Reason and Faith: Themes from Richard Swinburne [Hardcover ed.] 0198732643, 9780198732648

Citation preview

Reason and Faith

Reason and Faith Themes from Richard Swinburne

E DI TED BY

Michael Bergmann and Jeffrey E. Brower

1

3

Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries # the several contributors 2016 The moral rights of the authors have been asserted First Edition published in 2016 Impression: 1 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this work in any other form and you must impose this same condition on any acquirer Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2015938932 ISBN 978–0–19–873264–8 Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work.

For Richard: A Model of Reason and Faith

Contents Preface Chapter Abstracts List of Contributors

I. Natural Theology Faith 1. The Idea of Faith as Trust: Lessons in Noncognitivist Approaches to Faith Jonathan L. Kvanvig 2. Working with Swinburne: Belief, Value, and the Religious Life J. L. Schellenberg

Theistic Arguments

ix xi xvii 1 3 4 26

47

3. Simplicity and Natural Theology Paul Draper

48

4. Swinburne’s Aesthetic Appeal Hud Hudson

64

Divine Power 5. Defining Omnipotence Dean Zimmerman 6. Law, Cause, and Occasionalism Alvin Plantinga

II. Philosophical Theology Atonement and Liturgy 7. Love and Forgiveness: Swinburne on Atonement Eleonore Stump

83 84 126

145 147 148

viii

CONTENTS

8. The Liturgical Present Tense Nicholas Wolterstorff

Immortality, Body, and Soul 9. The Rev’d Mr Bayes and the Life Everlasting Peter van Inwagen

171

195 196

10. What about Hylomorphism? Some Medieval and Recent Ruminations on Swinburne’s Dualism Marilyn McCord Adams

220

Index

245

Preface The past fifty years have been an incredibly fruitful period for work in the field of philosophy of religion, perhaps the most fruitful period since the Middle Ages. After reaching a low point in the first half of the twentieth century, with logical positivism declaring the death of traditional metaphysics and most, if not all, substantive philosophy of religion along with it, a corner was turned in the 1960s as a number of analytic philosophers began taking up the defense of traditional metaphysics and applying insights from this field, as well as from a number of others (including epistemology, ethics, philosophy of science, and philosophy of mind) to issues in philosophy of religion. Although a number of figures have contributed to the contemporary revival of interest in philosophy of religion, few have done more to advance its development than Richard Swinburne (Emeritus Nolloth Professor of the Philosophy of the Christian Religion, Oxford University). His pioneering work in the field is distinguished, not only for its commitment to articulating and defending a comprehensive set of positions, but also for the way in which it builds on and contributes to contemporary work in other fields, such as metaphysics, epistemology, and philosophy of science. Indeed, his corpus stands out for its remarkable breadth and systematic engagement with issues in both natural theology (dealing with what can be known about God and his relation to the world independently of any particular religious tradition or revelation) and philosophical theology (reflecting critically on the doctrines associated with particular religious traditions). This volume is a collection of new essays in natural and philosophical theology that were first presented at a conference in honor of Swinburne. This conference was co-organized by the editors of this volume and hosted at Purdue University in September 2014, which coincided with the year of Swinburne’s eightieth birthday. It brought together a number of leading figures in contemporary philosophy of religion, many of whom

x

PREFACE

also played a key role in the post-1960 renaissance in the field.1 The resulting essays provide a sampling of the life and vigor this field now enjoys. All ten of the essays collected here touch on themes in Swinburne’s work, and several critically engage specific ideas and methods that his work has helped to make popular. The first six essays address topics familiar from natural theology: the nature of faith (Chapters 1 and 2), theistic arguments (Chapters 3 and 4), and certain issues associated with divine power (Chapters 5 and 6). The next four essays address topics bearing on philosophical theology: atonement and liturgy (Chapters 7 and 8), and immortality, bodies, and souls (Chapters 9 and 10). We hope these essays will be of interest not only to those who work in the field of philosophy of religion, but also to those who work in other fields on which their content bears. Work on this volume was made possible by a generous grant from the John Templeton Foundation, as well as by financial assistance from the Center for Philosophy of Religion at the University of Notre Dame, the Society of Christian Philosophers, and Purdue University. We are grateful to each of these institutions. We are also grateful to the following individuals: the volume contributors, the participants at the conference (including, especially, the commentators), our graduate student assistant, Mark Satta, and—for sage advice at various points along the way—Susan Brower-Toland, Jan Cover, Patrick Kain, Peter Momtchiloff, and Michael Rea. Finally, we all owe a debt of gratitude to Richard Swinburne for his enormously productive and inspiring philosophical career, which provided the motivation for the project that resulted in this volume and the conference that preceded it. It is to him that the volume is dedicated. Michael Bergmann and Jeffrey E. Brower, August 2015

1 Videos of the original conference presentations, as well as replies by commentators and questions-and-answers from the audience, are available online at: .

Chapter Abstracts CHAPTER 1: JONATHAN L. KVANVIG, “THE IDEA OF FAITH AS TRUST: LESSONS IN NONCOGNITIVIST APPROACHES TO FAITH” Various noncognitivist approaches to faith have been developed recently, treating faith, for example, in terms of hope or some other affective state. One such approach identifies faith with trust. This chapter examines such approaches, focusing on the latter, and concluding that the identification of faith with trust is a mistake. At the same time, it articulates and defends a more generic affective approach to faith, claiming that it does a better job than standard cognitive approaches to faith in helping us to understand important examples of faith, both religious and mundane. The faith of Abraham is discussed as an example of religious faith, and the faith of ordinary people in pursuing their ideals is discussed as an example of mundane faith. CHAPTER 2: J. L. SCHELLENBERG, “WORKING WITH SWINBURNE: BELIEF, VALUE, AND THE RELIGIOUS LIFE” This chapter is a tribute to Richard Swinburne’s pioneering work in Faith and Reason. It seeks to learn from his insights that what is most fundamental in the religious life are the purposes to which one becomes committed, and that belief is subservient to these. As will be shown, when its conative and evaluative features are sufficiently developed, and the value of certain religious states of affairs being realized is sufficiently emphasized, religious faith is able to flourish not only without religious beliefs, but also without the non-doxastic cognitive states most commonly discussed in this connection, and moreover without the epistemic beliefs Swinburne himself regards as necessary for it. A central conclusion is that when analyzing and evaluating religious faith, philosophers will gain at least as much enlightenment from the ethics of religion as from the epistemology or metaphysics thereof.

xii

CHAPTER ABSTRACTS

CHAPTER 3: PAUL DRAPER, “SIMPLICITY AND NATURAL THEOLOGY” This chapter examines and critically discusses the role of simplicity in Swinburne’s probabilistic natural theology. After describing that role and the details of his theory of simplicity, the chapter challenges Swinburne’s view that the criterion of simplicity is a fundamental criterion for evaluating causal explanations, proposing instead that what is right about that criterion can be derived from a more fundamental criterion of (probabilistic) “coherence,” which measures the degree of fit between the known entailments of a hypothesis. It is further argues that if this proposal is correct, then Swinburne’s cumulative case for the existence of God is significantly weakened. CHAPTER 4: HUD HUDSON, “SWINBURNE’S AESTHETIC APPEAL” In his book The Existence of God, Richard Swinburne investigates an increasingly popular argument for theism that is grounded in the observation that the universe appears to be especially fine-tuned for embodiedfree-intelligent-sentient life. Swinburne’s discussion is intriguing and also unusual due to the role he assigns to aesthetic considerations. After motivating and reviewing a very plausible thesis known as skeptical theism, this chapter explains how that thesis undermines the most popular defense of a crucial premise in Swinburne’s (and others’) fine-tuning arguments. Then it introduces and defends a second skeptical hypothesis, aesthetic skepticism (which stands to claims about beauty and sublimity as skeptical theism stands to claims about value more generally), and shows how attempts to reformulate fine-tuning arguments with respect to beauty and sublimity are jeopardized by the same sorts of reflections that trouble other versions of this strategy for supporting theism. CHAPTER 5: DEAN ZIMMERMAN, “DEFINING OMNIPOTENCE” Swinburne’s analysis of omnipotence in The Coherence of Theism fails due to counterexamples like Plantinga’s McEar. Recently, Swinburne has suggested a fix in terms of the relative sizes of the sets of states of affairs

CHAPTER ABSTRACTS

xiii

that could be brought about by different candidates for omnipotence. The idea is a good one, though it requires some fleshing out and, in the end, it is not without its problems. An alternative response to McEar is explored, which makes use of the ability to prevent all other possible beings from exercising their powers. Finally, a rather different sort of counterexample is raised, involving future contingencies that depend upon the free actions of others; and a further revision is proposed to deal with it. CHAPTER 6: ALVIN PLANTINGA, “LAW, CAUSE, AND OCCASIONALISM” This chapter argues that there are such things as natural laws, and that they enjoy a sort of necessity. The necessity of natural laws is not a “broadly logical” necessity, but instead a sort of creaturely inviolability: natural laws are such that no creature (as opposed to the Creator) can change or abrogate them. The chapter goes on to consider three ways in which the laws are related to God: secondary causalism, decretalism, and as counterfactuals of (divine) freedom. The second and third suggest occasionalism. Since the very idea of creaturely causation is at least somewhat obscure, the chapter tentatively suggests that we should award the palm to occasionalism. CHAPTER 7: ELEONORE STUMP, “LOVE AND FORGIVENESS: SWINBURNE ON ATONEMENT” One kind of interpretation of the doctrine of the atonement, often associated with Anselm and recently defended by Richard Swinburne, supposes that the life, suffering, and death of Christ are somehow requisite to satisfy God’s justice or goodness or some other divine attribute. For God to forgive human beings for their wrongdoing, Christ’s life and death on the cross are needed. This chapter argues against that view. It presents a traditional account of love and forgiveness, and then examines the implications of this account for Swinburne’s view of atonement. The chapter concludes that Swinburne’s view of atonement is incompatible with this traditional account even when provision is made for retributive punishment.

xiv

CHAPTER ABSTRACTS

CHAPTER 8: NICHOLAS WOLTERSTORFF, “THE LITURGICAL PRESENT TENSE” A striking feature of Christian liturgies and hymns, especially those used on Christmas Day and Easter, is the use of the present tense when speaking of Christ’s birth or resurrection. This chapter explores the significance of “the liturgical present tense.” A theory that enjoyed wide acceptance among twentieth-century liturgical scholars was that these events are reactualized when the liturgy is enacted. The chapter argues that that is ontologically impossible, and that the liturgical present tense should instead be understood as a specific version of a widely used figure of speech—call it “the as-if trope”—in which one talks about someone as if they were what in fact they are not. The chapter concludes with suggestions as to why Christians use the liturgical-present version of this trope on Christmas Day and Easter. CHAPTER 9: PETER VAN INWAGEN, “THE REV’D MR BAYES AND THE LIFE EVERLASTING” The philosopher John Leslie and the cosmologist Don Page have recently considered the question whether an argument parallel to the so-called Doomsday Argument can be used to refute the thesis that human persons are immortal. The Doomsday Argument is (roughly) a Bayesian argument for the conclusion that one ought to assign a much lower probability than most of us unreflectively do to the hypothesis that our species will have a long future. This chapter first reviews the Doomsday Argument. It then considers the question whether “Doomsday-style” reasoning can be used to show either that (a) we should assign a probability of 0 to the hypothesis that we shall have a post-mortem existence of infinite duration, or (b) that we should assign a very low probability to the hypothesis that we shall have a post-mortem existence that is long in comparison with a normal human lifespan. CHAPTER 10: MARILYN MCCORD ADAMS, “WHAT ABOUT HYLOMORPHISM? SOME MEDIEVAL AND RECENT RUMINATIONS ON SWINBURNE’S DUALISM” This chapter uses a comparison and contrast of Swinburne’s dualism and medieval hylomorphisms to expose some strategic theoretical choices that

CHAPTER ABSTRACTS

xv

Swinburne has made. Among other things, he forwards a thin view of substance, a reduction of natural kinds to res cogitans and res extensa, and an isolation of teleological explanation to the plans and choices of Divine and human rational agents. Where Nagel wants to make a theory of everything the goal of human inquiry, Swinburne’s dualism forfeits any attempt to find a natural explanation of mind–body interactions. For Swinburne, what explains everything is Divine providence. The chapter closes with the suggestion that alternative choices—such as those embraced by medieval hylomorphisms—might make it easier to avoid marginalizing the biological, explain the unity of human being, and solve the problem of evil.

List of Contributors M ARILYN M C C ORD A DAMS , Rutgers University and Australian Catholic University P AUL D RAPER , Purdue University H UD H UDSON , Western Washington University J ONATHAN L. K VANVIG , Baylor University A LVIN P LANTINGA , University of Notre Dame and Calvin College J. L. S CHELLENBERG , Mount Saint Vincent University E LEONORE S TUMP , Saint Louis University P ETER

VAN

I NWAGEN , University of Notre Dame

N ICHOLAS W OLTERSTORFF , Yale University D EAN Z IMMERMAN , Rutgers University

I

Natural Theology

Faith

1 The Idea of Faith as Trust Lessons in Noncognitivist Approaches to Faith Jonathan L. Kvanvig

1 Introduction A, perhaps the, standard approach to faith treats it as a type of belief. In recent times, however, various noncognitivist approaches to faith have been developed, treating faith, for example, in terms of hope or some other affective state. One such approach identifies faith with trust: Here [when discussing faith in a person] the crucial feature would seem to be trust, reliance on the person to carry out commitments, obligations, promises, or, more generally, to act in a way favorable to oneself. I have faith in my wife; I can rely on her doing what she says she will do, on her remaining true to her commitments, on her remaining attached to me by a bond of love (Alston 1996: 13).

This quotation identifies a certain kind of faith with trust and thus introduces the possibility of thinking of faith in noncognitive terms. It is this approach to faith that I will focus on here, using it as a platform for talking some about faith and some about philosophical methodology in thinking about faith (among other things). I will point out the controversies involved in current thinking about trust, controversies which should make us wonder what to make of identifying faith and trust. I will then show that, for what I will call faith worth having, identifying it with trust of any kind is mistaken. I will begin by motivating the move away from cognitivist portrayals of faith, and then turn to the trust view and the wide variety of opinions

THE IDEA OF FAITH AS TRUST



about what trust involves. One suggestion, in the face of such variety, is to conclude that the search for the nature of the thing in question may have a false presupposition: that there is such a nature. I consider this possibility and the change in philosophical methodology it counsels, arguing for a focus on what I take to be the most important and central examples of faith when theorizing. Such faith I term ‘affective faith,’ and use it in the sections that follow to assess the prospects for the view that faith involves trust. I conclude that what is fundamental to affective faith is not trust, but rather something much more active and sustained.

2 Making Room for Noncognitivism What kind of thing is faith? We find the claims of the cognitivists that it is a type of belief (Locke 1698), that it is a type of some related nondoxastic cognitive state such as acceptance (Alston 1996; HowardSnyder 2013), or that it is a kind of knowledge (Plantinga 2000). In the last few decades, however, new approaches have been championed, claiming that it is a feeling, that it is an attitude or affective state of a certain sort (Clegg 1979), that it is a special kind of trust (Audi 2011; Schellenberg 2009), that it involves hope (Pojman 1986; Sessions 1994), that it is a preference of a certain sort (Buchak 2012), or that it is a kind of practical commitment or disposition toward certain patterns of action (Kvanvig 2013; Tennant 1943). Are there good reasons to abandon the standard cognitive approach? That is a difficult and complicated question, one which I will not attempt to settle here. But some quotes of disaffection might help us see why one may find attractive an alternative picture of faith: ‘Unbelief,’ like ‘theology’ is a product of the Greek mind. . . . Where the word ‘belief ’ is used to translate the biblical Hebrew and Greek it means not the ‘belief that’ of Plato, but ‘belief in,’ a matter not of cognitive assent but of faith, trust, and obedience. . . . Plato’s ‘theology’ is not in fact an accurate apprehension of traditional religion. It is the self-conscious intellectual’s translation of that religion into terms that he can understand (Bellah 1991: 216).

To those inclined toward cognitive construals of faith, these remarks do not help much. The quote claims that faith is a matter of belief-in, not belief-that, but it also says that belief-in is a matter of faith, among other things. It brings to mind the idea of being on a trip with one foot nailed



JONATHAN L . KVANVIG

to the floor. But if we drop the word ‘faith’ from this list, the idea is that belief-in is more a matter of trust and obedience, and it is belief-in that is playing surrogate here for the idea of faith. In that light, consider another passage: While belief now refers to a state of mind, a disposition to assent to a set of propositions, even within the early Christian intellectual tradition historically it had as much or more to do with love, loyalty, and commitments akin to pledging one’s allegiance to a person as Lord or to a cause or to entering into a covenant such as marriage. The Latin word credo (apparently a compound of cor, cordis ‘heart’ and -do, -dere, ‘to put’ derived from the proto Indo-European root for placing one’s heart upon something, *kred-dhē) means ‘I set my heart’ upon the entity or doctrines in question. Even for scholastics such as Aquinas . . . credo meant to pledge allegiance to, to give one’s self and one’s loyalty. The Latin terms most closely expressing today’s meaning of belief and opinion, opinio (‘opinion, belief, supposition’) and opinor (opinari, to be of the opinion, to believe) played an almost negligible role in Christian thought (McKaughan 2013: 107–8).

The idea I want to pursue that is at least hinted at in these quotes is that when we ask about the nature of faith, we are not asking about necessary and sufficient conditions for it, though of course that is a topic that will arise in the inquiry. We are asking more about its essence (if it has one), about species and genera, that of which it is a particular instance and what makes it be the particular instance it is. When we want to know what it is to be a person or a human being, we will find it frustrating to be told that it is a featherless biped or a thing inclined to don attire. Even if these claims are true, they do not give us what we want. Just so with accounts of faith. We want to know what faith is, not what it co-varies with, even if necessarily. Thus, the standard cognitivist approach tells us that faith is an instance of belief, and then goes on to describe which particular instance it is. So if faith is an instance of a cognitive state of believing, we should expect an account of faith in terms of a mental state of one sort or another. Both quotes signal dissatisfaction with such an approach. Whatever faith is, to these authors, it is not primarily or fundamentally a cognitive mental state. That leaves open the possibility that various mental states, including cognitive ones, are involved in having faith, but the project is one of understanding the nature of faith, and for that we must look first at what is fundamental and primary: what is it that faith is an instance of, and what makes it the particular instance of that phenomenon?

THE IDEA OF FAITH AS TRUST



Attention to the history of Christian doctrine reinforces such discontent with cognitive approaches. Very early in this history, doctrinal disputes arose, and the language of anathema was quickly adopted for those with non-standard views. See, for example, the seemingly hyperbolic language of the Athanasian Creed: “Whosoever will be saved, before all things it is necessary that he hold the catholic faith. Which faith except every one do keep whole and undefiled; without doubt he shall perish everlastingly. . . . This is the catholic faith; which except a man believe truly and firmly, he cannot be saved.” The lesson here is clear. The faith is defined in terms of propositional content, and salvation requires believing truly and firmly the members of the set of propositions in question. Antagonism to such doxastic requirements for faith and salvation results immediately when we consider heroes of faith: Abraham, Moses, Job, and even the Apostles. It is mind-boggling to think of them as satisfying the cognitive standard embraced. Imagine any of these at the beginning of their lives of faith endorsing any of the languages of the Athanasian Creed: [W]e worship one God in Trinity, and Trinity in Unity; Neither confounding the Persons; nor dividing the Essence. . . . our Lord Jesus Christ, the Son of God, is God and Man; God, of the Essence of the Father; begotten before the worlds; and Man, of the Essence of his Mother, born in the world. Perfect God; and perfect Man, of a reasonable soul and human flesh subsisting. Equal to the Father, as touching his Godhead; and inferior to the Father as touching his Manhood.

Whatever status we want to give to such orthodoxy, it is essential to distinguish the life of faith and the central, salvific role faith plays in such a life from the alethic concerns about what is true and what is not that dominate the creeds. In short, no reasonable approach to our topic can fail to distinguish salvific adequacy from alethic adequacy when it comes to purely cognitive issues. Once this distinction is made, it becomes clear that what is central to faith, the kind of faith intimately connected with salvific adequacy, must be somewhere else or something else than something purely cognitive. Thus arises the idea of searching for a less cognitive, or noncognitive, account of faith, one more attuned to what exemplars of faith display, both in religious and other contexts. Once we see things in this light, there is a natural progression from the doxastic picture of faith to the trust view. The first step is to become



JONATHAN L . KVANVIG

dissatisfied with claims such as that Christian faith is to be identified with, e.g., believing that God exists and is good. Whereas the doxastic view identifies faith with some sort of belief-that, the more natural locution at least in Scripture for the kind of belief involved in faith is belief-in, not belief-that. And believing in someone or something seems quite similar to, if not synonymous with, trusting in that person or thing. Our question, then, is how promising an approach to faith this trust view is.

3 The Trust View and the Controversies When we ask what trust itself is, we get a stupefying variety of opinions. Russell Hardin identifies trust with belief: “The declarations ‘I believe you are trustworthy’ and ‘I trust you’ are equivalent” (Hardin 2002: 10). Annette Baier (1986) thinks of trust as involving dependence on the goodwill of another, while Karen Jones (1996) claims that trust is constituted by an attitude of optimism about the goodwill of another. James Coleman (1990) characterizes trust as a kind of action that one places or not, depending on decision-theoretic calculations of the ordinary variety, and Richard Holton (1994) thinks it is crucial that an account of trust include normative, reactive attitudes involving dispositions to feel betrayed and grateful, depending on whether one’s trust is violated. How could such disparate approaches occur if people are actually looking at the same phenomenon? One answer to that question is given by Thomas W. Simpson (2012): The ways that the word is used are simply too various to be regimented into one definition. Sometimes ‘trust’ is naturally understood as referring to a sort of affective attitude (‘I will trust my husband, I will not be jealous’); at other times to a conative one (‘Come what may, I will trust you to the end’); and at yet others to cognitive ones (‘I know you are an honourable woman, so I trust you’). Sometimes it is not a mental state but action which is described as trust (‘The patrol followed the scout, trusting him to spot any ambush’). Similarly, it is used in situations where the motivation to trustworthiness is dramatically varied: love, or mutual gain, or moral considerations may all count as reasons not to betray someone’s trust. These all support the inductive argument against the plausibility of analysing of trust. Counter-examples can be given so easily because there are so many ways the word may permissibly be used, and so it would be foolish to seek a single definition (Simpson 2012: 553–4).

THE IDEA OF FAITH AS TRUST



And what of Simpson’s solution? He recommends abandoning the project of analysis and substituting the methodology of genealogy in its place: An alternative way of thinking takes the value of trust as of first importance. A genealogical approach addresses the concept obliquely, by asking why we might have the concept that we have, given some broad facts about how we live and the projects we pursue. The claim is that if we can give a ‘role description’ for an important concept that looks and feels very similar to a notion that we actually operate, explaining what is needed of a concept in order to do a particular job, then that sheds light on the content of our actual one. Rather than trial by thought-experiment, it instead accommodates the vagueness and conflicting intuitions that surround difficult and abstract notions. No doubt all the advocates of the ‘trust is this’ claims I critiqued above have replies to my counter-examples. But the sorts of replies on offer will tend to consist in argument over the cases, and ultimately a discovery that we use the word ‘trust’ to describe different things on different occasions. . . . What would be genuinely revealing is an explanation of why the term permits this variability. As well as yielding a clearer grasp on the content of the concept, such an approach also makes it entirely perspicuous where its value comes from. Two results for the price of one enquiry is no bad thing in these straitened times (Simpson 2012: 555).

