practice of basingbeliefs on experiences counts as cognitive only if it "exhibitsthe same generic structure as the former."41 T his requires thatits tests be analogous to those of the sense-experience doxastic practice, although they need not be in terms of senseexperiences, which would be an objectionable type of chauvinism. Therefore, it would be most inconsistent of him tocharge an objector with being an objectionable chauvinistw h en it is claimed tha t the religious-experience do xastic prac

tice fails to count as cognitive because its tests are quite dis-analogous to those of the sense-experience doxastic practice.Alston, at this point, might try to fudge the issue in the way

321



Religiom-exfefknee arguments

in which Wainwright did when he claimed that the tests ofthe two practices are "similar" yet different in "nature." Hecould say that the tests exhibit the same "generic" or "basic"structure but still are disanalogous in all the respects thathave been pointed out. But this is the same as the Wainwright posit ion, one that only Bob would find unobjectionable. Both authors begin by professing a version of languagegame fideism with analogous tests but, when pressured bysignificant disanalogies, make an unannounced retreat tolanguage game fideism with any old tests.

Alston also attempts to neutralize the challenge posed bythe disagreement among different religious traditions overthe correct tests, which supposedly has no analogue in thesense-experience doxastic practice, thereby rendering it aninvidious disanalogy because i t undermines the analogicalargument. His first response is to claim that experientiallybased disagreement among people undermines the reliabil i tyof their doxastic practices of belief formation only if theyshare a co m m on set of tests or ove rriders . Since this isn 't thecase in the disa gre em en ts am ong peo ple within differentworld religions, their experientially based disagreements donot undermine the reliability of their respective practices:

Since . . . each of the major world religions involves (at leastone) distinct doxastic practice, with its own way of going fromexperiential input to beliefs formulated in terms of thatscheme, and its own system of overriders, the competitorslack the kind of common procedure for settling disputes thatis available to the participants in a shared doxastic practice.Here, in contrast to the intra-doxastic practice cases, myadversary and I do not lack something that we know perfectlywell how to go about getting. Hence the sting is taken out ofthe inability of each of us to show that he is in an epistemicallysuperior position.41

Ag ain, we find A lsto n com m itting the fallacy of thin kin g th atif he can give a categoreally based explanation for a disanalogy between the religious- and sense-experience doxasticpractices, it re nd er s the disanalogy harm less. This sho uld be

322



The analogical-arguer' s response

cal led the "Als to n fa l lacy." H e un w i t t i ng ly is g iv ing u s ano the r good news-bad news joke . Wha t he f a i l s to see i s tha t

th i s d i sana logy concern ing a sha red se t o f t e s t s o r ove r r ide r su n d e r m i n e s h i s a n a l o g ic a l a r g u m e n t , w i t h it s a n a l o g i c a l -based jus t i f i ca t ion fo r accord ing pa r i ty o f t r ea tmen t to thet w o p r a c t ic e s , in p a r t i c u l a r fo r e x t e n d i n g t h e p r e s u m p t i v einference ru le of the la t te r to the former. We r ight ly doubt thecogn i t iv i ty of r e l ig ious exp er i en ces , in pa r t , bec aus e th os ew h o m a k e c l a i m s b a s e d o n s u c h e x p e r i e n c e s c a n n o t a g r e ea m o n g t h e m s e l v e s w h a t a r e t h e p r o p e r t e s t s f o r t h e i r v e r i d i -ca l ity, fo r w hic h the re is no ana log ue in the sen se -ex pe r i en cedoxas t i c p rac t i ce .

Als ton has a second way o f neu t ra l i z ing the d ive r s i ty o ft e s t s d i sa na lo gy tha t is p re se n te d in two s t age s . T he f i rs ta rg u e s t h a t t h e r e a r e actual cases in sc ience in which there i sa c o m p a r a b l e d i s a g r e e m e n t o ve r w h i c h t e s t s a r e a p p l i c a b l ewi thou t de t r imen t to the ep i s t emic l eg i t imacy o f the r iva lsc ien t i f i c p rac t i ces . He g ives the example o f the d i spu te bet w e e n b e h a v i o r i s t s a n d p s y c h o a n a l y s t s o v e r t h e n a t u r e a n d

p r o p e r t r e a t m e n t o f n e u r o s i s :

He re we are not so ready, or should not be so ready , to judg ethat it is irrational for the psychoanalyst to continue to formclinical beliefs in the way he d oes w itho ut hav ing non-circularreasons for supposing that his method of forming clinicaldiagnoses is a reliable one. Since we are at a loss to specifywhat such non-circular reasons would look like even if themethod is reliable, we should not regard the practitioner asirrational for lacking such reasons.43

H e n e x t i m a g i n e s a counterfactual s i tua t io n in w hich the rei s a d ive r s i ty o f sense -exper i ence doxas t i c p rac t i ces tha t a reana logous to the d ive r s i ty o f the r e l ig ious doxas t i c p rac t i ces .T h e s e c o m p e t i n g s e n s e - e x p e r i e n c e d o x a s t i c p r a c t i c e s " c o ns t rue wha t i s p resen ted in v i sua l pe rcep t ion in a r ad ica l lyd i ffe ren t way. "4 4 In o n e the ap pa re n t ob jec t i s des c r ibe d inW h i t e h e a d i a n t e r m s a s a c o n t i n u o u s s u c c e s s i o n of a c t u a loccas ions , in ano the r a s an Ar i s to te l i an subs tance , and in ye t

323




another as a Cartesian indefinitely extended medium that ismore or less concentrated at various points. These "alternatives . . . give rise to difficulties for the rationality of e ng ag ingin the sense-experience doxastic practice that are quite parallel to those arising from the actual diversity of religions forthe rationality of engaging in the Christian doxastic practice."45 In each of these analogous cases, the people can "justifiably engage" in their doxastic practice despite not beingable to show its epistemic superiority to its competitors. "Itmay seem strange that such incompatible positions could be

justified for different people, but this is just a special case ofthe ge neral po int that incom patible prop ositions can be justified for different people if what they have to go on is suitablydifferent."46

In the first place, it must be pointed out that what is atissue is not whether one may "justifiably engage" in the religio us do xastic practice or wheth er it is "rationa l" to do so , forthere can be pragmatic reasons for doing so, as we shall seein the next chapter. What is at issue is whether the practice isa knowledge-yielding or cognitive one.

Second, both the actual and counterfactual examples ofrival doxastic practices are irrelevant to the point at issue,since they are not doxastic practices for forming experien-tially based beliefs. They involve instead the activity of formulating scientific or philosophical theories to explain certain empirical facts that are common to the rival theories.These rival practices are incommensurable exactly because

the y accept the sam e empirical facts. There is no disa gre em en tamong their practitioners about the existential language-entry assertions that are to be made on the basis of senseexperience, nor over the tests for determining their veridi-cality; and, thereby, these cases are not analogous to thedis ag ree m en t am on g rival religious-experience doxastic practices, in which the disagreements concern both the correctlangu age-e ntry respo nses to religious experience inpu ts andthe test for determining their veridicality.

The Aristotelian, Cartesian, and Whiteheadian, for example, d o participate in the same sense-experience doxastic prac-

324



The aMtOgical-arguer's response

tice. They make the same physical object assertions whensubject to the same sensory input. Where they disagree isover the way in which such assertions are to be analyzed.Even Berkeley, who thought physical objects were congeriesof ideas in different minds, insisted that it was correct to saythat we perceive tables, chairs, and the like. He was not challenging our ordinary doxastic practice of asserting physicalobject propositions on the basis of sense experience butinstead offering a philosophical analysis of what these propositions mean. The same is true for the other philosophers in

question. Thus, their disagreement is over the correct philosophical analysis of physical object propositions. Whethertheir activity of philosophical analysis is a cognitive one is acan of worms that will be left unopened, but, it might benoted, one does not have to be a Richard Rorty to view theirfailure to agree upon the proper method to employ in doingphilosophy as calling into question its cognitivity.

Alston's analogy with the dispute between behavioristsand psychoanalysts over the nature and genesis of neurosisfails for the same basic reason, the dispute now concerningthe proper scientific rather than philosophical theory toadopt. Again, we find considerable agreement in the sensory-based language-entry assertions of the rival theoristsconc erning w h e n there is a case of neu rosis, such as som eon ewho is a compulsive hand washer or is unable to walk acrossan open area.47 Without such agreement about the empiricalfacts, their theories would not be incommensurate theories.They offer different theories about the cause and cure ofthese agreed upon neuroses. And, most important , they arein basic agreement about the tests that are relevant to resolving their disagreement. That Freudians have been lax if notunconscionable about putting their theories to empirical testand have been too quick to take certain clinical data as confirmatory of their causal theories of the etymology of neurosesdoes not show that they disagree with their behaviorist rivals

over the proper tests. We must distinguish between themethods and tests that inform the scientific doxastic practice,the ones that are actually accepted by scientific practitioners,

325



Religious-experience argumen ts

and the correctness with which some practitioners followthem . Ano ther can of w orm s that I will leave un op en ed concerns whether psychoanalysis, as Ricoeur and Habermasclaim, is not to be viewed as a scientific theory at all. For athorou gh discussion of these issues one can do no better th anto read Adolf Grunbaum's monumental The Foundations ofPsychoanalysis.

" N E A R E R M Y G O D T O T H E E "

The conclusion to be drawn is that language-game fideismwith analogous tests is in serious difficulty and that theresponses to the various disanalogies that challenge it by thedoxastic practice theo rists or lan gu age -gam e f ideists d o littleto neutralize th em . B ut the case is even wo rse for this v ers ionof fideism that has been made out so far. We have yet tounearth a deep disanalogy between sense and religious experience that will totally destroy the analogical premise of itsanalogical argument. This "big disanalogy" will prove to bethe shipwreck of this defense of cognitivity, a time for Alston,Gu tting, Sw inb urn e, an d W ainw right to join their fellow an alogical arguers on the deck for a few heart-felt choruses of"Nearer My God to T he e," tho ug h Bob, I kn ow , will be u n impressed, remaining below deck sipping a Lablatt 's, his lastwords being "This is just aboot l ike any other shipwreck."

Necessarily, any cognitive perception is a veridical perception of an objective reality. It now will be argued that it isconceptually impossible for there to be a veridical perceptionof God (the undifferentiated unity, etc.), from which it follows by modus tollens that it is impossible that there be acognitive religious experience. My argument for this is ananalogical one that, like those for the cognitivity of religiousexperiences, takes sense experience to be the paradigmaticmember of the analogy. A veridical sense perception musthave an object that is able to exist when not actually per

ceived and be the common object of different sense perceptions. For this to be possible, the object must be housed in a

326



"Nearer my God to Thee"

space and time that includes both the object and perceiver. Itthen is shown that there is no religious experience analogueto this concept of objective existence, there being no analogo us dim ension s to space an d t ime in which God , along withthe perceiver, is housed and which can be invoked to makesense of God existing when not actually perceived and beingthe common object of different religious experiences. Becauseof this big disanalogy, God is categoreally unsuited to serveas the object of a veridical perception, whether sensory ornonsensory.

In arguing that it is impossible for there to be a veridicalreligiou s exp erien ce of an objective reality, I am no t en ga gin gin an objectionable form of chauvinism by requiring that thesort of objective existence enjoyed by the objects of veridicalsense experiences, physical objects, hold for all objectiveexisten ts. I am h ap p y to gr an t tha t there are objective realitiesthat do not occupy space and/or time nor any analogousdimensions, such as the denizens of Plato 's nonspatiotempo-ral heaven; and God might very well be among these objectively existent abstract entities. What is impossible is thatthere be a veridical perception of one of th em , ev en of th eintellectual sort described by Plato in the Phaedrus, accordingto which we "s ee" them with our m ind's eye as we m ight seebigger-than-life Macy Day Parade balloons - "Look, comingup Broadway, is triangularity and behind it is the Idea of theGood in itself. Only those who have veridically "perceived"them are fit to be philosopher-kings. It is notorious thatwould-be philosopher-kings disagree among themselves asto the nature of these "balloons" and the proper methods forcorrectly apprehending them, which parallels the disagreements between different religious-experience doxastic practices as to the true nature of the object of these experiencesan d the correct me th od s an d tests for gaining a veridical perception of it. "Catch one of them and you catch all of them "

So far we have been working with the analogical arguer'sview of the sen sory tests as forming a cluster, such that s om e

of them might have a religious experience analogue and others not. This is a fund am entally m istake n view of these tests .

327




Tests 5-11 are an organic unity of conceptually interrelatedtests that presuppose the same concept of a veridical senseperception of an objective reality, namely, that of a commonspace-time in which the subjects and objects of veridicalsense experiences causally interact in a mostly deterministicmanner and that permits these objects to exist when notactually perceived and be common objects of different perceptions, as well as to be individuated by their positions inspace-time.

The agreement test 's requirement that the different obser

vations be m ade u nde r norm al or stand ard co nditions, w hichencompasses the requirements of tests 8-10 that the observerbe physiologically and psychologically normal and properlypositioned in space and time, are fillings in for the caused-in-the-right-way requirement of test 11. Our concept of the sortof objective reality that is the object of a veridical sense experience sup plies an explanation of the fact of agreem ent un de rstandard conditions in terms of a common cause - an objector event tha t is the initial m em be r of the severa l causal cha insthat terminate in these different experiences. The notion of acom mo n cause also und erlies tests 6 an d 7, wh ich thereby arebranches of the same conceptual trunk as these other tests.This com m on cause is also invoked to acco unt for the law likecontinuity between the contents of a single person's successive sense experiences as well as the coherence between theway in which different observers are perceptually appearedto at their different spatiotemporal perspectives. Thus, tests

5-11 all involve the same worldview - that of a commonspace-time receptacle in which the objective accusatives ofveridical sense experiences are the common causes of thenom ic-type coheren ce am on g the sen se experiences of the differently positioned perceivers.48

The ultimate ground of individuation for the objectiveempirical particulars in this worldview is supplied by spaceand time, it being a necessary truth that no two empirical

individuals of the same kind spatiotemporally coincide. It isthe space-time receptacle that creates the possibility of therebeing cou nterex am ples to the principle of the identity of indis-

328



"Nmrer my God to Thee

cernibles when restricted to fully general or descriptive properties. Any such property admits of the possibility of multiple instantiations at different regions in this receptacle. Andthis in turn creates the possibility of distinguishing betweennumerically and qualitatively identical particulars. And as aconsequence we are able to distinguish between perceptionsthat are of numerically one and the same particular and thoseof particulars that are only qualitatively similar. In the lattercase there are noncoincident particulars that are hooked upwith different perceptions via different causal chains.49

Before comp leting my analogical arg um en t by d em on strating that none of the features of this worldview of objectiveempirical particulars could have any analogue in regard torehgious experience, I will a t tempt to enrich and deepen ourunderstanding of this view by showing why its objectiveparticulars m ust be ho use d in both space an d time. While noone would question the need for a temporal dimension,some would deny that objective particulars of this kind must

be located in space as well. A persuasive case for this notbe ing necessary has been m ade ou t by P. F. Straw son in hishighly original and imaginative chapter on "Sounds" in Individuals. Through a detailed criticism of his effort I hope toshow that such particulars must be located in space as wellas time.

The question whether objective empirical particulars mustbe in space is given an epistemic twist by Strawson in termsof his general principle that we cannot have a concept of acertain type of individual unless we have experientiallygrounded criteria of identity for individuals of that type.Accordingly, it turns into the question of whether criteria ofidentity for objective empirical particulars could be formulated in terms of sense experiences that are not representative of things as being in space.50 He puts aside visualexperiences, since they seem to be representative of space, atleast two-dimensional space. Kinaesthetic and tactual experi

ences also seem to be representative of space, and there canbe no doubt that in combination they are. Olfactory andgustatory experiences, neither alone nor in conjunction, are

329




representative of space, but Strawson chooses not to workwith them since he thinks there is little hope of establishingthe possibility of a world of objective smells or tastes.

Jt is auditory experience that he thinks hoJds out the bestchance for success. He argues, convincingly in my opinion,that auditory experiences alone are not repre senta tive ofspa ce. O ur ability to locate a sou nd in space on the basis of anauditory experience is dependent (in some way that is leftunspecified) u po n correlations betw een auditory experiencesand other kinds of experiences of a visual, kinaesthetic, or

tactual sort. This ability often depends upon having knowledge of a sound's cause, but such knowledge is not basedupon auditory experiences alone.

It is Str aw son 's aim to sh ow th at criteria for objective so un dparticulars can be form ulated exclusively in terms of a ud itoryexperiences, thus establishing the logical possibility of therebeing a no-space world composed of objective sounds. Thesesounds would resemble commonsensical sounds in that bothcan occur when not actually heard. (A tree that falls in a forest devoid of sentient beings does make a loud noise, unlessit falls in to a sn ow ban k o r pile of leaves.) Th ey differ fromcommonsensical sounds only in not being in space.

According to Strawson our purely auditory criteria of identity for objective sound particulars must enable us, first, todistinguish numerically between sounds, even in cases inwhich sounds are qualitatively the same, and, second, to re-identify sounds, even in cases in which there has been a

lapse in our perception of them. The latter is crucial to ourascription of a m ind -ind ep end en t status to the sou nd s of theno-space world. No mere regularity in uninterrupted auditory expe riences w ou ld suffice. If w e were continually to h eara repet i t ion of an A-B-C-D sequence of sounds, we wouldbe able to distinguish numerically between an occurrence ofone A-type sound and an earlier or later occurrence of thissame type of sound, but we would have no grounds forholding that these sequences of sounds could occur evenwhen not actually heard.

33



"Niarer m^ God to Thee"

To m a k e s e n s e of a s o u n d o c c u r r i n g w h e n u n p e r c e i v e d , t h en o - s p a c e w o r l d o f s o u n d s m u s t b e g i v e n s o m e d i m e n s i o nana logo u s t o spa ce s o t h a t we can s ay whe re a s o u n d i s wheni t f ai ls to be he ar d . "W e m u s t ha ve a d i m en s io n o th e r th ant h e t e m p o r a l i n w h i c h t o h o u s e t h e a t p r e s e n t u n h e a r d s e nsory par t iculars , i f we are to g ive a sa t i s fac tory sense to theidea o f the i r ex i s t ing now unperce ived , and hence to the ideaof re iden t i f ica t ion of pa r t i cu la r s in a pure ly a ud i to r y w or ld . "5 1

T h i s w o u l d m a k e t h e n o - s p a c e w o r l d o f s o u n d s a n a l o g o u s t oou r o rd i n a r y s p a t i a l w o r ld f o r wh ic h

the crucial idea is that of a spatial system of objects, throughwhich oneself, another object, moves, but which extendsbey ond the limits of one's o bservation at any m om ent. . . .This idea obviously supplies the necessary non-temporaldimension for, so to speak, the housing of the objects whichare held to exist continuously, though unobserved.5 2

T h e n e c e s s i t y fo r a n o n t e m p o r a l d i m e n s i o n in w h i c h u n p e rce ived s o u n d s a r e hous e d can be e s t ab l i shed i n a way t ha tS t r awson doe s no t exp l i c i t l y cons i d e r, a l t hough he s a y sth in gs tha t sug ge s t it . Th i s con s i s t s in sh ow in g a fatal f law inthe most l ike ly way of making out a case for the logica lposs ib i l i ty o f the re be ing ob jec t ive sound par t i cu la r s tha toccur on ly in the d imens ion o f t ime , th i s cons i s t ing in theapp l i c a t i on of a we l l - f oun ded i n d u c t i ve g e n e r a l i z a t i on t o acase in w hic h the re is a ga p in pe rce p t io n . I t w or ks in th efo l lowing way.

Le t u s su p p o se t ha t o n n n u m b e r of o c c as io n s a n A - B - C - Dsequenc e o f s o u n d s i s hea rd and no sound o f any one o fth es e type s i s ev er he a r d excep t in su ch a se qu en ce . Th i sobse rved r e g u l a r i t y j u s t i f i e s t he i nduc t i ve g ene ra l i z a t i on t ha teve ry A -B s e qu en ce of s o un ds i s f o l lo we d b y a C - D seq u e n c e . O n o c c a s i o n n+i, a n A - B - E - D s e q u e n c e is h e a r dr a t h e r t h a n a n A - B - C - D s e q u e n c e . T h e q u e s t i o n i s w h e t h e rC occur red on th i s occas ion even though no t ac tua l ly heard .

It i s c l a im ed tha t ou r ab ov e ind uc t iv e genera l i za t io n jus t i f ie san a f f irma t i ve a ns w er , an d t he r e by a c a se ha s bee n m a d e ou tfor the poss ib i l i ty of sounds exis t ing unperceived even though

33i




not housed in any nontemporal dimension analogous tospace.

The obvious rebuttal to this arg um en t is that the n + 1 occasion, the one one which sequ ence A- B -E -D was heard, constitutes a counterinstance to the inductive generalization thatevery A-B sequence is followed by a C-D sequence. Howmight this seeming counterinstance be neutralized? Thereare two ways in which seem ing co unterinstance s to empiricalgeneralizations are met. It will be seen that neither is of anyavail in this case.

The first way is to show that a perceptual error occurred:On occasion n + i, C was misheard as E, or the observer onlythou gh t she hear d E, it be ing C that wa s really hea rd. We candetermine that a perceptual error occurred only if we havecriteria for determining what objectively is the case. But it isjust such criteria tha t ar e lacking in th e abo ve case: We can no tappeal to the empirical generalization that every occurrenceof A -B is followed b y C - D to establish th at a perc eptu al erroroccurred on occasion n + 1 since this w ould rend er th e " em

pirical" generalization unfalsifiable by experience. It is thisgeneralization that is supposed to give us a criterion fordetermining what objectively is the case, but it seems to facea counterinstance.

The second way of meeting a seeming counterexample isto show that the observer was not in a position to observewh at was there to be perceiv ed; for instan ce, sh e or the objecthad moved away, she was looking in the wrong direction,her view was blocked, and the like. The fact that on someoccasion I saw a burning log and later saw ashes in place ofthis log does not refute the generalization that a log burnsdo wn continu ously into ash es, prov ided th at I had m ove daway in the interim or ha d m y view of the log blocked, or th elike. It is apparent that all of these explanations invoke aspatial dimension in which are housed coexistent objects,one of which is an observer capable of moving in relation tothese other objects. Since the above no-space world contains

no nontemporal dimension that houses coexistent sounds,this sort of explanation cannot be offered. Without such a

332



Nearer my God to Thee"

dimension, there is no way our inductive generalization canbe saved from the seeming counterinstance.

If the above line of reasoning is correct, it bears out Straw-son's claim that a no-space world oH objective sounds mustcontain some nontemporal dimension in which coexistentsounds are housed. But for exactly what features of spacemust analogues be found in this nontemporal dimension?According to Strawson it is not simply that of distance, sothat one sound can be said to move nearer to another or befurther away from one sound than it is from another, but that

of an observer, who is one sound among others and canmove in relation to these other sounds. This would make theanalogy with the way in which material objects fill spacerather precise in that the observer is a material object or atleast has a body that is one object among others and canmove in relation to these objects. This would enable us todispose of the seeming counterinstance on occasion n +i bysaying that the observer - the hea rer - had m ove d aw ay from

the place in the nontemporal dimension at which sound Coccurred and therefore wasn't properly positioned to hear C,which is a species of the second way of meeting a seemingcounterinstance to an empirical generalization.