Simpson is well aware that his case for an alternative methodology is not exactly compelling. But I think he is onto something important here. We can engage philosophically in an analysis of anything in the dictionary, but we do not, and the reason we do not, when justifying our practice, is because we presuppose the value or importance or significance of what we focus on in our philosophical explorations of them. It is also true that much of what passes for linguistic analysis is simply bad faith: we are not attempting lexicographical improvement over what the dictionary says, but are instead attending to the phenomenon that our language indicates, however much it happens to cloud our view. We then say things like, “I don’t care what the ordinary concept of knowledge is, I don’t care what ‘knows’ means; I’m interested in knowledge, not ‘knowledge’,”1 in much the same way we have come to expect scientists to disavow any interest in the ordinary meaning of terms like “mass” and “motion” when they attempt to give theories about the physical world. In each such case, the process of inquiry presupposes the importance of the topic 1

Compare here Keith Lehrer (1974).



JONATHAN L . KVANVIG

in question, and one way to get past the remaining vestiges of the disease of ordinary language philosophy is to argue explicitly concerning what is valuable and worth theorizing about prior to offering a theory of it. In the process, we may find ordinary language to have a distorting effect and we may be in a position to do what philosophy has always claimed some promise of doing: to improve our lives by improving our understanding of ourselves and the world we inhabit. But even if Simpson is onto something important, the intrusion of genealogy into the story is unfortunate. Genealogical accounts, by the use of the term, ought to be relying on some history or supposed history, and this is their flaw. Looking again at Simpson’s description, there are two things to note here—one good, one bad. The good is the value-driven heart of the approach, and the bad is the bad metaphysics. The bad part is a direct result of including the historical part of what is claimed for the genealogical approach, hoping perhaps for something like what Bernard Williams is looking for in his discussion of truth, when he claims that the genealogical method is broadly naturalistic in that it “helps to explain a concept or value or institution by showing ways in which it could have come about in a simplified environment containing certain kinds of human interests or capacities” (Williams 2002: 21). That is, the historical aspect of the genealogical account is supposed to point us to the fundamental reality the importance of which motivates our theorizing, and do so in a way (perhaps) that lets us see how to defend our favorite metaphysical picture in the process. But no historical or historically tinged mythology will accomplish either of these tasks, at least when considered purely abstractly. That is the stuff of which the genetic fallacy is made. Perhaps in particular cases, genealogy might be useful to add to our value-driven approach, but the general idea that the value-driven approach is itself part and parcel of genealogy is certainly wrongheaded. And it is wrongheaded in a way that favors value-driven approaches over genealogical ones, since the latter require immense backfilling in order to justify the relevance of the historical dimension they tout.

4 Value-Driven Inquiry and Affective Faith So, to repeat, the methodology I favor here focuses on what is important in a given domain. In this domain, I thus focus on what I take to be an

THE IDEA OF FAITH AS TRUST



important, and perhaps fundamental, kind of faith—‘fundamental,’ in the sense of being common ground between religious and mundane examples of the phenomenon that is central to a flourishing life. This kind of faith emerges when we examine different possible approaches to life.2 We can begin with the idea that there can be both unified approaches to life and approaches that involve disunity or disconnectedness. In the latter category are lives that instance the Humean view of causation: just one damn thing after another, with no attempt on the part of the individual in question to be doing anything beyond coping with whatever comes one’s way. Such patterns of life are difficult to sustain, but it is not uncommon to see lives that display them in significant temporal chunks. In contrast to such patterns of behavior are approaches that pursue unity, that aim at connecting the multiplicities in experience into some sort of plan or purpose or understanding of things, and the boundary of such a search involves plans, purposes, and goals that are all-encompassing. The development of plans and purposes typically arises out of conative or affective aspects of human life,3 such as negative emotional experiences and positive emotional experiences. Among the negative emotional experiences are fear, horror, regret, guilt, worry, sorrow, shame, anger, misery, meaninglessness, and despair; among the positive affective states are joy, compassion, awe, wonder, and experiences of beauty and the sublime.4 Such aspects of human life can provoke an interest in finding meaning or in developing a plan or purpose or goal that reaches beyond merely coping with each particular episode in life as it comes. In some cases, the call of the goal is merely attractive; in other cases, it takes on the guise of the mandatory. In either 2 In the next few paragraphs, I draw on material from Kvanvig (2013), where I first began investigating the kind of faith in question. 3 I will here use the terms ‘conative’ and ‘affective’ interchangeably, thinking of the mental life of a person as divided into two parts. Conation is sometimes defined as one of three parts of mental function, distinguished from both cognition and affection, but I am perplexed by attempts to sort the conative from the affective, so will not abide by the strictures of this definition here. 4 Some might think that the experiences of beauty and the sublime should be classified with the cognitive rather than the affective. Since nothing I say here turns on the issue, I will bypass extended discussion of the issue by simply noting the attraction of views that treat normative and evaluative experiences and judgments as essentially involving a pro-attitude or con-attitude. So by classifying these experiences with the conative and affective, my intention is to signal the way in which these attitudes can function in the story of faith in much the same way as the emotions cited.



JONATHAN L . KVANVIG

case, these features of human life need not prompt the kind of unification that interests me, but the possibility exists, and is often realized, of responding to such by adopting a pattern of life that involves longerterm projects, goals, and plans. Such responses involve decisions, whether in the most self-conscious, deliberative sense, or in the hardly discernible way in which, for example, one chooses to follow one’s standard route from home to work. The point is that the response counts as human behavior, motivated by the affective states and experiences in question. It is on these affective origins of the pursuit of ideals of one sort or another that I want to focus initially, for these unifying responses, I want to suggest, are responses that involve faith of one sort or another. To adopt a longer-term project or goal or plan involves a kind of hope that success may be possible, or at least a decision not to give in to feelings of hopelessness, and a kind of self-trust and reliance on the structure of the universe and the society in which one hopes to flourish, regarding the accomplishment of some plan. None of these underlying attitudes or dispositions needs to be brought to consciousness in deliberate reflection on the experiences that prompt them, nor need they be part of any fully deliberate approach taken to the motivations for them. Ordinary human experience in response to the kinds of motivations I want to focus on range from the spontaneous to the fully deliberative. But what is important about the responses in question is that they involve a setting of a direction for the individual, one which a person may faithfully pursue. Faith, in this sense, is an orientation of a person toward a longer-term goal, an orientation or disposition toward the retaining of the goal or plan or project in the face of difficulties in its realization, one prompted by affections of various sorts and involving complex mental states that are fundamentally affective even if they involve cognitive dimensions as well. Such faith is a disposition to act in service of an ideal, whether that ideal is a political one such as equality and liberty for all people, or a sacred one, such as being part of the Kingdom of God. A plan, purpose, or goal is developed, and the culmination of this process involves a commitment by the individual to such a plan, and in following through on such a commitment the person displays the kind of faith that I am describing. People can be faithful to their commitments, or not, and when faithful, their actions express the disposition to act in service of an ideal. Though there is much to be said on the issue of what is

THE IDEA OF FAITH AS TRUST



fundamental to such a disposition, let me merely register my view without arguing for it here. Central to such a disposition is that it involves an ideal, and central to the notion of an ideal for a person is, not a certain cognitive stance toward it, but rather a felt attraction regarding it. Thus, even if the disposition in question involves both affective and cognitive features, I will here talk of the disposition as being fundamentally affective rather than cognitive. Moreover, even if the disposition in question involves cognitive features, the breadth of possibilities here is wide indeed. Dispositions are well known for being realized in a wide variety of underlying states. Being fragile is compatible with being made of glass but also being made of steel, for example. And when we speak of dispositions toward action, the same point holds. One can be disposed to flee danger, and express this disposition in a given instance, both in the presence of beliefs regarding the range of situations that are scary and in the absence of such beliefs. One can pursue something worth achieving, and be disposed to do so, whether or not one believes that success is likely. In the same way, one can pursue a relationship with a person, and be disposed to do so, even when one suspends judgment about whether that person actually exists. In each such case, behavior and the underlying dispositions that lead to the behavior depend on underlying combinations of cognitive, affective, and conative states, and any given cognitive state can be replaced by some other cognitive state if one varies the other states sufficiently. It is for this reason that this approach to faith is hostile to cognitive views of faith, since a cognitive view of faith must insist that the cognitive element can be pinned down precisely. To the contrary, the kind of faith worth having is dispositional and the behavior that expresses the disposition, as well as the disposition itself, admits of a wide variety of possibilities concerning its constitutional basis, not only in terms of cognition but also conation and affection. A mundane example will allow better focus on such affective faith. Suppose a young Little League pitcher gives up a game-winning home run and experiences the typical despondency for having done so. One reaction is to adopt a goal of becoming a better pitcher and never having to feel that way again. Such a reaction can generate an orientation or disposition toward various efforts at becoming better, in hopes of doing so (or at least some aversion to the idea that efforts of any sort are hopeless), and display a kind of self-trust or self-reliance and perhaps



JONATHAN L . KVANVIG

some trust of others who may be recruited to help in the project. Our youngster makes a commitment to a certain kind of future. It might be an intense commitment or more casual in its firmness, but when he carries through on this commitment, he will be properly characterized as being faithful to it, or pursuing his goal faithfully. In a short motto, we grasp the noun form of what (this kind of) faith is from the adjectival and adverbial form involving being full of faith: faith is what faithful people have and faith is whatever it is that underlies and explains patterns of behavior and activity that are engaged in faithfully. Such an identification of affective faith with faithfulness is also well suited to explaining important and central examples of religious faith, such as the faith displayed by Abraham in leaving Mesopotamia for Canaan.5 The narrative in Genesis 12 is sparse—we are told only that the Lord said, “Go,” and, “So Abraham went.” Later accounts of the matter attribute his behavior to faith: “By faith Abraham, when he was called to go out into a place which he should after receive for an inheritance, obeyed; and he went out, not knowing whither he went” (Heb. 11:8). The explanation here is quite natural: Abraham is faithful to the divine command to leave for Canaan; his doing so is “by faith.” It is also a fundamentally noncognitive feature of Abraham that undergirds his behavior. We do not know the details, but some possibilities come to mind easily: perhaps he experiences a felt obligation to a divine command, or perhaps a deep attraction to a life guided by a divine being and being the recipient of divine favor and blessings; perhaps these were mixed with fear of divine displeasure and with disobedience. What we can be confident of, however, is that the source of his behavior and the attitude that underlies it are just as fundamentally affective as they are in the case of the Little League pitcher. In both cases, affective faith involves faithfulness to an ideal. What we should resist are interpretations of Abraham’s faith that are overtly cognitive. As Abraham’s life with God progresses, he receives specific promises about what his future will be like, and (we are told)

There is some uncertainty about the details here as to when the first call came to Abraham. Genesis 12 has the initial response involving leaving Harran, not Ur, whereas Stephen, in Acts 7, has the call coming to Abraham to leave Mesopotamia before he was in Charran (Harran). These uncertainties are irrelevant to our discussion, so I will ignore them here. 5

THE IDEA OF FAITH AS TRUST



believes God regarding these promises. But his initial acts of faith are not described in terms of belief but instead in terms of following the command of God. It is the following that is essential here, not the believing, since we can easily make sense of his behavior as a display of faith even without any belief in the promise of God (for example, it might have been only a deep hope). How things develop is one thing, where things start is quite another. One might resist this conclusion that the initial faith of Abraham is typical of saving faith, maintaining that it would not have been adequate faith without the more full expression of Abraham’s faith involved in the Akedah. The idea here is that the real substance which is faith is expressed in the binding of Isaac and the willingness to sacrifice him, and prior events come to be classified correctly as examples of faith only in light of this later event. Moreover, if seen in this way, we might then prefer more cognitive accounts of faith if more cognitive accounts fit better with this crowning achievement of faith on Abraham’s part. On this matter, Eleonore Stump writes: And so this is indeed a test of Abraham, as the narrative says. God’s command to sacrifice Isaac tries the measure of Abraham’s commitment to the goodness of God. The way in which Abraham dealt with Ishmael makes the form of this test the right one for him, too. For Abraham to treat Isaac in the same way as he treated Ishmael is for Abraham to commit himself whole-heartedly to the belief that God is good. . . . [I]t is precisely Abraham’s willingness to believe in God’s goodness, even against strong temptations to the contrary, that makes him the father of faith. When Abraham passes the test, he passes it just because he believes that God is good and will not betray his promises. . . . (Stump 2010: 298–300).

There is much to say here about how different a perspective I have on the life and times of Abraham, but the following will have to suffice. First, even if the story of offering to sacrifice Isaac is the primary jewel in the crown that is Abraham’s faith, it is not the substance of it. Whatever is central to Abraham’s faith was already present in his leaving Mesopotamia at the command of God, and so if faith requires believing that God is good and keeps his promises, we would have to find some sign of that in the early relationship between God and Abraham. As pointed out already, there is none, and any attempt to treat the initial ascriptions of faith as honorific— correct only insofar as they are a poor shadow of the real thing yet to come—takes liberties with the text that are simply unjustified.



JONATHAN L . KVANVIG

Moreover, any account of the Akedah passage that leaves certain options open cannot be complete. One such option left open by Stump’s account is that nothing in her account explains why God’s command to sacrifice Isaac is essential to the situation. As far as I can tell, her account leaves open Abraham’s deciding to do so unprompted, perhaps vowing to sacrifice Isaac on the model of Jephthah and then have the same degree of belief in God’s goodness as she attributes to Abraham in the actual story. Stump makes much of the contrast between Jephthah and Abraham, arguing that the usual reasons for taking Abraham to be the father of faith (having to do with his willingness to kill his only son) are reasons for giving the title to Jephthah instead, since (according to one traditional interpretation) he carried through in his vow to sacrifice his child. There are several reasons why this argument is less than compelling. For one thing, the father of faith needs to be earlier in the chronology rather than later, and Abraham wins on that score. Second, it is crucial to the story that this episode is instigated by God, and thus that the moral of the story (from the authorial perspective) is something about God himself and what he is like. As I see it, the story is a central component in the development of the ethical monotheism distinctive of the Abrahamic traditions, and culminating in the explicit denunciation of child sacrifice in the prophetic literature. Such an emphasis on authorial intent does not help us understand the experience of Abraham in responding to God’s command, but when the question is one of who deserves the title “father of faith,” it is the perspective of the author or final redactors that are central, not the experience of the characters in the stories themselves. I should also point out what I regard as excessive optimism in Stump’s account of the sacrifice event. Stump notes the duplicity and doublemindedness that has characterized Abraham’s relationship to God and his promises prior to the command to sacrifice Isaac, and she emphasizes as well his continued double-mindedness toward God’s promise in the taking of Keturah as his wife (or concubine) after Sarah’s death and fathering six children with her (as insurance of progeny should Isaac and Rebecca continue to be childless, which they have already experienced for twenty years). But somehow, Abraham manages, in the Akedah, to achieve purity of motive and full clarity of belief regarding what God is like and how sure his promises are. I find such an explanation overly optimistic and going beyond what we have textual evidence for, and if we pay attention to the point made earlier that a primary moral of the story

THE IDEA OF FAITH AS TRUST



is about God’s self-revelation of his character to the father of the nation he has chosen, we can see why there is no need to posit such an astonishing change of character in Abraham. What has characterized Abraham from the beginning in his relationship with God is constant here: when God says go, Abraham goes; when God says do, Abraham does. It is the following, the obedience, that is the heart of the faith of Abraham, not some imputed story about motives and beliefs that are somehow both ideal and out of character in the most difficult following of all. It is, in a word, a rendering of Abraham’s faith in terms of his faithfulness that is crucial to the story, which is the fundamental point made about the affective faith characterized here. One might worry that this account places too great a weight on the connection between having faith and being faithful. Robert Audi argues against a close connection between the two, claiming that one can have faith without displaying faithfulness: It should also be stressed that even if a person’s attitudinal faith (faith in) is highly comprehensive, as is faith in democratic institutions, the person might not have the character trait of faithfulness. This may not be widely realized because we so often speak of faith in the context of considering faith in God, and that kind of attitudinal faith is commonly presupposed to imply some degree of commitment to being faithful to what God, at least in the eyes of the person in question, commands or requires. Nonetheless, someone could have faith in God or great faith in other people, yet not be able to keep faith with them and thus lack faithfulness toward them. Having faith in others is mainly a matter of how we view them, and how we are disposed to respond to them, in terms of what we value. Being faithful to others is more a matter of how we act toward them in terms of what they value (though it is not wholly a matter of what they value, since people can be clearly mistaken about their own good). One way to see this difference is by considering the relation of each kind of faith to trust. If I have faith in you, then I trust you, at least within a certain domain of conduct. If I am faithful to you, you may properly trust me; and if I am faithful to an ideal, I may, in the main, be trusted to live up to it. Neither case of trust implies the other. We can have faith in people, such as foreign heads of state, with whom we have either no relationship or one that does not call on us to keep faith. One way to put the contrast is this: faith is fulfilled when its object meets certain expectations—roughly, fulfills the trust—of the subject; faithfulness is fulfilled when the subject—the faithful person—meets certain expectations, or certain hypothetical expectations, of the object (Audi 2011: 296–7).

Audi makes two central points against the idea that faith is always tied to faithfulness. One idea relies on thinking of faith as connected with trust,



JONATHAN L . KVANVIG

relying on the idea that if one has faith in something, one trusts. In the next section, I will argue that there is no such connection, so we can focus here on the other point, which is that faith is an attitude of a certain kind while faithfulness is a disposition of character. As Audi says, “Having faith in others is mainly a matter of how we view them, and how we are disposed to respond to them, in terms of what we value. Being faithful to others is more a matter of how we act toward them in terms of what they value.” Consider first the claim that having faith (in others) is “mainly a matter of how we view them.” While this may be true of special kinds of faith, it is not true of faith in general. For one thing, it is primarily cognitive, and hence unable to explain the kind of affective faith displayed by our Little League pitcher and Abraham. Abraham has faith, according to the author of Hebrews. In what does this faith consist? Is it mainly a matter of how he views God? There is no textual evidence for such a claim, and it has a bizarre ring to it. Perhaps if we thought Abraham’s faith was a kind of trust, and agreed with cognitive positions on trust, such as Hardin’s view that trusting someone just is believing that they are trustworthy, one might come to the conclusion that Abraham’s faith was a matter of viewing God as trustworthy. The central problem is that such an account of Abraham’s faith is—how shall I put it?— unfaithful to the text. Abraham hears God’s command and leaves. He does so by faith. So what is involved in the faith in question? The only textual evidence of what such faith involves is that it provides the explanation of his behavior. It involves a displayed disposition to follow through and the disposition has an affective source, of the sort that provides impetus to the agency of the individual in question. So whatever we might wish to say about some other possible kinds of faith, the distinction Audi is drawing between faith and faithfulness does not apply, or at least does not obviously apply, to the affective faith involved in our examples of Abraham and the Little League pitcher. Even if we resist Audi’s explanation of the distinction between faith and faithfulness, the point he makes remains. Faith, as a disposition, can occur unrealized in faithful behavior. That point, though, is not at issue here, since the view of faith being developed and endorsed here relies on what we might call the “Hebrews inference,” the inference from faithful behavior to underlying faith. That inference is all that is needed in order to maintain the account of Abraham’s faith noted previously.

THE IDEA OF FAITH AS TRUST



What of the second point, however, the one relying on the link between faith and trust? I turn to this issue in the next section.

5 The Inadequacy of Trust-based Accounts of Affective Faith Consider, first, the insight of Simpson, arising out of Baier’s important early work on trust: whatever trust is, it involves, or is built on top of, reliance. Simpson claims that reliance is Ur-trust, and that a genealogical investigation of trust shows how various human interests will build various uses of the term ‘trust’ on top of Ur-trust. Baier holds a similar view, claiming that we distinguish between reliance and trust, but that the latter is a special case of the former. Examples of reliance without trust are easy to find. Spies often find themselves in the position of having to rely on informants, even though they do not trust them. The simple reason such examples can be found is that, while trust can be betrayed, reliance cannot: when we rely and things go badly, feelings of disappointment and regret are normal and appropriate, but feelings of betrayal are not. It is for this reason that no account of faith can rest comfortably with only the remarks about trust contained in the Alston quote from the introduction, where trust is glossed in terms of “reliance on a person . . . to act in a way favorable to oneself.” Bernie Madoff relied on his victims to act in ways that would be favorable to him—that is what con-artists are good at. But they do not trust their victims, and feelings of betrayal, while appropriate in the case of violations of trust, are not appropriate when scams are uncovered and victims do not play along. This same difficulty arises in another prominent account of trust deployed in service of understanding faith. Richard Swinburne claims that [t]o trust someone is to act on the assumption that she will do for you what she knows that you want or need, when the evidence gives some reason for supposing that she may not and where there will be bad consequences if the assumption is false, . . . [where to act on an assumption is] to do those actions which you would do if you believed the stated assumption strongly (Swinburne 2005: 144).

Once again, shysters and con-artists do not trust their victims, but they rely on them for exploitative purposes. In Swinburne’s terminology, they



JONATHAN L . KVANVIG

act on the assumption that their victims will do for them what they know is wanted or needed. Even though their accounts do not fully characterize the distinction between trust and reliance, both Swinburne and Alston agree that trust is a form of reliance, and with this assumption there is nearly universal agreement. Of course, the plausibility of such approaches requires adopting a suitable account of reliance. Accounts of reliance can go wrong by demanding too much in the way of vulnerability: reliance does not require likelihood of harm, nor does it require some objective chance of harm. One can be unknowingly immune to the flu, for example, and still rely on one’s flu shot to protect one from the disease. Neither does reliance require the possibility of harm, for one can rely on God’s goodwill even if it is impossible for God not to have goodwill toward one. Moreover, accounts of reliance can err in another direction, by taking a significant chance of harm to be sufficient for reliance. I do not rely on the venomous snake in the grass (not to bite me) of which I am unaware just by walking across the yard. So reliance-based accounts of trust need to be careful not to undermine this approach to trust simply by adopting a ham-fisted account of reliance. I think this issue can be finessed easily enough, but will not take the time here to do so, instead urging that we treat it as a platitude that trust involves reliance when the latter is properly finessed. If we think of trust as a species of reliance, I think we know enough about trust to understand why such accounts of faith are bound to be inadequate. At least when it comes to the affective faith I am focusing on here, trust-based accounts are thoroughly wrongheaded. What is distinctive about affective faith is that it is active rather than passive. Abraham is faithful to the command of God to leave, the Little Leaguer is faithful to his resolution to never let that happen again. And, time and again in Scripture, the model of commitment to the Kingdom of God is described in terms of being a follower. These examples of faith are not properly characterized in terms of passivity. And yet pure passivity, of the sort displayed by the successful achievement of Stoic apatheia, can be an expression of trust. Such an attitude toward the universe as a whole can display one’s trust in the created order and in whomever or whatever is responsible for that order. But it is simply too passive to characterize the kind of affective faith involved in the examples of our Little Leaguer, Abraham, or followers in general. So the first point to note about the view that faith involves trust is that the view will have to be modified to

THE IDEA OF FAITH AS TRUST



accommodate the fact that some instances of trust do not help us understand certain types of faith. In short, we cannot identify faith with trust; at most, we can only take faith to be a special kind of trust. Even that much cannot be sustained, however. Consider people who have trust issues, people who fail to develop basic trust in early childhood. Such people will have difficulty trusting, and some will never be able to trust much at all. They will, nonetheless, have to rely on others and the world around them, and they will typically come to realize that even though trust is optional, reliance is not. Could Abraham have been such a person, consistent with the remarks made about him? I do not see why not. He would still have been relying on God in the process of leaving, and his leaving would have been by faith. But it would have been questionable whether his leaving would have involved trust. It is also noteworthy, as documented in Stump (2012), the degree of doublemindedness in Abraham’s following of God. It is not really clear to what degree he believed the divine promises, as opposing to holding a cherished hope that the path he was following would get him what he wanted—children or heirs if progeny were not on the cards. He brings Lot, a member of his family, with him as a potential heir, even though told to leave his family; he makes Eliezer his heir even after being given the promise of his own offspring; and even after Sarah’s death, he fathers more children with Keturah, apparently as an insurance policy against the continuing barrenness of Isaac and Rebecca. Perhaps the preponderance of the evidence here suggests that our question about Abraham’s level of trust is that it is not a hypothetical possibility that he did not really trust much, but rather the actual situation as described in the story. Such a failure of trust is fully compatible with an exemplary faith. What makes Abraham the father of faith, on this understanding, is the stunningly broad and deep array of circumstances in which his disposition to act in service of an ideal is realized. A disposition can be present and yet fragile, realized only in the best of circumstances.6 The divine demands on Abraham are extensive and significant. His faith is deep and secure, expressed even under the most trying circumstances, and the exemplary character of that faith is not threatened at all by the supposition that it involves a significant degree of failure of trust. 6

Think here of the situationist challenges to virtue ethics, especially the empirical data suggesting strong fragility of virtues we might have expected to see in ordinary humans.



JONATHAN L . KVANVIG

We can see the distance between faith and trust more easily in the case of mundane faith. Consider those who followed Caesar across the Rubicon. What did they see in him that led to them crossing the Rubicon with him? Never mind those who followed blindly, or as mercenaries, or as conscripts. No emperor survives without a loyal band of devoted aids. And what did they see in him? What they “saw” was not something they cognitively perceived at all: it was their affections at work, an admiration, or hope, or wish, or affection for the man or the ideal he represents, that motivates. And what does their commitment to, their embracing of, an ideal, which in this case is a person and the prospects and hopes he embodies, involve? Their commitment is expressed in the crossing of the Rubicon. So the question to ask of such followers is not what cognitive states they were in at this crucial moment of, perhaps, January 10 in 49 BCE if we want to know whether his followers had faith in him; it is, rather, whether they crossed the Rubicon with him. And that they could have done, with the same affective features that led to their initial commitment to the vision for Rome embodied in this Julian band while finding trust difficult and any specific beliefs beside the point.

6 Metaphysics and Epistemology Is this line of thought one gigantic failure that confuses metaphysics and epistemology? The confusion charge can be generated as follows: you talk of knowing who has faith and who does not, and you answer that question by looking to see whether the person in question is faithful. But that is just an epistemological matter of trying to discover who has faith and who does not, not the metaphysical issue of what faith itself is. So you can have your inference to faith from faithfulness if you like, but we should still retain the cognitive picture of faith we know and love. To see why this line of concern is misplaced, let us begin with a more telling Christian example. Consider the Johannine perspective here on the attraction of Christ: “and we beheld his glory, the glory as of the only begotten of the Father, full of grace and truth.” It is clear that what is fundamental to the Johannine perspective here is not cognitive at all. It is seated in the affections, involving a complex aesthetic and emotional experience of a sort that resonates with many of us. To try to paint it as fundamentally a cognitive experience involving belief that God exists and is good is simply too cognitive for understanding the experience of

THE IDEA OF FAITH AS TRUST



beheld glory. As I prefer to characterize this, the experience and the attachment to an ideal which results, are fundamentally affective. So the source of Johannine faith is clearly affective. But the source is not the faith, for if John had beheld the glory in question, and then decided to return to his ordinary life, he would have been in the same faithless category as the rich young ruler, who with sadness went back to his wealth when hearing the call to a different life. For faith to be present is for something else to be found here. And notice, what more is needed is not a better understanding of God and his goodness, but instead the orientation of a person in service of one ideal rather than some other. What is needed is a disposition to follow and for that disposition to be displayed in action. The disposition alone, undisplayed, perhaps may be counted as the weakest of faith, and the general defeasibility of dispositions may require us to categorize some persons as people of faith in spite of lack of faithfulness. But such are the extremes, at best, leaving open the barest possibility of faith without faithfulness, where the disposition to the relevant behavior is present but always and everywhere masked or finked. But when we find faithfulness to an ideal, displaying behavior that is an expression of a disposition whose source and identity are found in the affective origins of the attraction of the ideal, we can not only infer the presence of faith, but have located its nature as well.

7 Conclusion We may sum up our discussion this way: some kinds of faith might involve trust, just as some kinds might involve belief or some closely related cognitive feature. In fact, it might be typical or normal or expected that even the affective faith I am considering involves trust as well as belief or some closely related cognitive feature. But its core lies elsewhere, and the central explanatory role that affective faith can play both in understanding various religious and secular examples of faith, not to mention the obvious way in which it allows for an explanation of how and why faith is a virtue, should lead us to conclude that at least one central kind of faith should not be understood in terms of trust.7 7 This publication was made possible through the support of a grant from Templeton Religion Trust. The opinions expressed in this publication are those of the author and do not necessarily reflect the views of Templeton Religion Trust.



JONATHAN L . KVANVIG

References Alston, William P. (1996). “Belief, Acceptance, and Religious Faith,” in Jeff Jordan and Daniel Howard-Snyder (eds), Faith, Freedom, and Rationality. Lanham, MD: Rowman and Littlefield, 3–27. Audi, Robert. (2011). “Faith, Faithfulness, and Virtue,” Faith and Philosophy, 28 (3): 294–309. Baier, Annette C. (1986). “Trust and Antitrust,” Ethics, 96 (2): 231–60. Bellah, Robert N. (1991). “Religion and Belief: The Historical Background of ‘Non-Belief ’,” in Beyond Belief: Essays on Religion in a Post-Traditionalist World. Berkeley: University of California Press, 216–30. Buchak, Lara. (2012). “Can It Be Rational to Have Faith?,” in Jake Chandler and Victoria S. Harrison (eds), Probability in the Philosophy of Religion. Oxford: Oxford University Press, 225–46. Clegg, Jerry S. (1979). “Faith,” American Philosophical Quarterly, 16: 225–32. Coleman, James. (1990). Foundations of Social Theory. New York: Cambridge University Press. Hardin, Russell. (2002). Trust and Trustworthiness. New York: Russell Sage Foundation. Holton, Richard. (1994). “Deciding to Trust, Coming to Believe,” Australasian Journal of Philosophy, 72: 63–76. Howard-Snyder, Daniel. (2013). “Propositional Faith: What It Is and What It Is Not,” American Philosophical Quarterly, 50 (4): 357–72. Jones, Karen. (1996). “Trust as an Affective Attitude,” Ethics, 107 (1): 4–25. Kvanvig, Jonathan L. (2013). “Affective Theism and People of Faith,” Midwest Studies in Philosophy, 37: 109–28. Lehrer, Keith. (1974). Knowledge. New York: Oxford University Press. Locke, John. (1698). Essay Concerning Human Understanding. London: William Tegg. McKaughan, Daniel J. (2013). “Authentic Faith and Acknowledged Risk: Dissolving the Problem of Faith and Reason,” Religious Studies, 49: 101–24. Plantinga, Alvin. (2000). Warranted Christian Belief. Oxford: Oxford University Press. Pojman, Louis. (1986). “Faith Without Belief?,” Faith and Philosophy, 3 (2): 157–76. Schellenberg, John L. (2009). The Will to Imagine: A Justification of Skeptical Religion. Ithaca, NY: Cornell University Press. Sessions, William Lad. (1994). The Concept of Faith: A Philosophical Investigation. Ithaca, NY: Cornell University Press. Simpson, Thomas W. (2012). “What Is Trust?,” Pacific Philosophical Quarterly, 93 (4): 550–69.

THE IDEA OF FAITH AS TRUST



Stump, Eleonore. (2010). Wandering in Darkness: Narrative and the Problem of Suffering. Oxford: Oxford University Press. Stump, Eleonore. (2012). “Divine Simplicity,” in Brian Davies and Eleonore Stump (eds), Oxford Handbook to Thomas Aquinas. Oxford: Oxford University Press, 135–46. Swinburne, Richard. (2005). Faith and Reason, 2nd edn. Oxford: Oxford University Press. Tennant, Frederick R. (1943). The Nature of Belief. London: Centenary Press. Williams, Bernard. (2002). Truth and Truthfulness: An Essay in Genealogy. Princeton, NJ: Princeton University Press.

2 Working with Swinburne Belief, Value, and the Religious Life J. L. Schellenberg

1 Richard Swinburne supervised my D.Phil. at Oxford back in the late 1980s. His supervision was exemplary in every way, and I am deeply grateful to him for his guidance and support, which were amply forthcoming despite his disagreement with my atheistic conclusion. In this chapter I express my gratitude, and also pay tribute to Richard and to his penetrating work in philosophy by building on his implicit response to a problem I myself would ordinarily treat differently: the problem of how religion, so often infused with barren certainties or fertile hatred today, could be realized more rationally and less harmfully in the future. One way to deal with this problem is by developing new forms of religious faith. This is the direction my own work has recently taken. Another way to do it is by seeking to give to traditional religion an interpretation that the future can more easily bear. This is an enterprise to which Swinburne can be seen as having contributed in his important book Faith and Reason (2005). And it is one that I propose to take further here. A feature shared by Swinburne’s work and its proposed extension in this chapter is a certain attenuation of the emphasis ordinarily found among faith’s defenders on propositional beliefs and their rational status. Swinburne has argued, it is true, that the existence of God, as traditionally conceived, is more probable than not, and that the same holds for central Christian doctrines. This links him very firmly in many minds with the rational defense of traditional religious belief. It would therefore

WORKING WITH SWINBURNE



be natural for those who know of this to suppose, before considering what he has to say on the subject, that propositional belief would have a very significant role in his account of faith. But Swinburne’s account of faith is very general—he is concerned with the most fundamental notion in this neighborhood: that of a religious life.1 And when it comes to the shape of the religious life, including the Christian life (which it has been his central aim to defend), we see him arguing that what is most important are the purposes one finds valuable and to which one becomes or remains committed, and that belief is subservient to these. Moreover, to serve Christian purposes, belief need not be straight-out belief that some Christian proposition is true. What is needed, and what the Christian should seek reason to support, is rather the belief that no religious creed distinct from the Christian creed specifies a path more likely to reach goals as valuable as those of Christianity. At this point a light may seem to go on for those who remember that Swinburne’s approach to the defense of God’s existence and various Christian propositions is probabilistic and Bayesian: in Faith and Reason he is simply adapting Bayes for the purpose of talking about action, and applying Bayesian decision theory! So if belief is attenuated in anything he says, this is only because of the nature of such an approach, which in the right circumstances will allow for rational action even when one’s degree of belief that one’s goal will be achieved by a certain means is low. But Swinburne himself denies that he is employing Bayesian decision theory. Though some of his own central concepts—goals, values, probability—are formalized in it, he explicitly says, in Faith and Reason, that “it will not be helpful to use Bayesian decision theory in this book” (2005: 82, n. 24). That is because, in his view, the formal apparatus of such an approach is not easily or naturally applied in the religious case, and more reliable judgments of rationality than it would deliver are, in this domain, independently attainable. Thus when it comes to the nature and rationality of faith, there is no need to stay with Bayes to build on Swinburne. Indeed, I will be arguing that Swinburne’s own emphasis on “value first” should lead us further beyond Bayes and the associated

1 The sort of faith we have here seems much the same as what Robert Audi distinguishes as “global faith,” which is, he says, what we have in mind when we speak of someone as “a person of faith.” See Audi (2011: 53, 57–8).



J . L . SCHELLENBERG

concepts of probability and propositional belief than Swinburne himself is prepared to go. In making this argument—so as to work with Swinburne in this piece— I will assume that all of the other arguments I (and others) have ever made that, to my mind, tell in favor of the falsehood of Christian beliefs are actually unsound, and properly seen as such by those who investigate. Against that background, I will argue that Christian faith—a species of the fundamental religious faith Swinburne has in mind, which comes down to a religious form of life or religiousness—can authentically be realized entirely without the metaphysical beliefs usually thought to be required for it, and also without the epistemic beliefs Swinburne requires, so long as certain important evaluative and conative conditions are satisfied. I will also try to get clear about the shape it might have when this is the case, noting the consequences of my results for a current movement in philosophy of religion emphasizing the possibility and importance of non-doxastic propositional faith. I will conclude by delivering a minisermon on the besetting sins of philosophers—including both Swinburne and myself—that can prevent us from seeing how a flourishing faith without the cognitive states of which we are so fond is possible. I should underline that my focus here will be largely on the nature of religious faith. And so when, for example, I speak of what a Christian “can do” in the absence of certain beliefs, I will generally have in mind what can be done while authentically exhibiting Christian faith, not what such a person can do rationally. But having got clearer about this, we will also be in a position to point out, near the end of the chapter, some important consequences for the rational evaluation of faith.

2 Let us begin by seeing to what extent and in what way Swinburne himself thinks a life of faith can manage without propositional beliefs. The one who is on a religious way and so exhibits religious faith, he says, is pursuing the goals of religion, which centrally include salvation or deep well-being both for oneself and for others (2005: 160).2 Belief ’s role, 2 Swinburne does not always make as clear as he might that to have faith one need not actually be acting in one way or another; being disposed to act in the relevant way is enough. But that this is his view is confirmed here and there—see, for example, Swinburne (2005: 141).

WORKING WITH SWINBURNE



according to Swinburne, is to function as a guide of action (2005: 9–15), and this in religious contexts as much as in any others. He recognizes that there are, in the world today, various creeds specifying quite different means to the end of salvation. Which one, if any, should a person with that end in view follow? Without some belief about the relative intellectual status of his options, so he thinks, one could hardly be expected to act at all. In having Christian faith as opposed to Buddhist faith or Hindu faith or some other sort of faith, one has to form a belief about the Christian creed that permits one to follow the Christian path to salvation rather than one of the others. What belief, exactly, must one form in such cases? Assuming (for the moment, and for simplicity) that salvation has the same value however construed by the world’s religious traditions, the one who has Christian faith must believe that what the Christian creed taken as a whole directs one to do—the total set of actions entailed by the conjunction of all its items—is at least as likely to lead to salvation as what any alternative creed commends (2005: 197, 225). One’s creed must also, Swinburne (2005: 197) says, be regarded as more likely to do so than at least one other option (if this were not the case, then its selection could not be favored over doing nothing, which he regards as a “way” of proceeding here)—but for simplicity’s sake I shall ignore this second requirement hereafter. Swinburne takes all this to imply that when having Christian faith, the Christian believes that the Christian creed taken as a whole (the conjunction of all its items) is at least as likely to be true as any other religious creed taken as a whole (2005: 225). Do any of these beliefs, on his view, add up to believing straight out that the Christian creed is true? Swinburne regards belief as a contrastive notion (2005: 5) with the relevant contrast expressible, at least by mature believers, in terms of probability: in believing that p, one believes p to be more probable than any alternative to p (2005: 7). Normally the alternative with which p is contrasted is its negation, not-p, but sometimes— and here the religious case is of course illustrative—the contrast will be with a range of alternatives, and Swinburne argues that in the latter situation one should be said to believe that p even if one does not believe it to be more probable than its negation (and so does not regard it as more probable than the disjunction of those alternatives): in that sort of situation one may be said to believe that p if one believes that p is more



J . L . SCHELLENBERG

probable than each of the alternatives (2005: 6–7).3 So, to begin to answer our question: one may, on Swinburne’s view, be said to believe straight out that the Christian creed is true if one believes it to be more probable than not, but also if one believes only that it is more probable than each of a range of religious alternatives. This generates an important interpretive conclusion. We have seen that the minimum required for faith in the way of belief, on Swinburne’s view, is less than this. On his view, one may have Christian faith even if one does not believe either that the Christian creed is more probable than not or that it is more probable than each alternative creed, for one might believe that it is at least as likely as any alternative—and do this while not believing that it is more likely than each of them (2005: 197–8). It follows that, on Swinburne’s view, one may have Christian faith without believing that the Christian creed is true. Indeed, at one point he suggests that one may do so while, strictly speaking, believing that the Christian creed is false—for one might believe it to be at least as likely as any alternative, while believing it less likely than the disjunction of alternatives (2005: 227). Here we see Swinburne’s open door to the idea that faith—and even religious faith—can in the relevant sense be non-doxastic. This point is somewhat obscured by his inclination to call the belief had by a person of Christian faith, even in the case where she has the minimum required for Christian faith: weak belief in the Christian creed (2005: 225, cf. 197). But it follows from everything else he has said that this label is not, strictly speaking, appropriate. One who believes the minimum required for faith might not in any Swinburnian sense believe that the Christian creed is true. Nonetheless, she has a belief. She must have a belief, on Swinburne’s view, an epistemic comparative belief, if she is to act in pursuit of her goal. We need now to consider whether that is in fact the case, or whether we and Swinburne should move even further beyond belief about religious and related propositions, when thinking about faith, than he does in Faith and Reason. To smooth the way and also confirm what we have heard from him about the central importance of the conative and evaluative dimensions of a religious life, let me conclude this first section of the chapter by offering three broad comments. First, as I have argued

3

See also Swinburne (2001: 36).

WORKING WITH SWINBURNE



elsewhere, one cannot understand religion without taking account of its soteriological features. Philosophers who treat claims such as the claim that there is a God as contributions purely to metaphysics do not rightly see themselves as dealing with a religious claim. And religion’s soteriological features are of course tightly linked to the goals and evaluations of religious people. Second, as any up-to-date (or even quite out-of-date) religious studies textbook will show, work in that field richly supports the claim that religion at its heart features practices of various kinds. This has been so, as far as we can tell, ever since religion originated in human evolutionary history many tens of thousands of years ago. And although people’s beliefs about the world will here be relevant, one certainly cannot understand a practice without asking about the goals and evaluations underlying it. Third, and on the basis of the first two points, if our understanding of religious faith is to be grounded in basic truths about religion, rather than in doctrinal proliferations and prejudices of recent times, which have so focused our attention on religious truth claims and the appropriateness or inappropriateness of believing them, we would do well to put our emphasis on the goals, evaluations, and practices at the heart of it. A corollary of my third comment is that respecting such an emphasis is more important than such things as finessing the distinction between “faith that” and “faith in,” which is found in ordinary language and has recently been debated obsessively by philosophers (including my former self).

3 With these points as part of our background, we can see, I think, that Swinburne is on the right track. But what should we say about his view that, in religious faith, certain beliefs—comparative epistemic beliefs, specifically—remain necessary? If we endorsed his general conception of belief, we might find it easier to go along with this, since for Swinburne all beliefs are epistemic in nature: believing always involves giving one proposition a higher status than some other. I do not agree with this, and have indicated as much and why elsewhere (Schellenberg 2005: 62–4). That said, I do not plan to make it an issue here. Instead, I will look for a reason supporting Swinburne’s view about comparing one’s own creed with alternatives that does not depend on his general conception of belief.



J . L . SCHELLENBERG

We may be able to find one in the following passage: To pursue a way in order to achieve the goals of religion, someone needs to believe that it is at least as probable that pursuit of that way will attain those goals as that pursuit of any other way will. . . . It would not merely be foolish or irrational but logically impossible to pursue a religious way in order to obtain a certain goal, if you believed that pursuing that way would make you less likely to obtain the goal than you would be otherwise (Swinburne 2005: 197).

A little argument is, if not intended, at any rate strongly suggested here: (1) If you believe that pursuing a certain way in order to achieve the goals of religion will make you less likely to obtain those goals than you would be otherwise, then you cannot pursue that way in order to obtain those goals. Therefore: (2) To pursue a way in order to achieve the goals of religion, you need to believe that doing so will make you at least as likely to obtain those goals as you would be otherwise. Unfortunately, the conclusion of this argument does not follow. One way to avoid the belief that pursuing your path will make you less likely to obtain your goals than you would be otherwise is to go all the way over to the other side of believing that it is at least as likely that your path will get you where you want to go as that any other will. But it is not the only way. A sufficient condition for not believing that p is believing that not-p. But this is not a necessary condition. For one might avoid believing that p simply by not having any belief on the matter at all. It follows that you might avoid believing that the pursuit of a certain way in order to achieve the goals of religion will make you less likely to obtain those goals than you would otherwise be simply by not having any belief on the matter at all. But is it really possible in the present case to avoid the relevant belief in this way? Can you pursue a religious path in order to obtain a religious goal while not having any belief at all in some way specifying the likelihood of your action’s success compared to that of other actions? If not, then Swinburne can simply say so explicitly, thus providing an extra premise that will help to make valid the argument his text suggests. I want to explore the view that this really is possible and that such a premise would be false.4 The first and perhaps most obvious point to be 4

In doing so, I have been helped and encouraged by Kvanvig (2013).