The problem is how we can derive from the properties of asound alone a nontemporal dimension having the aboveanalogous features to space. There are three properties ofsound: pitch, volume, and timbre. Only pitch and volumedetermine a serial ordering, and either of them could beselected to determine this dimension. Strawson elects towork with pitch, no doubt because there are rich spatialmetaphors concerning pitch; for instance, we speak of thepitch of one note as being closer to that of another than it isto that of some third note, of one note becoming closer inpitch to an oth er, of higher an d lower pitch es. Such analog iesbetween pitch and space will be familiar to students of theSchillinger system of musical composition in which melodies

are represented on graphs.Strawson concocts the following description of a possible

sound world in which pi tch alone determines the nontempo-

333



Religious-experience Ktgummts

ral dimension in which objective sound particulars arehoused. Auditory experience contains a sound, called the"m aster so un d, " w hich is un ique in i ts t imbre and continuity.The sound varies in pitch, and when it is at certain pitchesoth er sou nd s are hear d, so m e of wh ich have the kind of unitypossessed by a musical composition. For instance, at pitchlevel L of the m aster s ou nd , a un itary m usical com position Mis he ard , w hile at level L ' a different composition M ' is heard.W hen the ma ster so un d rem ains at L, composition M can behe ard in its en tire ty, as well as repetitions of it. As the m ast ersound gradually moves up or down in pitch from L, there isa gradual decrease in the volume of M, and as its pitch approaches close to L, there is an increase in the volume of M.

The master sound serves as the surrogate in the soundworld for the body of the observer in the spatial world, itspitch being the position from which the observer hears othersounds, and changes in i ts pitch representing her change ofposition in th e pitch dim ensio n. Th is can serve as an e xplanation of why certain sound sequences cannot be heard at certain pitches of the master sound, that is, at certain positionsin the pitch dimension. In saying that the pitch of the mastersound gives the position of the observer, we do not mean toimpute self-consciousness to her.53 It is we, the storytellers,who talk about her position and change of position, not theobserver who does.

The sou nd s or so un d s equ ence s that are heard at the different pitch levels of the master sound qualify as objective par

ticulars because there are criteria for reidentifying them whenthey are no t contin uo usly hea rd, as well as for disting uish ingbetween qualitative and numerical identity within the pitchdimension itself and not just within the time dimension.Imagine a case in which the master sound is at pitch-level Land M is being heard:

Then suppose the master-sound changes fairly rapidly inpitch to level L' and back again to L; and then M is heard oncemore, a few bars having been missed. Then the sounu-

334



'Nearer my God to Thee"

particular now being heard is reidentif ied as the same particular instance of M. If during the same time, the master-sound

had changed not from L to L' and back again to L, but from Lto L", then even though M may be heard once more, a fewbars having been missed, it is not the same particular instanceof M that is now heard, but a different instance.54

In the final sentence Strawson commits himself to the possibility of the principle of the identity of indiscernibles beingcou nterexa m pled within the pitch dim ens ion, which is to say

that we can draw the distinction between qualitative and numerical identity within this dimension. Just as the samepurely descriptive property can be instanced at two or moreplaces in space at the same time, the sam e so un d universal orpattern of universals can be instanced at two or more places(pitches) in the pitch dimension at the same time.

Th ere is a m inor flaw in Straw son 's acc ou nt. It is clear thatthere is a need for some kind of clock. This comes out in hisclaim that the same particular instance of M is heard only ifthe master sound changes "fairly rapidly" from L to L' andback again. He has not provided the observer with any wayof determining when such a change occurs rapidly enough.This difficulty can be circumvented by imagining a JimmyDurante-type of observer who has his own inner clock orstandard of temporal congruence that is marked by his continual hearing of "boop-boop-bee-doop."

Has Strawson's ingenious account of his wonderful world

of sounds shown that there could be objective empirical particulars that are not in space? Two different kinds of arguments will be advanced to show that he has not. The firstchallenges Strawson's unstated assumption that there areexclusively auditory criteria of identity for sounds in virtue ofwhich it can be determined both when a sound persists andw h e n it is distinct from som e coexistent so un d . The second isbased upon there being crucial conceptual disanalogies be

tween the way in which physical individuals occupy spaceand the way in which sounds occupy the pitch dimension

335




that renders the latter dimension logically incapable of housing objective particulars.

Strawson's account assumes that we have purely auditorycriteria for the persistence of a sound as well as for distinguishing between coexistent sounds. This is quite dubious.W hat criteria w e have are not only imp recise bu t also seem torequire a reference to the causal source of a sound, and thusinvoke knowledge that is not gained from auditory experience alone. If w e have a co ntin uo us a nd phenom enologicallyinvariant auditory experience, do we hear numerically one

and the same sound particular throughout this t ime? Theonly way of answering this question is by reference to thecausal source of our auditory experience. At the old SavoyBallroom, there was a large ba nd stan d on which two bandscould be seated. The music went on continuously; when oneband stopped playing, the other immediately began to play.I remember once hearing Count Basie 's band playing the"Two O'Clock Ju m p," and w he n the y end ed with a sustainedchord, without any perceptible break, Cootie Williams's bandbegan to play this chord. On this occasion I heard two phenomenologically similar sounds, the first of which was produced by Basie's band and the second by Williams's. Mycriteria of identity are based on the causal source of my auditory experience, but obviously my knowledge of such causesis not based on auditory experience alone.

A similar problem arises for sustained auditory experiencein which there is phenomenological alteration. Strawson as

sumes that the master sound persists even though it changesin its pitch. But this is gr atu itous since w e ha ve n o criteria fora sound 's persisting that are ind ep en de nt of i ts causal source.The glissando we hear remains one and the same soundparticular throu gh ou t if it is pr od uc ed by som eon e sliding h isfinger up a violin or a guitar string. But the glissando wouldbe composed of numerically distinct successive sounds if itwe re instead pro du ced by a very ra pid p laying of a chromaticscale on a piano.

It is also dubious that there are purely auditory criteriafor distinguishing between simultaneously heard sounds. It

336



"Nearer my God to Thee"

would seem that our ability to distinguish between the notesof a heard chord ultimately depends on our knowledge ofsome nonauditory feature of the situation in which it is produced. Jf this be so, Strawson has no right to assume that themaster sound is numerically distinct from the sounds that arealleged to accompany it at its different pitches.

I suspect that the reason Strawson failed to see the dependency of our criteria of identity for sounds upon their causalsource is that he leaned too heavily upon the radio analogyby which he first introduced the master sound. The master

sound is said to be analogous to "the persistent whistle, ofvarying pitch, which, in a wireless set in need of repair,sometimes accompanies the programmes we listen to."55 Theway in which we hear different sound sequences at differentpitch levels of the master sound is compared to gradually"tu nin g-o ut o ne station and tuning-in ano the r. . . . On ly . . .instead of a tuning -kn ob being gradually turn ed, we have thegradual alteration in the pitch of the master sound."5 6 Ourability to distinguish between the persistent whistle and theprog ram s that ac com pan y it, as well as to distinguish be twe endifferent pro gra m s, is based on kno w ledg e of ho w thesesounds are produced.

The preceding arguments against Strawson's wonderfulworld of sounds are not decisive because our criteria of identity for sounds are so imprecise. Some arguments basedupon crucial conceptual disanalogies between space and thepitch dimension will now be advanced. And they are decisive. What will make these disanalogies crucial is that theyconcern the very grounds of individuation for objective empirical particulars.

The first crucial disanalogy to be argued for is that whereasit is possible for the identity of indiscernibles, when restricted to purely descriptive properties, to be counterexam-pled within the spatial dimension(s), it is not possible for it tobe counter^xampled within the pitch dimension. In other

words, the distinction between qualitative and numericalidentity has a possible application within the spatial dimension^) but not within the pitch dimension. Strawson would

337




deny that there is this disanalogy. In the lengthy quotationcited, he claim s that if M is hea rd at pitch-level L of the m aster

sound, which then changes to pitch-level L " at which M isheard "it is not the same particular instance of M that is nowheard, but a different instance." Although these two token-ings of the sound-type M are qualitatively identical, they arenumerically distinct because they occur at different places(i.e.,pitches) in the pitch d im en sion . It will be illum inating tosee why there could not be any such use of the distinctionbetween qualitative and numerical identity within the pitchdimension.

The tok ening of M at L is su pp os ed to be diverse from thatat L " because they occur at different positions in the pitchdimension, which would make the grounds of individuationfor sounds quite analogous to that for coexistent spatial individua ls. But w her eas S trawson tells us how to determ ine theposition in this dim ension of the ma ster soun d, h e says nothing about how w e are to deter m ine the position of a non -m aster sound.57 If the position of the master sound at any giventime is to be determined by its pitch, we should expect thesam e to hold tr ue for all so un ds . Since, ex hypothesi, the token-ings of M at L and L ' ' have all their purely descriptive properties in common, they have the same pitch (or are in thesam e key), and the reb y are at the same position. If we sho uldpersist in holding th at, neve rtheless, there are two tokeningsof M, we would be working with a principle of individuationfor sounds that is radically disanalogous to that for spatial

individuals, for which it is necessarily true that sortally similar diverse individuals are not spatiotemporally coincident.Even if we were to hold that only the master sound's posi

tion is determined by its own pitch, the position of everyother sound being determined by the pitch level of the master sou nd from w hic h it is he ar d, the principle of the identityof indiscernibles still could not be counterexampled withinthe pitch dimension. The tokenings of M are supposed toconstitute a cou nterex am ple to this principle. Although theyhave all their purely descriptive properties in common, theynevertheless are diverse because they occupy different posi-

338



Nearer my God to Thee

tions, this being due to their being heard at different pitchlevels of the master sound. They do not, however, have alltheir purely descriptive properties in common, since theyhave different relational properties; one has the property ofbeing heard at pitch-level L of the ma ster so un d wh ile theother had, instead, the property of being heard at pitch-levelL". It is no t only the m aste r sou nd that can be tok ene d o nlyonce within the pitch dimension, but every sound type aswell.

The disanalogy between a spatial dimension and the pitch

dimension consisting in the principle of the identity of indis-cernibles being capable of counterexample within the formeralone is a con sequ enc e of an even mo re fund am enta l con ceptual disana logy betw een the two. The pi tch dim ension , unl ikea spatial dimension, is an empirically determined dimensionin that i ts parts or positions are generated or determined bydifferences in the empirical properties, namely, pitches, theyinstantiate. In other words, the principle of the identity ofindiscernibles is necessarily true of positions in the pitchdimension but not of places in space. Whereas diverse positions in the pitch dimension cannot have all their purelydescriptive properties in common, diverse places in spacecould, for instance, the places occupied by two qualitativelyindiscernibles d iscs. This, of cours e, is no t to say that div erseplaces never differ in their properties, for one place mighthave the property of being occupied by a disc and anotherplace not have this relational property. Since the principle of

the identity of indiscernibles is not necessarily true of theparts of space, it is not necessarily true of the objective particulars in space.

This more fundamental disanalogy between space and thepitch dimension explains why only the former can accountfor the ultimate grounds of individuation of coexistent objective em pirical partic ulars . To acco unt for the ultim ate g ro u n d sof the diversity between such particulars, we must invoke

some realm or dimension(s) within which they can simultaneously occupy different positions. If this realm is an empirically determined one, the positions in i t are themselves

339




objective empirical particulars; for instance, the positions inthe pitch dimension are not universals - pitch types - butobjective particulars that instantiate these types. But then wewill hav e failed to accoun t for the g ro u nd s of ind ividu ation ofthese objective particulars. Obviously, we are launched on avicious infinite regress if we try to account for their individuation in terms of their occupying different positions in someother empirically determined dimension(s). Since the spaceof the receptacle is not an empirically determined realm,being nothing but pure extension, it can serve as the groundof individuation of objective empirical particulars. This is insubstantial agreement with Plato's somewhat poetical account of the receptacle:

Hence that which is to receive in itself all kinds must be freefrom all characters . . . that which is duly to receive over itswhole extent and many times over all the likenesses of theintelligible and eternal things ought in its own nature to befree of all the characters. For this reason, then, the mother andReceptacle of what has come to be visible and otherwise sensible must not be called earth or air or fire or water, nor any oftheir compounds or components.58

This completes my account of the concept of an objectiveempirical particular, that is, one that is a suitable object of averidical sense experience. Hopefully, it was established thatfor such an object to be capable of existing when not perceived and being a common object of different perceptions,as well as being individua ted in a way th at m akes roo m for adistinction betw een n um erical an d qualitative identity, it ha dto occupy both space and time. It will now be argued thatno ne of these features of the conce pt of an objective em piricalparticular have any religious-experience analogues, therebycompleting my analogical argument.

W hereas it is necessary th at the a pp are nt object of a veridical sense experience must be housed in the dimensions ofspace and time, there are no analogous dimensions in whichthe ap pare nt object of a religious experience c ould be ho us ed .The ap par ent object of a unitive religious experience, the un -

34



Nearer my God to Thee"

differentiated unity, obviously, does not occupy any anal

ogous dimensions to these; and, al though God could, and asargued in the first three chapters ought, to be conceived asin t ime, he does not occupy any analogous dimension tospace, such that we could be said to draw nearer or furtherfrom God in this dimension, "Nearer My God to Thee" is ametaphor.

Because these objects are nondimensional, they will bedisanalogous to empirical particulars in several importantrespects. First, they will have radically different grounds of

individuation. Whereas empirical particulars are individuated by their position in nonempirical dimensions, they areno t. They are indiv idua ted instead by their prop erties - tho sethat they have both essentially and uniquely across worlds -in a way that satisfies the principle of the identity of in-discernibles. As a result, the numerical-qualitative identitydistinction has no application to them. That their grounds ofindividuation, in virtue of their nondimensionality, are soradically disanalogous to those for the objects of veridicalsense experiences does not disqualify them as objective particulars, but it does disqualify experiences of them as veridical perceptions of objective p articula rs. I am no t be ingchauvinistic about the concept of objective existence.

An other invidious consequen ce of their non dime nsionali tyis that no analogous explanation can be given for how theycan exist unperceived and be common objects of differentperceptions to that which was previously given for empirical

particulars. Whereas we could explain our failure to perceivean empirical particular, as well as our perceiving numericallyone and the same empirical particular, in terms of our relationship to it in som e nonem pirical dim ension , no such analogous explanation can be offered for our failure to perceiveGod and the like, or our perceiving numerically one and thesame God. This means that it is impossible in principle todistinguish between, for example, mystical experiences thatare of numerically one and the same undifferentiated unityand the like and those that are of merely qualitatively similarones . Stace's argument for the numerical identity of the pure




souls of different mystics clearly reveals this. He argues thatthe undifferentiated unities experienced by different mystics in introvertive mystical experien ces are qualitatively indiscernible. He then infers, by appeal to the principle of theidentity of indiscernibles, that they are numerically identical. This is the "deep" it-is-time-to-start-singing-hymns dis-analogy.

Similarly, how is Gutting going to decide when two religious experiences of a very powerful and loving nonhumanperso n had by tw o peo ple at on e t ime or by one person at tw o

different times are of num erically o ne an d the same being oronly qualitatively similar ones? People in our epistemic circumstances - we human beings - are not able to vouchsafethe uniqueness of the object by its phenomenological traits.To kn ow that one is exp eriencin g G od, a nd not just an y oldvery powerful, loving, nonhuman person, of which therecould be several, it is necessary to know that it completelyand solely de term ine s every feature of th e world. But thisrequires knowing the negative fact that there does not existan y being oth er tha n it that de term ine s an y of these features.Swinburne, who restricts his principle of credulity to posit ive-seeming experiences,59 thereb y wi nd s up being inconsistent when he applies it to of-God experience since they are,in part, negative for this reason.

A third consequence of the dimensionality disanalogy isthat whereas the worldview into which empirical particularsenter offers an explanation for the organic unity of the sen

sory tests, there is no analogous view into which the apparent object of religious experiences enters that explains howthe various religious-experience analogues to these tests (assu m in g that the re are any ) are intercon necte d. It w as thecommon cause theory, based upon a receptacle that housesbo th objects an d sub jects of sens e experience in a m ostly n om -ically determined system, that ultimately explained why tests5-11 worked together in harness, but no such explanationcan be given for the interconnection of religious-experiencetests. They form a mere heap, with no rhyme or reason whythey should go together and serve to mutually reinforce each

342



Nearer m y God to Thee"

other. No doubt Alston will be ready to explain away thisdisanalogy, as well as the preceding ones, in terms of thecategoreal differences between God and empirical particulars, but it would be just another instance of the Alstonfallacy.

Summation. Th e case for religious experiences being cognitive is exceedingly weak, and thus they cannot serve as anepistemological or truth-directed reason for believing thatGod exists. Even though it is impossible to have a veridical

religious experience, that is , nonsensory perception of God, itdoes not follow that an of-God-type experience could not becaused by God in the right way and thereby qualify as somekind of nonperceptual apprehension of God. It just wouldn'tbe analogous enough to a veridical sense experience so as toqualify either as a veridical perception of God or as evidenceor epistemic warrant for believing in the existence of itsapparent object. Because we lack a truth-directed rea son forbelieving in the existence of the apparent object of theseexperiences - God - it does not follow that there could be noreason for accepting them as revelations of an objective reality; for there are, in addition to truth-directed reasons, pragmatic ones concerning the benefits of so believing. To trustsuch expe riences as objective revelations m ight p rov e to be ofthe greatest possible benefit to the experient, ena bling him tolive a richer and more meaningful life. The prediction test, asapplied to religious experiences, concerned the moral and

spiritual growth of the experient; and, if it should be the casethat suc h desirable gro wth occurs only if the experient trus tshis experience as an objective revelation, then h e migh t h av ea valid pragmatic reason for so believing. This is the topicthat will be our concern in the next chapter.

343



Chapter 9

Pragmatic arguments

So far only epistemological or truth-directed reasons havebeen considered for believing that God exists. It is now timeto consider pragmatic reasons based on the desirable consequences of belief. Th ere is the big qu estio n, w hich will no t becon sidere d, of wh eth er ind ivid uals, a s well as society at large,are better off having religious faith. While religion and thefaith it insp ires h as a mixed track record, hav ing been am on g

the causes of man's greatest and basest deeds, we have nowa y of dete rm inin g w heth er thin gs overall wo uld have gon e,or will go, better with it than without it. My interest, however, is not with questions of empirical fact but with theconditional question "If overall there were to be desirablecon seq uen ces of believing that God exists, wo uld that justify,that is, constitute a sufficient reason for, one's believing?"

The desirable consequences can take different forms andthereby serve as the basis for different types of pragmaticargument . They can be prudentiqlly desirab le if they satisfythe needs, wants, or interests of the believer. Or they mightbe desirable on moral grou nd s either by enabling us to engagein the practice of morality, that is, following moral rules evenw h en d oin g so sacrifices our ow n self-interest, or m aking thebeliever and /or h er society mo rally better. K ant's moral argum en t for believing is an exam ple of the former: The de m an dsof morality to follow the categorical imperative make sense

only if we assume that there exists a God who will bring itabout that those who do so are rewarded with happiness ,something that rarely happens in this life. Because this justi-

344



Pragmatic argum ents from prudence

fication of faith take s us so dee ply in to the n at u re of mo rality,it will not be possible to consider it in this book, nor am I upto tackling it. Given that it is rational or reasonable to promote one's own interests as well as those of morality, bothsorts of con seq ue nce s can serve as a justification or sufficientreason for believing.1 Pragmat ic arguments f rom prudencewill be discussed first, then those from morality.

P R A G M AT I C A R G U M E N T S F R O M P R U D E N C E

The most influential of the prudential-type pragmatic arguments for faith is Pascal's "wager" in his Pensees. His w age rbegins w ith this piece of negative th eolo gy : "If the re is a G odhe is infinitely incomprehensible, since having neither partsnor limits, he has no prop ortion to us : w e are the n, incapab leof kno w ing either wha t he is, or w he th er h e is." Because G odis infinite and we are finite, we can know neither that he

exists nor what he is like.2

No t all neg ative theolo gians w ou ldagree with the former; as we saw in Chapter 6, Anselmthought that he could use a largely negative conception ofGod as a being than which a greater cannot be conceived asthe basis for a successful ontological argument.

That we cannot know what God is l ike seems inconsistentwith his wager's claims about how God will treat believersand nonbelievers in respect to rew ards and pu nis hm en ts . Wecould call these claims "linkage propositions" since they concern how God is related to the created world, especially itsfree pe rsons . W ha t is offered for ou r belief in the wa ge r is no tthe bare proposition that God exists but an entire creed thatincludes these linkage propositions reporting how God willrespond to various free deeds on our part . Obviously, Pascalis not being true to his initial severe version of negativetheology. Maybe the best way to interpret his negative theology is that we cannot know in this life, short of a mysticalexperience, God 's true positive nature or essence. The linkagepropositions do not describe this but only some of his relational properties.

345



Pragmatic arguments

Strictly speaking, the option that is offered to us in thewager is not to believe that God exists, along with otherpropositions in the Christian creed, but to do everything inour pow er to self-induce su ch a belief. Pascal recognizes thatbelief is not, to use later terminology, a basic action, something that w e can do at will, and instea d claims th at it can b eself-induced by various basic actions. Pascal sug ges ts th at w elearn how to do this by imitating the ou tw ard actions of th osew ho cured themselv es of disbelief. "Follow their cou rse, th en ,from its beginning; it consisted in doing all things as if theybelieved in them , in using holy w ater, in hav ing m asse s said,etc." Pascal's causal recipe does no t requ ire u s to act hypocriteically, since w e are enjoined only to act as if w e believed , no tto pretend to believe. Pascal is probably right that if we become an active member of a religious community of believers, we shall wind up believers. For the sake of brevity, inw hat follows I shall speak of choosing to believe, w h en w ha tis meant is choosing to perform those actions that are likelyto induce belief.

He also says that we must make a choice. "Yes, but youmust wager." He does not explain why one cannot place nobet at all, such as one migh t do at a racetrack w h en they placeno wager at all on a given race. His reaso n, no do ub t, is base don the nature of the linkage propositions in the Christiancreed, namely, that in terms of afterlife payoffs, God treatsthe agnostic the same way he treats the atheist. Thus, tomake no choice at all and thereby wind up an agnostic has

the same consequences as if one chose to disbelieve.Depending on how the linkage propositions are filled in,we get different versions of the wager. In one way, Godrew ards all and only believers w ith an u n en d in g life of infinitefelicity in a Christian-type heaven, the big utility, agnosticsand atheists merely ceasing to exist upon death. In anotherhe also assigns agnostics and atheists to an end less existencein hell - the big disutility. And, in a third, which is not anextant version of Christianity, he punishes with hell allagnostics and atheists, believers m erely ceasing to exist u p o ndeath. The first is the best-possible case version, the third

346



Pragmatic arguments from prudence

the wo rst-possible case version, and the second the com binedversion. Some have claimed that the third and second versions rest on a linkage proposition that is incompatible withGod's goodness - that he would endlessly torture nonbeliev-ers . E ven if this is so, the re still rem ain s the best-possible caseversion, which is the version that can be used for wimpswho cannot get themselves to worship and adore a CharlesBronson-type dei ty.3

Th ese thre e vers ion s spell out w hat is to be gained an d lostby the w ag ere r on the different possible outco m es of the avail

able options, namely, to believe or not to believe that Godexists. But payoffs are only one factor in determining what isthe smart bet; otherwise one would always bet on the long-shot. The probability of the different outcomes must also betaken into consideration. In accordance with his doctrine ofGod's total epistemological inaccessibility, Pascal makes the apriori assumption that "the chance of gain and loss areeq ua l." The p robab ility that God exists is taken to be .5. Wh atPascal is doing in effect is to render probability considerations otiose in his wager.