WORKING WITH SWINBURNE



made is that one might find the relevant probabilities indeterminate, and if I am in doubt about how likely my creed is compared to others, then clearly I lack a belief on what the relevant likelihoods are. However, in that case I might still have the belief that my creed is not obviously less likely than others, and Swinburne could easily transition to the view that this at least is required. Indeed, such a claim is close enough to his implied view to count as a version of it here. So our first point will not carry us very far. A stronger case for the possibility we are exploring will come from a closer look at how powerfully dominating one’s religious purposes and evaluations might be, especially when attended by relevant emotions. Years ago when I was a student, I heard a religious studies professor say, “Religion involves both a way of living and a way of seeing.” It is natural to suppose—and I supposed then—that the way of seeing referred to must involve religious beliefs (in particular, religious metaphysical beliefs). Or, after reading Swinburne’s Faith and Reason, we might suppose that it would have to at least involve comparative epistemic beliefs about religious propositions of the sort he emphasizes. But as Swinburne’s own work suggests, the “way of seeing” referred to might also or rather feature certain broadly evaluative beliefs about things religious: beliefs about such things as the extraordinary value of religious goals, the meaningful depths of experiences encountered along the way, how certain religious actions or dispositions are the actions or dispositions one should be doing or cultivating, and so on. And although a “way of living” will certainly involve actions or action dispositions, it too can be taken more broadly: in particular, if one’s way of living is shot through with emotions, then the former could hardly be described properly without reference to the latter. Now, and this is important, when a religious person loses religious beliefs as usually construed, if she does, she might keep these other beliefs (evaluative beliefs are often very firm, and resistant to being lost), and keep important components of her emotional life. And it seems that if she does keep these things, she might continue in religious practice and the pursuit of religious goals as before, too, and this without any intervening epistemic calculations—provided she does not acquire the belief that religious claims are false, or that some other path to a comparable set of goals is more likely to reach it (after all, the premise of the little argument stated earlier is very plausible). Here, as you can see, some



J . L . SCHELLENBERG

weaker conditions related to the epistemic condition emphasized by Swinburne are allowed—negative conditions involving an absence of certain beliefs about religious propositions or their relative likelihoods, and perhaps also the belief that the former beliefs are absent: “Well, at least I don’t believe that there is no God!” We might add that the religious person should be expected often to entertain religious propositions—to have them pass before her mind—in the contexts of religious experience and practice, but this too is a relatively weak condition if belief is what was emphasized before. Similarly for the point that such a religious person would be disposed to believe various religious or epistemic propositions—that is, be ready to come to or return to belief of them in appropriate circumstances—which, as Robert Audi has demonstrated, is quite distinct from believing them dispositionally (Audi 1994: 419–43). It will be good to work here with a more concrete example. So imagine a Christian—let us call her Esther—who, because of her contact with skepticism about Christian doctrines, over time slowly and involuntarily loses her Christian beliefs. She no longer believes, for example, that God was incarnate in Christ, though she does not disbelieve this either (and it is important to her that she does not). Her honest response to the evidence has left her in doubt on the matter, not knowing what to think. She does not have the concept of “credences,” but even if she did, she would not be able to make use of it in her own case; if you asked her about this, she would reply that she has no clue how much credence to give to any of the relevant propositions. Now it would be hard—and in any case unrealistic—to imagine that Esther is not at some level troubled by being in religious doubt. Her former believing way of being a Christian, which did feature the usual propositional component, is gone. But in the circumstances this only prepares her for a new way of being a Christian. Her Christian goals and Christian evaluative beliefs, so much a part of her, take on a new salience, and together with the attendant emotions, which are revealed as very strong indeed, they enable her to continue to engage in Christian practice. Perhaps we need to reverse the order of goals and evaluations to see how this could occur. So think about Esther’s evaluations. Though she no longer believes the doctrine of the Incarnation, her very entertaining of it during the time when doubt came has led her to see afresh what a wonderful and glorious idea it is: that God in compassion and humility

WORKING WITH SWINBURNE



would take on the form of a human being—this thought when she entertains it evokes powerful and rich emotions, and she thinks it would be simply the most tremendously good thing for it to be true. Certain associated thoughts Esther reacts to similarly. Take, for example, the thought that God became a human being so that human beings might transcend their limitations and become like God, and the related thought that she herself, by seeking to become more like Christ through Christian practice, can move ever deeper into an intimate relationship with God. Here we see a little more of what Swinburne would call a Christian creed, and part of it concerns what we need to do to obtain what Swinburne would call the Christian version of salvation. These things Esther continues to do, led into a continuation of Christian practice including the pursuit of Christian goals despite the loss of Christian beliefs, because of the freshly experienced power of her Christian evaluations and emotions. Do we have to suppose that she has any thoughts at all about alternative creeds, let alone their relative probabilities of being true, to explain Esther’s continuing practice? It seems not. Swinburne’s idea that one cannot follow a particular creed in pursuit of this or that religious purpose without thinking it to have a certain chance of being the right one when compared with alternatives, though plausible in the abstract, does not seem confirmed by concrete examples.5 Indeed, it is disconfirmed. And, ironically, it is his own important claim about the centrality of evaluations to religious faith, considered further, that leads to this result. Esther may not ever have thought about alternative creeds, only (say) about the problem of evil. Or she may have considered alternative creeds only as additional sources of doubt about Christian doctrines, not as alternative ways to salvation. Here we need to notice that this idea of salvation, too, has been treated quite abstractly by Swinburne and tends to be thus treated by most philosophers of religion, including myself in previous work. Esther is not likely to think of it in the same way—and this even if intellectual virtue is everywhere manifested by her mental 5 Or at least not by religious examples. Perhaps in mundane cases—or in some subset of mundane cases—we will find that it is impossible to pursue a goal by a certain means without believing that that means is at least as likely to allow one’s goal to be obtained as any alternative. It seems plausible that the move from the mundane to the transcendent—from particular purposes to the most general and most emotionally powerful governing purpose of one’s life—might affect whether consideration of alternatives must come into practice. But I have no space to discuss this issue further here.



J . L . SCHELLENBERG

and other behavior: one need not think in all ways as a philosopher in order to be intellectually virtuous! In fact, here too we can draw on information about the possible conative and evaluative features of faith to help us see the point more clearly. Precisely because of how strongly she responds, emotionally and evaluatively and in other ways, to Christian goals and Christian ideas, when Esther thinks of salvation it is only the Christian version of salvation that is before her mind, and when Esther thinks of how salvation is to be obtained it is only the Christian means that she entertains. These are the things she values. Indeed, she might regard it as absurd if someone were to suggest to her that there are alternative non-Christian ways of reaching her goal, ways that have nothing to do with the incarnation of God in Christ! But does this sense of absurdity itself not suggest that Esther believes those alternative ways to be less likely to lead to her goal than the Christian way to which she is devoted? Actually, if she thinks of them at all, she must regard them as being quite irrelevant to her own goal— understandably followed, perhaps, by those whose hearts are otherwise directed, but irrelevant to her own practice. The comparative judgment of “less likely” that the critic refers to here may follow from the irrelevancy claim, but there is no reason to think that Esther must make that inference and believe the result. (At most, she might be disposed to do so in the Audi sense.) Even if she did believe it, that belief would no more play a role in her faith or its cause than the belief that going to the Rotary Club twice a week is less likely to lead to salvation for Esther than the path she pursues, or that the action of petting a dog or kitten every day has that status. It may help to add plausibility to what I have said about Esther if we consider an analogy. So think about someone who together with others is pursuing the purpose or goal of ending war. You ask her whether she really believes that war will cease, or that her way of pursuing this goal will work or make a useful contribution to achieving that end, and she says she honestly does not know whether she believes either of these things—from which you rightly infer that she does not believe either of them. (Of course, others in her camp may feel differently; I am not suggesting that everyone who pursues her purpose will feel similarly any more than I am suggesting that every religious person is secretly like Esther.) But she adds, and with great fervency, that it would obviously be hugely important and a tremendously good thing if war were

WORKING WITH SWINBURNE



made to cease, and if her way of pursuing this goal would turn out to make some contribution to the achievement of that end. This is what she focuses on, and this is what keeps her going in the important project of trying to end war. Now it may occur to you as you listen to her that she must not believe that war will not be made to cease; and when you see that she and her people are pursuing their goal in their own particular way (say, through the strategy of nonviolent demonstration, which they have lovingly honed over many years), it occurs to you that she must not believe that among the clear alternatives is a better way of pursuing their goal. And you will naturally and rightly think that these negative conditions help to explain the nature and continuation of her activist practice. But seeing how much she is devoted to the ideal of the cessation of war brought about through nonviolent protest, you also see that you do not have to suppose that she has thought much about these things (certainly you do not have to think that she has made relevant epistemic calculations) in order to explain her behavior. Being moved to action by what we might as well call love in this context, love of an end and of a certain means of pursuing it, as also in Esther’s religious context, is a bit like being subject to the principle of inertia in physics: the movement continues unless some external force—in this case, coming to believe that the relevant view is false or that some alternative practice is more likely to be successful—can effectively be brought to bear. Perhaps it will be replied that my inertia analogy can be used against me: if the only reason Esther continues to practice is inertia, then teleology has been given up—Esther is no longer on the Christian path for the sake of any religious goal. How could she be, if she no longer even implicitly gives some credence to such propositions as that her engaging in Christian practice will lead to her achieving her religious goals?6 This sort of objection has some rhetorical effect because “doesn’t give some credence to p” can easily lead one to “gives no credence to p” and thence to “gives to p a credence of 0.” But though the latter move may psychologically follow quite naturally, its content does not follow logically. Indeed, the negation of that content follows. In particular, Esther gives no credence to the relevant propositions only in the sense of not assigning any probability to them at all, not even 0, because she is in

6

I owe this objection to Paul Draper.



J . L . SCHELLENBERG

doubt—in the familiar state of not knowing what to think. Another problem with the objection is that it tempts us to construe religious practice entirely in terms of behavior segregated from such things as emotions and desires—a temptation I have already suggested we ought to resist. Esther is not just robotically producing certain behaviors. Her behavior is part of a bundle of dispositional states including also these other things. And these other things keep her purpose—to achieve a deep relationship with God—very much alive and make it the case that she produces those behaviors that Christians regard as promoting Christlikeness in part for the sake of that purpose. The fuzziness of the relevant evidence is not enough to kill it. Clear evidence against the relevant propositions would be required to do so, just as it would be in the case of the anti-war activist and in many other examples we can think of—such as the following. A woman may continue to pursue the success of her marriage, a goal to which she is deeply devoted, by loving her wife or husband the best she can even when circumstances have placed the achievement of her goal very much in doubt. Her devotion may suffice to make this true unless and until she has strong evidence that her goal will not be achieved—say, her spouse leaves her and marries someone else. And, of course, while in the grip of such devotion, talk of other people with whom she might achieve a successful marriage and the thought of comparing relevant likelihoods here will simply leave her cold. But perhaps the position I am defending can in some other way be shown to be false or in need of revision. Here is one more way of trying to accomplish this result. Even if Esther can get by without Christian beliefs, metaphysical or epistemic, someone may say, surely she still needs some form of faith that the relevant propositions are true—in this case, a nondoxastic form of propositional faith of the sort that various writers, including Bill Alston (1996), Robert Audi (2008), Dan Howard-Snyder (2013), and a certain Schellenberg fellow (2005, ch. 6), have been championing recently. Thus the relevant cognitive states cannot altogether be avoided, and the move beyond religious belief cannot be taken nearly as far by the Christian apologist as I am suggesting. Here I am reminded of an episode from the year 2005. The second edition of Swinburne’s Faith and Reason came out that year, and so did my Prolegomena to a Philosophy of Religion. We traded copies of our books, and I recall that one of Richard’s comments on mine was to the

WORKING WITH SWINBURNE



effect that the concept of faith-that—one I had analyzed but that he himself had not employed—deserved a place in such discussions. He seemed to think that he should have brought it up. Naturally, I agreed! But after reading his book again while preparing for this chapter and thinking about it even more carefully than before, I am not so sure I do anymore. So to Richard I say: Don’t be too certain that faith-that is of any great importance in these discussions! Let us first push the discussion of your more fundamental issues further. Perhaps it will turn out that, in these matters, the pioneer is closer to getting it right than the new kids on the block. Here is why that might be the case. Religious faith—the fundamental set of dispositions identical with a religious form of life or personal religiousness—can be variously instantiated, as my own examples show. It would be overly narrow to suppose that it must include beliefs of the sort in question. But I now think that it might well be another sort of narrowness or dogmatism to suppose, as many do, that the religious life must include propositional faith. And this whether the latter be regarded as entailing a voluntary and nonbelieving sort of mental assent (as I once maintained, allowing that in many religious minds belief takes the place of propositional faith) or more irenically as capable of being instantiated believingly or nonbelievingly, through some disjunction of acceptance, assumption, trust, and relevantly similar states in the latter case (the views of Alston, Audi, and Howard-Snyder seem to fall here). Think again of Esther. To understand how she evinces religious faith in the new dispensation of doubt we do not need to introduce the idea that she has some continuing disposition of acceptance or assumption or trust or imaginative assent in relation to religious propositions or comparative epistemic ones, any more than we have to think of her as believing them. What is constant in her are religious goals and evaluations and emotions, as well as dispositions to act in pursuit of those goals. Of course she is entertaining religious propositions all the time, and it is hugely important to note that she entertains them with a powerful and rich pro-attitude, which all of the participants in the movement I have mentioned would regard as one of the essential components of propositional faith. But for propositional faith as they understand it, a cognitive component tied up with the relevant propositions is also needed. Esther, by their lights, needs to be in some way disposed to take a stand on the truth of some such propositions, but



J . L . SCHELLENBERG

I cannot now see that she is. The goodness these propositions represent is so much on her mind that, lacking the belief that they are false (and perhaps also occasionally reminding herself of this fact), she feels no need continually or frequently to represent to herself their truth. Even if in some moment of the complex, shape-shifting phenomenon that is an instantiation of religious faith she passes mentally into a truth-valued state in connection with this or that religious proposition, as distinct from her constant pro-attitude and the actions it makes her disposed to perform, I see no reason why, every time such a state arises, it must be the same one. Perhaps once it is mental assent, another time assumption or trust (whatever exactly a truth-valued version of the latter would be), and so on. Here I anticipate the reply that we can say she has a non-doxastic attitude of propositional faith toward religious propositions even if it is manifested in different ways at different times. But for this to be the case, we would still have to be thinking of a continuing disposition, and it is not at all clear that Esther must have one of this sort. Another reply that Howard-Snyder in particular might be inclined to try, given his latitudinarian approach to the question as to what might count as “a positive cognitive stance toward p,”7 is that some relevant part of Esther’s thinking about Christian propositions being doubtworthy might itself provide an instantiation for the relevant cognitive attitude. Notice how it is very important to her that these claims not be discernibly false—that they be only doubt-worthy. But this move seems a bit strained, fed more by the assumption that Esther must have propositional faith, which generates a willingness to be excessively latitudinarian, rather than by any plausibility in thinking of what Esther has got as a cognitive attitude toward the relevant Christian propositions that could fill the bill here. What if we dropped that assumption, recognizing that there can be faith without propositional faith, much as there can be propositional faith without belief? What if we admitted that the really fundamental faith concept is Swinburne’s, which, as I have noted several times, is more general, and can be instantiated variously? Then we could tighten the requirements for propositional faith and make our understanding of it more appealing and plausible, allowing that the cognitive component of propositional faith, even if

7

See Howard-Snyder (2013: 367).

WORKING WITH SWINBURNE



non-doxastic, should at least be belief-like (manifested by a positive attitude toward the relevant proposition rather than just one that is not negative—e.g., the assessment “not discernibly false”). Then, also, there would be no question as to whether Esther has propositional faith. A more specific response can be made here too—one more sensitive to the details of the Esther case. When we look closely we can see that Esther’s doubt (and any associated state including it) is the result of losing her old form of Christian faith, which did have a propositional element, namely, belief. Doubt signals the end of her old faith; it is much less naturally construed as a cognitive part of the new. Of course in a sense it permits the new. For it allows the value Esther places on her practice and its ideas (and her love for these things) to take on a new salience for her and allows the practice to continue. But this is a causal relation. Doubt remains part of the causal background to the new faith here rather than a cognitive part of it. If that doubt changed to disbelief, her new faith would probably end. But this relationship too can be understood causally. As for evaluative matters: with Esther, the epistemic part of our evaluation ends when (hopefully!) we conclude that her move from belief to doubt and the loss of her old believing form of faith were epistemically appropriate. Her new way of being a Christian, which follows after, requires us to shift our attention to other, more value-oriented considerations. And this itself suggests that her new way of being a Christian is more radically new than any attempt to stretch propositional faith to fit could allow. The upshot? Although the faith-that/faith-in distinction is well entrenched in the ordinary language of today, we should not impose it on our most fundamental analyses of religious faith. This is not safe, since the latter is often much more concerned with the good than with the likelihood of truth. More to the point: if I love the goal I am pursuing, and this certainly—perhaps preeminently—includes religious goals, then that may explain why I pursue it, and a reference to propositional attitudes of the sorts and with the objects we have been considering will not be needed to understand this. And of course if my religious goal is worth loving that much, something a Christian could hardly deny when it comes to a Christian goal like Esther’s, then the religious faith that grows from my pursuit of it may also be rationally defensible in ethical terms (more on this in a moment), and this without questions about the justification of religious metaphysical beliefs or relevant epistemic beliefs ever coming up.



J . L . SCHELLENBERG

4 Our extension and revision of Swinburne’s emphasis thus allows for a form of Christian practice that is grounded primarily in a sort of love, and remains thus grounded even in the absence of religious propositional beliefs or faith and comparative epistemic beliefs or faith related thereto. If in the future religious skepticism were to gain an even deeper hold, this sort of Christian practice might remain possible. And so we can see Christians like Esther as experimenting with a form of Christianity that may be much more widely realized one day. In conclusion, with an eye on these thoughts, I want to deliver the minisermon promised at the beginning. This was to be about the besetting sins of philosophers—including both Swinburne and myself—that can prevent us from seeing how a flourishing faith is able to exist without the cognitive states of which we are so fond (states such as belief that one’s creed is more probable than others or faith that one’s creed is true). Our sins, in this regard, are many, and I shall have to restrict myself to the more important, leaving aside such venial sins as our tendency to overdo conceptual fastidiousness, and the way we still so often bow to ordinary language. The first of the two graver sins I have in mind I will only mention, since it is amply illustrated in earlier parts of this chapter and will, I think, immediately be understood. This is a tendency to what might be called overintellectualization. Sometimes in our analyses of religious faith, we veer, I think, into a description of faith for philosophers—a form of faith that is full of cognitive states and truth-valued attitudes, whose rationality therefore, naturally enough, depends largely on whether these states and attitudes are appropriately formed, with due sensitivity to the facts. Simply by thinking about Esther and others like her, we can, I suggest, see the inadequacy of such an approach if it is intended to produce an analysis of faith with a fully general application—or even if it is intended, more narrowly, to cover rational forms of faith. But the sin I want to focus on is a somewhat different one. We philosophers of religion commit this sin whenever we forget that philosophy of religion includes not just the metaphysics of religion and the epistemology of religion, but also the ethics of religion.8 When in its grip, 8

One who has not forgotten is John Bishop. See his defense of the view that the proper evaluation of religious faith is moral in Bishop (2007).

WORKING WITH SWINBURNE



we think about religious faith, but are not in a good position to notice—as Swinburne to his credit has done—that religious faith is fundamentally about what states of affairs are most deeply valued, and the goals and action dispositions and emotions formed in relation thereto. Of course cognitive states are not just inadmissible. For one thing, they appear wherever religious evaluative beliefs are found. But we should not assume, as we may be inclined to do as metaphysicians or epistemologists, that religious faith must include serious attention to the truth-values of claims about the world, and that some such claims must be believed or mentally embraced through some belief-like state. So what might we learn if we repented of this sin? Well, the ethics of religion, taken most generally, may be seen as involving these two things: the application of insights from moral philosophy in philosophical work on religion, and also the reciprocal endeavor on the part of philosophers to understand what religion might have to contribute to moral philosophy. Let me make one relevant point from each side. Applying insights from moral philosophy, we might recognize that the philosophical evaluation of religion should be just as much about how moral virtue is or can be exemplified religiously as about religion’s capacity for intellectual virtue. It is easy to get hung up on whether a Christian justifiedly believes, say, the doctrine of the Incarnation, or on the metaphysical intricacies involved in the latter, and to fail to notice that some Christians do not believe the doctrine at all while managing nonetheless to be related mentally to it in a way that is religiously and morally approvable. Esther, for example, places great value on the ultimate humility and compassion she associates with that doctrine, in effect holding that it would be good for these character traits to penetrate to the very heart of reality. Moreover, in connection with this evaluation, she commits herself to making it the case that these traits penetrate to the heart of her own reality—to the heart of the life that she lives. And with the deep emotions that attend this evaluation and this purpose, she displays the fact that this is already to some extent the case. How could her religious life, if it includes these features, fail to be both religiously and morally approvable in a very deep way? But, by one-sidedly applying insights from metaphysics and epistemology, largely forgetting about ethics, we will fail to notice this important point. How about what religion may contribute to ethics? The point I want to draw out here has to do with the way in which a religious life, quite



J . L . SCHELLENBERG

without metaphysical propositions believed or taken as true on faith, offers itself as a candidate for the center of a good life—one that may seem especially appropriate when we consider how short, so far, has been the journey of intelligence on our planet and how much further our species may have to travel in the future of evolution. My description of Esther’s life, which again functions paradigmatically, does not reflect a non-cognitivism of the sort we have seen before in the philosophy of religion, which allowed no place for such a thing as the doctrine of the Incarnation taken literally, but rather involves a shift of emphasis: a shift to the goodness reflected in that doctrine taken literally and to the religious purposes that the entertaining of it may subserve, and away from a bare consideration of its truth. And this sort of emphasis will perhaps more easily be noticed by moral philosophy as potentially forming the center of a good life. For moral philosophy to take fresh interest in religion in this way, it will be necessary to bring out and make conspicuous one more point about purposes. A religious life of the sort in question offers itself to ethics not just as providing ethical benefits deriving from the pursuit of distinctively religious purposes, but also as especially (and perhaps uniquely) conducive to the fulfillment of a variety of aims that, on independent grounds, we can see are deeply appropriate for beings like us, finding our way at an early stage of evolution. Swinburne is right to associate the religious life with religious purposes, but the religious life may gain some of its attractiveness from how it serves many other purposes too—such as the purpose of a fully expanded imagination, or of carrying out a wholehearted commitment to the human good, or of making a fully appropriate responsive gesture to what ought to be true in respect of our own lives, our religious experiences, the human species, and the universe, even when it is very much in doubt whether it will be true. Sermons are supposed to end with uplift, even if they began with admonishment. Perhaps reflection on what all these points have to say about the promise of a new or renewed ethics of religion will be sufficient to make for uplift among philosophers—and also for stimulation and provocation! If so, do not thank me. We have been working with Swinburne. Thank him instead.9 9

Here I would also like to thank, for their help with this chapter, Michael Bergmann, Jeff Brower, Paul Draper, Dan Howard-Snyder, and Trent Dougherty.

WORKING WITH SWINBURNE



References Alston, William. (1996). “Belief, Acceptance, and Religious Faith,” in Jeff Jordan and Daniel Howard-Snyder (eds), Faith, Freedom and Rationality. London: Rowman and Littlefield, 3–27. Audi, Robert. (1994). “Dispositional Beliefs and Dispositions to Belief,” Noûs, 28: 419–43. Audi, Robert. (2008). “Belief, Faith, and Acceptance,” International Journal for Philosophy of Religion, 63: 87–102. Audi, Robert. (2011). Rationality and Religious Commitment. Oxford: Clarendon Press, 2011. Bishop, John. (2007). Believing By Faith: An Essay in the Epistemology and Ethics of Religious Belief. Oxford: Oxford University Press. Howard-Snyder, Daniel. (2013). “Propositional Faith: What It Is and What It Is Not,” American Philosophical Quarterly, 50: 357–72. Kvanvig, Jonathan. (2013). “Affective Theism and People of Faith,” Midwest Studies in Philosophy, 37: 109–28. Schellenberg, J. L. (2005). Prolegomena to a Philosophy of Religion. Ithaca: Cornell University Press. Swinburne, Richard. (2001). Epistemic Justification. Oxford: Clarendon Press. Swinburne, Richard. (2005). Faith and Reason, 2nd edn. Oxford: Clarendon Press.

Theistic Arguments

3 Simplicity and Natural Theology Paul Draper

Richard Swinburne is without doubt the greatest natural theologian of the twentieth century. When I read his masterpiece, The Existence of God, in graduate school, it felt like the scales fell from my philosophical eyes, revealing the best approach to religious epistemology. Thus, it is both a great pleasure and a great honor to contribute a chapter to this volume celebrating Professor Swinburne’s remarkable career. My project is to examine and critically discuss the role of simplicity in Swinburne’s probabilistic natural theology. After describing that role and the details of his theory of simplicity, I challenge Swinburne’s view that the criterion of simplicity is a fundamental criterion for evaluating causal explanations, proposing instead that what is right about that criterion can be derived from a more fundamental criterion of “coherence.” I close by exploring the implications of my proposal for Swinburne’s natural theology.

1 Swinburne on Simplicity and Natural Theology Natural theology, says Swinburne, “claims that the most general features of the universe show that there is a God” (2012: 107). Swinburne understands God to be, most fundamentally, an everlasting person who is omnipotent, omniscient, and perfectly free from the influence of nonrational desires. Other important divine attributes like perfect goodness

SIMPLICITY AND NATURAL THEOLOGY



and omnipresence allegedly follow from these core attributes, which Swinburne takes to be essential properties of God. Theism, the hypothesis that there is a God, purports to provide a causal explanation of, among other things, the most general features of the universe. Thus, whether or not those features show that theism is true depends on how well theism satisfies the criteria we ought to use to assess the probability of any causal explanation. According to Swinburne (2001: 80–3), there are four such criteria, two a posteriori and two a priori. The first a posteriori criterion is explanatory power. A hypothesis has explanatory power to the extent that it entails or makes probable data that is otherwise unexpected. In other words, the higher the inductive probability of the data conditional on the hypothesis (conjoined with relevant background information) and the lower the inductive probability of the data given the falsity of the hypothesis (again conjoined with relevant background information), the more probable an explanatory hypothesis will be.1 The second a posteriori criterion is fit with background information. For many scientific theories, this will be a matter of how well (compared to its rivals) a hypothesis meshes with established hypotheses that explain data outside of its range. The first a priori criterion is smallness of scope, which I will call “modesty,” where the scope of a hypothesis is a measure of how much the hypothesis purports to tell us about the world. Clearly, quite apart from a posteriori considerations, the less a hypothesis says that might be false (that is, the more modest it is), the more likely it is a priori that the hypothesis is true. The second a priori criterion is simplicity, which, as we shall soon see, has five distinct facets according to Swinburne, spanning considerations of parsimony, intelligibility, and elegance. Together, modesty and simplicity determine the intrinsic or a priori probability of a hypothesis, though Swinburne argues (and as we shall see, this is crucial to his case for theism) that simplicity should be given much more weight than modesty. Lest anyone be tempted to make do without the criterion of simplicity, Swinburne is quick to point out that, 1 My notion of ‘explanatory power’ is slightly different from Swinburne’s. For Swinburne, a theory’s explanatory power with respect to the data is the ratio of the inductive probability of the data given the theory to the prior probability of the data. The difference here is only terminological. The criterion I call ‘explanatory power’ is Swinburne’s criterion, though he sometimes divides it into two criteria (e.g. 2012: 104). He just would not call it ‘explanatory power.’



PAUL DRAPER

no matter how well a hypothesis satisfies the first three criteria, there will always be other hypotheses that satisfy those criteria at least as well. Yet we often judge confidently and correctly of various explanatory hypotheses that they are probably true. Therefore, there must be a fourth criterion and simplicity is a natural candidate, since it appears to play an important role in theory choice in science and all other factors that seem to play such a role (with the possible exception of beauty) can be subsumed under the other three criteria. So simplicity plays a crucial role in Swinburne’s probabilistic natural theology partly because theism is a causal explanation and simplicity is one of the four fundamental criteria that, according to Swinburne, should be used to assess the inductive probability of such explanations. Thus, in order to understand his natural theology, it is essential to understand what Swinburne has to say about the so-called “problem of simplicity,” which I take to be the two-part question of what, exactly, the relevant sort of simplicity is and why possessing simplicity of that sort makes a hypothesis more likely to be true. Concerning the first part, it turns out that simplicity of the relevant sort is, according to Swinburne, a rather complex thing, involving not just multiple but also diverse facets. Specifically, one hypothesis or one formulation of a hypothesis can be simpler than another, according to Swinburne (1997: 24–6), for any of the following five reasons: (1) It postulates fewer entities. (2) It postulates fewer kinds of entities. (3) The terms (including any mathematical terms) it contains are easier to understand in the sense that fewer other terms must be understood in order to understand them (e.g., because the properties or entities to which those terms refer are more readily accessible to observation or experience). (4) It consists of fewer laws or laws that relate fewer variables. (5) It contains fewer mathematical terms. Usually, simplicity is understood either as ontological parsimony, which makes simplicity a property of a hypothesis, or as some sort of linguistic elegance, which makes simplicity a property of a particular formulation of a hypothesis. Swinburne’s five facets of simplicity encompass both parsimony and elegance. Just the first two facets are quantitative and qualitative ontological parsimony. The last two are nomological and

SIMPLICITY AND NATURAL THEOLOGY



mathematical elegance. Since every hypothesis has multiple formulations, some more elegant than others, Swinburne says that to assert that one hypothesis is simpler than another (when the difference is one of elegance instead of just parsimony) is really to assert that the simplest formulation of the first hypothesis is simpler than the simplest formulation of the second. But what makes the simplest formulation of one hypothesis more elegant than the simplest formulation of another hypothesis? For Swinburne, this is not just a matter of syntax. For example, compare the hypothesis that all Asian crows are black and all non-Asian crows are black to the equally modest hypothesis that all Asian crows are black and all non-Asian crows are white. The former can be formulated more elegantly as the statement that all crows are black. This suggests that the former hypothesis is simpler. But what if a new predicate is introduced, namely, ‘blite’ where something is blite just in case it is either Asian and black or non-Asian and white? Then the second hypothesis can be formulated quite elegantly as the hypothesis that all crows are blite. Does this mean that the second hypothesis can be formulated as simply as the first? No, says Swinburne, because of his third facet of simplicity, which is a sort of intelligibility. The term ‘blite’ cannot be understood without understanding the term ‘black’ (not to mention ‘white’ and ‘Asian’), while the term ‘black’ because of the accessibility to observation of the property it denotes, can be understood without understanding the term ‘blite.’ Thus, even if we take advantage of the word ‘blite,’ the hypothesis that all Asian crows are black and all nonAsian crows are also black can on Swinburne’s theory be formulated more simply and thus is, according to Swinburne, intrinsically more probable than the hypothesis that all Asian crows are black and all non-Asian crows are white. Swinburne’s approach to the second part of the problem of simplicity is, to my mind, less satisfying. Swinburne needs to explain why being simple (or having a formulation that is simple) in any of the five ways he describes makes a hypothesis more likely to be true; but all he says is that his criterion of simplicity “cannot be established empirically, and hence it must be an a priori principle. We do not use it because it follows from some more obvious a priori principle. Hence it must be a fundamental a priori principle” (2001: 102). If he said the same thing about modesty (and removed the assumption that more fundamental a priori principles



PAUL DRAPER

must be more obvious), I would agree for two reasons. First, modesty is just one thing: it does not have multiple and diverse facets. Second, the criterion of modesty is supported by our modal intuitions: it seems to be a necessary truth that, if we set aside all data and background information, then saying less about the world will, other things being equal, make a hypothesis more probably true. Simplicity, however, is not like modesty in either of these two ways. First, even if Swinburne’s criterion of simplicity can be justified a priori, the fact that it is so complicated is good prima facie reason to doubt that it is fundamental. To convince us otherwise, Swinburne would need to say more. For example, he could try to specify some precise general notion of simplicity from which all five of his facets can, in spite of their diversity, be derived. Second, the simplicity criterion is not supported by our modal intuitions: it does not seem to be true of necessity that simpler hypotheses or hypotheses that can be formulated more simply are more likely to be true. And non-modal intuitions about whether simplicity is evidence of truth vary widely among both philosophers and scientists. So our intuitions do not help to justify a principle of simplicity as much as they help to justify a principle of modesty. Still, I do agree with Swinburne (2001: 100–2) that any justification of his criterion of simplicity must be a priori. Therefore, it is worth investigating whether there is some more fundamental a priori criterion from which Swinburne’s criterion of simplicity or whatever is right about that criterion can be derived.

2 Coherence, Uniformity, and Simplicity I will approach this question by asking a more general question: is Swinburne’s list of four criteria for assessing the probability of a causal explanation and more specifically his list of two criteria for evaluating the intrinsic probability of such an explanation complete? To appreciate why a negative answer to this question is warranted, reflect for a moment on the idea of intrinsic probability. What intrinsic features of a hypothesis could affect its probability? In other words, what features could affect its probability independently of all data and background information? As we have seen, modesty—the inverse of scope—is one such feature. The less a hypothesis says about the world, the less likely it is to say something false and so the more likely it is to be true. So the intrinsic or

SIMPLICITY AND NATURAL THEOLOGY



a priori probability of a hypothesis depends at least in part on the size of its content, on how much it says. What else could it depend on? One answer should be uncontroversial. Just as there might be good or bad fit between hypothesis and data or between hypothesis and background information, so too there might be good or bad fit between the various “parts” (i.e., logical consequences) of a hypothesis. Let us call this new criterion for assessing the probability of an explanatory hypothesis ‘coherence.’ This criterion, unlike the criterion of simplicity, is supported by our modal intuitions: it seems to be true of necessity that the intrinsic probability of a hypothesis depends not just on how little it says, but also on how well what it says fits together. Coherence admits of many degrees. At one end of the spectrum are necessary falsehoods, which have contradictory parts and so are maximally incoherent. Notice that such statements entail everything and so are maximally immodest as well.2 At the other end are necessary truths. A necessary truth is maximally coherent because each of its parts entails every other part. Also, since such hypotheses tell us nothing about the world that could be false, they are not just maximally coherent, but maximally modest as well. Two equally modest statements, however, may not be equally coherent. This is because there may be inductive support relations between the parts of a hypothesis, support relations that hold independently of data and background information. For example, consider again the hypothesis that all crows are black. As previously mentioned, this hypothesis is logically equivalent to the hypothesis that all Asian crows are black and all non-Asian crows are black. Now compare this hypothesis to a second one, namely, that all Asian crows are black and all non-Asian crows are white. The two 2

One might object that I am measuring immodesty by the number of logical consequences a statement has. This will not do, since every statement has infinitely many logical consequences. That is not, however, what I am doing. When I say that one statement has more content than another, I do not mean that it has a greater number of logical consequences than the other statement, because I use ‘more’ in the “all and then some” sense instead of in the “greater number” sense. Notice also that my using ‘more’ in that sense does not imply that one statement must be entailed by another statement in order to be more modest than that other statement because “all and then some” can apply, not just to whole propositions, but also to the proper parts of a proposition. (My colleague Dana Tulodziecki helped me to see this point.) Thus, for example, since ‘purple’ has all of the content that ‘colored’ has and then some, it follows that the statement “this ball is colored” is more modest than the statement “that ball is purple” even though each has infinitely many logical consequences and even though neither statement entails the other.



PAUL DRAPER

hypotheses are equally modest, but not equally coherent. The first half of both hypotheses states that all Asian crows are black. This supports the second half of the first hypothesis, which states that non-Asian crows are also all black, while it counts against the second half of the second hypothesis, which states that non-Asian crows are all white. Thus, the hypothesis that Asian and non-Asian crows are all black is more coherent and thus intrinsically more probable than the equally modest hypothesis that all Asian crows are black and all non-Asian crows are white. Not surprisingly, it can also be formulated more elegantly in all natural languages (which do not include words like ‘blite’). The next question, especially given my last remark about elegance, should be obvious. What is the connection if any between simplicity as Swinburne understands it and coherence? If there is no connection, then presumably Swinburne should add a criterion of coherence to his list and make whatever adjustments to his natural theology such an addition calls for. Alternatively, and this is the position I wish to explore, it may be that, while simplicity in Swinburne’s sense and coherence in my sense are clearly distinct properties, something close to Swinburne’s criterion of simplicity can be derived from the more fundamental criterion of coherence. I submit that this position is defensible, as long as it is safe to assume that basic inductive support relations can hold a priori between the parts of a hypothesis when data and background information are set aside.3 Given that assumption, hypotheses that attribute real uniformity to the world are to that extent more coherent and thus more probable intrinsically than hypotheses that attribute real variety, either at a time or over time. Swinburne worries that principles like this are empty, because every hypothesis attributes uniformity to the world in some respect (2001: 84).

3 One might challenge this assumption on the grounds that it is not true of necessity that the universe is uniform to any particular degree or within any particular range of degrees. This challenge fails for two reasons. First, even if nature is not necessarily uniform, it still might be a necessary truth that, in the absence of evidence showing otherwise, uniformity is more probable than variety. Indeed, I argue for precisely this thesis in “A Modest Solution to the Problem of Induction” (unpublished). Even if that thesis is false, however, the almost universal and irresistible a priori intuition that induction is reliable and hence that uniformity is, in the absence of evidence to the contrary, more likely than variety supports the assumption in question. This is so even if that intuition is not a modal intuition. Contrast this with the far less widespread and more easily resistible a priori intuition that simplicity is evidence of truth.

SIMPLICITY AND NATURAL THEOLOGY



Thus, as we have already seen, the hypothesis that Asian crows are all black and non-Asian crows are all white says that all crows are “uniformly” blite, where again something is blite if it is Asian and black or non-Asian and white. It is clear, however, to all who are not caught in Goodman’s grip, that simply inventing a word like ‘blite’ does not transform variety into uniformity. In other words, the hypothesis that all crows are blite does not attribute real uniformity to the world, where an instance of uniformity (or variety) is real only if its being an instance of uniformity (or variety) instead of an instance of variety (or uniformity) does not depend on the linguistic machinery or conceptual scheme used to describe or apprehend it. Swinburne himself, as we have already seen, gives preference to hypotheses employing predicates that refer to properties that are more accessible to observation. Hypotheses employing such predicates exhibit his third facet of simplicity. Thus, when Swinburne says at one point that we should expect nature to be uniform in the simplest respect (2001: 84), I suspect he is identifying “simplest respect” with the respect most accessible to observation (or other sorts of experience). In claiming, however, that this sort of simplicity increases the probability of a hypothesis being true, Swinburne is assuming, as I am, that our faculties are in touch with reality when we judge there to be uniformity in the surfaces of a group of Asian and non-Asian black objects (or in the way those surfaces appear to us) and variety in the surfaces of a group of Asian and non-Asian blite objects. Of course, all this is at odds with a popular view about the significance of Goodman’s paradox. A number of philosophers—e.g., Skyrms (1999, ch. 4)—take the lesson of that paradox to be that all uniformity and variety is linguistically or conceptually relative. There is, however, nothing more fundamental to cognition than our sense of similarity and dissimilarity (both at a time and over time). Thus, while I admit that the successful cognitive relation that human beings have to the extra-linguistic and extra-conceptual world is difficult to explain, the implicit denial by Goodman (and others who take his paradox very seriously) that we have such a relation to that world should strike realists as preposterous. Whether or not I am right about that, however, my main point here is that Swinburne is no less committed to “uniformity realism” than I am. Let us return, then, to the crucial question. Can Swinburne’s criterion of simplicity or something sufficiently close to it be derived from the



PAUL DRAPER

criterion of coherence, at least on the assumption that the coherence of a contingent hypothesis is largely a matter of how much real uniformity that hypothesis attributes to the world? By a criterion of simplicity being sufficiently close to Swinburne’s, I mean close enough to fit with scientific practice just as well as Swinburne’s criterion does. Because Swinburne distinguishes five facets of simplicity, this question requires five answers. Let us begin with Swinburne’s first two facets of simplicity, each of which involves a sort of ontological parsimony. Swinburne’s second facet is qualitative parsimony. Hypotheses that postulate fewer (presumably real) types of entities are to that extent simpler. This follows from the criterion of coherence, because the more kinds of entities a hypothesis postulates, the more variety it attributes to the world. Further, if one accepts the identity of indiscernibles, then multiplicity requires multiple types and so qualitative parsimony, Swinburne’s second facet, entails quantitative parsimony and so the latter need not be regarded as a distinct facet. Moreover, even if one denies that indiscernibles are necessarily identical, as long as one accepts that all or even the vast majority of indiscernibles are as a matter of fact identical, one can account for the success of a principle of quantitative parsimony in terms of a principle of qualitative parsimony and so ultimately in terms of the criterion of coherence. We have already discussed Swinburne’s third facet of simplicity. I am not sure that I would want to rely in every case on accessibility to observation or experience to distinguish different degrees of naturalness in properties, but Swinburne’s third facet is supported to some extent by the criterion of coherence. Hypotheses that employ artificial or highly theoretical predicates like ‘blite’ and ‘grue,’ which are less accessible to observation and so must be defined in terms of predicates that refer to observable properties, are certainly more apt to attribute real variety to the world even if, grammatically, they attribute uniformity to it. According to Swinburne’s fourth facet of simplicity, the fewer the number of laws that a hypothesis postulates, the simpler it is. Swinburne’s favorite example from science compares Kepler’s theory of planetary motion with its three laws (and six propositions indicating a mean distance from the sun for each planet) to Copernicus’s theory, which, like Ptolemy’s geocentric theory, employs at least forty laws. Now, it might be tempting to think that postulating fewer laws makes Kepler’s theory intrinsically more probable than Copernicus’s because it makes it

SIMPLICITY AND NATURAL THEOLOGY



more modest. This, however, is a mistake if one thinks of Kepler’s and Copernicus’s theories as each specifying equally precise positions in space over time for each of the six planets known to them. (I ignore the moon here.) If that is the correct way to interpret their theories, then they are equally modest. The question, then, is whether the greater number of laws in Copernicus’s theory makes that theory prima facie less probable than Kepler’s even if it does not make it less modest. This may not be an ideal case for testing our intuitions about Swinburne’s fourth facet. One complicating factor is that Kepler’s theory is more probable for other reasons, most notably the fact that it fits the data (available at the time) much better than Copernicus’s theory. Another is that, individually, Copernicus’s laws exhibit Swinburne’s fifth facet of mathematical simplicity to a higher degree than Kepler’s. Swinburne, of course, is aware of both of these complicating factors. Yet he thinks it is clear enough that having more laws is a theoretical disadvantage for Copernicus’s theory—that if the two theories satisfied all other criteria equally and exhibited all other facets of simplicity equally, any right-thinking person would without hesitation conclude that Kepler’s theory is more likely to be true precisely because it has fewer laws. I disagree. Greater nomological elegance is positively correlated with greater intrinsic probability because it is correlated with greater modesty. In cases like this, however, in which we have equal modesty, a difference in nomological elegance does not imply a difference in intrinsic probability. I do share Swinburne’s intuition, however, that Kepler’s theory is intrinsically more probable than Copernicus’s, but I submit that this is due to a difference in coherence. To appreciate this point, one needs to examine what the two theories actually say about the solar system. Kepler’s theory is more probable intrinsically than Copernicus’s simply because his laws together imply more uniformity in the motion of actual heavenly bodies than Copernicus’s laws do. This may be one reason why fewer laws are required to state Kepler’s view, but it is also true that Copernicus is doomed to have a large number of laws by his misguided view that all natural motion must be circular motion at a constant speed. Granted, circular motion at a constant speed is more uniform than any other motion except rectilinear motion at a constant speed, but as Copernicus recognized, the planets clearly do not circle the sun and they do not travel at a constant speed relative to the center of the sun



PAUL DRAPER

or to any other fixed point in space. This forced him to employ Ptolemy’s device of multiple epicycles of various sizes for each planet (though not the same number for each planet) and varying rates of constant speed for the real or (in most cases) imaginary motion around his several dozen heavenly circles. Part of Kepler’s genius was that he was willing to give up the ancient ideal of circular motion at a constant speed, focusing instead on attributing as much uniformity as the data allows to the actual paths of the planets around the sun. Swinburne’s fifth facet of simplicity focuses on hypotheses whose laws are mathematical equations (or descriptions). As he did in the case of his fourth facet, Swinburne equates simplicity here with a sort of elegance: other things being equal, the more elegant an equation is in the sense of having fewer terms, the more probable intrinsically it is that the equation is true. Given his third criterion, however, he also takes the simplicity of an equation to depend on how intelligible the mathematical entities and relations in the equation are, where one entity or relation is more intelligible than another if the second cannot be understood without understanding the first (but not vice versa). Once again, I submit that the more fundamental criterion of coherence accounts for the plausibility of Swinburne’s criteria. One reason for this is that, when uniformity can be described in the language of mathematics, fewer terms are, as a general rule, required to describe it. For example, constant speed can be described by the very elegant equation d = kt, while uniform change in speed can be described by a less elegant but still very elegant equation d = kt2. Further, if one were to change the ‘2’ in this equation to a larger integer or a non-integer, then, contrary to what Swinburne says, the real reason that the equation will be less likely to be true is not that smaller integers are more intelligible than larger integers nor is it that integers are more intelligible than non-integers. It will be true for the more fundamental reason that the equation will not attribute as much real uniformity to the world. Swinburne asks why, in the eighteenth century, Newton’s law of gravity, F = mm’/r2, was far more likely to be true than the law F = mm’/r2.000 . . . (100 zeros) . . . 1. Given the inevitable limitations to the accuracy of our measurements, both theories yielded the data available at the time equally well. Swinburne’s answer is that the integer 2 is a simpler number than the non-integer 2.000 . . . (100 zeros) . . . 1. But again, there is a more basic answer,

SIMPLICITY AND NATURAL THEOLOGY



namely, that Newton’s equation attributes more uniformity to the (changes in) motion of physical objects solely under the influence of the force of gravity than the second equation does because it says that the inverse of the force of gravity attracting two objects of (unchanging) mass increases at a constant rate as the distance between the centers of those two objects increases. It would seem, then, that nothing more than the criterion of coherence (combined with the thesis that basic inductive support relations can hold a priori between the parts of a hypothesis) is needed to account for the cases of theory choice in science that Swinburne uses to defend or illustrate his criterion of simplicity. In addition, if the criterion of coherence is fundamental, then we can explain why Swinburne’s multifaceted criterion of simplicity fits scientific practice to the extent that it does. Finally, the criterion of coherence, unlike Swinburne’s criterion, is strongly supported by intuition. What all of this strongly suggests is that the criterion of coherence should replace the criterion of simplicity on Swinburne’s list of fundamental criteria for evaluating the probability of explanatory hypotheses. This would be a substantial change, because the two criteria are not equivalent for two reasons. First, one hypothesis might attribute greater uniformity to the world and so be more coherent than another, even though it is not simpler in any of Swinburne’s five ways. For example, the hypothesis that Swinburne believes in God right now and will hold that same belief a second from now is more coherent but not simpler in Swinburne’s sense than the hypothesis that Swinburne believes in God right now but will lack that belief a second from now. Indeed, Swinburne is actually committed to the position that the second hypothesis is the simpler of the two because asserting that a property (like the property of holding a certain belief) is absent is, according to Swinburne, “always simpler” than asserting that one is present (2010: 19). Second, there may be cases in which a hypothesis that attributes more uniformity to the world than another hypothesis is mathematically less elegant (or postulates more laws) than that other hypothesis. In other words, it may be that some of Swinburne’s facets cannot be deduced from coherence, but instead are only correlated with it (or in the case of nomological elegance with modesty). If so, then any criterion of intrinsic probability based on those facets has the status of being a good rule of thumb that admits of exceptions.



PAUL DRAPER

3 Implications for Swinburne’s Natural Theology As suggestive as the results of the last section are, more investigation is needed to settle once and for all the issue of whether or not coherence should replace simplicity. Suppose, however, for the sake of argument, that it should. Does this make a difference to Swinburne’s probabilistic natural theology? I argue in this third and final section of my chapter that it does. To set the stage for my argument, I need to add a little to what I said in Section 1 about Swinburne’s natural theology and the role that simplicity plays in it. Swinburne claims that the probabilities of theism and its rival explanations do not depend on fit with contingent background knowledge, because the data to be explained here are facts about the universe’s most general features. Thus, the probabilities of theism and its rivals depend solely on (i) how well they “predict” the data—that is, on how (antecedently) probable the assumption that they are true makes the data—and (ii) on their intrinsic probability, the latter of which depends in turn on simplicity and modesty. Swinburne largely ignores considerations of modesty, however, because he thinks that far more weight should be given to simplicity than to modesty. Accordingly, there are two main components of his natural theology, one a priori and one a posteriori. The a priori component consists of a case for theism’s being simpler and so intrinsically more probable than all of its rivals4 and vastly simpler and so vastly more probable intrinsically than the vast majority of those rivals. The a posteriori component consists of a case for theism predicting the data much better than any rival whose intrinsic probability is not negligibly low. Notice that the a posteriori component of his case depends on the results of the a priori component and so ultimately depends for its success on his views about simplicity. If he could not, based on considerations of simplicity, rule out most rivals to theism as having negligibly low intrinsic probability, then he would have to compare how well theism predicts the data to how well a plethora of rivals predict the data, and practically speaking that would be impossible.