Since the probability of winning and losing are the same,Pascal holds that one should bet solely on the basis of theutility or payoff of the different outcomes. When the probability of w in ni ng an d losing a w age r are equ al, one shou ld b eton the basis of payoffs - the odds. If you bet on God, that is,get yourself to believe tha t God exists and follow th e re qu isitereligious course of life, and win, there is an infinite payoff

according to the best possible case and combined versions,while if you lose there is only at best a finite loss, somesacrifice of worldly pleasures and advantage. However, ifyou bet against God and win, there is only a finite gain atbest consisting in not being hampered in seeking yourw orld ly intere st by th e restrictiveness of the Christian wa y oflife, while, if you lose, on the first version your loss or gainis finite, and on the second and third versions the loss is

infinite. If we are to bet solely on the ground of self-interest,we should bet on the existence of God, since, as Pascal putsit, w e are riskin g a single life - ou r worldly life - w ith an e qu al

347



Pragmatic arguments

chance of gaining an infinity of otherworldly lives. It wouldbe like betting that a true coin will come up heads when wewin an infinite number of dollars if it does and lose only onedollar if it doesn't. Thus, it is prudentially rational to believethat God exists, since doing so advances one's own self-interest. Let us call this version the "w ager sa ns proba bility."

Some very bad objections have been made to this version.One holds that it is internally inconsistent, because, according to its own conception of God, God would not rewardwith heaven someone who believed in him for only pruden

tial reasons. Rather, he would take great delight in assigningsuch selfishly m otivate d ind ividu als to hell. The pr op er replyis that Pascal is not advocating that the person who choosesto believe wind up with a prudentially motivated belief, onlythat, given her presently fallen nature, she must be so initially motivated. What the person attempts to do by gettingherself to believe is in effect to change her character, her network of values and motivations, so that she will eventuallywind up with a nonprudentially based belief in God. God

shou ld judge so m eon e on the basis of how she wind s u p , nothow she began. Thus, there is no internal inconsistency.

Some have objected that the first (third) version of thewager is an instance of the best (worst) possible case fallacy.Since the only possible way to realize (avoid) the best (worst)possible outcome is to believe that God exists, we ought tobelieve that G od ex ists. In many ca ses this m od e of re aso nin gis fallacious. For example, if it is urged that since there issome possibility, however small, that if we engage in recombinant DNA experiments, we will realize the worst possibleoutcome - the extinction of life - we ought not to engage insuch research.4 W hat is fallacious abou t this arg um en t is thatit is ba sed on ly on consid erat ion s of the utility of the differentpossible outcomes, not their expected utility, in which theexpected utility of the outcom e of som e option is the pro du ctof its utility, as determined by the chooser, and the probability of it being realized if this option is selected. Pascal, how

ever, is not committing this fallacy; for, while he can grantthat we ought to work with expected utilities rather than just

348




utilities when probabilities can be assigned to the differentpossible outcomes, his negative theology entails that no suchprobability assignments can be made in respect to the existence of God and thereby to the different possible outcomesin his wager. Thus, we cannot do other than to work withconsiderations of utility alone.

This reply does no t en d the m atter, for the critic can chargethat Pascal has no right to m ake the a priori assu m ptio n thatthe probability that God exists is .50. Pace his negative theology, it might be claimed that the amount and kind of evil wefind in the world renders it highly improbable that Godexists. To meet the se objections th e following "rational choicewager" can be given:

1. It is logically possible that God, as conceived of by theChristian creed, exists;

2. The probability that Go d, so conce ived, exists is gre aterthan zero [from 1];

3. It is prudentially rational for a person to choose thatoption among those open to her that will maximize her expected gain, in which the expected gain of a given option isthe sum of the expected utilities of its possible outcomes[premise];

4. Believing that this Go d exists m axim izes one's expe ctedgain [based on the Christian creed]; and

5. It is prudentially rational for a person to choose to believe that this God exists [from 3 and 4].

There is nothing to argue with in regard to 3, since it doesnot make the bold claim that the only rational choice for aperson to make is the one that she thinks will maximize herexpected gain, as would be claimed by contemporary rationalchoice theorists. Step 4 holds on the best and worst possiblecase versions of the wager, as well as the combined version.On the former there is an infinite utility, an infinitely longlife of infinite happiness, realized when one's bet on God is

won, and since the probability of realizing this outcome isgreater than zero (is a number n such that o < n =s 1), it follow s that the expected utility of this outco m e, alon g with the

349



Pragmatic arguments

expected gain of the option to believe that God exists, is infinite, given that the product of an infinite number and anynumber greater than zero is infinity. The expected gain forthe option of not believing that Go d exists m us t be only finiteat best, since the expected utility of winning one's bet againstthe existence of God is finite, given that it has a finite utilityan d a probability oil -n. Similar con sideratio ns hold m uta tismutandis for the other two versions.

There are decisive objections to this rational choice wager,some of wh ich also app ly to the w ager s an s proba bility. In the

first place, from the fact that it is logically possible that Godexists, it do es no t follow that th e pro du ct of the probab ility ofhis existence and an infinite number is infinite. In a fair lottery with a denumerable infinity of tickets, for each ticket itis true that it is logically possible that it will win, but theprobability of its do ing so is infinitesima l, an d the p ro du ct ofan infinitesimal and an infinite number is itself infinitesimal.Thus , the expected gain of buying any ticket is not infinitebut infinitesimal. There is at least a denumerable infinity oflogically possible deities w ho rew ard and pu ni sh b elievers inthe ma nn er described by the three versions of the wag er. Forinstance, there is the logically possible deity who rewardswith infinite felicity all and only those who believe in himan d s tep on only one sidewalk crack in the course of their life,as well as the two-crack deity, the three-crack deity, and soon ad infinitum. And similarly for the other two versions.

It might be urged that it is still prudentially rational to bet

on one of these rew ardin g and/or pu nis hi ng d eities, since, bydoing so, one assures that it is logically possible that one willrealize the big utility and/o r logically imp ossib le that o ne realize the big disutility. The same response can be made onbehalf of the wager sans probability version when it is confronted with the infinitely-many-deities objection. The problem is that among the logical possibilities is that there existsthe antitheistic deity - the one who rewards with the bigutility all and only those who believe in no theistic Godand/or pu nish es all and only those wh o believe in som e theistic G od.

350



Pragmatic arguments froMi prudence

The only way of dealing with the problem posed by theinfinitely many possible deities, as well as the possibility ofthe antitheistic God, is to resort to epistemological considerations that give favored status to some member or finitenumber of members of the former. This way out is not available to Pascal, given his negative theology, but there is noreaso n w hy th e defender of a wager argum ent m ust accepthis negative theology.5 If we come out with a finite numberof theistic Gods with equally good epistemic credentials, thepr ud en tia l thin g to do is to bet on any on e of them or as ma ny

of them as you can consistently bet on or combine into ahigher synthesis, thereby becoming the "religious hustler"who spends the weekend racing around town at tending asmany different religious services as possible, producing ahigher synthesis when possible by, for example, not eatingpork on Fridays. The expected gain on any of these finitelym an y th eistic optio ns is infinite, since no w the probability ofwinning is no longer infinitesimal, while that of nonbelief isonly finite. A n d, if it sh ou ld tu rn out that the probability tha tone of these theistic Gods exists is greater than that for any ofthe others, i t would be reasonable to supplement rationalchoice theory w ith the ad hoc proviso that w he n two or mo reop tion s ha ve infinite expe cted utilities, the rational thing is tochoose the one that has the highest probability of realizingthe big utility (and/or avoiding the big disutility).6 I will notcomment on whether one of these deities is in fact in anepistemically privileged position, especially since no conclu

sion was reached in my discussion in Chapters 1-8 of theepistemological reasons for believing or disbelieving in theGod of Western theism.

The epistemologically reinforced version of the wager isnot w ith ou t difficulties. First, Pascal is being pr es um pt uo usin filling in our utility assignments for the different possibleoutcomes. As Richard Swinburne correctly pointed out,"Pascal as su m ed that all m en w ould evaluate in the same w ay

as he the various outcomes. But in fact, rightly or wrongly,m en w ill p u t different valu es from ea ch other on the differentou t comes . "7 Pascal seems to be assuming that because the

351



Pragmatic arguments

afterlife in a Ch ristian heav en is on e of infinite bliss or h ap piness without end that all of us will assign it an infinite utility,but we can imagine H um ph rey Bogart res po nd ing to this bysaying , "Listen, Blaise, there are othe r thing s in life tha n infinite bliss, like doing right by a pal."

M any peop le, in particular, wo uld reject the value h e placeson our worldly life vis-a-vis an eternal life of infinite bliss ina Christian he ave n. H e states that in op ting to believe w e risklosing only the finite - this worldly life - with the equalchance to gain an "infinity of life infinitely happy." He holds

that "when there is the finite to hazard in a play where thechances of gain and loss are equal, and the infinite to gain"one has to be irrational not to play. Consider an existentialistwho believes that the most important thing, that which hasthe highest utility as sign m en t in his sche m e of values , is thathe lives this worldly life, the only one that he knows for surehe has, as authentically as possible. If he squanders this lifeby living in a m an ner tha t is no t expre ssive of his tru e self, herisks losing everything. The loss, given his pecking order, isan "infinite" on e, that is, the totality. It m us t be rem em ber edin this connection that there is no neat correlation betweenthe payoffs of the various outcomes and the utility that willbe placed upon them by different persons. For instance, if apers on h as only two dollars and ne ed s it to secure food so asto survive, she might well refuse to gamble this sum even ifshe has an equal cha nce to win a n infinite or as-large-as-you-please number of dollars. This person would not give the

highest or infinite utility assignment to the winning outcomebut would give the lowest possible utility assignment to thelosing outcome. My own ethical intuitions accord with theexistentialist's in this matter. People morally ought not tolive their worldly lives in a way that they deem inauthentic,no matter what possibilities it might open up for otherworldly gains or benefits.

The question, then, is whether the worldly life required ofthose who accept the wager is an authentic one. For some itis and for others not. There are those free spirits who thinkthat they would not be true to themselves if they were to

352




adopt the authoritarian orientation that is required of anactive, believing member of most of the world's leading religions. The restrictions that this would entail in regard to howthey must think, feel, and act would be anathema to them.

On the other hand, there are those for whom the religiouswa y of life is au then tic a nd self-fulfilling. Th ey are th e so rtw ho w ould w ant to p u rs u e this religious w ay of life eve n if itdid not lead on to the big afterlife payoff. They agree withPascal tha t "you will ga in by it in this life." In the grea t e xta ntreligions of the world, which are the only live or practically

possible religious options for the vast majority of people,worldly means and afterlife ends are of a piece, forming anintegral unity. In Christianity, for example, heavenly existence is just a more intense and refined version of thewo rldly religious life in wh ich on e's relation to God , since no tencumbered by the body, is an unceasing beatific or mysticvision of his true positive nature. Only people who find theChristian way of life attractive would find an unending survival in a Christian heaven especially valuable. Their reason,accordingly, for believing that God exists is not a pragmaticone based on its being a means to some desirable end, butthat they want to be people of faith, the religious way of lifehaving an intrinsic value for them.

Thus, Pascal's w age r is not really a w ag er, since the pe op lewho will accept it do not see themselves as gambling at all.According to their scheme of preferences, they are not riskingsomething finite, their worldly life, for the chance of gainingsome infinite otherworldly reward, since the religious way oflife is the one that has the greatest value for them. Pascal's"wager" turns out to be nothing but a pep talk to those whosuffer from a kind of weakness of will. They find the religiousway of life attractive (and thereby the afterlife to which itsupposedly leads) but can't get themselves to make the requisite commitment because of counteracting traits of theirpresen t character. The y are in the same situation as those w h o

want to change their character but finds it difficult becausethey must overcome the counterbalancing force of formedhabits. They need a pep talk so that they can strengthen and

353



Pragmatic arguments

make dominant their second-order intentions to change theirfirst-order motivations and reasons for acting.

N ot only is there an integral unity to the m ea n s- en d s relationsh ips described by the linkage prop ositions of the creed sof the great religions, there ought to be such a unity. It waspointed out that for the wager to work it was necessary tolimit the field of entrants to at least a finite number of the-istic deities. It w as suggested that this m igh t be accom plishe dby appeal to pragmatic considerations. But the requirementthat the this-worldly-means and the-otherworldly-ends form

an integral unity also is a way of narrowing the field, sincemany of the logically possible theistic creeds do not satisfyit. For instance, the one-crack-deity creed does not, since itenjoins us to follow a worldly life in which we step on justone sidewalk crack in the course of our life as a means togaining entrance into a Chirstian-type heaven . Herein m ean sand ends fall apart; the connection between them is a purelyaccidental one that is secured by the causal efficacy of anexternal agent. Anyone who followed such a creed would beacting in a totally irrational manner, and not just becauseof any lack of adequate epistemological support for it. Alot more will be said about this integral unity requirementwhen we discuss moral justifications of faith, to which wenow turn.

PRAGMATIC ARGUMENTS FROM MORALITY

According to this mode of pragmatic justification of faith,one could have a sufficient reason for believing if doing sowere to promote the cause of morality by rendering thebeliever and/or her society morally better. It could be contended, for example, that if the subject of a religious experience dev elops in a morally desirable ma nn er as a result of he raccepting the experience as an objective revelation, therebybelieving th at God exists, she is justified in believing tha t G odexists. Since some moral good results from her so believing,she has at least a prima facie moral justification for he r belief,

354



Pragmatic arguments from morality

even thoug h, as argued in the previo us chap ter, she does nothave adequate epistemic warrant for it. The onus is on thenaysa yer to sh ow tha t ther e are defeating con ditions, na m ely,that by so believing, our believer brings about some moralevil that outweighs the moral good that is realized.

W. K. Clifford wou ld c on ten d tha t the good o ur believerrealizes by her epistemically unsupported belief is outweighed by the evil that results from her allowing herself tohave such a belief. He holds that "it is wrong always, everyw he re, an d for an yo ne , to believe any thing up on insufficient

evidence."8

Let us call this "Clifford's principle." His argument for it is based on an act-utilitarian moral theory, although it is not explicitly stated. His basic contention is thatwhile such a belief m ight m axim ize utility in the sh ort ru n, inthe long run its overall consequences are horrendous. Thereason is that by allow ing ourse lves on e such belief, howeverseemingly trivial and harmless, we become infected with anincurable case of credulity and dishonesty that will eventually infect the entire community, with disastrous long-rangeconsequences. The following quotations, which for maximum effect should be read aloud while "Pomp and Circumstance" is played in the background, defend this "plaguetheory" of epistemically unwarranted belief:

Our words, our phrases, our forms and processes and modesof thought, are common property, fashioned and perfectedfrom age to age; an heirloom which every succeeding genera

tion inherits as a precious deposit and a sacred trust to behanded on to the next one, not unchanged but enlarged andpurified, with some clear marks of its proper handiwork. Intothis, for good or ill, is woven every belief of every man whohas speech of his fellows. An awful privilege, and an awfulresponsibility, that we should help to create the world inwhich posterity will live.9

Whoso would deserve well of his fellows in this matter will

guard the purity of his belief with a very fanaticism of jealouscare, lest at any time it should rest on an unworthy object, andcatch a stain which can never be wiped away.10

355



Pragmatic arguments

That duty is to guard ourselves from such beliefs as from apestilence, which may shortly master our own body and then

spread to the rest of the town. What would be thought of onewho, for the sake of a sweet fruit, should deliberately run therisk of bringing a plague upon his family and his neighbours?"

In like manner, if I let myself believe anything on insufficientevidence, there may be no great harm done by the mere belief;it may be true after all, or I may never have occasion to exhibitit in outward acts. But I cannot help doing this great wrongtowards Man, that I make myself credulous. The danger to

society is not merely that it should believe wrong things,though that is great enough; but that it should become credulous, and lose the habit of testing things and inquiring intothem; for then it must sink back into savagery."

The credulous m an is father to the liar and the cheat; he livesin the bosom of this his family, and it is no marvel if he shouldbecome even as they are. So closely are our duties knit together, that whoso shall keep the whole law, and yet offend

in one point, he is guilty of all.13

I assume that ten minutes have passed and that the readerha s just beg un to pull back together ag ain after rolling a rou ndon the floor in hysterical convulsions. But my purpose inprese nting this series of quotatio ns is no t just to enterta in bu tto protect myself against the charge of creating a strawmanby attributing the plague theory of epistemically un w ar ra nt edbelief to Clifford. Like many an act utilitarian trying to fendoff a desert-island unkept-promise counterexample, Cliffordhas greatly exaggerated the deleterious consequences ofallowing ourselves even a single epistemically unwarrantedbelief, however tr ivial and disconnected from the workadayw orld . To pu t it mildly, his plag ue the ory ha s very du bio usempirical credentials. It would be most unnecessary for theUnited States to mount a "Just Say No to EpistemicallyUnwarranted Beliefs" campaign. Imagine the TV commer

cials that Clifford would make for it: We see little Johnny

356



Pragmatic argum ents from m orality

happy in his totally unfounded belief that the Yankees willwin the World Series followed by a grown-up Johnny, nowthe "liar and the cheat," who is a PR nan for Exxon assuring the nation that everything possible is being done to cleanup the oil spill, and then we see some super-straight farmfamily in Iowa who, immediately upon hearing this spiel,travels to downtown Ames to hustle phony Rolex watches.

That Clifford's principle is not supportable by act utilitarianism does not show it is false. Co unte rexam ples, ho w ev er,can be given to it consisting of cases in which people, in vir

tue of entering into a trust relationship, have a moral duty totrust each other and therefore to believe certain things aboutothers regardless of whether they have adequate epistemicsupport for these beliefs. For instance, spouses have such amo ral du ty to trust each other an d believe in the faithfulnessof the other person, even if their belief is not supported bythe results of an empirical inquiry; they ev en h av e a du ty no tto perform such a n inquiry. To perform it is already to s tan dou tside of the special relation ship. I believe that family m embers have a similar moral duty in regard to their beliefs abouteach other.

While Clifford's principle faces such trust-relationshipcounterexamples, an epistemically unwarranted belief thatG od exists is not am on g them ; and th us it is op en to Cliffordto restrict his universal principle in such a way that it willno t apply to trust-relationship cases bu t will apply to all oth ercases of belief. William James in "The Will to Believe" produces counterexamples even to this restricted version of Clifford's princip le, the most exciting of which is an epistem icallyunwarranted belief that God exists. His intent, however, isno t just to prod uce c ounterexam ples bu t to spell out the n ecessary and sufficient conditions under which one has a moralright to believe without adequate epistemic justification.

These necessary and sufficient conditions will be marshaled by James to m oun t a pragm atic arg um en t from mo ral

ity for epistemically unfounded faith. It is a substitutioninstance of the following argument form:

357



Pmgtmtic arguments

6. D oing X he lps to bring about Y;7. It is morally desirable that Y; therefore,8. One has a prima facie moral permission to do X.

Notice that the moral permission in 8 is only prima facie,since it faces possible d efeating con dition s. It is desirab le tha tI keep my promises, and my giving a revolver to someonew ho m I pro m ised a revolver to brings it about that I ke ep m ypromise. But I have only a prima facie permission to handover the revolver, since this person could have gone insanesince I made my promise and would use the gun to kill aninno cen t per son . O ur duty no t to do that which will result inan innocent person being killed takes precedence over ourm oral duty an d/or permission to kee p our prom ise.

James's pragmatic argument from morality plugs into thisargument form as follows:

6'. By believing that God exists, I help to bring it aboutthat I and/or my society become morally better;

7 '. It is m orally desirable that I and /or my society bec om e

morally better; therefore,8'. I have a prima facie moral permission to believe thatGod exists.

Notice that the conclusion is weaker than might appear necessary, but it will suffice for James's purpose of counteringClifford's moral prohibition against epistemically unfoundedbelief under any circumstances, save for those that occurwithin a trust relationship. It will be seen that there are sev

eral possible defeating conditions for the prima facie moralpermission in 8' that James did not consider.We shall now follow in detail the route James takes in

developing his counterexamples to Clifford's principle, evenin its res tricte d fo rm, an d th e se"t of necessary a nd sufficientcon dition s for the m oral right to belief w ithou t ad equ ate epis-temic support that fall out of them. Since all of these examples involve what James calls a "genuine option" to believe,we must begin with his account of a genuine option.

He begins by offering two definitions. We are "to give thename of hypothesis to anything that m ay be propose d to ou r

358



Pragmatic arguments from m orality

belief and to call the decision between two hypotheses anoption. "H James's choice of "hypothesis" to characterize the

object of a decision is unfortunate, for not all options involvea decision between hypotheses in the sense of choosing oneof them as a working hypothesis or, more dubiously, tobelieve one of them. We need a more generic term to characterize the object of choice, and I suggest that we use "Proposition," understood in the sense of a proposal. This termwill be written with a capital " P " so as to disting uish it fromthat which refers to what is the bearer of a truth-value; thelatter will be written as "proposition" with a lower case "p."A proposal is best described by a predicate infinitive - "toF" - rather than by a noun "thaf '-clause. My terminologybegs no question against ]ames, since it leaves it openwhether among the Propositions that could be subject to ourchoice is to believe some proposition (or hypothesis).

A genuine option is on e tha t is live, momentous, a nd forced. "Aliving option is one in which both hypotheses [Propositionsaccording to my terminology] are live ones" (WB 3) in the

sense that each is a "real possibility to him to whom it ispro po sed " (WB 2).An option is m om ento us w he n it is either un iqu e (it 's yo ur

only chance to do wh at is pro pose d) or significant (w het he ror no t you elect to d o wha t is prop osed will vitally affect you rfuture life). It should be clear from what is required for anoption to be live and m om en tou s that a gen uine o ption m us tbe relativized to a person at a time, since what is living ormomentous can vary from person to person and from onetime in a person's life to another. This relativization of a genuine option to a perso n at a t ime will beco m e imp ortan t w h enwe devise a way out of a difficulty in James's argument.

James's account of when an option is forced is unclear.After offering some examples of unforced options, he saysthat in a forced option "there is no standing place outside ofthe alternative" (WB 3). This rather unhelpful remark is followed by the puzzling sentence "Every dilemma based on a

com plete logical disjunction , w ith no possibility of not ch oo sing, is an option of this forced kind." What is puzzling about

359



Pragmatk arguments

this is that even when we have an option between Propositions that are mutually exclusive and exhaustive - to do F orno t to do F - it is no t clear wh y w e are forced to choose one ofthe m . Wh y can't we refrain from ch oosing either alterna tive?Judgin g by som e of the thing s James later says abou t forcedoptions, the basis of a forced option is that if one does notchoose one of the Propositions, then, given the circumstances that obtain, one wind s up doing wha t the other P roposition proposes, even when one does not actually choose todo th e latter. We could call the P roposition in a forced op tio nthat becomes actual if no choice at all is made the "negativealternative," the other the "positive alternative." We put inthe qualification "in the circumstances that obtain" so as topreclude the positive alternative becoming actualized whennot chosen. The option to join Dr. Nansen's expedition, touse James's example, is not forced if we will wind up joiningthe expedition regardless of what choice is made, as wouldbe the case if the good Dr. Nansen were to shanghai us.

At the outset, all of James's examples of options, including

genuine ones, involve a decision between proposed actions,that is, pieces of intentional behavior. He next attempts toshow that there can be a genuine option to believe a proposition. He has a rebuttal to the objection that there cannot besuch an option since belief is not an action - something wecan do intentionally, at will, voluntarily, and the like. Thisproblem will be considered later, and for the time being weshall assume that there is no problem about choosing tobelieve.

A genuine option to believe must also be forced, live, andm om entous . An agent A, at a time T, has a genu ine op tion tobelieve or not to believe a proposition p just in case:

9. The circumstances are such that if A does not choose tobelieve p, she will not believe p;

10. To believe p and to not believe p are each real possibilities for A at T; and

11. Whether or not A believes p will vitally affect her future life.

360




Condition 9 says that only if A chooses to believe p will shewind up believing p : Not only does she not already believe p,

no external power such as a hypnotist or mad surgeon will"shanghai" her into believing p. Con dition 10 requ ires that Ahave an open mind about the truth of p. Con dition 11 can besatisfied only if it is as su m ed that w ha t A believes concerningthe truth of p will ha ve imp or tan t conse que nce s in her future,this req uiring that he r belief conc ernin g p be deter m inativ e ofher behavior.

Although James does not explicitly state this, a case of a

ge nu ine o ption to believe a pro pos ition is a necessary bu t n ota sufficient condition for a counterexample to Clifford's principle. The needed additional conditions can be gleaned fromw ha t James say s, as well as from his examples, w hich w ill bediscussed later. The first condition which must be added is

12. It is impos sible at T for A to decide on epistemologicalgrounds the truth of p.

Here the impossibility can be of the in-practice or in-principlesort. This is strongly suggested by these remarks:

Our passional natu re not only lawfully may, but must decidean option between propositions, whenever it is a genuineoption that cannot by its nature be decided on intellectualgrounds. (WB 11)

In concrelo, the freedom to believe can only cover living

options which the intellect of the individual cannot by itselfresolve.*5 (WB 29)

The following sort of possible counterexample to Clifford'sprinciple, involving a genuine option to believe, is ruled outby 12. Im agin e tha t at T one of the so ns of the Rose nbergs hasa genuine option to believe or not to believe that his parentswe re not guilty of espion age : He is und ecided about the truthof this proposition; his belief option is forced; and what hebelieves will determine much of his future behavior, forinstance, how he relates to people and institutions. Even if

361



Pragmatic arguments

w e assum e that at T he has no evidence on e way or the othe rfor this pro po sition , 12 still pre clud es h is bein g m orally justi

fied in choosing to believe what he wants, since its truth canin practice be de term ined at T by empirical inquiry - checkingrecords an d traces, examining witne sses, and so on . An d thisprohibition holds regardless of how beneficial the consequences of believing this proposition might be. James is farfrom the "cheap pragmatist" that he is depicted as being byRussell and others, for he holds that we are never free toshirk our responsibility to investigate a question in a thorough, scientific manner when it is amenable to such aninvestigation. In fact, James goes too far with Clifford's principle, since he fails to note that trust-relationship cases are acounterexample to it.

An othe r requirem ent that James seem s to place on a counterexample is that

13. It is w ithin th e po w er of A to help make p true (or false).

The textual supp ort for this additional req uirem ent, w hich , itwill turn out, is not needed, is this:

There are, then, cases where a fact cannot come at all unlessa preliminary faith exists in its coming. And where faith in afact can help create the fact, that would be an insane logicwhich should say that faith running ahead of scientific evidence is the "lowest kind of immorality" into which a think ingbeing can fall. (WB 25)

In truths dependent on our personal action, then, faith basedon desire is certainly a lawful and possibly indispensablething. (WB 25)

Condition 13, in conjunction with the conceptual truth thatcausation cannot go backward, entails that p must be aboutthe future. Th e Ro senb erg case also fails to m eet this r equ irem en t; for no th in g th at is do ne at T or at a later time can br in git abo ut or pr ev en t it from b eing the case tha t the R ose nb erg scom mitted espion age prior to T as defined by the pre-T law s.

362



Pragmatic argtiinetfts from m orality

Th ere i s on e fu r the r n ece ssa ry con di t io n tha t imbt teSJames ' s d i scuss ion , espec ia l ly h i s examples , namely,

14. I t i s moral ly des i rable that p be t rue .

Since A is su pp os ed t o be morally jus t if i ed in ch oo s in g tobe l ieve p so tha t she can he lp to make p t rue , i t mus t be tha ti t i s moral ly des i rable that p be the case , the re by f i t t ing th eear li e r a r gu m en t fo rm.

It i s Jam es ' s co n ten t ion tha t an y case tha t sa ti s f ies co nd it ions 9 -14 above cons t i tu tes a counte rexample to Cl i ffo rd ' sp r inc ip le . Such a case wil l be ca l led a " spec ia l ge nu in e op t io nto be l i e ve" so as to d i s t in gu ish i t f rom a "ge nu in e op t io n tobel ieve ," which need sa t is fy only the f i rs t three of the s ixcon d i t i ons . It n ow wi ll be sh ow n tha t J am es ' s c ou n t e r ex amples are in fac t specia l genuine opt ions to bel ieve .

His f irs t case is the option to bel ieve or not to bel ieve thepropos i t ion tha t you wi l l l ike me:

Whether you do or not depends, in countless instances, onwhether I meet you half-way, I am willing to assume that youmust like me, and show you trust and expectation. The previous faith on my part in your liking's existence is in such caseswhat makes your liking come. But if I stand aloof, and refuseto budge an inch until I have objective evidence, until youshall hav e do ne s om ethin g ap t . . . ten to on e yo ur likingnever comes. (WB 24)

Yet an o t he r case invo lves the spec ia l g en u i ne o p t io n to be l i ev ethat God exis ts . F i rs t , James gives a pragmat ic analys is of th isp r o p o s i t i o n . H e fi n d s a p r a g m a t i c c o m m o n d e n o m i n a t o r fo ra l l r e l ig ions in these two propos i t ions :

First . . . the best things are the more eternal things . . . thething s in the univ erse that thro w the last ston e . . . an d saythe final word.The second affirmation of religion is that we are better offeven now if we believe her first affirmation to be true. (WB25-6)

363



Pragmatic arguments

We shall consider only the first affirmation as constitutingthe religious hypothesis, the second affirmation being not atenet of any religion but rather a proposition about thedy na m ics of religious faith, su ch as migh t be m ad e by a socialscientist or an apologist for religious faith.

The first affirmation, basically, seems to say:

R. Good will win out over evil in the long run.

This se em s to be wh at is me ant by the "th e best thing s . . .throw the last stone . . . say the final word." The idea is thatthe universe is morally good in that, in spite of its undeniableevils, i t will have a morally good outcome, a just and goodor de r even tua lly will come to prevail, if w e ma ke the requisitemoral effort. There are forces within the universe that willsupport our most noble moral efforts so that they will cometo a successful denouement. There are interesting affinitiesbetween James's theology and later process theologies.

It remains to be shown how each of these options constitute s a special gen uin e option to believe. Cond ition 14 is sat

isfied in both cases, since, supposedly, it is good that goodwin out over evil in the long run and that you like me.Condition 13 is satisfied, since the agent has some power tohelp make these propositions true, although the extent of herpower differs radically in the two cases. She can do little tomake the former true, no matter how altruistically she acts,b ut h er actions are almost the sole determ inan t of the trut h ofthe latter. This difference results in a difference in the way 12is satisfied in the two cases. In both cases I cannot have

evidence for the proposition before I choose to believe it, butafter I cho ose to believe that yo u will like me I hav e goo devidence that you will, since I then know that I shall act in afriendly w ay an d that this is very likely to succeed. H ow eve r,after the agent chooses to believe R, she still has virtually noevidence for it. Thus, in the you-will-like-me case, the agentdoes not wind up having an epistemically unjustified belief,and thereby there is no counterexample to Clifford's principle. But it does violate the spirit if not the letter of it, since

364



Pragmatic argum ents from morality

it also means to proscribe choosing or getting yourself to believe without epistemic justification.

Th ere is a further reas on for think ing that the you-will-like-m e case fails to satisfy 12. I could ha v e goo d in duc tive ev idence that I am likable, and this could consti tute adequatejustification for believin g tha t you w ill like m e. But Jam es canrespond that whether you will l ike me hinges on what Ichoose to believe in this matter (since what I believe determ ines how I act tow ard yo u), and I can not kn ow w hat I shallchoose before I choose. The latter is a conceptual truth and

does not rest on James's dubious Libertarian doctrine thatchoice is acausal and thus unpredictable.It only remains to be shown that both cases involve a gen

uine option to believe, thereby satisfying all six conditions.We can cook the circumstances so that failure to choose thepositive option results in the actualization of the negativeoption - no t believing the proposit ion in ques tion. N ot everyone will find it a live option to believe R, but this poses nodifficulty since we can follow Jam es in restrictin g ourse lves tothose for whom it is live:

If we are to discuss the question at all, it must involve a livingoption. If for any of you religion be a hypothesis that cannot,by any living possibility be true, then you need go no farther.I speak to the "saving remnant" alone. (WB 26)

It is over the momentousness of the options that serious

problems arise. The momentousness of a belief option depends not on the momentousness of what the proposi t ionforecasts but the momentousness of the agent 's believing orno t believing it. W hile R pred icts so m eth in g of great imp ortance to most people, it is not obvious that it is of greatimportance to the agent's future life whether or not shebelieves it; for, it could be claimed, she can behave altruistically eve n if sh e d oe s no t actually b elieve R, just as I can actin a friendly m an ne r eve n if I su sp en d belief in the pro po sition that you will like me. Because belief does not seem to

365







Pragmatic arguments

that their efforts will succeed and thus that the proposition ist rue. Accordingly, we must add to the six preceding conditions this new one:

15. A knows that she will act so as to help make p true ifand only if she believes in advance that p is true.

The reason why it is required that the agent know this factabout herself is that it must be possible for the agent, as wellas other people, to morally justify her choosing to believe anepistemically unsupported proposition if we are to have acounterexample to Clifford's principle.18

There still remains the objection that it is conceptuallyimpossible for there to be an option to believe a propositionsince we cannot believe at will, voluntarily, intentionally,and so on. James's response consists in an argument for theempirical proposition that "our non-intellectual nature doesinfluence our convictions" (WB 11). This is a flowers-that-bloom-in-the-spring response, since this psychological truthhas nothing to do with the case. Granted that there are non-

rational causes of an emotional and passional sort for all ourbeliefs, how is this supposed to show that we can choose tobelieve? If anything, it proves the contrary, since we cannotcontrol our passions and emotions at will.

It is surprising that James did not deploy his theory of thewill from The Principles of Psychology to meet the belief-is-not-an-action o bjection. By an inten tional act of attention or concentration we can get ourselves to be conscious in a certainway. Once we are conscious of what a proposition p reportswith sufficient intensity over a long enough time, we willhave acquired the belief that p since "ou r belief an d atten tionare the same fact." '9 Following Spinoza, James argued that ifan infant were to be conscious of nothing but the image of acandle, this would amount to a belief in the existence of thiscandle.

Many would find this account of belief unacceptable andthus would reject any attempt to deflect the belief-is-not-an-

action objection that appealed to it. What, then, is the properresponse to this objection? In the first place, its claim that it

368




is conceptually imp ossible to believe at will shou ld be cha llenged. To be sur e, it w ou ld be very q ue er to spe ak of believing at will, voluntarily, on purpose, carefully, and so on, butthat might be due, not to such talk violating any rule oflanguage, but our never having had the occasion to so speaksince we haven't yet learned the knack of controlling ourbeliefs at will. At one time in history it might have soundedequally queer to speak of controlling the frequency of ourbrain waves at will but now that we have learned the knackthrough biofeedback training such talk is unobjectionable.

Possibly something similar could happen in regard to ourlearning how to control our beliefs at will.For the purpose of constructing a moral justification of

epistemically unsupported belief, James need not involvehimself in the issue of whether belief is, or could be, a basicaction. He can avail himself of Pascal's strategy of holdingthat there is a causal recipe consisting of basic actions bywhich belief can be self-induced, in particular the belief thatGod exists. Although James and Pascal have radically different conceptions of God, they agree that in principle we cannot have an adequate epistemic justification for belief in God.Therefore, the methods that our causal recipes for self-inducing faith p rescribe are non rationa l on es , for instan ce, "takingthe holy water, ha ving m asses said ," and the l ike, as o ppo sedto examination of arguments and evidence. James's claimthat all of our beliefs are caused, at least in part, by nonrational factors can be viewed as a justification for no nra tion ally

self-inducing a belief in p w he n rational m eth od s can be of noavail.It should no w be clear wha t mu st be don e to protect James

against the belief-is-not-an-action objection. His special genuine option for A at T to believe or not to believe p must bechanged to a special genuine option for A at T to try bynon rationa l m ea ns to get herself to believe p or no t to try th is.The latter will have to satisfy the same seven conditions for a

special genuine option to believe, with the exception of thefirst three, which are modified to account for the differencein the objects of choice:

369



Pragwmtfc arguments

9'. The circum stances are such that if A do es not choose totry to get herself to believe p, she will not believe p;

10'. To try to get herself to believ e p and not to do this areeach real possibilities for A at T;

11'. W hether or no t A tries to get herself to believe p willvitally affect her future life;

12. It is impossible at T for A to decide on epistemologicalgrounds the truth of p;

13. It is within the power of A to help make p true (orfalse);

14. It is morally desirable that p be true; and15. A knows that she will act so as to help make p true ifand only if she believes in advance that p is true.

Let us call an option that satisfies these conditions a "specialgenuine option to self-induce a belief.

It is not only James's sufficient and necessary conditionsfor a counterexample to Clifford's principle that must bereconstructed in this m an ne r so as to get aro un d the belief-is-not-an-action objection, but this principle as well. If I cannotbelieve at will, there cannot be a moral imperative enjoiningme not to believe without adequate epistemic justification.The principle, accordingly, must be revised to say that it ismorally wrong under any circumstances to induce an epis-temically unwarranted belief.

Since our concern is with whether one can be morally permitted (or justified) to believe or get oneself to believe thatGod exists when lacking adequate epistemic support , wem us t consider wh eth er a pe rso n w h o satisfies all seve n of th econditions for having a special genuine option to self-inducea belief in R is m orally pe rm itted to do so . It is no t difficult toimagine a perso n w ho is so psychologically co nsti tuted thatsh e satisfies all of thes e con ditio ns . To simplify m at ter s, let u simagine that there is available to her a belief-in-R inducingpill, this being representative of any nonrational method forself-inducing a belief in R, which are the only methods that

can be of any avail, given that R is a prediction of what willcom e to pas s in the indefinite future.

37




It is clear tha t James w ou ld say that in thes e circum stance s

ou r agen t is m orally pe rm itted (or ha s a m oral justification) totake the pill, since by doing so she brings about somethingmorally desirable, namely, that she acts in an altruistic orgood-making fashion.20 Herein he is plugging into the previous valid argument form, according to which our agent has aprima facie moral permission to do that which helps to bringabout s om eth ing that is mo rally desirable. It will be arg ue d atsome length that there are defeating conditions for her primafacie moral permission to take the pill, that by taking the pill

under those circumstances, she brings about an evil thatoutweighs the good she promotes. But first it will be shownthat to accord her such permission violates the moral principle of universalizability.

The moral principle of universalizability states that if itis morally permissible (required, forbidden) for X to do Y incircumstances C, then it is morally permissible (required, forbidden) for anyone to do Y in circumstances C. Let us imagine two perso ns, A and A', w ho are exactly alike save for on efeature of their psychology. A, being short of courage, willnot act so as to help make R true unless she first believes thatR will beco m e tru e, wh ere as this is not true of A ', the psychologically stron ger m em ber of the pair. It thereby tu rn s o utthat A, but not A', is m orally perm itted to take the belief-in-Rinducing pill. And this seems to violate the principle of universalizability. The reply is that their circumstances are notthe same since A satisfies condition 15 w hi le A' does not. But

is this a morally relevant feature of the circumstances? I thinknot. It seems wrong to accord a moral privilege to someonebut not to another on the grounds that the former is a psychologically weaker person. This question of moral permission must not be confused with the quite different questionof w he th er it is morally permissible to give the wea ke r m ember of the pair more help and consideration than the strongerone , to which the answer is yes.

My argument that the good realized by taking the pill isou tw eig he d by the evil that results m akes use of a highly no rmative concept of personhood, according to which freedom

371



Pragmatic arguments

of the will is esse ntial. After I ha ve d ev elo pe d th is concep t, itwill be utilized for the purpose of showing why, on balance,it is imm oral to take the p ill, thereb y can celing the prim a faciemoral permission to do so.

It is m y firm m oral conviction th at the re is an ab solute valueto personhood. (If the reader is unable to go this far, I trustthat he at least will grant that it has a very great value, whichwill suffice for my argument.) This means the following:

P. It is alw ays w ron g to bring it abo ut that a pers on beco m esless tha n or less of a pe rson or that a potential perso n becom es

something less than a person.Before pointing out what it is to be a person and what is involved in treating som eon e as a perso n, i t m ust be noted thatneither capital punishment nor abortion is ruled out by P. Inkilling a fetus one does not bring it about that an individualwho is potentially a person becomes something less than aperson, for one simply terminates this individual's existence.As both Kant and Hegel have correctly argued, the adminis

tration of capital punishment under certain conditions involves treating the criminal as a person. Robert Brandom haspointe d ou t to m e a resemblance betw een P and the C hristiandoctrine of the absolute value of the soul. Martyrdom is notruled out, for the martyr loses her life but not her soul: Thepers on w ho is executed b y the state loses his life but do es n otbecome something less than a person, since, assuming thereis no survival of dea th, he do es not becom e any thin g at all. Palso does not rule out treating human beings paterna listically,for not all human beings qualify as persons. In some casesthe justification for so treating them is that it helps themdevelop into persons.

What, then, is it to be a person? This is a normative question, and according to my moral convictions a necessary andsufficient condition for personhood is having free will. Tohave free will is to behave as a morally responsible agent. Ishall explicate the latter no tion in term s of the moral resp on

sibility game, the playing of which is a necessary and sufficient condition for being a person.

372




This gam e is de term ine d by two kind s of rules, ontologicaland sociological. The former specify the conditions underwhich a person is morally responsible for an action, the latterwhat is involved in treating someone as a morally responsible agent. The ontological rules must be given first, since totreat someone as a person (i .e. , as a morally responsibleagent) is to view th eir actions as satisfying the se rule s. W hilethe ontological rules have this priority in the order of analysis, there is, as Annette Baier has shown me, a psychologicalsense in which the sociological rules have priority in that by

treating individuals as pers on s, w e help to bring it abo ut thattheir actions satisfy the ontological rules.The first ontological rule states:

R,. A player is morally responsible for an act only if hecould have avoided doing it .

For the pu rpo se of m y arg um en t, I nee d no t commit m yselfto any particular analysis of "could have avoided." In Chapter 4 I tried to refute Lockean objections to R, and will notdiscuss the issue again.

Although no analysis will be offered of R,, some necessarycon ditions for m oral respon sibility can be spelled ou t in ter m sof the following rules:

R2. A playe r is m orally re spo nsib le for an act only if h e d idit as a rational agent.

R3. A playe r perfo rm s an action F as a rati on al agen t only if:

(a) he knows what he is doing;(b) he has good reasons for doing F;(c) his rea son s are at least a necessary cau se of his do in g F;

and(d) he has no reasons that are both necessary causes of his

doing F and not good reasons for doing F.

While i t is reasonable to make rational agency a necessary

condition for moral responsibility, R3 ne ed s further explanation and justification, especially since some of its conditionsare mooted.

373



Pragmatic arguments

Cond ition (a) is unp roblem atic, b ut (b) is not. A reaso n r isa good reason for an agent A to do an action F just in case it is

true both that A is justified in accepting r (even if r is in factfalse) and r is logically relevant to his doing F (even if r is notthe best reaso n). The co nce pt of a goo d re aso n is a pa rt of ou rnorm ative concept of a rational age nt. T here are som e sorts ofthings that a rational agent would not believe (only a madman would believe that sort of thing ), as well as some sortsof motives, intentions, and purposes that no one could haveand be a rational agent. The concept of insanity is in part anormative one.

Some examples will help. My reasons for reaching for aglass of water are that I desire a drink and believe that thereis a glass of water in front of me and that water quenchesthirst. These constitute good reasons for my actions in that Iam justified in accepting them and they are logically releva nt. If, instead, m y reas ons w ere that I desired to pr ev en tthe "heat death" of the universe and believed that there is aglass of water in front of me and that drinking water causes

entropy to decrease, my reasons would not be good, since Ihave no justification for the latter belief reason. If, instead,my reason were that I believed that the sky is blue, myreason s again wou ld no t be good, this time beca use m y beliefreason, although justified, is logically irrelevant.

Let us consider some possible counterexamples to (b). Aperson pinches someone and when asked why she did i treplies "For no reason at all." This is the same as saying "Because I wanted to." But this gives a logically relevant reason,namely, that she enjoys pinching people and believed thatthere was a pinchable subject in front of her. By telling uswhat sorts of things she enjoys and desires, she tells us a lotabout what kind of person she is. Another possible counterexample is the case of weakness of the will. The incontine nt pe rson shou ld be held morally respo nsible for his actionsince he fails to act on what he considers to be the best orproper reason. But the incontinent person, at the time he

acts, does have a good reason, namely, self-interest, eventhough at other t imes he judges that moral reasons should

374



Pragmatic a rguments from m orality

take priority o ver tho se of self-interest. To be a good rea so n isnot to be the best reason or even a morally acceptable re aso n.

While conditions (c) and (d) permit a reason to be a cause,no attem pt is m ad e to assimilate the con cept of a cause to tha tof a reason or vice versa, which would be wrong-headed,since reasons and causes play different logical roles. Theseconditions only point out that a rational action must becaused in a certain way, and the rationale for these conditions can best be seen by considering cases that run afoulof them.

Consider first a case in which a person's reasons (e.g., hisdesires, wants, intentions, and beliefs) for doing F, a l thoughconstituting good reasons for his doing F, are not necessarycauses for his do ing F. C hap ter 4 considere d such a case, tha tof the conscious puppet, Pinnochio, who, by coincidence,with good reason endeavored to move his l imbs just whenStromboli caused them to move by pulling on some wire.This is a case of causal overdetermination, and since Pinno-chio's reasons are not necessary causes of his actions, he isnot morally responsible for these actions.

Notice that (c) is consistent with the identity thesis; for, ifmy desires, wants, intentions, and beliefs are identical withvarious neurophysiological states and events, then if the latter are necessary causes of my actions, so are the former byLeibniz's law of the indiscernibility of identicals. Furthermore, (c) allows there to be causes for a rational act otherthan the agent 's reasons; for instance, neurophysiologicalevents that are numerically distinct from these reasons.

The need for condition (d), which rules out any reasoncause, wh eth er nec essary or not, am on g the causes of a rational act that is not a good reason, can be seen from this case.I intentionally reach o ut for the glass of w ate r, b ut am on g thereason causes are the above good reasons plus my belief thatVerdi wrote Ernani. Since the latter is not a good reason formy action, being logically irrelevant, it destroys the rational

ity of my action.That a rational act must be caused in a certain way, which

is what R2 and R3 require, has interesting analogies with

375



Pragmatic arguments

causal theories of memory and perception. In recent literature there are many examples of seeming memories (perceptions) that do not count as a memory (perception) becausethe causal chain linking them with their apparent object isnot of the right sort, being too kinky. We can imagine similartypes of science fiction fantasies in which a person has goodreasons for her action but ones that are not linked with heraction by the right sort of causal chain.