I restrict “rivals” to live options and so do not count as a rival the hypothesis that nothing exists, which Swinburne believes is simpler than theism. 4

SIMPLICITY AND NATURAL THEOLOGY



Swinburne’s theory of simplicity, then, does not just play a role in his natural theology—it plays the starring role. What, then, would become of Swinburne’s case for theism if the criterion of simplicity were replaced with the criterion of coherence? I believe it would be seriously weakened, though how it would be weakened depends on whether Swinburne’s view that simplicity should be given much greater weight than modesty could be replaced without loss of plausibility by the view that coherence should be given much greater weight than modesty. If it could, then the a posteriori component of Swinburne’s case would be weakened much more than the a priori component. If not, then both components would be seriously weakened, though the a posteriori component would be weakened for a different reason. Or so I will argue in the remainder of this chapter. Suppose, then, on the one hand, that simplicity is replaced with coherence and coherence is given far more weight than modesty. In other words, suppose that theism’s intrinsic probability (or at least how that probability compares to its rivals) depends almost entirely on its coherence. Then it is arguable that the intrinsic probability of theism is, like Swinburne claims, much greater than at least many of its rival explanations of the universe. Polytheistic hypotheses, for example, postulate many gods and so postulate variety that theism does not. Quasitheism postulates a finite God, a God that is not omnipotent or that is not omniscient or that has non-rational desires. Such a God’s power, knowledge, or motivational structure exhibits much less uniformity than a theistic God’s. To postulate, for example, that God can bring about some contingent future states and not others is less coherent than claiming that God is omnipotent because the ability to bring about some such states is evidence for the ability to bring about other such states. For similar reasons, omniscience is intrinsically more likely than limited knowledge. Further, postulating a being that is everlasting and that uniformly lacks all non-rational desires is more coherent than postulating a being that exists at some times and not others or that variously has some non-rational desires and not others. Again, an object’s existing at some times increases the probability that it exists at other times and an object’s lacking some non-rational desires is evidence that it lacks other non-rational desires. It is less clear that the most coherent inanimate explanations of the universe are less coherent than theism, but it is also less clear that they are less



PAUL DRAPER

simple, and so the switch from simplicity to coherence does not weaken the a priori part of Swinburne’s case. The a posteriori part does not fare as well, however. One reason for this is that at least one of the general features of the universe to which that case appeals, namely, temporal order or regularities of causal succession, is itself a sort of uniformity. On his substances–powers– liabilities account of the laws of nature, such uniformity consists in the fact that “all substances [fall] into very few kinds with the same powers and liabilities as each other” (Swinburne 2004: 162). Swinburne claims that the fact that nature exhibits such uniformity must be an enormous coincidence if theism is false, while it is expected given theism (because it is necessary for embodied moral agents like human beings to exist and there is reason to expect God to make such agents). We are supposing, however, that the intrinsic probability of the statement that such uniformity exists depends almost entirely on its (high degree of) coherence instead of on its (low degree of) modesty. On this supposition, such uniformity is not intrinsically very improbable. Further, the denial of theism conjoined with relevant background information (which is very sparse) does not lower its probability. Therefore, our supposition implies that such uniformity is not very improbable given the denial of theism and hence that Swinburne’s criterion of explanatory power is not satisfied. More precisely, if such uniformity is not very improbable given the falsity of theism, then it is mathematically impossible for it to be many times more probable given theism than given its denial and so it cannot raise the odds of theism being true many-fold. (By the “odds of theism being true” I mean the ratio of the probability of theism to the probability of its denial.) In short, it is only weak evidence for theism. Suppose, on the other hand, that simplicity is replaced with coherence, but coherence is not given much greater weight than modesty. Then, even if theism is more coherent than any individual polytheistic or quasitheistic hypothesis, there are so many of these hypotheses that their disjunction must be vastly more modest and thus far more probable intrinsically than theism. Thus, on this second supposition, the a priori component of Swinburne’s case for theism falls apart. And this undermines the a posteriori component of his case as well because non-theistic personal explanations of the universe can no longer be ignored when evaluating the probability of the various data given the falsity of theism, making the task of assessing that probability unmanageable.

SIMPLICITY AND NATURAL THEOLOGY



So whether or not coherence should be given much more weight than modesty, if I am right that the criterion of simplicity derives whatever force it has from the criterion of coherence, then Swinburne’s natural theology is significantly weakened; and if considerations of coherence do not far outweigh considerations of modesty, then it completely collapses.5

References Skyrms, Brian. (1999). Choice and Chance: An Introduction to Inductive Logic, 4th edn. Stamford, Connecticut: Cengage Learning. Swinburne, Richard. (1997). Simplicity as Evidence of Truth. Milwaukee: Marquette University Press. Swinburne, Richard. (2001). Epistemic Justification. Oxford: Clarendon Press. Swinburne, Richard. (2004). The Existence of God, 2nd edn. Oxford: Clarendon Press. Swinburne, Richard. (2010). “God as the Simplest Explanation of the Universe,” European Journal for Philosophy of Religion, 2 (1): 1–24. Swinburne, Richard. (2012). “Bayes, God, and the Multiverse,” in Jake Chandler and Victoria S. Harrison (eds), Probability in the Philosophy of Religion. Oxford: Oxford University Press.

5

I am grateful to Mike Bergmann, Jeff Brower, Nico Kirk-Giannini, Brad Monton, John Schellenberg, and Dana Tulodziecki for very helpful comments on earlier drafts of this chapter.

4 Swinburne’s Aesthetic Appeal Hud Hudson

1 Some Comments on Swinburne’s Fine-Tuning Argument I intend to examine a suggestion I encountered when reading Swinburne’s several contributions to the design argument for the existence of God (Swinburne 1990, 2003, 2004, 2005), an idea he introduces, sketches, approves, but does not himself pursue at length, an idea that I have very much enjoyed exploring and that I hope others will come to enjoy, as well. The suggestion combines two topics of considerable interest: first, aesthetics (in particular, the beautiful and the sublime); and second, fine-tuning arguments. As I will construe the suggestion, it is the thesis that considerations of beauty and of sublimity can strengthen the force of the fine-tuning argument (or, since the definite article is not really appropriate, can improve some existing fine-tuning arguments, or, better still, can form the basis of an altogether new fine-tuning argument). After giving this suggestion a long, hard think, I tentatively conclude (in opposition to Swinburne) that it is not a successful one. In this chapter I come bearing gifts—nearly a dozen of them—designed to make my job of recommending a negative verdict on the thesis more difficult by conceding a series of controversial points along the way which, if true, would surely improve the case for an affirmative verdict. Unfortunately, even after giving away a great deal of ground, I cannot see a way to make good on the tantalizing suggestion in question. I wish it were otherwise, though, and I would certainly welcome attempts to revive the thesis after I bring my critical remarks to a close.

SWINBURNE ’ S AESTHETIC APPEAL



Some introductory matters: On Swinburne’s preferred terminology, a design argument moves from premises about a “general pattern of order in the universe or provision for the needs of conscious beings to a God responsible for these phenomena” (Swinburne 2004: 153); the former (pattern-of-order) arguments are the teleological ones, and the latter (provision-for-needs) arguments are the providential. A fine-tuning argument, then, is a special instance of a teleological, design argument grounded in the specific pattern of order described by a wide variety of facts—facts about our world’s fundamental objects (e.g., their types, numbers, and distribution at some time), its fundamental forces, their relative force strengths, and so on—facts about the actual world that locate us within the remarkably small range of possible worlds whose cosmic conditions are such that if they were actual, it would be physically possible for there to be embodied agents who are sentient, intelligent, and free. Accordingly, we may note that Swinburne departs from other popular treatments of arguments in this neighborhood by thus restricting his use of these central terms, by drawing conclusions explicitly about God (rather than about one or more supernatural designers or, perhaps, a multiverse), and by indexing the fine-tuning to embodiedfree-intelligent-sentience (rather than the more familiar index simply to life). Also, Swinburne’s version of the argument is explicitly Bayesian, and thus he needs to take a stand on the prior probability of theism, a problematic commitment that other theorists (some of whom also traffic in probability judgments) have attempted to sidestep and that still others (who take design to be perceived rather than inferred) can avoid altogether (Manson 2003). Facing this commitment, he places tremendous weight on the relation between metaphysical simplicity and intrinsic probability, and thereby secures himself a substantial head start with a generous prior on theism. Gifts 1 and 2: I will challenge neither the Bayesian approach to the argument nor the thesis about simplicity and intrinsic probability that yields the surprisingly high prior. Gift 3: I will simply accept the suitably Swinburnean version of the fine-tuning thesis, namely, that the range of cosmic conditions that permit embodied-free-intelligent-sentient beings is exceedingly small when compared to the total range of cosmic conditions for which we can reasonably make such determinations. I here borrow and adapt a phrase from the work of Robin Collins, whose presentation of relevant



HUD HUDSON

bits and pieces of the argument strike me as especially clear and accessible (Collins 1999). Whereas I suspect much of the literature on the thesis of fine-tuning (for life) simply carries over to a thesis of fine-tuning (for embodied-sentient beings), I hereby spot (without further comment) all the extra complexities that come with adapting the fine-tuning claim to accommodate the addition of freedom and intelligence, as well. Fine-tuning arguments usually feature a pair of premises that have roughly the following form (here formulated in a way to reflect Swinburne’s particular commitments): Let O = the cosmic conditions permit embodied-free-intelligentsentient life Let G = God exists Let F = contemporary science and the fine-tuning thesis as described in Gift 3 (1) P(O/~G&F) is abysmally low (2) P(O/G&F) is impressively high (Swinburne opts for “moderately high” here) Hereafter, I will be concerned just with premises that take this form in the fine-tuning arguments. One popular and natural sketch of a general defense of the lowprobability premise proceeds in this way: if atheism is true, then the cosmic conditions are left to chance, and if the fine-tuning claim is true, such chances are abysmally low. (Occasionally, the evidence for the finetuning is dramatically displayed, explained, and then some number like 1-in-10-to-the 60th is trotted out. Give or take a bit—that’s small.) But this is all a bit too quick, I fear. Gift 4: I will note (but not endorse) the complaint that there are many ways to go atheist, and they do not all speak with one voice on the probability of such relevantly permissive cosmic conditions, given atheism and the fine-tuning. Some versions of atheism do indeed leave it to chance, but some versions leave us with nondivine and well-intentioned supernatural designers aplenty to divide up the work of setting favorable cosmic conditions. Gift 5: I will note (but not endorse) the worry that if actuality happens to be fragmented into one or another sort of multiverse whose partitions vary independently with respect to their cosmic conditions, the chances of such permissive cosmic conditions (at least initially) seem much rosier.

SWINBURNE ’ S AESTHETIC APPEAL



Gift 6: I will note (but not endorse) the critique that the move from “determined by chance” to “abysmally low likelihood” invokes reasoningwith-ranges (rather than with point-values) and that such reasoning illegitimately presupposes a privileged manner of ordering the infinitely many point-values to be found within each of a finite number of ranges. Gift 7: I will note (but not endorse) the objection that proper reasoning (i.e., with point-values rather than ranges) either runs afoul of the normalizability problem or invokes mysterious infinitesimals. (For more discussion on these criticisms and their rejoinders see Hudson (2006 and 2008), White (2000), and Swinburne’s own strategies for responses in Swinburne 2004.) That is, in what follows, I will not take issue with the low-probability premise, despite its vulnerabilities. Instead, I wish to focus on the moderate-to-high-probability premise—that it is moderately to very likely that the cosmic conditions permit embodied-free-intelligentsentient life, given theism and the fine-tuning. One popular and natural sketch of a general defense of the highprobability premise proceeds in this way. If theism is true, then the cosmic conditions have been fixed by a being who is omnipotent, omniscient, and perfectly good, features that ensure the ability, awareness of the best means, and motivation to select cosmic conditions that will permit the tremendous goods arising from embodied-free-intelligent-sentient life. Thus, it is likely that if God exists, that is what he has done. (To be fair, Swinburne’s own defense is much more nuanced, a point to which we will return. And to be honest, most other defenses of this premise in the literature are often breezy, underdeveloped lines of reasoning not much more informative than this quick gloss—see Manson 2009.)

2 Skeptical Theism A recent and significant literature has emerged under the label of “skeptical theism,” and a lot appears to be at stake. Many a theist is willing to concede both that there exists some evil, which God would have prevented had it not been for the existence of a compensating good or some other morally justifying reason for its permission, but also that we are not aware of any compensating good or other morally justifying reason for the permission of that evil.



HUD HUDSON

Such a theist is then vulnerable to an atheistic strategy which exploits a bridge premise of the following form: “If there were a compensating good or some other morally justifying reason for that evil, then we would be aware of it and we would recognize it as such.” Once this bridge premise is defended, the admission of ignorance can be invoked to infer that there is no compensating good or other morally justifying reason for the evil, after all. Hence, atheism. Skeptical theism provides a satisfactory block to the most promising way of establishing that bridge premise, and this is responsible for much of its appeal, but it is also independently immensely plausible. Skeptical theism has various costumes, but here I will formulate it as a conjunction of three theses (cf. Bergmann 2001 and 2009; Howard-Snyder 2009): We are in the dark about whether the possible goods and possible evils we are aware of are representative of the possible goods and possible evils that there are, and We are in the dark about whether the necessary connections we know of between the obtaining of possible goods and the obtaining of possible evils are representative of the entailments of this type that there are, and We are in the dark about whether the amount of possible good or possible evil we recognize in a state of affairs is representative of the total amount of possible good or possible evil it manifests.

(Note that “representative” is here elliptical for “representative with respect to the property—figuring in a God-justifying reason for permitting evil.”) Consequently, argue our skeptical theists, we have no good reason to think that if there were a compensating good or some other morally justifying reason for the world’s horrific evils, we would be aware of it, or—if it were somehow an object of mere awareness—we would recognize its function as a compensating good or morally justifying reason. This seems to me a clear-headed, honest, and humble recognition of our epistemic state. Perhaps that seems unfortunate for us, but the conjunction just identified with skeptical theism strikes me as overwhelmingly plausible and not really very negotiable. It is some comfort to the theist, I suppose, that this apparent deficiency in our epistemic state effectively undermines one very popular manner of arguing for atheism, but (unfortunately) it comes with some hidden costs for the theist as well. (A fuller discussion of skeptical theism together with a sustained exploration of an additional, unexpected, and truly alarming cost can be found in Hudson 2014.)

SWINBURNE ’ S AESTHETIC APPEAL



3 The Consequences of Skeptical Theism for the Fine-Tuning Argument We surely have a fair amount of knowledge about value. For example, we certainly know that embodied-free-intelligent-sentient beings and certain interpersonal relationships that are possible in a community of such beings are goods, magnificent goods even. Still, moving from that (defensible) claim to the judgment that therefore it is likely that an omnipotent, omniscient, perfectly good being charged with the task of determining values for the cosmic conditions will aim at ones which permit embodied-free-intelligent-sentient beings is just the sort of all-in judgment that the skeptical theist reminds us we are not in an epistemic position to make. It is an inconstant and uneasy skeptical theism which professes ignorance about how much we know about just which things are possible goods, about just how good they in fact are, and about the necessary conditions of their realization in a wide variety of scenarios, but gratuitously grants an exception when it comes to pronouncing on what God is likely to aim at in creating a cosmos. Once again, we are in the dark about whether the possible goods in question are representative of the possible goods that there are and about what far greater goods may of necessity be forfeited or far greater evils may of necessity be guaranteed, if the particular goods of embodied-free-intelligentsentient beings are realized. Those claims are beyond our ken, and whereas this interrupts no-see-um inferences against compensating goods and morally justifying reasons for permitting evils, it also throws up obstacles to establishing the likelihood of permissive cosmic conditions relevantly like ours, given the power, knowledge, and goodness of God together with the fine-tuning. Skeptical theism, then, is a double-edged sword, cutting against popular theistic considerations and popular atheistic considerations alike. Despite my first seven gifts, then, at least one premise in a great many fine-tuning arguments is in jeopardy—at least insofar as its defense depends on such all-in judgments about value barred by the commitments of the skeptical theist. Perhaps, then, it is best to pursue other defenses of the high likelihood that the cosmic conditions permit embodied-freeintelligent-sentient life, given theism and the fine-tuning . . . or, perhaps, it is worth looking at aesthetic properties that need not be considered in terms of the value their instantiations contribute to the world’s value.



HUD HUDSON

4 Swinburne’s Aesthetic Appeal Why do I say—“looking at aesthetic properties that need not be considered in terms of the value their instantiations contribute to the world’s value”? Because, the skeptical theist worry just articulated will be no less problematic even if we (correctly) acknowledge all the additional value contributed by beauty and sublimity to a world suited to embodiedfree-intelligent-sentient life. That is, we could spot ourselves the knowledge that beauty and sublimity are superb goods (or even, if carried away, that they are the best of the created goods), and this would put us in a position to declare that God has a reason (perhaps an extremely strong reason) to ensure that the world of embodied-free-intelligentsentient beings is an aesthetically excellent one, featuring animal and plant beauties in an evolutionary chain of diverse wonders, sublime mountain ranges, vast oceans, rich sunsets, and galaxy clusters. But at what price—all things considered? We are utterly in the dark at how to answer. Given skeptical theism, we have no idea whether God’s admittedly very strong reasons to produce cosmic loveliness are overwhelmingly trumped by stronger reasons to secure an even greater value inconsistent with such loveliness or to prevent some more significant disvalue guaranteed to accompany it. If that line of reasoning captures how aesthetics is supposed to improve fine-tuning arguments, it is a failure. When aesthetics enters the picture only as a means to increasing overall value, as part of a package we would expect will likely be found (and observed) in the world on account of God’s perfect goodness, the battle is already lost. But there is another alternative. As a way of appreciating the alternative, consider what we might call a maximizing consequentialist theory of aesthetic permissibility (modeled on a maximizing consequentialist theory of moral permissibility): Necessarily (an action is aesthetically permissible if and only if and because the consequences of the action are at least as aesthetically good as the consequences would be of any alternative available to its agent). Gift 8: I will assume objectivism about aesthetic properties. (I, myself, cheerfully endorse this thesis, but it will make things harder for me momentarily, rather than easier.)

SWINBURNE ’ S AESTHETIC APPEAL



Note that an agent might be devoted to always doing what is aesthetically obligatory, choosing in favor of aesthetic obligations whenever they conflict with moral obligations or legal obligations or personal desires or whatever. Of course, maximizing may be an excessive view of aesthetic permissions (as it is of moral permissions), and plausibly, especially on occasions of divine action, there may be no alternative available to God which maximizes, if for every aesthetic excellence there is always a better. Perhaps, then (as argued in Swinburne (2004, ch. 6), with respect to God’s production of value in the world) it would be better to hold that there are aesthetically best kinds of action available to God (even in the absence of the possibility of maximizing). Thus, if God were aesthetically perfect in addition to being perfectly good, we would expect God’s agency to be constrained by the performance not only of one of the morally best action kinds, but by one of the aesthetically best action kinds, as well. Accordingly (with perfect goodness as our model), consider the trait of perfection with respect to aesthetics, where aesthetic permissibility is something short of a requirement to maximize: “S is perfectly aesthetically good” =df (i) S never does anything aesthetically wrong, (ii) S possesses an unsurpassable set of aesthetic virtues, (iii) S is never aesthetically blameworthy, and (iv) S is inclined (but not necessitated) to create and sustain aesthetic value and is inclined (but not necessitated) to prevent and eradicate aesthetic disvalue. Gift 9: Let us suppose that in addition to the divine attributes of perfection with respect to power, and perfection with respect to knowledge, and perfection with respect to goodness, God also manifests perfection with respect to aesthetics (thus understood). For the record, I do not yet feel sufficiently moved to accept this last thesis. But the point of introducing it at this juncture is to identify a way that someone could argue from God’s essential nature to reasons for divine action. That is, just as God’s perfect goodness can be invoked (à la Swinburne) as a means to demonstrate God’s good reasons to fine-tune the universe for embodied-free-intelligent-sentient life, so too God’s perfection with respect to aesthetics can be invoked as a means to demonstrate God’s good reasons to fine-tune the universe for beauties and sublimities of a particular sort (without simply thinking that their



HUD HUDSON

contribution to overall value earns them their place in the cosmos, but rather on the grounds that by his nature God aims at aesthetic excellence in its own right). Gift 10: To forestall the new atheistic arguments that are surely brewing at this suggestion of an addition to the divine attribute list, let us also grant (for the sake of argument) a harmony thesis—namely, that aesthetic permissibility (and even aesthetic maximizing, when it is possible) amongst God’s alternatives never requires a failure to satisfy a moral obligation. In other words, let us suppose that whenever God acts, at least one of his morally permissible alternatives is also an aesthetically permissible alternative. So, where do we go from here? On to the formulation of a new (Swinburnean-style) fine-tuning argument, Bayesian in flavor with a generous prior on theism (starting points which will once more go unchallenged in accordance with Gifts 1 and 2). Just as Swinburne identifies particular magnificent goods that we actually observe in the world (e.g., embodied-free-intelligent-sentience) and argues that it is moderately to highly probable that a perfectly good God will finetune the universe with respect to those particular goods, so too we may identify particular magnificent aesthetic excellences that we actually observe in the world in the hopes of arguing that it is moderately to highly probable that a perfectly good God will finetune the universe with respect to those particular aesthetic goods, as well. Which aesthetic excellences? Well, here are some promising candidates noticed by Raphael, Gabriel, and Michael in Goethe’s Faust: Raphael I The Sun intones, in ancient tourney II Its aspect gives the angels power With brother-spheres, a rival song, Though none can ever solve its ways Fulfilling its predestined journey, The lofty works beyond us tower With march of thunder moves along. Sublime as on the first of days

Gabriel III And swift beyond where knowledge IV The sea foams up, widespread ranges, and surging Earth’s splendor whirls in circling flight; Against the rocks’ deep-sunken base, A paradise of brightness changes And rock and sea sweep onward, merging To awful shuddering depths of night. In rushing spheres’ eternal race.

SWINBURNE ’ S AESTHETIC APPEAL



Michael V

And rival tempests roar and shatter, VI There flames destruction, flashing, searing, From sea to land, from land to sea, Before the crashing thunder’s way; And, raging, form a circling fetter Yet, Lord, Thy angels are revering Of deep, effective energy. The gentle progress of Thy day.

The Three VII Its aspect gives the angels power, Since none can solve Thee nor Thy ways; And all Thy works beyond us tower, Sublime as on the first of days. —Translation by George Madison Priest

Gift 11: Accordingly, our new fine-tuning argument will require a rather highly specialized fine-tuning claim, to wit, that the range of cosmic conditions that permit beauty and sublimity of the sort observed in the actual world by our angels is exceedingly small when compared to the total range of cosmic conditions for which we can reasonably make such determinations. I suspect the case to be made for this member of the family of fine-tuning claims is rather less solid than the case to be made for, say, fine-tuning (for sentient life), but it does not strike me as at all implausible—after all, we do not have much hope of a sea that “foams up, widespread and surging” if, say, the nuclear weak force had been somewhat stronger and all hydrogen had burned to helium in the first few moments. Besides, once again, it makes my task more difficult to grant this more specialized fine-tuning claim (and I continue to be as generous to the opposition as I can). Finally, our new fine-tuning argument, like most of its cousins, will require versions of the two probability verdicts noted earlier: Let B = the cosmic conditions permit the beauties and sublimities noted by our angels Let G = God exists Let F = the fine-tuning thesis for these beauties and sublimities as described in Gift 11 (1) P(B/~G&F) is abysmally low (2) P(B/G&F) is impressively high Our new low-probability premise is subject to the same style of defense and to the same four criticisms that were mentioned and then forgiven in



HUD HUDSON

Gifts 3–6. They are hereby forgiven again. (1) is really no better and no worse than its predecessor. (Well, maybe a bit worse. The Swinburnean case for the low-probability premise presupposes that a certain degree of order in a universe is a prerequisite for beauty, yet there are interesting considerations in favor of sublimity-in-chaos and of beauty-despitechaos worthy of further reflection.) The action, then, is once again to be found in the new high-probability premise—that it is very likely that the cosmic conditions permit beauty and sublimity of the sort noted by our angels, given theism and the finetuning (for such beauty and sublimity). One natural sketch of a general defense of the new high-probability premise proceeds as before: if theism is true, then the cosmic conditions have been fixed by a being who is omnipotent, omniscient, and aesthetically perfect, features that ensure the ability, awareness of the best means, and motivation to select cosmic conditions that will permit beauty and sublimity of the sort noted by our angels. Thus, it is likely that if God exists, that is what he has done. And—take note—since we are not invoking any all-in judgments about value, skeptical theism has no battle to wage here.