Now for the sociological rules. The first of these specifieshow one can opt into the moral responsibility game.

R4. O ne opts into the gam e by declaring that on e wan ts tobe held morally responsible for one's actions by the otherplayers or by being willing to hold them an d be held by themto be morally responsible.

Rarely does one opt into the game by uttering some explicitperformatory sentence, such as "I hereby declare that I wantto be held morally responsible for my actions." Rather, onebecomes of age and just finds oneself playing the game byholding others morally responsible and in turn being willing,at least on some occasions, to be held morally responsibleby them.

Since one can be held to be morally responsible only byanother person, R4 has the logical consequence that

R5. A player must hold the other players to be morallyresponsible for their actions.

It should be app are nt w hy t he principle of reciprocity app lieshere. Since a necessary cond ition of pe rso nh oo d is being heldmorally responsible and only a pers on can be held or hold on emorally responsible, I must invest the other game playerswith personho od by holding them morally responsible so thatthey in return can invest me with pe rso nh oo d by hold ing m emorally respo nsible . A con sequ ence of rules R4 and R5 is thatno on e can be a per son except in the society of othe r pe rso ns .

The next group of rules specify what is involved in treatingsomeone as a morally responsible agent. The first of theserules is

376



Pragmatic arguments front morality

R6. To treat someone as being morally responsible for anaction requires that her action be viewed as satisfying th«requirements specified in rules R2 and R3.A consequence of R2 is that we cannot treat someone as a morally resp ons ible ag ent un less we treat her as a rational ag ent .The requirements for treating someone as a rational agentcan be derived from Ry Since R3 requires that a rationalagent have good reasons for her actions that are causallyoperative in the appropriate way, it should follow that wecannot treat someone as a rational agent if we either denythat her reasons for acting are good or induce her to performan action without her having good reasons for it. Thus, weget these rules:

R7. To treat someone as a rational agent requires thather reasons for performing an action not be dismissed orundermined .

R8. To treat someone as a rational agent requires that shenot be induced or caused to perform an action for which shelacks a good reason.

Rj forbids us to say of a rational agent's professed reasonsfor acting that they are not the real reasons, a mere coverupor rationalization by her "conscious" mind for the true "unconscious" reasons. We are all too familiar with these devicesfor explaining aw ay, unde rm ining , som eone's professed reasons . However, R7 does permit us to supplement a rationalagent's reasons and to offer causal explanations, for instance,neurophysiological ones, that do not make use of her professed reasons.

If I were to induce someone to perform an action withouthe r havin g good reason s at the time she acts, I wo uld bringit abou t th e she fails to act as a rational age nt. T his, c ertainly ,is not to treat her as a rational agent. Notice that R8 doe s notentail the further rule: To treat someone as a rational agentrequires that only rational methods be used to induce her toperform an action . Later it will be show n th at som eon e can beinduced by nonrational methods to perform an action forwhich she has good reasons.

377



Pragmatic arguments

There is no thing in R;, o r R8, or any of the other rules of thegame, nor even in principle P itself, that rules out treating ahuman being as a non pers on . But to treat some one in a waythat violates a rule of the game must be justified by ampleevidence that she cannot act in a way that satisfies the onto-logical rules R ,-R3. It could ev en be a former ga m e play er th atgets so treated. This indicates a need for a further rule specifying how a player can be eliminated from the game:

Rg. A player m ay be elimina ted from the game by the othe rplayers when there is good evidence that her behavior cannot satisfy the requirements specified in rules Rj-R3.There are less radical moves than total elimination from thegam e. A player could be view ed as lacking mo ral responsibility for some but not all kinds of act for which she was form erly held mo rally resp ons ible. As a con sequ ence, there is aneed for a restricted version of Rj.

R10. A player may no longer be held morally responsibleby the oth er play ers for a kind of action F w he n th ere is goodevidence that she cannot perform F-type actions in a waythat satisfies the requirements specified in rules Rj-Rj.

Ru les Rg an d RI0 bring out some important features of ourconcep t of a pe rso n. R, sh ow s th at it is a phasal pro pe rty (i.e.,one that can apply to an individual at some but not all timesin that individual's history), R10 that it admits of degrees,one 's degree of personh ood de pen din g upo n the num ber anddifferent kinds of action for which one can be held morallyresponsible.

The reader may w on de r w hy Rg and R10 do not allow forreflexive instances - self-elimination from the game or self-restriction on de gre e of pe rso nh oo d. It is becau se I basicallyagree with Sartre's dictum that the one thing we are not freeto do is to give up our own freedom.21 When rendered interms of rules of the game it becomes

R n . N o play er can op t out of th e game;

R u . No player can request that other players no longerhold her morally responsible for actions of kind F.

378



Pragmatic a rguments from morality

There can be no voluntary commitment . An interest ing consequence of these rules is that it can never be part of our justification for treating a former game player as a nonpersonthat she has consented to such treatment, which does notmean that we cannot have justification.

These rules are controversial and thereby require somedefense. To some it might seem odd that we can opt into butnot out of the moral responsibility game. The reply to this isthat there is nothing absurd about certain kinds of noncance-lable contracts and com pac ts. A mo re serious objection is tha tI can have exactly the same justification for eliminating myself from the game as for eliminating another player. Thejustification for my eliminating a person from the game isthat I ha ve good em pirical ev ide nc e that he r be ha vio r fails tosatisfy the on tological rules, bu t I can ha ve exactly the sa m esuch evidence for my own behavior failing to satisfy theserules.

The answer to this objection will consist in establishing

first , although there is no epistemic asymmetry between anagent and observer in regard to the agent 's past actions, the reis in regard to her future actions; and, second, this epistemicasymmetry results in a justificatory asymmetry between theelimination of oneself and others from the game, since eliminating someone from the game concerns her future behavior.

Let us call someone an agent in respect to an action if shedoes this action intentionally. It is obvious that the very samekind of evidence that can be appealed to by an observer tojustify his claim to know what some agent did in the past canbe appealed to by the agent in support of her claim to knowthat she did this. But this evidential sym m etry betw een ag en tand observer does not hold for the agent 's future actions invirtue of the following conceptual truth: It is impossible foran agent to know in advance of her forming an intention todo some action (which could coincide with her beginningto do it) that she will do that action. A special case of thisis it is conceptually impossible for an agent to deliberate andat the same t ime know what she wil l decide to do." Anobserver, ho w ev er, faces no such conce ptual bars from kn o w -

379



Pragmatic arguments

ing in advance what an agent will choose or intentionally do.The inductive argument that she can give to support her

knowledge claim can be of no avail to the agent who has yetto make up her mind; for, to try to decide what to do is notto try to inductively infer what someone will choose do.2 3

Deciding is not predicting. A decision, unlike a prediction,is not justified by appeal to an inductive argument but bygiving good reasons for bringing abou t w hat one ha s decide dto do.

It remains to be shown that this epistemic asymm etry between agent and observer in respect to the agent's futureactions results in a justificatory asym m etry betwe en first- andthird-person elimination from the game. As pointed out, anecessary part of the justification for eliminating a playerfrom the game is possession of good empirical evidence thatthis player's past actions did no t satisfy th e req ui rem en ts ofthe ontological rules. A player can have such evidence inregard to her ow n p ast actions. So far there is no asy m m etry .However, there is a prospective dimension to eliminating

someone from the game, since it results in her not being heldmorally responsible for her future actions. Obviously, thisrequires justification in the form of a good inductive argument that her future actions will not satisfy the ontologicalrules. But, in accordance with the stated epistemic asymmetry between agent and observer, the agent who has yet todecide how to act in the future cannot then use any suchinductive argument to support her claim to know how shewill act in the future, since she cannot know in advance ofher decision how she will decide to act. Whether or not theagent acts in a morally responsible way in the future - in away that satisfies the ontological rules - depends, at least inpart, upon what choices she will make. Thus, in advance ofthese choices, she is conceptually barre d from kn ow ing w ha tshe will do, although she faces no such bar from knowing inadvance on inductive grounds the choices and intentionalbehavior of others. Therefore, if a person were to opt out ofthe game, she would have to justify this move by appeal toinductive knowledge of her future lack of moral responsibil-

38




ity; but it is exactly this sort of inductive knowledge that shecannot have in advance of her decisions regarding her futureconduct.24 So far only R „ ha s been d efend ed, bu t it shou ld beapparent that exactly the same sort of argument can be givenin support of R,2, which only deals with a more restrictedcase of opting out of being held morally responsible.

This completes my rather sketchy analysis of personhoodin terms of playing the moral responsibility game. If I treata person in a way that violates a rule of the game, I impugnher personhood, since to be a person is to play this game.

Thereby, I violate P, the absolute (or very great) value ofpersonhood. It can be shown that to take the belief-in-Rinducing pill is to treat myself in violation of the rules ofthe game: And this is sufficiently immoral so as to defeat thegood that might be realized by taking the pill. But first Iwould like to dispose of a bad-personhood-based argument.

The agent's moral justification for taking the belief-in-Rinducing pill is that this constitutes a sufficient and necessary

condition for her acting altruistically in the future. But, theargument points out, this results in her future altruisticactions failing to satisfy R,, since she cannot avoid actingaltruistically. This involves a loss of or radical restrictionup on her per son ho od , which violates P. T hu s, i t is immoralfor her to take the pill.

This argument lacks force because it rests on the dubiouspremise of causal incompatibilism. Even if causal incompati-bilism is gra nte d, the argu m ent can be ans w ered . O ne way isto assume that the agent can take a pill that negates theeffects of the belief-in-R inducing pill. Thereby she is notpermanently locked in on a future of altruistic acts, since shehas it within her power to reverse the effects of this pill bytaking an antidote pill. Another strategy is to hold that abelief in R is only a necessary cause of the agent acting altruistically. This would in no way weaken the force of the Jame-sian arg um en t for her being m orally perm itted to take the pill.

No w for a good-personhood-based argum ent based on A'sviolating P by tak ing the pill. According to Jam es's a rgu m en t,our hypothesized agent is supposed to be morally permitted

381



Pragmatic arguments

to take the belief-in-R inducing pill, since her belief in R,given the proposition 15-type fact about her psychologicalm ak eu p, is at least a nece ssary co ndition of her acting altruistically so as to help make R true. First it must be establishedthat A's belief in R constitutes part of her reasons for actingaltruistically, that is, so as to help make R true.

When a rational ag en t's belief is pa rt of th e cau se of h eraction, this belief constitutes part of her reasons for this action. A must be considered to be a rational ag ent for thefollowing r ea son s: (1) A is supposed to be morally permitted

to take the belief-in-R ind uc ing pill; (2) b ut o nly a p ers on canbe morally permitted to do something; therefore, (3) A is ape rso n; (4) a nec essa ry con dition for be ing a pe rso n is play ingthe moral responsibility game; (5) a necessary condition forplay ing the ga m e is that a player's actions satisfy ontologicalrule R2 requiring that she act as a rational agent; therefore,(6) A is a rational agent. Since A is a rational agent and herbelief in R is part of the cause of her acting altruistically, itfollows tha t he r belief in R con stitute s par t of her re aso ns for

acting altruistically.It sho uld b e obvio us th at A's belief in R fails on both co un ts

to be a good reason for her acting altruistically. First, A, exhypothesis, lacks an y ep istem ic justification for R. Second ,an d m ore im po rtan t, R is logically irrele van t as a justificationfor her acting altruistically so as to make R true. It would beabsurd to give as one's reason for acting so as to make aproposition true that it is in fact true or will turn out to bet rue . It would be crazy to work to bring about an economicdepression because one believes that an economic depression will occur. But this is exactly the sort of logically irrelevant reason that A h as for acting so as to help m ake R true . Alogically relevant reason for so acting would be that it ismorally desirable that R becom e tru e. That o ne's rea son s foracting be logically relevant is a variant on the cited requirement that an acceptable prudential justification for doing Xso as to bring abo ut Y involves an integral unity between X,the m eans, and Y, the end . T hus, A violates her personhood

382




by taking the belief-in-R inducing pill, since she causes herself to perform actions for which she lacks good reasons.

How serious is this moral transgression against P? Is itserious enough to defeat our age nts prim a facie m oral permission to take the pill based on the good that will resultfrom doing so? If we take P to be an absolute moral rule, theanswer is yes. Even if we do not go so far as to accord anabsolute value to personhood , and thereby an absolu te s tatusto P, a case can be ma de out for an affirmative an sw er, p rovided we are willing to grant that there is very great value

to personhood, and thereby that it takes a very pow erfuljustification to act in a way that violates P. The case goes asfollows.

Given the very extensive na ture of the actions and dispositions caused by taking the pill and the extent to wh ich theyare constitutive of A's character and personal i ty, she in effectopts out of the moral responsibility game, which violates R 1V

By taking the pill she causes m ost of her future actions to failto satisfy the R3 requirement of rational age ncy, sin ce she willnot have good reaso ns for these actions.25 We can, of course,modify the case so that the future beh avio r that is caused byher taking the pill is not so extensive, thereby a llowing A stillto be a game player in respect to s ome of her actions. Nevertheless, she would still violate P, since by constricting therange of her rational, and therefore morally responsible, behavior, she lessens her degree of pe r sonhood . She makesherself less of, if not less than , a perso n. This violates P, and

thereby is a very serious moral transgression , eve n if P is notaccorded absolute status. Thus, there appear to be sufficientdefeating conditions for A's prima facie moral permission totake the pill when she has a special genuine option to self-induce a belief in R; and, thereby, James 's argument for aperson in such circumstances being morally justified in self-inducing an epistemically unjustified belief in R fails.In general, James has failed to show tha t one can have a sufficient moral reason for self-inducing an epistemical ly u ns upported belief. 26

383



Pmpnaim wptf»ent$

The battle is not over yet, since by a slight revision in theconditions of James's special genuine option to self-induce abelief, a more defensible moral justification of faith can becon structed. A ccording to the former it is numerically o ne an dthe same proposition that the agent can help to make trueand that, given her psychological makeup, she must firstbelieve before she will work to make it tru e. A more intere sting option results if the latter propos ition is distinct from theformer. This results in this revised version of James's argument in which propositions p and q are distinct:

9'. The circumstances are such th at if A does not choose totry to get herself to believe p, she will not believe p;

10'. To try to get herself to believe p and not to do this areeach real possibilities for A at T;

11'. Whether or not A tries to get herself to believe p willvitally affect her future life;

12. It is impossible at T for A to decide on epistemologicalgrounds the truth of p;

13'. It is within the power of A to help make q true (orfalse);14'. It is morally desirable that q be tru e; a nd15'. A knows that she will act so as to help make q true if

and only if she believes in advance that p is true.

Before using this revised version of James's argument to morally justify som eone's acquiring an epistemically un w arr an tedbelief that God exists, a further condition will be added to it

so as to get aroun d a devastatin g objection to the justificationbased on the case of the special gen uine op tion to self-inducea belief having to do with its rendering the agent a nonra-tional agent in virtue of giving her a logically irrelevant reason for her "desirable" action.

Let p be the proposition that two dino saurs w ere rutt ing onthis very spo t five m illion years ago and q be the propo sitionR. Let us assu m e that A at T is so psychologically cons titu tedthat she can act so as to help make R true only if she firstbelieves p. According to the revised version of James's argument , A is morally permitted to take the belief-in-p inducing

384



Pragmatic argument® rom morality

pill, since this br ings it ab ou t th at she will behave in a morallydesirable manner. Obviously, her belief in p is not a logicallyrelevant reason for her subsequent efforts to help make Rt rue , and thereby she brings it about that she becomes lessthan or less of a person, for the reasons already given. Andsince she thereby violates P, this constitutes a defeating condition of her prima facie moral permission to take the pill.

To protect the revised version of James's argument fromthe logically irrelevant-reason-type of counterexample, thefollowing condition must be added:

16. The belief in p is a logically relevant reason for trying tomake q true.

This also meets the desideratum of means and end formingan integral unity. The connection between means and endshould not rest on some purely accidental, external factor,such as the psychological quirks of the agent.

Let us now see how this revised version of James's argu

ment, with the additional condition 16, can be used to givea moral justification for acquiring an epistemically unwarranted belief that God exists, understood in the orthodoxway, rather than in terms of James's phony substitute for it,R.27 There are many people who are so psychologically constituted that they can act so as to help make R true only ifthey first believe that theism is true. Let us use "G" as ageneric name for a theistic creed, such as the Christian one,

according to which God, an absolutely perfect being, is thecreator of and sovereign over the universe and also the re-warder of morally good persons. The latter linkage proposition is very important in their system of motivation, sincethey can't get themselves to sacrifice their own interestunless they believe there is some reward for doing so. Bygetting themselves to believe G, which they will have to doby some nonrational method since we are assuming that wecould not have adequate epistemological warrant for G, theyhelp to bring about something morally desirable, namely,that they act in a good-making fashion so as to help make Rbecome true. Herein their belief in G is a logically relevant

385



Pragmaticarguments

reas on for their acting so as to mak e R true, since it is p rud en -tially rational for one to maximize his own self-interest.Therefore, these people seem to be morally permitted to takethe belief-in-G inducing pill, since doing so helps to bringabo ut som ething morally desirable and there are no app aren tdefeating conditions.

There is an oddity to this moral defense of faith - that thereare moral reasons for certain people, namely, those whosepsychology fits 15', acting prudentially by taking the pill.This creates the problem of who can give this defense. The

concerned persons, being prudentially motivated, cannotconsistently give moral reasons, based on the moral desirability of their future good-making actions, for their takingthe pill. It w ou ld ha ve to be som e third pe rso n w ho gives thismoral justification of their prudentially motivated act of taking of the pill. W he the r this justificatory asy m m etry be tw eenthird persons and the concerned persons is vicious is amooted issue that will not be further pursued, being betterleft in the hands of the Derek Parfits.

An even m ore serious worry is w heth er the m oral defenseof faith based on the revised version of James's argument,like that based on the special genuine option to self-induce abelief, violates the moral principle of universalizability.Again, i t turns out that someone's moral permission to perform som e action, tha t of nonrationally indu cing an epistem-ically unwarranted belief, depends on her psychologicalquirk s, w heth er she satisfies the psychological requ irem entsspecified in 15'. Is this a morally relevant feature of the circum stance ? I think tha t it is no m ore m orally relevant t ha n isthe fact that the agent suffers from cowardice and therebycannot try to make something come to pass without priorbelief that she will succe ed, th ou gh I hav e no pow erful argument to support my intuit ion in either case. I t seems wrongthat a prudentially motivated person (or one lacking in self-confidence) should have moral permission to do somethingthat a morally motivated person (or one possessed of self-

confidence) is denied. Why should the more highly principledpe rso n (or stron ger pers on ) be mo rally pena lized in this way?

386



Epilogue

This completes my critical evaluation of pragmatic arguments, both from prudence and morality, for having faith.While grounds were given for doubting that any of themsucceed, no knockout was scored. At best some nasty cutswere opened that might lead to a TKO. That no decisiveresult was reached should come as no surprise, since, asAristotle pointed out, one cannot expect the same sort ofconclusive results in reasoning about normative issues as canbe had in other areas.

E P I L O G U E

The skeptical outcome of my discussion of pragmatic arguments for faith dovetails with the sceptical outcome of myearlier discussions of epistemological arguments for faith.Since I completely eschewed inductive arg um en ts, n o definite conclusion can be drawn regarding the rationality offaith. Only the hypothetical conclusion can be drawn that ifthe only available arguments were the epistemological andpragmatic arguments examined before, faith would lack anyrational justification. Such an outcome would be welcomedby a wide range of Kierkegaardian types who completelyeschew any attempt to give an "objective" justification offaith. I resonate to their view of faith as a subjective passionthat outstrips our reason.

387



Notes

INTRODUCTION

For the seminal discussion, see Peter T. Geach, "IntentionalIdent i ty," journal of Philosophy 64 (1967), pp. 627-32.Belonging to the same religious community must not be constru ed as ne ce ssa ry but only as part of a sufficient con ditionfor two users of "God" to be coreferers, since some personoutside a given rel igious community, such as an anthropolo

gist or philosopher writing about the God referred to by themembers of this community, can refer to the same God asdo the members of this community. These descript ive andcritical uses of "God" are secondary, since they are parasiticupon the uses of "God" by members of an established relig ious communi ty.

CHAPTER 1

For an insightful discussion of this point, see Peter Geach,Providence and Evil (Cambridge University Press, 1977).Clem ent D ore , "Plantinga on the Free Will De fense," Review ofMetaphysics 24 (1971), p. 692.St. Anselm gave this sort of argument for God's noncom-posi teness in Monologion cha p. 17. He implicitly assu m ed thatan individual is composed of the properties she instantiates,but this assumption, unl ike the assumption that the individ

ual is co m pos ed of her different instanc ings of various pro pe rties, is dubious and thereby undercuts this motivation for thedoctrine of the divine simplicity.

389



Notes to pages 25-55

4. Eleonore Stum p and No rma n Kretzm ann, "Absolute Simplicity," Faith and Philosophy 2 (1985), pp. 356-7.

5. Ibid., p . 357.6. Ibid., p p . 379-80.7. William Ma nn , "Simplicity and Prop erties: A Reply to Mo r

ris," Religious Studies 22 (1986), p. 352.8. It will not do, as Leibniz is som etim es interprete d as do in g, t o

say that God is not essentially benevolent in this sense butinstead essentially freely chooses to follow the principle ofperfection. If he freely chooses to follow this principle, insome possible situation in which he exists he does not sochoose; but if he essentially freely chooses to follow this principle, then in no possible situation in which he exists does henot choose to follow it.

9. The pickle th e Leibnizian God is in is qu ite different from th atof Buridan's anorexic ass, in which there is a choice betweenequally good alternatives. The latter, but not the former, canbe solved by allowing a benevolent be ing to make an arbitrarychoice betw een equally good alternatives whe n the re is a go odreason to choose one of them. In the former case God wouldhave to choose the lesser of the alternatives available to him.

CHAPTER 2

All quotations are from Saint Augustine, Confessions, t ran s .E. B. Pu sey (Literary Guild of Am erica, Ch icago, 1948). N umerals in parentheses indicate sections of Book 11.The exceptions to this are terminal events that are the beginning or end of a state or process, but such terminal events areconceptually dependent upon the temporally elongated statesand processes that they bound.The reader is reminded of the qualification in note 2 concerning so-called terminal events and moments of time.Eleonore Stump and Norman Kretzmann, "Eterni ty," Journalof Philosophy 78 (1981), p. 432.Ibid., p. 444.Ibid., p. 446.William Alston's "Divine-Human Dialogue arid the Nature ofGod," Faith and Philosophy 2 (1985), pp. 5-20, presents a clear

and forceful defense of this position.Stum p and Kretzm ann, "Eterni ty," p . 452.

390




CHAPTER 3

1. By the use of the plural en di ng on "pr op osi tion s," it is beingassumed that successive tokenings of the same A-sentencewith the same m ean ing exp ress different A -propositions. Thisway of speaking squares better with ordinary usage concerning when people say the same thing than does the Aristotelian-scholastic-Lukasiewicz-Prior stipulation according to whichsuch tokenings always express the same proposition, thoughone that can change in truth-value over time. If I successivelytokened "It is now t?l" I w ou ld not be said to have said th esame thing - expressed the same proposition - on each occasion. In fact, I w ou ld b e said to have cha nge d in regard towh at I believe, which sh ow s that my successive tokeningsexpre ss different a nd inco m patib le propos itions . Fu rther, if Isaid at one time "Jones is now seated" and at a later time"Jones is not now sea ted ," I wo uld not be thoug ht to be deny ing w ha t I formerly said . For a full discussion of this, as wellas other intricacies and subtleties concerning A-propositions,see chap ter 2 of my Language of Time (Routledge & K egan P aul,London, 1968).

2. A. N. Prior, "The Form alities of O m niscie nce ," Philosophy(1962), p p. 114-29. These arg um ent s are repeated by A.Kenny in "Divine Foreknowledge and Human Freedom," inAquinas: A Collection of Critical Essays, ed. A. Kenn y (AnchorBooks, N ew York, 1969), p p . 255-70. Both articles are reprinted in Readings in the Philosophy of Religion, ed. B. Brody(Prentice-Hall, Engelwood Cliffs, N.J., 1974). Future references are to the pagination in the Brody volume, since it ismore widely available than the Kenny.

3. A. N . Prior, "Than k Go odn ess That 's Ov er," Philosophy (1959),p. 17.

4. Prior, "The Form alities of Om nisc ienc e," p. 415.5. Prior's use of "th ere are " sho uld be replaced by "the re could

be," since if it were always raining, as is possible, then everytokening of "It is now raining" would in fact express a trueproposition.

6. There is an ot he r arg um en t for 4 in Prior's "The Formalities ofOm niscien ce" in which it is m ad e to appe ar that the m ere factthat a timeless being's kn ow ing cannot be located in time barsthe being from being omniscient, in fact, from knowing any-

391



Notes to pages 62-y

thing at all. "It seems an extraordinary way of affirming God'somniscience if a person when asked what God knows now,

must say 'Nothing' , and when asked what He knew yesterday, must again say 'Nothing', and must yet again say 'Nothing' w hen asked w hat God will kn ow tom orro w " (p. 416).This rhetorical flourish derives its effectiveness from the factthat since most of the time we are not talking about a timelessbeing we might forget that we are doing so on this occasion.Since such a being is extraordinary, it should not surprise usthat what we must say about it is extraordinary.Geach, Providence and Evil, p . 40.

See Hector-Neri C astane da, " 'H e': A Stu dy in the Logic ofSelf-consciousness," Ratio 8 (1966), pp. 130-57; idem, "Indicators and Quasi-indicators," American Philosophical Quarterly4 (1967), pp. 85-100; idem,"On the Phenomeno-Logic of theI," Proceedings of the XlVth International Congress for Philosophy,vol. 3, pp. 260-6; idem, "On the Logic of Attributions of SelfKnowledge to O the rs ," Journal of Philosophy 65 (1968), pp . 4 39 -56; idem, "On the Philosophical Foundations of the Theory ofCommunication: Reference," Midwestern Studies in Philosophy2(1977), pp. 165-90; and idem, "Perception, Belief, and theStructure of Physical Objects and Consciousness," Synthese 35(1977), p p . 285-351.The same claim is implicit in the atte m pt by S tum p and Kretz-mann in "Eternity" to refute the omniscience-immutabilityargument by appeal to a nontemporal type of simultaneityrelation, called "ET-simultaneity," in which a timelessly eternal God stan ds to every even t in time . In virtue of hav ing thisrelation to a given eve nt, G od is able to observe it "as tem porally pre sen t" (p. 439). H e "kn ow s w ha t is actually ha pp en in g

as it is ha pp en in g" (p. 457). Th ere is the s am e am biguity inthese claims as in Geach's in regard to ho w w e are to interp retthe tempo ral indexical references with in th e oratio obliqua construction; an d, again, if prem ise 4 is to be challen ged, w e m ustinterpret them as expressing both the speaker's and God'sindexical reference.Robert Coburn, "Professor Malcolm on God," AustralasianJournal of Philosophy 41 (1963), pp. 155-6.For a helpful a ccou nt of G od's tim eless c aus atio n, see W. E.

Mann, "Simplicity and Immutability in God," International

Philosophical Quarterly 23 (1983), p p . 2 67- 76.

392



Notek to pages 69-95

12. Stum p and Kretzm ann, "Eternity," p. 435.13. Ibid., p . 439.14. Ne lson Pike, God and T imelessness (Schocken B ooks, N ew

York, 1970), p. 90.15. The core po rting thesis is the brainc hild of Michelle Beer. See

her "Temporal Indexicals and the Passage of Time," Philosophical Quarterly 38 (1988), pp. 158-64.

16. John Perry, "Frege on D em ons trative s," Philosophical Review86 (1977), pp. 474-97, and idem, "The Problem of the Essential Indexical," Nous 13 (1979), pp . 3-2 2. Vie ws sim ilar toPerry's are to be found in the writings of Tyler Burge, JosephAlmog, and Jeremy Butterfield, to mention only a few.

17. Perry, "The Problem of the Essential Inde xical," p . 16.18. Perry, "Frege on D em ons trativ es," p p. 481-2 .19. For a fuller treatm ent of thes e pro blem s, see my article "P rop

osi t ions, Judgments, Sentences, and Statements ," in Encyclopedia of Philosophy, ed . P. Ed w ard s (Collier-Macm illan, N ewYork, 1967), pp. 494-505.

20. He doe s this in Hector-Neri Ca stane da, "O m niscience an dIndexical Reference," Journal of Philosophy 64 (1967), pp. 203-10. Norman Kretzmann's ar t icle , "Omniscience and Immutability," appeared in the same journal one year earlier. Botharticles have been reprinted in Readings in the Philosophy ofReligion, ed. Brody, and all future references a re to the p agi nation in this book.

21. Kretzm ann, "Om niscience and Im m utability," p. 366.22. Cas taneda , "Omniscience and Indexical Reference," p. 378. I

have changed Castaneda's numbering of this proposition

from "(4a)" to "6" so as to fit the format of my chapter andsimilarly for his other numbered propositions.23. Ibid., p . 381. In w ha t follows the qualifica tion "(or w ou ld

know)" will be drop pe d since, as will be se en , it is not n ee de d.24. Ibid.25. Ibid.26. In correspondence Castaneda has expressed his wh olehea rted

approval of this amendment, since it was his intention allalong to refute the timeless version of the omniscience-immutability argument.

27. A good discu ssion of this is to be found in Robert Carr-Wig-gin, "God's Omnipotence and Immutabil i ty," The Thomist 48(1984), pp. 44-51.

393




CHAPTER4

1. H . J. M cCloskey, "God and Ev il," Philosophical Quarterly 10(i960), p. 97.

2. Ibid ., p . 105.3. Ibid ., p . 112.4. J. L. Mackie, "Evil and O m ni po ten ce ," Mind 64 (1955), p. 1.

This article is reprinte d in Readings in the Philosophy of Religion,ed. Brody. Future references will be to the pagination in thisbook.

5. M ackie, "Evil and O m nip ote nc e," p . 160. M ackie, likeMcCloskey, has a short memory span, for at the end of hisarticle he creates a bogus paradox of omnipotence thatrequires of an omnipotent being that he be able to create freecreatures whose wills are so free that not even he can controlthem , result ing in there being so m ething that he cannot do -control the free will of these creatures. What Mackie fails tonotice is that the latter act description is contradictory according to the Libertarian intuitions of the theist and thus is notsom ething that an om nipoten t being mu st be able to do, giventhe restricted 51 definition of "om nip ote nce " that he accepts inthe earlier part of his discuss ion. H is parad ox of om nip oten ce,far from being an important new discovery, is only a varianton the timeworn paradox of the stone discussed in Chapter 1,bein g based o n the same abus e of ou r concept of om nip ote nce .

6. Ibid ., p . 161.7. Ibid ., p . 163.8. Ibid., p p . 163-4. My italics.9. In fairness to Mackie it m us t be stressed tha t at th e time h e

wrote, no adequate formulation of the FWD had yet beengiven and that Mackie was instrumental in eliciting thesemore adequate formulations.

10. M ackie, "Evil and O m ni po ten ce ," p . 164.11. This point is well m ade by Terertce Pen elhu m in "Divine

Goodness and the Problem of Evil," Religious S tudies 2 (1967),p p . 95-107-

12. N elson Pike gives this re sp on se on behalf of the theist in"Hume on Evi l ," Philosophical Review 72 (1963), pp. 180-97.

13 . Pe ne lhu m argue s effectively for this in "Divine G oo dn ess ."

394



•


14. Not all theists accept this. Robert M. Ad am s, in "M ust GodCreate the Best?" Philosophical Review 81 (1972), pp. 317-22,

argues against it, claiming that God's grace permits him tocreate less perfect peo ple th an he m ight ha ve . I am co nvincedthat his view is based on a misinvocation of the traditionaldoctrine of grace but will not argue for it here.

1$. The exception is M cCloskey, "G od an d Ev il," p. 114, w howrites as follows: "The real alternative is, on the one hand,rational agents with free wills making many bad and somegood decisions on rational and non-rational grounds, and'rational ' agents predestined always ' to choose' the right

things for the right reasons - that is, if the language of automata must be used, rational automata. Predestination doesnot imply the absence of rationality in all senses of that term.God, were he omnipotent, could preordain the decisions andreasons upon which they were based; and such a mode ofexistence would seem to be in itself a worthy mode of existence, and one preferable to an existence with free will, irrationality and evil." The proper response to McCloskey is notto challenge his moral intuitio ns, t ho ug h I wou ld want to, butto point out that they are not those of the theist. And since itis the internal consistency of theism that is at issue, it shouldbe the theist 's values that are operative.

16. Alvin Planting a, "The Free Will D efen ce," in Philosophy inAmerica, ed. Max Black (Cornell University Press, Ithaca, N.Y.,1965). This article is reprinted in Readings in the Philosophy ofReligion, ed. Brody; all future references to it will be cited inthe text of this chapter (as will be the case for all references toPlantinga's publications) and will refer to the pagination in

this volume.17. Alvin Plantin ga, God and O ther Minds (Cornell UniversityPress, Ithaca, N.Y., 1967); idem, The Nature of Necessity(Oxford University Press, Oxford, 1975); idem, God, Freedom,and Evil (Harper, New York, 1974), pp. 7-64; and idem, Self-Profile, in Alvin Plantinga, ed. Jam es Tomberlin and P eter vanInwagen (Reidel, Dordrecht, 1985).

18. Th is objection is forcefully m ad e by A nton y Flew, "DivineOmnipotence and Human Freedom," in New Essays in Philo

sophical Theology, ed. A. Flew and A. Maclntyre (SCM Press,Lon do n, 1955), p p . 150, 151, 153.

395




19. In GFE (pp . 31-2) Plantinga aludes to an argu m en t againstcausal compatibilism in which it is po in ted ou t tha t if an action

is causally determined then the conjunction of an earlier stateof affairs, which is not preventable by the agent at the time ofthe action, with certain causal laws m ake s it causally imp ossible for him to do otherwise. But this hardly refutes causalcompatibilism, since the compatibilist thinks a person couldhave acted otherwise even when it was not causally possibleto do so. An interesting argument could be developed fromPlantinga's remarks if it were added that the agent cannotprevent the causal laws in question from holding. Thus, twothings that he cannot individually prevent together entail thathe perform the action in question; and, supposedly, this hasthe consequence that he cannot prevent doing the action andthus does not do it freely. This unpreventability argumentwinds up begging the question, for the compatibilist believesthat a person can act otherwise than she is causally determined to act and thereby believes that she can prevent thecausal laws from holding. Also, it is dubious whether beingunpreventable is closed under deduction.

20. O thers of similar opinion are A nto ny Flew, "C om patibilism ,Free Will and G od ," Philosophy 48 (1973), p p . 237 -8; Richard L.Purtill, "Flew and the Free Will Defense," Religious S tudies 13(1977), pp. 477-83; and J. E. Barnhart, "Theodicy and the FreeWill Defense: Response to Plantinga and Flew," ReligiousStudies 13 (1977), p p . 439-53-

21. R. E. H oba rt (Dicken son Miller), in his classic form ulatio n ofcausal compatibilism, "Free-Will as Involving Determinismand Inconceivable Without It," Mind 43 (1934), pp. 1-27,writes, "In proportion as an act of volition starts of itself with

ou t cause it is exactly so far as the freedo m of th e ind ivid ual isconcerned, as if it had been thrown into his mind from without - sug gested to him by a freakish de m on ." Ano ther prom inent causal compatibilist, A. J. Ayer, in his Philosophical Essays(Macm illan P ress ; Lo nd on , 1959), p . 275, co ncu rs: "E ither it isan accide nt t ha t I choo se as I do o r it is not. If it is an accid ent,then it is merely a matter of chance that I did not choose

, oth erw ise; an d if it is m erely a m atte r of cha nce that I did n otchoose otherwise, it is surely irrational to hold me morally

responsible for choosing as I did. But if it is not an accidentthat I choose to do one thing rather than another, then pre-

396




sum ably there is som e causal explan ation of m y choice: an d inthat case we are led back to determinism."

Mackie, "Evil and Omnipotence," p. 165.Flew, "C om patibilism ," p . 238, raises a verificationist-typeobjection to Libertarianism and thereby to the FWD to theeffect that there is no way empirically to determine when anact is free, since its cause is a non ob servable self. But Libertarians from Kant to Campbell were not suggesting that thepenal code be formulated in terms of their kind of free acts:that would be disastrous because of the admittedly unverifi-able nature of such acts. Their point, rather, was that we can

consistently think of ourselves and others as performing suchacts and that this is required if we are to be coparticipants inthe moral responsibility game. For Flew's objection to touchthem, he would have to show that necessarily whatever isconceivable is verifiable. A nd I do no t think h e can sh ow this .Plantinga adds that if the atheologian does not grant theviability of this conception, then the real objection to theism isbased not on evil but on the unviability of this conception ofGod. "And if he insists that the theistic conception is impossi

ble just becau se it involves the idea of a per so n w h o is free bu tnot causally determined, then his real quarrel with theism isnot that God's existence is incompatible with that of evil; itis instead that God's existence is impossible simpliciter" (SP47). This overlooks the value of an atheologian's hav ing m orethan one argument against the existence of God; for while it istrue that any one of them, if successful, would do the job,each is subject to some doubt, and thus there is value in having converging independent arguments.

The minor exception is Flew, "Compatibilism," p. 243, who,in spite of being a professed "ordinary language" type ofphilosopher, makes the wildly counterintuitive claim that"there can be no ultimate and fundamental contradiction insuggesting that another man, or God, might, by direct physiological manipulations, ensure that someone performs whatever actions that other man, or God, determines; and that theactions of this creature would neve rtheless be g enu ine actions,such that it could alw ays be truly said that he co uld ha ve do n e

otherwise than he did."

Plantinga, for the most part, leaves room for my strategy,for he usually treats being causally determined as only one

397




among other sufficient conditions for an action not being free,among which he recognizes being supernaturally determined

by God's will. The exception to this is his careless remark that"if compatibilism is correct, the FWD fails" (SP 45). Herein hefails to realize that there are sufficient conditions for an actionnot being free other than its being causally determined.

27. Mackie, "Evil an d Om nipote nce ," pp . 164-5. Mackie errs, asw e shall soon see, in his wild claim that according to the FWD"sec ond ord er evils are logically necessary acco mp anim ents offreedom," but his objection is independent of this incorrectportrayal.

28. This acco unt is the same as 03 in Chapter 1

and is ado pted byPlantinga.29. Plantinga use s the more specific term "hu m an" ; but no harm

is done by working with the more generic, non-species-specific notion of a person. Plantinga also confines moral evilto what is wrought by human beings, while throughout thischapter the more generic notion of a moral evil as one freelywrought by a person has been used.

30. In SP Planting a calls them "co unterfactuals of freed om ,"which is inaccurate terminology, since not all of them havecounter-to-fact antecedents. It is especially strange that heshould give as an example of a "counterfactual of freedom" "IfEve were free to take the apple, then she would have freelydo ne so ." Wh at's the matter - d on 't theists believe in the Biblean y mo re? In the future I will sub stitu te "F -cond itional" for"coun terfactual of freedom" w he n I qu ote P lantinga.

31. I di dn 't realize, until I reread this cha pter, just how a pt myabbreviational terminology is.

32. A s w e shall see , in SP Planting a hin ts at a different ve rsion ofthe FWD based on a weaker account of omniscience in whichGod's excuse is the Reagan one of ignorance.

33. Robert M. A dam s, "Middle Kno wledge and the Problem ofEvil," American Philosophical Quarterly 14 (1977), pp. 109-17.

34. This is no t ho w Plantinga defines "strong ly actua lize," bu t aswe shall see, it is how he should have.

35. "Prio r" can m ea n either prior in time or prior in the order ofexplanat ion or determinat ion, depending on whether we areconceiving of God as having, respectively, omnitemporal ortimeless eternality. That F-conditionals have the truth-valuesthey do helps to explain or determine God's creative decision,

398




regardless of whether God's decision occurs in time. I wouldhope that the polemic against God's having timeless eternalityin Cha pters 2 and 3 ha s convinced the reader; but know ingthat philosophy is philosophy, I want to formulate my argument so that i t does not presuppose that God has omnitempo-ral eternality.

36. William W ainwright, "Freedom and O m nip ote nc e," Nous 2(1968), p p . 293-301.

37 . Ibid., p. 301.38. This analog y is elaborately dev elo ped in Nelso n Pike, "Plan-

tinga on Free Will and Evil," Religious Studies 15 (1979), pp.449-73. Another theological compatibilist attack on Plantingacan be found in James E. Tomberlin and Frank McGuinness,"God, Evil, and the Free Will Defense," Religious Studies 13(1977), pp. 455-75-

39. For readers who are too you ng to rem em ber, Gabby Hayeswas an ugly version of Willie Nelson, if such a thing is possible, who always played the hero's sidekick in the old RoyRogers Westerns.

40. Pike fails to see this bec aus e he gets misled by his analo gy intoconfounding possible with concrete persons, often speakingof the former as having, rather than containing, dispositions toperform various actions. For example, "The actual mouse,"which corresponds to John Wayne in my rodeo version of hismice analogy, "came already assembled - all God did wasallow it to be" by opening the door of the cage containing it(p. 460). Pike's "emanation" account of the instantiation relation is a piece of mystical mischief that further fudges thenumerical distinction between possible persons and theirinstantiator.

41. Ad am s, "M iddle K now ledg e," p. 109.42. Ibid ., p . 117.43. The just-in-the-n ick-of-tim e objection is raised by ou r indefati

gable theological compatibilist Nelson Pike, who writes:"Surely a being of this sort could create a world containingcreatures who freely do what is right but in which no creaturesucceeds in performing actions (perhaps even choosingactions) that are morally pernicious. In cases where a morallywrong action could be expected to result in the suffering andsorrow of other people, it would seem that interference onGod's part would, in fact, be morally required" (pp. 470-1).

399



Notes to pages 152-ji

44. Clement Dore persuasively prese nts this argu m ent in chapter5 of his book, God, Suffering, and Solipsism (Macmillan, Ne w

York, 1989).45. Joel Feinberg en do rse s Locke's cou nterex am ple in his Doing

and Deserving (Princeton University P ress, Princ eton , N .J.,1970), p. 185.

46. N elson Pike gives an excellent acc oun t of suc h cases in his"Ov er-Pow er an d G od's Responsibility for Evil," in The Natureand Existence of God, ed. A. Freddoso (Notre Dame UniversityPress, Notre Dame, Ind., 1983).

47. For a helpful disc ussio n of this see Fe inb erg, Doing and Deserving.

48. Again it is Pike w ho is the objector. He poin ts out that theresult of God's creating free persons without having middleknowledge "might be a world containing creatures who perform no morally right actions at all. In fact, the result mighteven be worse-a world containing creatures whose everymorally significant action is wrong and is such as to producean amount of suffering that would boggle even the most calloused imagination. What should we say of a being whowould risk actualizing a world of this description simply on

the chance that it might contain a balance of morally rightaction over morally wrong action? 'Perfectly good'? Hardly.'Reckless' would be more to the point" ("Plantinga on FreeWill and Evil," pp . 455-6).

49. A w rong m ove for the theist to ma ke in m eetin g this objectionis to argue that God did not take a very big risk, because,although he lacked middle knowledge, at least he knew thatit was probable that his creation of free persons would resultin a favorable balance of moral good over moral evil. The

probabilities of w hat w ou ld result from th e insta ntiatio n of theantecedents of F-conditionals must be determined a priori ,since there are no actual prior cases to which God couldappea l. And I do not see any w ay in wh ich i t could be determined a priori that this desirable outcome has a probabilitygreater than .5.

50. Adam s, "Middle Know ledge ," p . n o .51. Were som eone to den y thesis I on the gro un ds of a warr ante d

assertibility theory of truth, I would object that truth andwarranted assertabil i ty are not the same thing and prove mycase by pointing out that while there are brands of cigarettes

4




named "True" and "Fact," there never was and never will beone named "Warrantedly Assertable," for no one would buyit , not even John Dewey.

52. A dam s, "Middle Kn ow ledge," p. 112.53. Ibid ., p . 117.54. A da m s gives a pene trating crit ique of such analyses in his

essay.55. In this conne ction the reader is rem ind ed of prop osition 55 -

that it is impossible that an agent deliberate while knowingwhat decision she will make - or of an event for which herdecision is a causally necessary condition.

CHAPTER5

1. Da vid Lew is, On the Plurality of Worlds (Blackwell Publisher,Oxford, 1986), p. 2.

2. For Lew is, the Kan tian thesis that necessarily all times(spaces) are temporally (spatially) related holds intra- but notinterworld.

3. Lewis, On the Plurality of Worlds, p . 93.4. Ibid. Lewis ad ds that "th e 'actual at ' relation betw een w orl ds

is simply identity."5. Ibid ., p . 2.6. Lewis, Counterfactuals (Harvard University Press, Cambridge,

Mass., 1973), p. 183. This chapter is reprinted in The Possibleand the Actual, ed. Michael Loux (Cornell University Press,Ithaca, N.Y., 1973), pp. 182-9. Future references are to thepagination in the latter volume. That Lewis commits a category mistake in conjoining this definition with his insistencethat worlds are maximal spatiotemporal aggregates is notedby Robert Stalnaker, "Possible Worlds," Nous 10 (1976), pp.65-75, a n d Peter van Inwagen, "Indexicality and Actuality,"Philosophical Review 89 (1980), p p . 40 3-26 .

7. Lew is asse rts that othe r w orlds are no m ore abstract or linguistic entities than is the actual world in On the Plurality ofWorlds: "But w orlds, as I und erstan d them , are not like s toriesor story-tellers. They are like this world; and this world is no

story, not even a true story" (p. 7, n. 3).8. This ar gu m en t from th e indexicality of actua lity is no t Le w is's

only argument for actuality being world relative. There is also

401




an elaborate cost-benefit argument of the inference-to-the-best-explanation sort to the effect that a systematic analysis of

a number of concepts, including modality, causality, propositions, and properties, fares better under his theory of extrememodal realism than under any rival one that takes a possibleworld to be either a linguistic entity or an ersatz abstract entitysuch as a maximal compossible set of properties, propositions, or states of affairs. This argument forms the substanceof chap ters 1 an d 3 of On the Plurality of Worlds. Space will notpermit consideration of this argument, and thus Lewis's casefor extrem e m od al realism can survive the d em ise of the a rgument from the indexicality of actuality.