5 Aesthetic Skepticism But perhaps skeptical theism has a near relative which does deserve a place in this conversation. Consider the analogue of the kind of skeptical theism introduced in Section 2—here named “Aesthetic Skepticism”— and likewise presented as a conjunction of three theses: We are in the dark about whether the possible beauties and sublimities and the possible uglinesses and ridiculousnesses we are aware of are representative of the possible beauties and sublimities and the possible uglinesses and ridiculousnesses that there are, and We are in the dark about whether the necessary connections we know of between the obtaining of possible beauties and sublimities and the obtaining of possible uglinesses and ridiculousnesses are representative of the entailments of this type that there are, and We are in the dark about whether the amount of possible beauty and sublimity or possible ugliness and ridiculousness we recognize in a state of affairs is representative of the total amount of possible beauty and sublimity or possible ugliness and ridiculousness it manifests.

SWINBURNE ’ S AESTHETIC APPEAL



(Note that “representative” is now elliptical for “representative with respect to the property—figuring in an all-things-considered beauty or sublimity enhancing state of affairs.”) As before, this seems to me a clear-headed, honest, and humble recognition of our epistemic state. Perhaps a few further remarks will make you sympathetic with that claim and may entice you to join in my adherence to aesthetic skepticism, thus described. I am tempted to the first element in this thesis by the simple observation that the vast majority of candidate-states-of-affairs for beauty and sublimity are far too complex for me to entertain descriptions of or to manufacture imaginary depictions of them, and thus one reason for thinking that, say, the beauties of which I am aware are not representative of what there is to be had in this category is that not every candidate had an equal chance of showing up in my sample—most of them are permanently inaccessible to the likes of me. Moreover, one can argue that as with goods and evils, there are necessary connections to be recognized here as well. Here is a sketch of one argument on that theme that appeals to me: consider the aesthetic value of both the activity of carving a beautiful statue from a single block of marble and of the resulting artwork (i.e., of both the activity and its product). A perplexing question: Pre artistic activity, does the block already contain actual figures among its proper parts (patiently awaiting release) or only mere potential figures which are to be created by the sculptor’s hand rather than discovered by drawing back the layers of a marble curtain? The metaphysics of decomposition is highly relevant to giving an answer to this question. The Doctrine of Arbitrary Undetached Parts, for example, maintains that whenever there is a material object exactly occupying some region or other, then in every single one of that region’s proper subregions we will find a distinct material object which is one of our original object’s proper parts. Moreover, to call attention to these parts is to call attention to additional pieces of the world’s furniture and is not to be confused with talking about portions of the object, or about conceptual parts, or about potential parts, or about what would exist if the rest of the object were cut away. Once again, then, on the assumption that the block is itself occupying an extended region, any arbitrary subregion of that region hosts an object too. Thus, there is a vast menagerie of marbles sporting the figures of lions and tigers and bears



HUD HUDSON

all occupying subregions of the block prior to sculpting. Further let us suppose (as an extension of Gift 8 on objectivism in aesthetics) that although these entities are not themselves artworks, nevertheless, naturally occurring objects (such as pieces of marble) can exemplify aesthetic properties (such as beauty) even if no one ever apprehends them or appreciates them or even has any de re attitude whatsoever toward them. Well, then, in order to display a certain figure currently secreted away in the interior of the marble block, a sculptor will have to cut into the block and remove pieces of the marble currently obscuring the intended artwork—i.e., whichever figure the artist intends to transform into a statue by freeing it from its surroundings and exhibiting it in the proper sort of way. The cost—and perhaps you will think it only a minor cost— is the willful destruction of many aesthetically spectacular objects which will be broken into pieces in the process (and which will, in their scattered state, lose much of their aesthetic value)—for remember, our block contains billions of duplicates of every statue that will ever be carved (and many more besides), and any one of them may be exhibited only by way of the sacrifice of nearly all of the others. There is, of course, no guarantee that our sculptor will be successful and will present to view (as her accomplishment) one of the many beautiful figures within, but there is a guarantee that she will destroy many beautiful non-artworks in the attempt. The artist who sculpts from a single block of marble, then, always takes a (sometimes lamentable) risk. (This example is lifted from the discussion of the metaphysics of sculpture in Hudson 2012.) The point of all this is that (on the supposition that the Doctrine of Arbitrary Undetached Parts is necessarily true) the securing of a statue carved from a block of marble which is beautiful owing to its pattern or shape is necessarily connected to the sacrifice of fantastically large numbers of real and even more beautifully shaped objects (destroyed in the process of carving) as well as being necessarily connected (by being unavoidable-host) to an astounding range of aesthetic marvels and monstrosities alike (hidden away safely from view in its interior). And even amongst these everyday objects (of whose general whereabouts I am quite certain), I am utterly in the dark about whether the beauty I recognize in the statue before me is representative of the total beauty it (and its proper parts) manifest(s). Similarly, I am utterly in the dark about whether the mereological necessary connections just displayed are representative of the sorts of necessary connections

SWINBURNE ’ S AESTHETIC APPEAL



between beauties and uglinesses that there are. How would I know? . . . The world is a surprisingly interconnected place and necessity is a peculiar business! But—controversial metaphysics aside—we need not aim very high. Even if these reflections on the metaphysics of decomposition did not, in the end, give us any good reason to think that our samples were unrepresentative, aesthetic skepticism is the rather weaker thesis that we are in the dark about whether they are. And, again, that strikes me as a clearheaded, honest, and humble recognition of our epistemic state.

6 The Consequences of Aesthetic Skepticism for the Fine-Tuning Argument So, what is the upshot of endorsing aesthetic skepticism? Unfortunately, I think it effectively disrupts our new fine-tuning argument from beauty and sublimity in precisely the same way skeptical theism disrupts the fine-tuning argument from embodied-free-intelligent-sentient beings. To be clear, I do not contest the claim that we possess a fair amount of knowledge about aesthetic value. We know, for example, that the sorts of beauty and sublimity manifested in the actual world and praised by Goethe’s angels are aesthetic excellences, magnificent excellences even. Still, moving from that (defensible) claim to the judgment that therefore it is likely that an omnipotent, omniscient, aesthetically perfect being charged with the task of determining values for the cosmic conditions will aim at ones which permit the angelically praised beauties and sublimities in question is just the sort of all-in judgment that the aesthetic skeptic reminds us we are not in an epistemic position to make. We are in the dark about whether the possible aesthetic goods in question are representative of the possible aesthetic goods that there are and about what far greater aesthetic goods may of necessity be forfeited or far greater aesthetic evils (if we may use that term) may of necessity be guaranteed, if the particular beauties and sublimities of our world are realized. Those claims are beyond our ken, and consequently we are unable to establish the likelihood that the cosmic conditions will permit these particular delights (rather than others incompatible with them), given the power, knowledge, and aesthetic perfection of God.



HUD HUDSON

7 Changing the Observation and Relocating the Problem Recall that each of the arguments I have been critiquing takes as its central observation a premise asserting a specific state of affairs (e.g., the cosmic conditions permit embodied-free-intelligent-sentient life or a sea that foams up widespread and surging) and then adds a fine-tuning claim together with one high- and one low-probability judgment to a list of premises to yield the desired conclusion. Thus far, I have been arguing that the skeptical theist and aesthetic skepticism considerations will effectively block the defense of each of the high-probability premises in the mix. What if, however, those problematic observations were traded in for something that would avoid the complaint targeting the defenses of that high-probability premise? For example, in the original fine-tuning argument, let O* replace O, where O* = the cosmic conditions permit goods which are among the best contingent goods (all things considered) and in the new fine-tuning argument, let B* replace B, where B* = the cosmic conditions permit beauties and sublimities which are among the best contingent beauties and sublimities (all things considered). Of course, the fine-tuning claims will now require some adjusting to say that the ranges of cosmic conditions that permit the best (or nearly the best) contingent goods and the finest (or nearly the finest) aesthetic excellences are very small when compared to all the possible sets of cosmic conditions there are, but that seems a perfectly reasonable assumption. Well, in that case, my objections to the high-probability premises simply vanish. Clearly, the probability of O* or of B*, given God and the relevant adjusted fine-tuning claim, will be as high as one could hope for, since O* and B* will fall directly out of the conception of being perfect with respect to goodness and perfect with respect to aesthetics (as those traits were developed earlier). Victory is only momentary, however, for although the high-probability premises are no longer vulnerable to skeptical theism or aesthetic skepticism, the price we pay for this advantage is to lose our entitlement to assert the observation premises in the arguments. Observations O and B are there for the

SWINBURNE ’ S AESTHETIC APPEAL



looking, but their proposed replacements, O* and B*, obviously are not, and (in fact) are not for reasons that are easily drawn out of skeptical theism and aesthetic skepticism as before; the argument has not really changed—it is just that the problematic lump has shifted its location under the carpet. Accordingly, we face a dilemma: with O and B we are entitled to the observation but not the high-probability premise, whereas with O* and B* we may defend the high-probability premise but not the observation. But is there not middle ground as of yet unexplored? Are there not observations that (like O or B) are there for the looking, but which yield instances of the high-probability judgment that may be successfully defended without implausible all-in claims about goodness or beauty or sublimity? Yes—there are observations which enjoy both these features. Let O** replace O*, where O** = the cosmic conditions permit some very good contingent states of affairs, indeed and let B** replace B*, where B** = the cosmic conditions permit some very beautiful and quite sublime contingent states of affairs, to be sure. No overreaching all-in judgments there. Granted, it is a bit harder to establish that embodied-free-intelligent-sentient life is a great good (than merely to note that it exists) and harder to show that raging sea storms are quite sublime (than merely to record that they occur), but it was no part of skeptical theism or of aesthetic skepticism to deny that we know that these particular states of affairs do have considerable value and do exhibit aesthetic excellence. So, let us take the observations O** and B** as given. Moreover, as before, the likelihood of the respective highprobability premises may be as high as you please and still defensible, for it is extremely likely that O** and B** will come out true (whether intelligent life or foaming seas or something else altogether turns out to serve as their truth makers) given God’s construal as perfect with respect to goodness and as perfect with respect to aesthetics. Again, though, the victory is short-lived, for the bump under the carpet has simply shifted location once again, for with O** and B** we can have the observation and the high-probability premise, but (and for



HUD HUDSON

the first time) not the adjusted fine-tuning claim indexed to our new observations. Suppose A = the range of all cosmic conditions. Now—what, exactly, is the relation between that range and L1, L2, L3, S1, S2, and S3? Where L1 = the range of cosmic conditions that permit embodied-freeintelligent-sentient life L2 = the range of cosmic conditions that permit some contingent state of affairs or other that is among the best (all things considered) L3 = the range of cosmic conditions that permit something or other of very high value S1 = the range of cosmic conditions that permit the beauties and sublimities noted by our angels S2 = the range of cosmic conditions that permit some contingent state of affairs or other that is among the most aesthetically excellent (all things considered) S3 = the range of cosmic conditions that permit something or other very beautiful or quite sublime. We have seen good reasons to answer “vastly larger” when the question pairs A with L1, L2, S1, or S2—and accordingly, the fine-tuning premises indexed to O, O*, B, and B* were left alone in the previous critical evaluations. But the pairing of A with L3 or S3 is much less clear (and it is these final pairings we would need to examine in order to evaluate the fine-tuning claims appropriate to the arguments featuring our new observations O** and B**). For example, what reason would we have to endorse the fine-tuning thesis featuring, say, ranges A and S3 (that is, to endorse the fine-tuning claim indexed to observation B**)? Recall familiar defenses of claims of this general sort: perhaps, for instance, if the nuclear strong force had been a touch weaker or stronger we would have had a world composed only of hydrogen gas, without enough stability or complexity to serve as a basis for embodied-free-intelligent-sentient life. But why also suspect that such a world would not permit some significant beauty/sublimity or other? Aesthetic skepticism is still on the table, and there is certainly a lot of room for complex shapes and patterns and beauty in such a world free of intelligent creatures. As with beauty/sublimity, so too with value:

SWINBURNE ’ S AESTHETIC APPEAL



skeptical theism should likewise surely make us hesitate at the judgment that we are in a good position to assume that only a small proportion of worlds will contain cosmic conditions permitting some significant value or other among their contingent states of affairs. On that wide-ranging and precarious assumption we are, as we have been saying, surely in the dark. It would appear, then, that we have reached a trilemma: pin your hopes on a particular state of affairs (as do O and B) and the highprobability premise is indefensible; indulge in an all-in claim about value or about aesthetic excellence (as do O* and B*) and the observation is indefensible; compromise with something modest in the middle (as do O** and B**) and the fine-tuning claim is indefensible. In conclusion, and on the strength of the foregoing discussion, it would appear that aesthetic skepticism is to our fine-tuning argument from beauty and sublimity what skeptical theism is to the fine-tuning argument from embodied-free-intelligent-sentient life—a powerful and serious threat to its success.1

References Bergmann, Michael. (2001). “Skeptical Theism and Rowe’s New Evidential Argument from Evil,” Noûs, 35: 278–96. Bergmann, Michael. (2009). “Skeptical Theism and the Problem of Evil,” in Thomas Flint and Michael Rea (eds), The Oxford Handbook of Philosophical Theology. Oxford: Oxford University Press, 374–99. Collins, Robin. (1999). “The Fine-Tuning Design Argument: A Scientific Argument for the Existence of God,” in Michael Murray (ed.), Reason for the Hope Within. Grand Rapids: Eerdmans, 47–75. Howard-Snyder, Daniel. (2009). “Epistemic Humility, Arguments from Evil, and Moral Skepticism,” in Jonathan Kvanvig (ed.), Oxford Studies in Philosophy of Religion, Vol. 2. Oxford: Oxford University Press, 17–57.

1 I would like to express appreciation to Michael Bergmann and Jeffrey Brower for inviting me to present this chapter at the “Faith and Reason: Themes from Swinburne” conference at Purdue University and to thank the audience at that gathering as well as audiences at the University of Notre Dame and Boise State University for their criticism and comments. I also thank Daniel Howard-Snyder, Neil Manson, and especially Roger White (who also served as commentator) for reading the chapter in an earlier draft and for their very helpful suggestions.



HUD HUDSON

Hudson, Hud. (2006). The Metaphysics of Hyperspace. Oxford: Oxford University Press. Hudson, Hud. (2008). “Précis of The Metaphysics of Hyperspace” and “Reply to Parsons, Reply to Heller, and Reply to Rea,” Philosophy and Phenomenological Research, 76: 422–6, 452–70. Hudson, Hud. (2012). “Reflections on the Metaphysics of Sculpture,” in Christy Mag Uidhir (ed.), Art and Abstract Objects. Oxford: Oxford University Press, 223–41. Hudson, Hud. (2014). “The Father of Lies?,” in Jonathan Kvanvig (ed.), Oxford Studies in Philosophy of Religion, Vol. 5. Oxford: Oxford University Press, 147–66. Manson, Neil A. (2003) (ed.). God and Design: The Teleological Argument and Modern Science. London: Routledge. Manson, Neil A. (2009). “The ‘Why Design?’ Question,” in Yujin Nagasawa and Erik J. Wielenberg (eds), New Waves in Philosophy of Religion. New York: Palgrave Macmillan, 68–90. Swinburne, Richard. (1990). “Argument from the Fine-Tuning of the Universe,” in John Leslie (ed.), Physical Cosmology and Philosophy. New York: Macmillan, 154–73. Swinburne, Richard. (2003). “The Argument to God from Fine-Tuning Reassessed,” in Neil A. Manson (ed.), God and Design: The Teleological Argument and Modern Science (London: Routledge), 105–23. Swinburne, Richard. (2004). The Existence of God, 2nd edn. Oxford: Oxford University Press. Swinburne, Richard. (2005). “Prior Probabilities in the Argument from FineTuning,” Faith and Philosophy, 22: 641–53. White, Roger. (2000). “Fine Tuning and Multiple Universes,” Noûs, 34: 260–76.

Divine Power

5 Defining Omnipotence Dean Zimmerman

1 Introduction When I first encountered Richard Swinburne’s The Coherence of Theism (1977), I was still an undergraduate. Although I thought of myself as an English major, I had begun to feel the pull of philosophy. In a couple of philosophy courses I had read Russell, Quine, and Chisholm; and I found Swinburne’s work admirably similar to theirs. The Coherence of Theism was deliberate and systematic, relentless in its pursuit of clarity. But it also took up many of the problems of philosophy of religion that had initially attracted me to philosophy. I doubtless understood little of the opening methodological chapters (how little I did not realize at the time). But when I reached the positive part of his program, which begins with exploration of the concept of omnipotence, I felt sure that I did understand his line of argument; and that what he was saying was not only philosophically ingenious, but also theologically important. My appreciation of Swinburne’s work helped to tilt the scales decisively in favor of philosophy. It is with gratitude (and more than a little nostalgia) that I revisit the pages that captured my attention nearly thirty years ago. I begin with some general remarks about the project of analyzing omnipotence. Then I follow a well-worn path of counterexamples and refinements that leads to one of the definitions Swinburne proposes in The Coherence of Theism. (Henceforth, I use ‘CT ’ to refer to Swinburne (1993), the revised version of the book.) The path is not step-by-step identical to Swinburne’s own, but it is driven by the same problems toward roughly the same stopping point. The definition fails due to counterexamples like Plantinga’s notorious McEar. Recently, however,

DEFINING OMNIPOTENCE



Swinburne has suggested a fix in terms of the relative sizes of the sets of states of affairs that could be brought about by different candidates for omnipotence. The idea is a promising one, though it requires some fleshing out. The devil is in the details; and, in the end, the resulting account of omnipotence is not without its problems. An alternative response to McEar is then explored. Finally, a rather different sort of counterexample is raised, and a further revision is proposed to deal with it.

2 The Analysis of Omnipotence What is the point of definitions or analyses of omnipotence? Two quite different descriptions of the project can be found in the literature. One aims for a kind of theological neutrality, and I will call it “Secular Omnipotence”: Project 1: Secular omnipotence There is a pre-theoretical notion of omnipotence, not essentially embedded in a theological tradition or beholden to religious desiderata, that can be made precise by sufficient philosophical reflection. It will then be an open question whether a deity, as described by this or that religion, has the property in question.1 If the result of an analysis that purports to be of this first kind is something extremely complicated, and the complexities could conceivably be resolved in more than one way, it will become doubtful whether there is anything precise and theologically neutral to be discovered in the conceptual vicinity of the word ‘omnipotence.’ Other attempts to define a notion of omnipotence wear their theological commitments on their sleeves. Some are explicitly aimed at the definition of the kind of power appropriate to a certain conception of divinity. Project 2: Divine power There is a degree of power ascribed to a Divine being by a particular religious tradition. The goal of analysis is to provide a philosophically illuminating account of this degree of power, showing it to be 1

Examples of philosophers self-consciously engaged in Project 1 include: Flint and Freddoso (1987) and Sobel (2004).



DEAN ZIMMERMAN

compatible with the other doctrines (about God and other matters) included within the religious tradition in question.2 Those engaged in this second project will likely reach different conceptions of Divine power if they are committed to different doctrines about God’s knowledge, goodness, freedom, eternity, and other attributes. Different accounts of the freedom of creatures are almost sure to make a difference as well. Arminians and Calvinists will disagree about whether God can simply decide to make a person do anything God wants her to do, while leaving her morally responsible for the action. Of course Project 2 will unravel if there is an incoherence in the doctrines that are being presupposed. It would also fail if there were no such thing as the maximal amount of power consistent with the other attributes ascribed to the deity. Here is a third project, one that most philosophical theologians will suppose must coincide with Project 2. Project 3: The power of the greatest possible being There is such a thing as a greatest possible being, and the goal of the analysis of omnipotence is to define the degree of power appropriate to such a being. The third project will fail if there is no such thing as the greatest possible being; or if there is, but this being has different degrees of power in different possible worlds. Project 3 is at least conceptually distinguishable from Project 2. Some religious traditions may not be so committed to greatest-possible-being theology as others. And the concept of a greatest possible being has a life outside any particular religious tradition, so one might pursue the question of the degree of power appropriate to such a being without adverting to premises from “revealed religion.” Philosophical theologians sometimes agree that, in principle, the two projects should reach the same end, while disagreeing about how useful it is to engage in Project 3 directly.3 2 In CT, Swinburne arrives at a notion of divine omnipotence ([D]) that answers almost entirely to non-theological desiderata; but he judges that it does not square with God’s impeccability and essential rationality, which require a shift to Project 2 (see CT: 165). As noted here, it appears to me that [D] is already compatible with these divine attributes; so, if I am right, Swinburne’s proposal could remain primarily secular in orientation after all. 3 For a general defense of the methodology of greatest-possible-being theology, see Leftow (2012: 9–12); for skepticism about the method, see Speaks (2014).

DEFINING OMNIPOTENCE



The first project, if it proves viable, could well come apart from the second and third. A being that satisfied a purely secular conception of omnipotence might be an amoral deity, able to do things that a certain kind of Christian God could not. Similarly, there may be trade-offs among great-making attributes, so that the degree of power of the greatest possible being would prove to be different from that of a being that satisfies a definition of secular omnipotence.4 I am happy to accept a greatest-possible-being constraint on my own theological speculations; so when pursuing Project 2, I believe I am pursuing Project 3 as well. But I take it to be an open and interesting question whether there are degrees of power that are comparable and that differ between the Christian God or a greatest possible being, on the one hand, and a being satisfying a secular conception of omnipotence, on the other. Some contemporary philosophical theologians believe that Project 1 can lead to a notion of omnipotence that the Christian God could satisfy—e.g., Flint and Freddoso (1987: 158). Other philosophers complete Project 1 and claim that God, as traditionally conceived, could not be omnipotent, in their sense of the term—e.g., Sobel (2004: 366–7). In what follows, the compatibility of omnipotence with impeccability will sometimes serve as a datum; insofar as one supports an aspect of the definition by such an appeal, one is engaged in Project 2. As shall appear, a bare appeal to libertarianism could also be used at those same points to support the same moves, and someone willing to rely on the assumption of libertarianism alone could fairly claim to be working within the strictures of Project 1.

3 The Range of Power Familiar descriptions of God’s power—“nothing is impossible with God,” God can “do all things,” etc.—suggest a definition of omnipotence 4

A slightly different project is important to Swinburne’s larger program within the philosophy of religion. Swinburne assumes that there is such a thing as “the simplest kind of person that there could be” (2004: 334), and that there is a certain degree of power that the simplest kind of person would have, one that can be worked out a priori by thinking about the concept of the simplest kind of person. Swinburne claims that the pursuit of this fourth project leads to the same property as the pursuit of Project 2, when 2 is undertaken with explicitly Christian constraints. The coincidence of Projects 2 and 4 becomes important for the success of Swinburne’s inductive argument for the existence of God.