9. van Inw ag en "Indexicality and Ac tuality," p. 407.10. See Lew is, Counterfactuals, p. 185.n . David Lew is, "A nselm an d Actual ity," Nous 4. (1970), p p . 17 5-

88; reprinted, with added postscripts, in idem, PhilosophicalPapers (Oxford University Press, New York, 1983), vol. 1.

12. See Lew is, On the Plurality of Worlds, pp . 92-3 , and Counterfactuals, p. 86.

13. Lew is, "A nse lm an d Ac tuality," pp . 184 -5. My italics.

14. Ibid ., p . 185.15. Robert M. A da m s, "Theories of Actuality," Nous 8 (1974),p p . 211-31; reprinted in The Possible and the Actual, ed. Loux,p. 205. Future references are to the pagination in the latter.

16. This point see m s to be missed by the cha rges of inconsistencyraised by Michael Loux in his introduction to The Possible andthe Actual, p . 47, an d W illiam L ycan in " The Trouble w ithPossible Worlds," in ibid., p. 289.

17. Lewis, "A nselm and Actuality," pp . 186-7.

18. Lewis, On the Plurality of Worlds,

p. 94.19. Lewis, Counterfactuals, p. 185.20. A da m s, "T heo ries of Ac tuality," p . 195. Lew is res pon ds in On

the Plurality of W orlds, p . 125.21. Allen H az en , in "O n e of the Truths abo ut Ac tuality," Analysis

39 (1977), p. 3, denies that there is any connection betweenthem: "Semantically, logically, the indexical theory is thetruth. This does not answer the metaphysical question aboutthe nature of possible worlds; it is » metaphysical question."

22. Lewis, "A nse lm an d Actuality," pp . 186-7.23. Ibid., pp . 184-5 .24. Ibid., p . 185.

402



Notes to pages 186-9?

25. This is the analogy tha t Lew is himself use s. It m ak es no difference to the argument if the analogy is drawn with some otherindexical term . It m ight seem to som e that the ana log y is closerif it is drawn to "here," since a possible world is a place. Thisis wro ng: A wo rld, i .e. , a maximal spatio tem pora l agg reg ate,although it contains places, is not itself a place.

26. See Lewis's postscript to "An selm an d Actuality" in his Philosophical Papers, p. 22, as well as "An selm and A ctua lity," p .185, and On the Plurality of Worlds, p . 94.

27. Lewis, On the Plurality of W orlds, p . 94.28. The following discussion builds up on van Inw ag en , "Ind ex-

icality an d A ctuality," pp . 418-1 9, from w hich I lea rn ed a g rea tdeal.

29. For a fuller discussion of these pro ble m s, the re ad er is referredto the account of A-propositions in Chapter 3.

30. Lew is fails to see th e ne ed to relativize trut h to a w or ld , as isdo ne in 9' , w he n he says th a t" 'This is the actual w orld ' is t ruewhenever uttered in any possible world" (my italics; "Anselmand Actuality," p. 186).

31. van Inw agen explores this strategy in "Indexicality and A ctuality," p. 422.

32. Lewis, "A nselm an d Actuality," p. 188. A similar rem ark ismade, but in an inaccurate way, in Counterfactuals: "My indexical theory of actuality exactly m irro rs a less con trove rsial do ctrine about t ime. Our present t ime is only one time amongothers. We call it alone present not because it differs in kindfrom all the rest, but bec ause it is the time w e inh ab it" (p. 184).But we do not inhabit just one time (although we mightinhabit only one world if counterpart theory is true). That we

inhabit more than one time is true whether we conceive ofourselves as enduring substances or bundles of successiveperson-stages.

33. My discussion here is indeb ted to A da m s, "Th eories of Ac tuality," esp. p. 160.

34. A parallel problem plagu es a simp le prop erty theory of being now or the present according to which "X is now" isanalyzed into "X has the property of nowness," in which thelatter means "X now h as the proper ty of no w nes s ." An examination of McTaggart 's vicious infinite regress argumentagainst any individual being pr ese nt is instructiv e in thi sconnection.

4 3



Notes topa$g$2m*-i6

CHAPTER 6

1. I do not claim that my use of "abstra ct enti ty" is the "official"or even the most common one in the history of philosophy.W hat is not a me re terminological point is w he the r an abstractentity in my sense is capable of causal agency and serving asthe accusative of prepositional attitudes. Peter van Inwagen,in a very penetrating article, "Ontological Arguments," Nousu (1977), denies that non spa tiotem por al entities such as"numbers, properties, propositions, sentence-types, sets" canbe objects of "love, ha te, w or sh ip . . . fear" (p. 380). T hisseems wrong, since a noun "thaf '-clause can properly serveas the accusative of a propositional attitude verb; e.g., I canfear that John is using drugs. But no spatiotemporal locationcan be assigned to that John is using drugs. St. Thomas lovedGod, and yet he thought of God as being nonspatiotemporal.Whether an abstract being can be a causal agent is more controversial, but plainly Anselm and the other medieval theiststhought that God was both nonspatiotemporal and capable oftimelessly bringing about effects in time.

2. Ross's version is pu t forth wit h great elab oration in his Philo-sophital Theology (Bobbs-Merrill, New York, 1969), pp. 173-81.

3. For a fuller discu ssion of this kin d of ex plan ation , see cha pte r2 of my Negation and Non-Being, A merican Philosophical Quarterly, mon ogr ap h no . 10 (1976).

4. St. Anselm's Proslogion, trans . M. J. Ch arlesw orth (NotreDame University Press, Notre Dame, Ind., 1979), pp. 169-71.

5. "M ," "L," and " - " are mo nadic operators that abbreviate,respectively, "it is logically possible that," "it is logically necessary that," and "it is not the case that."

6. No rma n Malcolm, "Anselm 's Ontological A rgu m en t," Philosophical Review 69 (i960), p . 48.

7. Ibid.8. Ibid., p p . 49-5 0.9. Ibid., p . 48.

10. Malcolm's failure to note that such lang uag e ga m es are actually played is due to a failure to take ord inary usa ge seriously ,

which is surprising in an "ordinary language" philosopher.For example, in support of his timeless conception of God heappeals to the Ninetieth Psalm: "Before the mountains were

4 4



Notes to pti$e8 216-40

brought forth, or ever thou hadst formed the earth and theworld, even from everlasting to everlasting, thou art God"(ibid., pp. 55-6). This is the conception not of a timeless but ofan om nitem pora l D eity. For a good critique of Malcolm on thispoint see Pike's God and Timelessness, pp. 183-4.

11. Malcolm's first arg um ent doe s not, because its prem ise 29says that it is impossible that God have duration. An omnitemporal God, although he can neither begin nor end, doeshave duration and thereby does not satisfy this premise.

12. S t. Anselm's Proslogion , trans. Charlesworth, p. 117.13. In "Anselm and Actuality," David Lewis interprets Anselm ,

incorrectly in my opinion, as requiring only the psychologicalpossibility of existence, bu t he retracts this interpretatio n in hislater postscript to this essay in his Philosophical Papers, vol. 1.

14. I hav e upd ate d Malcolm's example from "Anselm and Actuality," p. 45. Lewis claims that if persons A and B were to include the same properties on their list of desirable propertiesfor a secretary of defense, save for A alone including beingexistent, "any persons who satisfied A's description wouldnecessarily satisfy B's desc riptio n an d vice versa " Malcolmmust be wrong, because if there were to be a merely mythicalor fictional secretary of defense who satisfied B's description,he still would not satisfy A's.

15. My critique closely follows that of Lewis in "A nselm andActuality."

16. Th is is a som ew hat simplified version of Plantinga's argu m ent.17. Plantinga, The Nature of Necessity, p . 218.18. This type of exam ple is dep loye d against Plantinga's oncologi

cal argument by W. Tooley in "Plantinga's Defence of theOntological Argument," Mind 90 (1981), pp. 422-8. See alsoPatrick Grim, "Plantinga's God and Other Monstrosities,"Religious Studies 15 (1979), pp. 91-7, for further examples ofproblematic properties that are supposed to satisfy conditions(i) and (ii).

CHAPTER 7

1. For an excellent survey and deva stating criticisms of suchscientifically rooted cosmological arguments, see Adolf Grun-baum, "The Pseudo-Problem of Creation in Physical Cosmol-

405



Notes to pages 240-%}

ogy," Philosophy of Science 56 (1989), pp. 373-94, and its seque l,"Pseudo-Creation of the Big Bang," Nature 344 (1990), pp.8 2 1 - 2 .

2. Richard Taylor, M etaphysics, 2d ed. (Prentice-Hall, EnglewoodCliffs, N.J., 1974), p. 105. The Leibniz quotation is from LeibnizSelections, ed. P. W iener (Scribner, New York, 1951), p. 346.

3. I have availed myself of William Rowe's mounting of both themain and subsidiary arguments in his Cosmological Argument(Princeton University Press, Princeton, N.J., 1975). All subsequent references to this book are cited in the text of thischapter.

4. Some causal loo ps occasion special difficulties. For exa m ple, aperson gains knowledge of how to build a time machine froma conversation with his future self, who has traveled back intime in order to impart th is know ledge to him; and his futureself 's knowledge is due to his r emem ber ing (a causal process)what he heard w hen he was the earlier self. Herein the re is anextra problem concerning how this knowledge was derived.But such extra difficulties are not endemic to every case of aclosed causal loop.

5. Tha t Clarke m ust ma ke this existential assum ption, as well as

assume PSR„ clearly brings out the a posteriori status of hisargument . In Part 9 of the Dialogues Concerning Natural Religion, Hume characterizes an argument that obviously isClarke's as being a priori." William R ow e, in correspondence,suggested that the meanings of a priori" and a posteriori"were different in Clarke's time th an they are now . These w erenot, as they are for us, epistemologically based distinctions,but rather concerned whether we are inferring a cau se from itseffect or vice versa. This issue is discussed in James P. Fergu

son, The

Philosophy of

Samuel Clarke and Us

Critics (VantagePress, New York, 1974), pp. 11-21.-H. The text of the debate is repr inted under the title The Exis

tence of God: A De bate betw een Bertrand Russell and FatherF. C. Coples ton ," in The Existence of God, ed. John Hick(Collier-Macmillan, London, 1964), p. 175.

7. Ibid.8. Paul Ed wa rds, The Cosmological Argument," Rationalist

Annual 1959 (Pem berron, L on do n); reprin ted in Readings in thePhilosophy of Religion, ed. Brody. The quo tation is from p . 78 ofthe latter.

406






Demonstration Clarke asks, "What is it that has from eternitydetermined such a succession of [dependent beings] to exist,

rather than that from eternity there should never have existed anything at all?" (my italics). Herein Clarke is demanding anexplanation bo th for wh y this infinite succession of d ep en de ntbeings exists and for why there exists at least one dependentbeing, which is Rowe's explanatory demand.

15. Rowe is well aw ar e of this difficulty w ith his arg um en t andexplores m an y different w ay s of revising th e PSR so tha t it willapply to fact D. He continues his discussion in "RationalisticTheology and Some Principles of Explanation/' Faith and Phi

losophy 4 (1984), pp . 357-69. It turn s o ut to be a very difficultthing to do, and I will not go into the issue here.16. For a dis cu ssio n of this pro blem , see Carl G. H em pe l, Aspects

of Scientific Explanation (Free Pr ess , N ew York, 1965), pp .293-5.

17. Hilary P ut na m , "Reductionism an d the Na ture of Psychology," Cognition 2 (1973), pp. 131-46; reprinted in Mind Design, ed. J. Haugeland (MIT Press, Cambridge, Mass., 1981).

18. For an inter estin g discussion of pragm atically deficient cos-

mological arguments, see Michael B. Burke, "Hume andEdwards on Why Is There Something Rather Than Nothing,"Australasian Journal of Philosophy 62 (1984), p p . 355-62. Unfortunately, Burke accepts at face value the HE principle: "I certainly agree with Hume that to explain the existence of eachmember of a collection of twenty particles is to explain theexistence of the 'whole twenty' " (p. 360).

19. For a full exp ositio n of Bergson 's the or y, see my "B ergson'sAnalysis of the Concept of Nothing," Modern Schoolman 51

(1974), pp. 269-300.20. For full de tails , see m y Negation and Non-Being.21. Th ere will be consid erab le com plexity in B, since an infinite

series can be ge nera ted by app lying certain truth-functionaloperations to any set of propositions, but there need not beany thin g vicious abo ut this; and if it sho uld turn out to be so,a requirement can be made that W be in conjunctive normalform.

22. This princip le, P, in a somewhat simplified form, states: If in

one world God determines whether a certain type of entityexists, then in every world in which God exists he determineswhether this type of entity exists.

4 8




CHAPTER 8

1. William Alston , "Psyc hoan alytic Th eory an d Theistic Belief,in Faith and the Ph ilosophers, ed. J. Hick (St. Martin's Press,New York, 1964), pp. 89-90.

2. William W ainw right, Mysticism (University of WisconsinPress, Madison, 1981), p. 82.

3. William Alsto n, "The Christian Lang uag e-G am e," in TheAutonomy of Religious Belief, ed. F. Cro sson (Notre Dam e University Press, Notre Dame, Ind., 1981), pp. 128-62.

4. Alston inclu des the existence of defeating con dition s or test s,but I am dr op pin g this requirem ent since I w an t to start w iththe weakest version of the analogical argument for the cogni-tivity of religious experiences.

5. Alston, "The Christian La ngu age -Ga m e," p. 143.6. Q uin n's excellent "In Search of the Fo un da tio ns of T he ism ,"

Faith and Philosophy 2 (1985), pp. 469-86, raises objections totaking the proposition that God exists as basic that parallelsome of the objections that I shall make to fideism.

7. W ittgenste in's claim that religious reas onin g "is an entirelydifferent kind of reasoning" seems to impute epistemologicalautonomy to the religious language game and serves as thebasis for his claim that the nonbeliever who denies that therewill be a Last Judgment does not contradict the theist whobelieves there will be. See his Lectures and Conversations onAesthetics, Psychology and Religious Belief, ed. C. Barrett (University of California Pre ss, Berkeley, 1967), pp . 53, 58.

8. C. B. M artin, "A Religious Way of K no w ing ," in New Essays inPhilosophical Theology, ed . A. Flew an d A. M aclnty re (SCMPress, London, 1955).

9. For a th or ou gh discuss ion of this issu e, see William A lston ,"Epistemic Circularity," Philosophy and Phenomenological Research 47 (1986), pp. 1-30.

10. See William Al ston , "Perceiving G od ," Journal of Philosophy 83(1986), pp. 655-65, for a discussion of this point. Alston overlooks a respect in which the two practices are asymmetrical.

Whereas the tests for the veridicality of a sense experience arebased solely on other sense experiences, those for the veridicality of a religious experience are not based solely on otherreligious experiences, since the latter are based on what we

409




hear other me mb ers of the religious co m m un ity say abou t the irreligious experiences, what we see inscribed in certain "holy"

books, etc. This will becom e clearer as we con sider these tests.11. Richard Sw inburne, The Existence of God (Clarendon Press,Oxford, 1979), p. 260.

12. William Alston, "Religious Diversity an d Percep tual Kno wledge of God," Faith and Philosophy 4 (1988), p. 434.

13. Ibid., p . 435.14. William Alsto n, "Referring to G od ," Philosophy of Religion 24

(1988), p. 121.15. Herein W ainwright is siding with Za eh ne r again st Suzu ki,

Stace, and M erton as to w he the r theistic and m onistic m ysticalexperiences are phenomenologically different, and as indicated earlier, this is an issue I want to avoid.

16. Gary Gu tting, Religious Belief and Religious S kepticism (NotreDame University Press, Notre Dame, Ind., 1982), pp. 149-50.

17. Alston, "The Christian La ng uag e-G am e," p . 159.18. Ibid. , p . 161.19. William Alston, "The Perception of G od ," Philosophical Topics

16 (1988), p . 50.20. Sw inbu rne applies his analysis to every type of of-God expe ri

ence, including perceiving worldly items as God-caused. LikeAlston, he makes too strong a claim on behalf of the latter:"That God is at work is no inference for these men but whatseems (epistemically) to be h ap pe n in g " (p. 253). This m ake s itappear as if these are recognitional seeing-as experiences andare occasions for nontheo retically ba sed lan gu age -en try assertions; but as shown earlier, the Quinn-type game players arewilling to give inductive backing to their assertion that the

, sunse t, for exam ple, is an expression of G od's love, wh ichentails that it is God-caused.

21. C. D. Broad, Religion, Philosophy and Psychical Research (Har-court, Brace, New York, 1953), p. 197.

22. Swinburne, The Existence of God, p. 262.23. Ibid.24. Gutting, Religious Belief, p. 255.25. Swinburne, The Existence of God, p. 255.26. Ibid., p . 270. The qualification "w he re he is" is pu zzlin g, since

a theistic religious experience is not of an embodied or spatially located being; and this is so even if God were to have(accidentally) the whole world as his body.

410




27. Wainwright, Mysticism, p. 130.28. For insightful acco unts of the ma nn er in which the Catholic

tradition dealt with the problem of deviously or ignoblycaused mystical experiences, see Nelson Pike, "On MysticVisions as Sources of Knowledge," and George Mavrodes,"D ecep tive Mystical Exp eriences," both in Mysticism and Philosophical Analysis, ed. S. Katz (Oxford University Press, NewYork, 1978).

29. W ayne Proudfoot, Religious Experience (University of California Press, Los Angeles, 1985), p. 139.

30. Alsto n, "Psychoanalytic Th eory ." His later writings are m orein agre em ent with W ainwright's and Sw inburne's on the issueof whether a veridical religious experience can have a naturalcause.

31. Wainwright , Mysticism, p. 131.32. This is in substan tial agre em ent with the "cognitive-w allop"

theory of mystical experiences of John of the Cross, accordingto which they are not a source of propositional knowledge butinstead serve to deepen our understanding of truths that aregained by nonmystical means. For a good account of thistheory, see Nelson Pike, "John of the Cross on the EpistemicValue of Mystic Visions," in Rationality, Religious Belief, andMoral Commitment, ed. R. Audi and W. Wainwright (CornellUniversity Press, Ithaca, N.Y., 1986), pp. 15-37.

33. The re is an aspect of the problem of religious dive rsity inaddition to that of a diversity of incompatible tests amongdifferent religious traditions; it concerns the fact that theap pa re nt object of religious experienc es is described differently by experients within these different traditions. Hospersand Flew were too quick to impugn the cognitivity of religiousexperiences on this ground; for the experients either couldhave been describing one and the same reality in differentways or could have been describing numerically distinct realities, God, nirvana, etc., being numerically distinct objectiverealities. The latter is not very plausible, since all of the experients purport to be describing some highest, ultimate reality,and there cannot be more than one such reality.

34. Swinburne , The Existence of God, p. 275.35. Wainwright , Mysticism, pp. 87-8.36. A lston , "The Perception of G od ," p. 42.37. Alston , "The Christian Lan guage -Gam e," p. 161.

4 11




38. Peter M oore gives an analogical ar gu m en t that is also am biguous between a language-game fideism with any old tests and

a language-game fideism with analogous tests, because it doesnot place sufficiently strong requirements on what counts asa "similar" or, in his case, "equ ivale nt" test. "W here th e objective validity . . . of any form of non-mystical ex perience is indoubt, recourse can be had to a variety of tests and checkingprocedures, although these do not always sett le doubts ordisputes. But mystics too have their own equivalents of suchtests and checking procedures for the assessment of mysticalexperienc e. Th ese are com prised on the one han d in the adviceand instructions given by mystics concerning the ethical,ascetical, and technical pre-conditions of mystical experienceand on the other hand in the detailed moral and psychologicalcriteria used in mystical traditions to determine the authenticity and value of mystical states and revelations" ("MysticalExperience, M ystical Doctrine , M ystical Tec hniqu e," in Mysticism and Philosophical Analysis, ed. Katz, p. 126).

39. It is intere sting that alth ou gh Wainw right defe nds the analogical arg um en t on the basis of analog ous tests, he is not willingto rest his case for the cognitivity of religious experience on it.In M ysticism, p. 102, he claims that even if it were not to work,due to religious and sense experience being radically dissimilar, religious experience could still qualify as cognitive inStace's "trans-su bjective " s en se . This wo uld b e to retreat to thesort of multivocalist account that underlies extreme language-gam e fideism - a very unattra ctive position. It sho uld bepointed out that there is no inconsistency in Wainwrightespousing at one and the same time both extreme language-game fideism and the analogical argument based on analogo us tes ts, this paralleling the p oint m ad e on p . 1 of thisbook that one can both espouse fideism and give argumentsfor God's existence. However, philosophers, l ike trapezeartists, usually eschew such safety nets, since they take awayfrom the bravura of their performance.

40. Alston , "Perce iving G od ," p . 655.41. Alston, "The Perception of G od ," p . 27.42. Alston , "Religious D iversity," p . 443.43. Ibid.44. Ibid., p . 444.45. Ibid., p . 445.

412



N&tes to pages 324-38

46. Ibid.47. No doubt their langu age-e ntry assertion s involve som e the o

retical elements, but they are common to both their theories.48. I wa nt to avoid comm itting myself on the vexing questi on

whether the apparent object of a veridical sense experiencem ust be a nec essa ry, sufficient, nec essa ry and sufficient, orjust plain old cause of the experience. I favor the last, since Ibelieve that what is necessary is that the object be a cause orpart of the cause of the experience in the right way. Ourworldview's notion of causal nexi gives us some start on thisnotion , as P. F. Stra w son h as po inte d ou t ("Cau sation in Perception," in Freedom and Resentment [Methuen, London,1974]), in terms of tautologies concerning perceptual perspectives and the masking of one object by another; but an essential part of this commonsense view is that it is open toexpansion and revision, as Grice has remarked, in the light ofscientific advancement. Determining what is the right sort ofcausal connection between object and experience, therefore,might have to await the coming of the scientific millennium.

49. The material in this para grap h is extensively dev elope d a nddefend ed in ch ap ter 2 of m y Negation and Non-Being.

50. The sense of objective particular that I am wo rking w ith isweaker than the one employed by Strawson. For him anobjective particular must not only be capable of existing whennot actually perceived but also be distinguishable by a subjectof experience from her own states of consciousness, which isto require self-consciousness on the part of this subject.Whether the possibili ty of self-consciousness depends on therepresentation of objects as being in space will not be considered. Strawson's answer to this question is indecisive, sincehe thinks that ou r con cep t of the self is no t sufficiently pre cis e.

51. Strawson, Individuals (Anchor Books, New York, 1963), p. 67.52. Ibid., p . 66.53. See no te 49.54. Strawson, Individuals, p. 70.55. Ibid., p . 68.56. Ibid., p. 69.

57. This oversight infects Jona than B ennett 's recon struction ofStrawson's sound world in his Kant's Analytic (CambridgeUniversity Press, 1966), p. 37. For an insightful discussion ofStrawson's sou nd wo rld, see M. R. Ayers, "Perception an d

413




Action," in Knowledge and Necessity, Royal Institute of Philosophy Lectures, vol. 3 (1968-9) (London, 1970).

58. Plato, Timaeus, 50E-51. Q uoted from F. M. C ornford, Plato'sCosmology (Humanities Press, New York, 1952).59. "The principle is so phra sed that ho w thin gs seem positively

to be is evidence of how they are, but ho w thin gs seem not tobe is not such evidence" (Swinburne, The Existence of God,p . 254).

CHAPTER 9

1. There is also the pragmatic arg um ent, sk etched in the finalparagraph of the preceding chapter, based upon the spiritualbenefits to the person who believes her religious experience isan objective revelation, b ut we w ill not c ons ider it since ev erything that will be said about the moral-development pragmatic argument can be extended to it.

2. This is a bit quick, since ou r finitude h as not prec lud ed usfrom gaining an understanding of infinity in mathematics. Itdoesn't take one to know one.

3. For a mo re com plete account, see Nicholas Rescher, Pascal'sWager: A Study of Practical Reasoning in Theology (Notre DameUniversity Press, Notre Dame, Ind., 1985).

4. St ep he n P. Stitch raises this objection to Pascal in his"Recombinant DNA Debate," Philosophy and Public Affairs 7(1978), pp. 187-205.

5. No t even Pascal himself ad he red to this neg ativ e theo logy , forin later portions of the Pensies he marshals historical evidencein favor of the reliability of the Bible, thereby according a

privileged epistemic status to Christianity. For a discussion ofthis, see Thomas V. Morris, "Pascalian Wagering," CanadianJournal of Philosophy 16 (1986), pp. 437-54.

6. This is the policy advo cated by William G. Lycan and G eor geN . Schlesinger in their "You Bet Your Life: Pascal's WagerDefended," in Reason and Responsibility, 6th e d. , e d. J. Fein-berg (Wadsworth, Belmont, 1988), pp. 82-90.

7. Richard Sw inburn e, "The Christian W ager," Religious Studies 4(1969), p. 221.

8. W. K. Clifford, "The Eth ics of Belief, in Lectures and Essays(Macmillan Press, London, 1879); reprinted in Readings in

414






22. These concep tual truths need tidying up along the linesspelled out in chapter 5 of my Language of Time.

23. This asym me try betwee n agent and observer is m ade possibleby the fact that the premises of the inductive argument permitbut do not make mandatory the dra w ing of the conclusion, asTeddy Seidenfeld pointed out to me.

24. That we are not permitted to opt out of the moral responsibility game does not entail that a game player must believe thatshe is free (whether or not she is), but only that she must actas if she is free, which she does by playing the game. Thisincludes her willingness to make this performatory utterance:"I hereby ask you to hold me morally responsible for myactions." This has an imperatival illocutionary force and doesnot logically or pragmatically entail "I believe that I shallbehave in a morally responsible wa y in the future," an y m orethan "I command you to shut the window" entails "I believethat the window will be shut."

25. When she perform s thes e future altruistic actions, she lacksgood reasons for doing them, and thereby does not qualify atthat time as either a rational agent or m orally responsible;however, at the time she takes the pill she is morally responsiblefor both taking the pill and these future altruistic actions thatresult from it. This shows that an agent's moral responsibilityfor a given act must be relativized to a time within her life. IfI intentionally take a drug to bring it about that at a time T inthe future I beco me a m ental baske t case, I am responsible atT for both taking the drug and my future state of mentalincapacitation. But subsequent to T, I am no longer morallyresponsible for these things.

26. N othin g that has been said so far entails that one is neve rmorally justified in self-inducing a belief by nonra tion al m ethods. In fact, I can imagine cases in which one has a sufficientmoral reason to do so. Consider a case in which A has adequa te epistem ic gro un ds for believing a prop osition p but forpsychological reasons is unable to believe it; and A can successfully perform some morally desirable action only if shefirst believes p. Imagine, to use Nancy Davis's favorite example, a diamond cutter who cannot successfully cut a certaindiamond unless she first believes that she is capable of do

ing so. She has ample inductive evidence from her past successes that she is capable of doing so but because of a lack of

416





Index

a c t u a l i t y, 3 6 , 1 7 9 - 9 8 , 2 7 7 ; see also G o d ,divine perfections of, absoluteactuality

Adams, Rober t M. , 32 ,113-14 ,129 ,131 ,134-6 ,138 ,149-50 ,169 , 170-5 ,183-4 ,233/ 3 95 1" 4, 401 n5 4, 403 r>33

Almog, Joseph, 393 ni6Alston, William, 2, 289, 292-4, 298-301,

305-6, 309, 312- 26, 343 , 390 n7Anselm, St . , 202,205-8, 212-13, 216-24,

345, 389 n3A-proposi t ions (see also indexical terms

and propos i t ions) , 57-60 ,6 2-3 , 65-6 ,68, 71-4 , 88-90 ,96

definition of, 59-60entai lm ent relat ions b etw een, 79, 82, 96,

193-4as expressible by quasi- indica tors , 82-9as reducible to B-propo si t ions, 6 0-2 ,67 ,

71-82 ,195-6tokening rules for, 59, 65-73,77

Aq uina s, St . Th om as, 150, 239Aristo tle, 24, 35,62, 95 ,11 9, 241, 272,

323-4atheological arg um en ts (see also ontological

dispro ofs), 36 -7 ,9 7, 198, 237def ined , 3-4,15based on benevolence and creat ion, 30-2based on creat ion and imm utabi l i ty,

Chapter 2, 35based on disappearance of t ime, 46based on evi l , Chapter 4based on necessary exis tence, 227-37based on omnipotence, 15,18-20, 394 n5based on omnipotence and benevolence ,

16,20based on omniscience and freedom, 16,

34

based on omnisc ience and imm utab i l i ty,Ch apt er 3, 6, 35, 56

based on sovere ign ty and benevo lence ,33-4

based on world-relat ive actual i ty,Chapter 5, 36

as pure and im pure , 16-7August ine, St . , 37-47, 58,67, 207, 272Ayer, A. J., 396 n2iAyers, M. R , 413 n57

Baier, Annette, 253, 373Barn hart, J. E , 396 n20Bayesian proba bility, 1-2Beer, Michelle, 393 ni5belief

as voluntary, 346, 360, 368-70and behavior, 365-8

benevolence , see God , d iv ine perfectionsof

Bennet t , Jonathan, 413 n57Bergson, Henri , 40,92, 273Berkeley, 215, 325blame, 153,155-7, "75Blumenthal , David, 161-2Boethius, 47, 207B-propositions (see also B-reduction) , 6 o- i ,

72-4 , 86 ,88-9defined, 60

Brai thwaite , Richard, 291Brandom, Robert , 372B-reduc t ion , 60- 3 , 67 ,71-82Broad, C. D ., 221, 307Burge, Tyler, 393 ni6Buridan's ass, 390 n9

Burke, Michael , 408 m 8Butterf ie ld, Jeremy, 393 ni6Butler, Bishop, 320

418



Index

Cam pbe l l , C . A. , 119, 397n2 3C a r n a p , Rudolf, 4Carr- W iggin , Robert , 393 n27

Cas t an eda , Hec to r-Ner i , 60 , 62 - 3 , 82 -9 ,94- 187

causa t ion , 53 , 15 4-6 ,16 1, 245-6 , 250 ,257 -8 , 315, 328, 375-6

Chisholm, Roder ick , 119Chr is t ian i ty, 46 , 57 ,109 , 293 , 300 -1 , 346,

353. 372Clark e , Samu el , 244 -8 , 252, 258-9 , 263 ,

265, 270, 282, 407 ni4Clifford, W. K., 355-8, 361-4, 368, 370Co bu rn , Rober t , 65-7 , 71 , 74 -5 , 78 , 83 , 89cogni t iv i ty, 287 -8 , 295 , 297 , 303-4 , 311 -13 ,

316, 325-6, 343compat ib i l i sm, 117 ,120

causa l , 115 , 118-22 ,148-9theologica l , 115 , 122-3 ,146-7- 149-50 .

164-5,3991143Coples ton , Fa ther, 251c o s m o l o g i c a l a rg u m e n t s , C h a p t e r 7 ,

238-42 , 280-1def ined , 239-42

cosmologica l a rguments , vers ions ofconcre te aggrega te , 242-58 , 279 , 281-$existential fact, 242, 258-77, 279, 281

world book, 242, 277-9, 281

Dan cy, Jona than , 166 -8 ,175 -6Dav idson , Dona ld , 29Davis , N ancy, 416 n26defe nse , 9 8-9 , 1 03-5 , 109-10 , 229 , 236

based on free wi l l, Ch apte r 4 ,1 06 , n o ,229-34

del ibera t ion , 167-8 , 365 , 379-81Dennet t , Danie l , 159De scar tes , 240, 324Dewey, John, 92 , 95Donnel lan , Kei th , 6Do re , Clem ent , 22 , 400 n44Ducasse, C. J . , 313-14

Eckhard t , Meis te r, 50Edwards , Paul , 253 , 255-6Epicurus , 99e terna l i ty, see Go d, d iv ine per fec t ions ofevil, 16, 30, 32, 98 -131 , 148 -5 2 ,155 -7 , 69 .

205, 228-37, 364mora l and na tu ra l 110 -12 ,151 -3 , 398 n29

exis ten t ia l e th ics , 352 -3exp lana t i on , 115 -16 ,166 , 204 -5 , 240-1 ,

243, 25 1-2 , 265 , 280-1based on comm on cau se , 254 -5 , 328

as nonagg lomera r ive , 252 - 3 , 278as nonextens ional , 255-6 , 267as not c losed under deduct ion , 267

pragmat ics of , 267-8as theoret ical ly relat ive, 240-1vicio us circularity of, 257, 259, 261-3 ,

265, 268-9 , 278-9

faith 1, 387Fein ber g, Joel, 400 nn4 5 and 47Fer gu son , Jam es , 406115fideism , 1, 289, 29 4-5 , 412 n39

based on l anguag e -game be ing p l aye d ,291

Flew, A nto ny , 117 , 395 n i8 , 397 nn2 3 an d

2 5 , 4 11 ^ 3 3forms of life, 10, 29, 178f r eedom (see also: def ens e, based on free

wil l ; l ibertar ianism; persons, asmo ra l ly respons ib le) , 116 ,118, 151-2 ,157, 230-5, 281, 371-2, 378

cance l i ng cond i t i ons , 120 -2 ,15 3 ,157 -65value of, 112, 118, 120, 152, 232, 372

free wil l def ens e, see def ens e, based on freewill

free wil l subjunctive condit ionals(F-condi t iona ls ) , see subjunc t ivecon dit io na ls , of free w ill

Frege , Go t t lob , 62 , 75- 6

Geach, Peter, 62-6, 389 niG o d (see also a theologica l a rg um ents ;

onto logica l d i sproofs )hard- and soft-core propert ies of , 7-8,

10 , 20, 178, 219, 300reference to (see also re ference), 4 -1 1 ,

300 -1Go d, ep is temologica l a rg um ents for be lief

'" . 3 . 343-4. 387cosmolog ica l a rgumen t s , see

cosmological a rg um en t sind uct ive ar gu m en ts 1, 241, 296, 387on to log ica l a rgu me n t s , see ontological

a rg u m e n t sr e l ig ious expe r i ence a rgu men t s , see

re l ig ious exper ience a rgumentsGo d, p ragm at ic a rgu m en ts for belief in ,

Chapter 9, 3, 343-4, 387from prudence , 344-54 , 387from mora l i ty, 344-5 , 354-87

G od , d iv in e per fec tions ofabsolu te ac tua l i ty, 35-6 ,179-81abs olu te per fec t ion , 8 , 20-3 , 29 , 34 -5 ,

54.

4*



index

God , divine perfections of (conf.)90 , 92-3 , 95 ,97 ,14 6, 203, 216, 219-20,223-5-233-5- 237

absolute perfection, paradox of, 23 -4,27 , 2 9 , 9 0 , 1 7 6 - 7

absolute s implic ity, 7-8, 23-9 , 40, 54 ,91 ,215

abstractness, 202, 205 -16aseity, 23, 25, 27-8, 54, 93benevolence, 15-6, 20, 24-3 2, 99-110,

130 ,142 -3 ,169 , 22 2-3 , 232-3 , 347creator, 30-2, 35-41, 47, 67-9,121,128,

139-43 ,179 -80 , 266eternal ity, 47-9, 5 1-2 , 69-7 1, 95eternality, as omnitemporality, 33, 58-9,

94 ,96 -7 ,1 67 ,17 7-8 , 215-6 , 398 n35eternality, as t imelessness, 35, 37, 39-41,

47-9- 51 -5- 58-9 / 68 -71 ,90 -1 , 93 , 97 ,167, 207-8, 2 15-1 6,391 n6, 392 n9, 398

n35freedom , 23, 280grace, 32, 149-50, 318, 395 ni 4immutabili ty, 16, 33, 35,38-9, 57-9, 86,

94-6midd le knowledge of , 13 1-6 ,1 38- 9 ,14 2 ,

1 4 7 , 1 4 9 , 1 5 1 , 1 5 5 , 1 5 7 , 1 6 6 - 9 , 1 7 4 - 5 ,232-5

o m n i p o t e n c e , 7 - 8 , 1 5 - 1 6 , 1 8 - 3 0 , 3 4 , 9 0 ,9 9 - 1 0 7 , 111 , 11 3 , 1 2 3 - 4 , 1 4 4 - 6 , 1 5 0 ,270, 394 ns

omniscience (see also God, divineperfections of, middle knowledge), 16,25-6 , 30 , 57 -9 ,6 7-7 , 74 ,86 -91 , 96 ,113 ,131-4 ,170 ,176-8 , 235

persona l, 4 , 24 , 37 ,47 , 52 -6 ,91 -2 ,95 ,178

sovereignty, 8 ,19, 31-4, 39,139-41, 215,231

sovereignty type-invariance of, 140-2,232, 283, 408 n22

triune, 55,216Goedel, Kurt, 245Gr ice H . P., 4131148Grim, Patrick, 405 ni8Grunbaum, Adolf 245, 326,405 niGutting, Gary, 298-9, 305-6, 309-14, 326,

342

Habe rmas, Jurgen, 326Hare, R. M., 291Hart-Honore Thesis , 156-7Haz en, Al len, 402 n2i

Heidegger, Martin, 92Heinlein, Robert, 246,258

Hegel , 372Hempel , Carl C , 115 -6 ,265 ,408 m 6Hick, John, 110

Hobart, R. E., 396 n2iHobbes, 116Hospers, John, 411 n33Hudson, W. D., 291Hu me , 2 ,116, 201, 252-3 , 265Husser l , Edmund, 92

identity, 188-9counterpart theory of, 185, 190,403, n32of eve nts , 26 -9, 375of indiscernibles, 382-9, 335, 337-9, 342,

375

as intentional, 8-10, 301as qualitative and numerical, 334-42as theoretical, 9, 29

incompat ibi l ism see also compatibilism),11 2 , 11 6 , 1 2 7 , 1 3 9 - 4 1 , 23 1 ,3 8 1 n i g

indexical terms and p ropo sitions (see alsoA-pro posi t ions) , 16, 62-4 , 73,18 1-9 7

de re theory of, 76 -82individuat ion

receptacle theory of, 2 71 -2,3 28 -9,3 39 -41lrenaeus, St. , 110

James, Wil l iam, 66 ,92 ,11 9, 288,305-6 ,309-357-71 , 381-6

Jeans, James, 110John of the Cross, 411 n32Judaism, 46, 57,1 09, 300-1

Kant, 45, 235, 238 , 271 , 344, 372, 397 rU3Katz, Steven, 286Kenny, Anthony, 391 n2Kierkegaard, 19, 291, 387know ledge , 6 5-6 , 80 , 87

de re and de dic to , 57-8 , 72, 75, 87 -8transitivity of, 85, 87-9

Kre tzmann , Norman , 25-6 ,48 , 51-5 ,6 9 - 7 1 , 8 2 - 4 , 3 9 2 n 9

Kripke, Saul , 6 , 7 3,8 0,1 91condi t ional excluded middle , 129,135-6,

174exc luded midd le , 129 ,13 5 ,137 ,140 ,1 42 ,

1 5 3 , 1 7 1 - 2

Le ibn iz , 30-1 ,109 ,123 ,125 , 243Lewis, C. I., 207, 277L e w i s , D a v i d, 3 6 , 1 2 9 , 1 8 0 - 9 8 , 4 0 5 n i 5

shorts of, 193

l ibe r ta r i an i sm, 117-22 ,127 ,134 ,148 ,154 ,160, 2 08 -1, 210 -11 , 365, 397 n23

42



Index

Locke, 152,159, 373logical atom ism , 2logical positivism, 2, 201Lukas iew icz , Jan , 62 ,175Lycan, Will iam, 402 ni6 , 414 n6

Mc Closkey , H. J ., 98-9 , 108-9, 395 n l 5McGuinness , Frank , 399 n38Mackie, John, 99-109,114, 118,120, 123-4McTaggart 's arg um en t for the unrea l i ty of

time, 403 n34Malcolm, Norman, 205, 208-16, 221manichaeanism, 37M ann , Will iam, 27-9 , 392 n nMartin, C. B., 298Mavrodes , George , 411 n28means -ends r e l a t i onsh ip

as an integral unity, 354, 382,385Mer to n , Tho mas , 286, 410 n i5midd le knowledge , see Go d, divine

perfections ofMol ina , 116 ,149,170Moore, G. E., 290, 313Moore, Peter, 412 n38Mo rr i s , Tho ma s , 414 n5mysticism, 47-55, 69, 216, 271,38 5-6, 388,

303-4, 315, 410 n i5as ineffable, 49 -51

negat ionincomp atibi l i ty theo ry of, 2 73-5

negative events , 260

object ivi ty, 311-4, 326-40; seealso religiousexp erience , object ivi ty of

om nip ote nc e, see Go d, divine p erfect ionsof

ominsc i ence , see God , divin e perfecHons ofonto logica l a rg um ents , Chapter 6 , 2 ,54 ,

238, 273, 345Anselm's version, 202, 217-24from God's abstractness, 202, 212-17from G od 's unpreven tabi l i ty ( the

Scotus-Ross vers ion) , 202-5Plantinga 's version based on necessary

existence, 202, 218, 224-7, 2^1< 233—5onto logica l d i sproofs , 227-37 , 23 " - 9 -

27 5 -6 , 281-4ord inary language phi losophy, 2O t t o , Rudolf, 286

Parf it , Dere k, 386Pascal , 220, 345-5 3, 369; seealso Pascal's

Wager

Pascal 's Wager, 2, 344-54Pen e lhum , Te rence , 109 , 3 9 4 n n n an d 13Perry, John , 7 5-82 , 88per son s , 92 -3 371-2 ,376, 378 , 382

as mora l ly respons ib le , 93 ,372-82rationality of, 373-5, 377, 382

Phillips, D. Z„ 291P ike , Ne l son , 7 1 -4 ,147 ,153 , 393 n i 4 , 394

n i 2 , 399 nn 38 , 40, an d 43 , 400 nn46and 48, 404 m o , 411 nn 28 and 32

P lan t inga , A lv in , 2 , m , 113 -69 ,1 71 -3 ,176, 224 -34 , 296

Plato, 24, 34, 47, 81,95,109, 126 ,171 , 202,272, 273 -4 , 27&> 29°< 3 27< 34°

Plot inus, 40poss ib le persons , 125-38 , 143-4 ,148, 154 ,

160 ,162 -4 , 166 -8 ,171 -76 ; seealsoposs ib le wor lds

poss ib le wor lds , 112, 125-6 , 180 -1 ,18 3,401 n8

access ib il ity re la tions be tw een , 226 -7prag ma t ic theory of t ru t h , 309 , 362 , 400 n5 ip ropos i t i ona l a t t i tudes , 61 -4 , 81 ,83 -5 , 96 ,

187 -8 ,196pro pos i t ions , 62 , 75 , 171

ident i ty of 61-2 , 73-4Pr ior, A. N. , 60 -2Proudfoot , Wayne , 315psychoanalys is , 323 , 325-6Purt i l l , Richard, 396 n20Pu tna m , Hi la ry, 6 , 268

quas i - ind ica tors , 82-9 , 187ques t i ons -begg ing a rgumen t s , 213Quine, W. V., 181Q uin n, Phil ip , 235, 279, 293, 301, 410 nap

rat ional choice theory, 348-51Reagan, Ronald, 132reference (see also Go d, referen ce to)

casua l theory of , 6 -7 ,10-1 , 300-1descript ive theory of , 5, 8by rig id de s igna to r s , 73 -4 ,1 88 -93 , ' 9 7

Reichenbach , Hans , 245rel igious diversi ty, 297-8, 316, 322-5, 327re l ig ious exper ience , C hap ter 8 , 286-7 , 293

as d i rec t exper ience of God 's pres ence ,285, 289, 300-1 , 305, 342

objectivity of, 326, 340-3tests for veridical i ty of, 29 8-300, 3 03- 28 ,

342re li g ious expe r ience a rgum en t s , Ch ap te r 8re l ig ious exper ience a rguments , f rom

analogy wi th sense exper iences , 288

4 2 1



Index

religious experience arguments, fromanalogy w ith sense exper ience, cont.)

based on language-gam e analysis , 291-4

based on any old tests, 295 -8, 322based on analogous tests, 298-317, 322

Rescher, Nicholas, 414 n3responsibili ty, 153,155-7, *75Ricoeur, Paul, 326Roger s, Fred, 112Rorty, Richard, 325Ross , James, 203, 242Row e, Wil liam, 248-51 , 258-77Russell, Bertrand, 81, 251-2, 362

Salmon, Wesley, 254

Sartre, Jean Paul, 92,119, 378Schlelermacher, 215Schlesigner, George, 414 n6Schmidt, Richard, 291Scotus , n 6 , 203, 242Seidenfe ld, Tedd y, 416 n23Sellars, Wilfrid, 66, 314sense experience, 286, 288, 294, 298, 326

tests for the veridicality of, 298-311,315-28 ,342

Slote, Michael, 162, 254Smart, Ninian, 286

sortal objects, 248-50, 254space, 327-31, 335, 337, 339-41Spinoza, 31, 280, 320, 368Stace, Walter, 286, 295, 314, 341, 410015,

412 n39Stalnaker, Robert, 401 n6Strawson, P. R, 320-40,413 n48S t u m p , Eleonore, 25-6, 48, 51-5,69-71,

392 ngSuarez , 149-50 ,170-1subjunctive conditionals, 128-9, 137<

154-5, 268

Taylor, Richard, 119, 243theodicy, 98-9,110, 229, 236t ime, 37,4 0-6 , 271-2 , 327-9, 331

as absolute, 38-9, 328-9, 333-41as closed, 24 5-6 , 257as dependent upon change, 38-9, 271-3the disappearance of, 45 -7measurement of, 42-5necessity of, 271-3as mind-d ependen t , 43 -5

Tomberlin, James, 399 n38Tooley, W., 405 m 8Turner, Donald, 166 -8,1 75 -6

Underhill, Evelyn, 286

universalizabilityof ethical claims , 371, 386of evidential claims, 288, 379-8 1

utili tarianism, 30, 32,1 10, 355 -6

van Inwagen , Peter, 181, 235, 401 n6 ,40 3n n 2 8 a n d 3 i , 4 0 4 m

veridicality (sec also: religious experience,tests for the veridicality of; se ns eexperience, tests for the veridicalityof), 287 -8, 295, 297, 30 3-4 , 313, 315,326, 328, 343

vicious infinite regress arguments, 105-6,197-8

Wainwright, William, 139-40,145-6, 289,298, 303-7, 309, 313-14, 316-1 8,320-3, 326, 407 m l

Whitehead, A. N., 40, 323-4Wilde, Oscar, 110will to believe , 357- 71 , 381 -7Winch, Peter, 291W ittgenstein, Ludw ig, 274, 291, 296, 409

7

Documents

_Gale, Richard - On the Nature and Existence of God.pdf