DEAN ZIMMERMAN

in terms of the range of things that an agent is able to do or to bring about. Swinburne’s analysis is an attempt to say something relatively precise about the extent of this range. It stands in a long tradition of such attempts, stretching back through Ockham and Scotus to Aquinas.5 These descriptions suggest that to be omnipotent is to have the power to do the broadest possible range of things; it is to come as close as a being can come to having the ability to do anything. If there is such a thing as “the closest one can come,” and if the range of things that must be doable in order to come that close can be described without explicit appeal to theological considerations, then Project 1 is viable. The starting point for those engaged in Project 2 is slightly different: to be omnipotent is to come as close as a thing can come to being able to do anything, given that it also has X, Y, Z . . . where ‘X,’ ‘Y,’ ‘Z,’ etc. stand in for the Divine attributes held fixed by theological commitments. The goal of Project 3 is the delineation of the range of things that could be done by something, given that it is the greatest possible being. Suppose the search for a non-theologically driven conception of omnipotence fails—there is no such thing as the closest one could come to being able to do anything, full stop. The failure of Project 1 need not undermine the significance of the other two. For the typical religious believer, what is important is that God is worthy of worship, that God created everything else, that God has power over every other being. What is important to the perfect-being theologian is the delineation of the kind of power appropriate to a greatest possible being. A plausible conclusion of Project 2 or 3 will satisfy these inquirers, even if Project 1 meets shipwreck. Suppose, on the other hand, that Project 1 were to succeed in producing some natural interpretation of “as powerful as a thing can be” which applied neither to God nor to the greatest possible being. Philosophers interested in the existence of God or a greatest possible being need not be overly concerned—so long as they were not somehow antecedently committed to there being a purely secular conception of omnipotence that must apply to God, or to the greatest possible being.

5 Brian Leftow regards the reflections on omnipotence in Augustine and Anselm as developments of an account of omnipotence in terms of strength rather than range—the latter notion only taking center stage with Aquinas and his successors. See Leftow (2009: 168–70).

DEFINING OMNIPOTENCE



The candidate definitions of omnipotence still standing at the end of this chapter strike me as fairly secular in inspiration, and yet also plausibly applicable to a God with the other traditional Divine attributes—so one may yet be able to complete Project 1 and make theological use of the result. A number of philosophers are skeptical about whether omnipotence can be spelled out in terms of the range of a being’s power. Famously, Peter Geach denies that such an analysis can succeed, preferring instead to characterize God’s power as “almightiness, or power over all things”; and he spells this out in terms of two conditions: (i) everything other than God has whatever power it has only in virtue of God’s will; and (ii) it is impossible for God to try to do something but fail (Geach 1977: 4–5). Of course, as a Catholic Christian, Geach believes a great many specific things about what God can and cannot bring about. Since many of these particulars do not follow from his definition of ‘almightiness’ (and he does not pretend to be able to derive them from it), his is the counsel of despair. If Geach’s concrete objections to proposals can be got round, there is no reason to agree with him—and his objections are far from being the final word on the question, as I hope to show.6

4 Act Theories vs. Result Theories One major division among attempted range analyses of omnipotence is between those that try to describe the range of actions that an omnipotent being should be able to do, and those that try to describe the range of states of affairs that an omnipotent being should be able to bring about. James Ross calls a theory of the first sort an “analysis in terms of doing”; a theory of the second sort is an “analysis in terms of choice” (1969: 202–21). Flint and Freddoso contrast analyses of omnipotence that refer to possible “tasks,” as opposed to analyses “stated in terms of an agent’s power to actualize or bring about states of affairs” (1987: 137). Kenneth Pearce (2011) introduces the labels “act theories” and “result 6

Wielenberg (2000) also objects to range theories as such. Leftow (2009) argues, more modestly, that a range analysis cannot be the full story: omnipotence is not merely a matter of the range of things one can bring about, but also a matter of how hard or easy it is to bring them about—and this contrast cannot, he argues, be captured entirely in terms of range.



DEAN ZIMMERMAN

theories” to draw the same distinction, and I shall use his terminology. All these authors prefer the second approach, developing result theories. Indeed, use of the result-theoretic terminology “bringing about a state of affairs,” and stipulations about the range of states of affairs appropriate to omnipotence, figure in most recent attempts to define the notion— including Rosenkrantz and Hoffmann (1980), Wierenga (1989), Leftow (2009), and Swinburne himself (CT).7 A result theory of omnipotence will, then, take something like the following form: (D*)

x is an omnipotent being = df for any state of affairs S (with certain exceptions), x can intentionally bring about S.

Result theorists refine this style of definition, largely by attempts to identify the class of exceptions. In this section, I focus on the main piece of ideology in the definition: the sort of “bringing about” that is relevant for defining omnipotence within a result theory. I am not entirely convinced by the reasons the defenders of result theories give for preferring them to act theories.8 And 7 I note that, in addition to the delineation of a range of states of affairs omnipotence must be able to bring about, Leftow’s definition requires that “all truthmakers of modal truths” depend upon God. Shortly, I shall follow Swinburne in denying that God brings about necessary states of affairs. But Leftow provides significant support for the idea that God must in some sense make necessary truths true. See Leftow (2009: 187–90), and, for the full story, Leftow (2012). 8 Ross (1969) has three sorts of objections to act theories. The first is that, in order to have the power to do every act, a being would have to have the power to perform (in his example) both Socrates’ sitting and Jezebel’s fornication with the heathen—and to do both of those, a being would have to be identical with both Socrates and Jezebel, an impossibility given their actual non-identity (202). Someone friendly to an act analysis should insist that ‘x’ (in Ross’s proposed definition) ranges over possible act types that could be performed by different persons—a not terribly ad hoc maneuver. Ross’s second objection to an act theory is that, upon refinement, it turns out to be susceptible to counterexamples; and Pearce and Swinburne raise similar difficulties. But Ross’s counterexamples are identical or closely analogous to ones that also arise for the most popular versions of result theories (including Swinburne’s), such as those discussed here using the example of “McEar” (Ross 1969: 206–10). Similarly, Pearce and Swinburne raise difficulties having to do with acts that require the past to have been a certain way (Swinburne CT: 155–6), the paradox of the stone, and McEar (Pearce 2011). All of these create very similar problems in the context of a result theory of the sort these philosophers favor; I suspect that the most promising proposals for fixing the latter could be translated into the terms of an act theory. Ross’s third objection assumes divine immutability, and turns upon the doubtful thesis that an act (but not the choosing or bringing about of a state of affairs) entails change in its subject (219–21). Flint and Freddoso object to act theories by pointing out that Smith’s inability to

DEFINING OMNIPOTENCE



one might worry whether all the powers of an agent can be understood in terms of the bringing about of states of affairs.9 In the present chapter, however, I shall simply follow Swinburne and the other result theorists in assuming that omnipotence is to be defined in terms of the bringing about of states of affairs—either because all action can properly be analyzed in terms of this notion, or because it is not the range of actions one can perform that is relevant to defining omnipotence, but rather the range of states of affairs one can bring about. The ability to bring about states of affairs that is required by omnipotence is not simply the ability to cause them; it is also an intentional bringing about, the kind of bringing about that constitutes an action. In ordinary contexts, one can be said to have the power to bring about all sorts of unforeseen consequences of one’s actions; for example: if, unbeknownst to me, raising my arm would signal an assassin to kill the president, someone might well say that I have the ability to bring about the president’s assassination. However, when the goal is to describe a kind of power that can purposefully be exercised, the range of one’s ability to bring things about should be restricted to what one can intentionally bring about. Some philosophers are skeptical whether intentional action always involves such things as intentions—at least, if intentions are understood

perform the task of “saying something which is (at the same time) being said only by Jones” should not count against Smith’s claim to omnipotence (1987: 138). Shifting to a result theory would indeed seem to prevent this sort of problem from arising. 9

As Robert Adams has pointed out to me, God’s power should certainly extend to mental acts, such as the decision to speak to someone 5 minutes from now—an act that does not, when it occurs, seem to result in the bringing about of any state of affairs distinct from the decision itself. The philosopher who would construe every act as a case of an agent’s bringing about a state of affairs could reply that the act of deciding is divisible into a bringing about and a state of mental preparedness—a set of dispositions including a readiness to speak if one believes that 5 minutes has elapsed, for example. But the action theorist must be careful. If every action must be divisible in this way, an infinite regress looms. It cannot be that each act of an agent bringing about a state of affairs S requires division into a further state of affairs S* and an additional intentional act of bringing about that is aimed at S*, on pain of a psychologically implausible regress of infinitely many intentional actions. Chisholm is willing to accept that, in causing something, an agent causally contributes to her causally contributing to . . . her causally contributing to the bringing about of something, for infinitely many iterations (1979: 370–2). But he develops a theory that distinguishes between different kinds of “bringing about” and does not lead to an infinite regress of intentional actions (see Chisholm 1976: 66–76).



DEAN ZIMMERMAN

as a special kind of mental state.10 Swinburne is solidly on the side of the believers in intentions as sui generis mental acts, often called ‘volitions.’ He equates having an intention and choosing (CT: 146); so when he talks of an intention, it is not a standing intention, but an “intention-inaction.”11 The discussion of intentions in CT is basically unchanged from that of the first edition (1977), so the chapter on the same subject in The Evolution of the Soul (1986) represents Swinburne’s later (and fuller) thoughts on the matter. Here, it becomes clear that Swinburne’s theory of action is of the same general type as those of Chisholm (1976), Ginet (1990), McCann (1998), and the many other defenders of volitions. According to these philosophers, there is a special kind of mental state essential to action, one that is not reducible to some combination of belief and desire or other “passive” mental states—i.e., states that could occur when a person is not intentionally doing anything. Ginet and McCann call them ‘volitions,’ while Chisholm speaks of ‘undertakings.’ Swinburne sometimes calls them ‘volitions,’ but prefers the term ‘purposings’—a substitution that is meant merely to avoid the implication that “every action is a quick decision subsequently efficacious” (1986: 94). With that qualification understood, I shall continue to use ‘volition’ for the special kind of mental state Swinburne posits. On all these theories, the contents of volitions are propositional. When a person raises her arm, she purposes or wills that her arm go up; or, according to Ginet and McCann, something more like that her arm exerts force in an upward direction. These propositional phrases represent the contents of the volitions. When the proposition that her arm goes up is true, one may also say that her arm going up is a state of affairs that occurs (or that “obtains,” to use Chisholm’s term). On such volition theories, then, the most basic kind of intentional bringing about, involved in all action, is very naturally described as a relation between a person and a proposition or state of affairs. The person wills or purposes that such-and-such proposition be true, or that such-andsuch state of affairs obtain. 10 Debunking “the Myth of Volition” was an important part of Ryle’s campaign against the ghost in the machine (Ryle 1949: 62–82). But the repudiation of volitions is not confined to behaviorists; one finds it frequently in Wittgensteinians (see, e.g., Melden 1961: 43–55) and occasionally even in the more metaphysically inclined action theorists (Taylor 1966: 57–85). 11 See the parallel discussion of intention in Swinburne (1993: 85, n. 1).

DEFINING OMNIPOTENCE



Swinburne prefers to talk of states of affairs rather than propositions. He does not, to my knowledge, develop a detailed metaphysics of states of affairs that could be plugged into the analysis of omnipotence in CT. As on Plantinga’s or Chisholm’s account of this ontological category, there seems to be a state of affairs for every proposition, whether false or true.12 What is less clear is whether Swinburne would follow Chisholm and Plantinga in accepting that some states of affairs obtain at one time and not at another time—i.e., whether they are repeatable.13 Swinburne eventually offers an analysis that makes use of the idea of a state of affairs being brought about at a time. If states of affairs obtain eternally, when they obtain at all, then the notion of bringing one about at a time is a little hard to understand. The locution is easily interpretable in terms of repeatable states of affairs, however, which obtain relative to some times and not to others. To bring about an event at a time is, on this interpretation, to cause a repeatable state of affairs to be true of that time. (In a manuscript for a second edition of The Coherence of Theism, Swinburne replaces the language of states of affairs with talk of the bringing about of events at times—an idea that can also be interpreted in terms of causing repeatable states of affairs to obtain at times.) Following Swinburne, I shall define a kind of omnipotence that is relative to a time; being omnipotent at one time but not at another is at least a conceptual possibility; and being omnipotent at a time requires having as much control as it is possible to have over what happens from then on. Since power is being understood in terms of the ability to bring about repeatable states of affairs, the power, at t, to ensure that a door closes 5 minutes later must consist in the ability to bring about a state of affairs along these lines: that something or other happen for 5 minutes (it matters not what) after which a door closes. This is an event that could happen over and over. I shall assume that “gappy” states of affairs like this—ones that happen entirely in virtue of what will happen, after a certain period of time—are available for any length of time one chooses.

12

Within Chisholm’s framework, propositions and states of affairs are not even to be distinguished; see Chisholm (1976: 114–17). On Plantinga’s theory, they are at least very similar, perhaps in a one-to-one correlation; see Plantinga (2003: 107–8). 13 Chisholm talks of states of affairs that can “recur”; Plantinga refers to “transient states of affairs.” Cf. Chisholm (1976: 128) and Plantinga (2003: 189).



DEAN ZIMMERMAN

A volition theorist could understand the notion of “bringing about a certain state of affairs” more or less broadly. Result theories have generally defined omnipotence by generalizing over the states of affairs that a being can bring about in a very broad sense—those that an agent can bring about directly, as part of what are usually called “basic actions,” or indirectly, by relying upon the consequences (causal and otherwise) of basic actions within the larger contexts in which they are performed. For now, while tracing the path that leads to the kind of definition Swinburne articulates in CT, I shall understand “bringing about” in the same broad way. However, in the end, a puzzle for such definitions will suggest that omnipotence should rather be defined in terms of what one can bring about most directly.

5 The Beginnings of a Range-Result Theory Many philosophers have offered definitions of omnipotence in terms of the range of states of affairs that a being has the power to bring about. As shall appear, such definitions encounter numerous puzzles and problems. Must the power of a truly omnipotent being extend to creation of round squares, unjust punishment of the innocent, causation of others’ free actions, suicide, or production of a stone too heavy for its creator to lift? Infamously, Descartes defended an extreme account of the range of things open to omnipotence. According to Descartes, omnipotence consists in the power to bring about any state of affairs whatsoever— including the existence of round squares and triangles with four sides, two-plus-two’s being five, etc. Few philosophically inclined theists have felt the need to attribute so much power to God as all that, and Swinburne is no exception. “It is no objection to A’s omnipotence that he cannot make a square circle. This is because ‘making a square circle’ does not describe anything which it is coherent to suppose could be done.”14 A first pass, then, at defining the range appropriate to omnipotence requires only the ability to bring about, by decree, possible states of affairs, ones that could be the case. The resulting definition of omnipotence would look something like this: 14

(CT: 153–4).

DEFINING OMNIPOTENCE

(D1)



x is an omnipotent being =df for any possible state of affairs S, x can intentionally bring about S.15

One obvious problem for such an account is that there are seemingly possible states of affairs that no one could possibly bring about, such as a certain billiard ball’s moving in a way that is not caused by anybody. Omnipotence should not require the ability to cause such things. Necessary states of affairs—for example, everything’s being male if it is a bachelor, or everything’s having three sides if it is a triangle—also pose a problem for this account. What is necessarily the case can hardly fail to be possibly the case. (“Is it possible for a triangle to have three sides?” “Possible! Why, I’ve seen it done! So of course it’s possible.”) Necessary states of affairs hardly seem to need bringing about (although many theists have felt that a full account of omnipotence must show that the necessary states of affairs are somehow grounded in God, so that even necessary truths are true because of God’s existence16). Perhaps, then, the focus should be upon just those states of affairs that are possible but not necessary—in other words, those that are contingent; they could obtain, but do not have to obtain. In that case, the range of things an omnipotent being must be able to bring about should be further restricted to just contingent states of affairs (CT: 156). The following modifications would take care of these problems: (D2)

x is an omnipotent being =df for any contingent state of affairs S that could possibly be intentionally brought about by someone, x can intentionally bring about S.

6 Excluding Power to Change the Past Explicating the concept of omnipotence is made more complicated by the controversies about God’s relationship to time. On the “Open Theist” picture of things, God has always and will always exist, acquiring new

15

(D1) corresponds more or less to Swinburne’s [B] (CT: 154)—except that in [B] Swinburne has not yet replaced talk about “logically possible actions” with talk of “bringing about states of affairs”; this modification occurs with his next proposal, [C] (155), which is roughly equivalent to (D4). 16 See, e.g., Leftow (2009: 187–90).



DEAN ZIMMERMAN

knowledge as time passes; but more traditional conceptions of God insist that God knows everything all at once—either because God exists in time and has always known what would happen in the future, or because God exists in some kind of “timeless eternity” and knows everything “all at once” from a perspective outside of time. If God is timelessly eternal, it would be wrong to imagine that God first did not believe that Moses was extending his hand toward the Red Sea, then came to believe this (when Moses stretched forth his hand), whereupon God decreed that the Red Sea should part. Rather, God believes, timelessly, that Moses is not stretching forth his hand at certain times, and that Moses is doing so at others; and God decrees, timelessly, that the Red Sea should be parted at some times and not at others. Similarly, on the complete foreknowledge conception of God’s relation to time, what will happen at each time, including what God will bring about at that time, is settled from the beginning; so God’s actions are chosen all at once and from eternity. Whoever is right about this controversy, it would seem to make sense to talk about the things that God can, at some point in time, still bring about; and the things that are over and done with, and for that reason no longer within anyone’s power at that time. Advocates of timeless eternity will think that God makes decisions about what to bring about at a particular time from some atemporal perspective; others, that God does so from a temporal perspective prior to creating anything else, long before the times at which we observe the effects of God’s decrees; others, that God makes choices within time, right when the effects of God’s activity come about. In any case, once we have reached a given point in history, we know, from our perspective, that it is now too late for certain things to be the results of God’s decisions, or God’s single grand decision—no matter the perspective from which God makes decisions about how to interact with creation. A familiar example from the literature on omnipotence is telling a joke for the first time. When a joke has already been told at noon, telling it for the first time should no longer be something omnipotence requires. One might try to get round the problem by restricting the relevant range of states of affairs to just the arbitrarily repeatable ones; but there would still be problematic examples, such as bringing it about that a certain joke is told 5 minutes after someone else has told it. If no one has told the joke recently, even the ablest comedian cannot bring this about—at least not here and now.

DEFINING OMNIPOTENCE



To take account of this sort of case, Swinburne (and many others, going back to the Middle Ages) have further refined the notion of omnipotence along these lines: (D3) x is an omnipotent being = df for any contingent state of affairs S that could possibly be intentionally brought about by someone, and for any time t, if it is possible, given the actual history of the world up to and including t, that S obtain, then x can intentionally bring about S after t. The history of the world, prior to t, must be understood as the so-called “hard past” relative to t. Suppose it is a fact that I saw Bob Dylan for the very last time in 2011 in Asbury Park. This would not (yet) be part of the “hard” history of the world, since (happily) my seeing Dylan again remains an open possibility. The fact that I saw Dylan in 2011 is hard, but that it was the last time I would ever see him is at most a “soft fact,” one that is not yet inevitable.17 But if a joke has already been told today, that is a hard fact about the past, and bringing it about that it is told for the first time now would not be required of omnipotence by (D3). Likewise, if the joke was not told 5 minutes ago, then omnipotence does not require the ability, right now, to tell it 5 minutes after its having already been told. (D3) requires that, if a being is omnipotent, it is omnipotent at every time. One of the virtues of Swinburne’s discussion of omnipotence is his allowance for the possibility of someone’s being omnipotent at some times and not at others—a conception of omnipotence that, as he points out, provides one way out of the puzzle about whether an omnipotent being could make a stone too heavy for anyone to lift. (D3) can be relativized to a time, generating a definition very close to Swinburne’s [C] (CT: 155): (D4)

x is an omnipotent being at t =df for any contingent state of affairs S that could possibly be intentionally brought about by someone, and for any time t, if it is possible, given the actual history of the world up to and including t, that S obtain, then x can intentionally bring about S after t.

In a forthcoming revision of CT, Swinburne develops an alternative way to set aside power to bring about something in the past. His new strategy

17

For more on this idea, see Freddoso (1989) and other papers in Fischer (1989).



DEAN ZIMMERMAN

deals with past-involving events—like a joke’s being told for the first time, or a promise’s being kept—by restricting the range of states of affairs that an omnipotent being must be capable of bringing about at t to just those that are hard relative to their times of occurrence. The concept of hardness is explained along these lines: a state of affairs is hard if and only if the truth-conditions for the proposition that it occurs during a period lie entirely within the period of its occurrence.18 This strategy brings to mind the appeal to “immediate states of affairs” in Flint and Freddoso’s definition of the hard past: [W]hat is temporally independent—or, to use Chisholm’s phrase, rooted in the present—at any given time can be specified in terms of the present-tense (or, as we prefer to say, immediate) states of affairs which obtain at that time. All nonimmediate, or temporally dependent, states of affairs that obtain at time t obtain at t only in virtue of the fact that the appropriate immediate states of affairs did or will obtain at moments other than t. (147)

Swinburne’s new strategy has the same basic goal and uses roughly the same tools as his earlier definition, [C]. The strategy in [C]—as in my (D3) and (D4)—was to say: restrict the states of affairs that, at t, must be accessible to an omnipotent being to just ones that are compatible with the occurrence of every hard state of affairs that happens at t and earlier. The newer strategy, pursued in Swinburne’s forthcoming book, says: restrict the states of affairs that, at t, must be accessible to an omnipotent being to just the ones rooted entirely in periods of time subsequent to t, and therefore compatible with any kind of hard past. A natural way to capture the basic idea behind Swinburne’s most recent proposal is to impose the following restriction upon the range of states of affairs that must be within the power of a being at a time, if it is to be omnipotent at that time: consider only repeatable states of affairs that are rooted in the times at which they obtain. In any case, the two strategies are obviously similar. Given the right assumptions, they may turn out to be equivalent.19

18

See Flint and Freddoso (1987) for a similar definition of a hard state of affairs, which they then utilize in constructing the notion of the hard past. 19 A different strategy for circumventing the same sorts of counterexamples is found in Rosenkrantz and Hoffmann (1980), in which the range of states of affairs relevant to defining omnipotence is restricted to states of affairs that are arbitrarily repeatable.

DEFINING OMNIPOTENCE



7 A Clutch of Counterexamples So far, perhaps, so good. However, as all students of the literature know, several problems loom for this sort of definition of omnipotence— counterexamples that led Geach to despair of any analysis along these lines. Bringing about states of affairs that could only be caused by a morally imperfect being—for example, lying and promise-breaking—should not be within the power of an essentially morally perfect God; and yet they would be within the range of things that could be done by a less perfect being, who would therefore be able to do things God could not. Do we want to say that only an amoral being could be truly omnipotent? If we are only pursuing Project 1, that might well be an acceptable conclusion—it would simply follow that the kind of power it is appropriate to ascribe to God is not secular omnipotence. Were these the only kinds of objections to (D4) it could stand as a suitable end point of Project 1; it need only be modified by those pursuing the other, religiously or normatively motivated projects. In CT, Swinburne’s final modification of the definition of omnipotence allows for an impeccable omnipotent being by not requiring that such a being be able to bring about states of affairs that it would be irrational to choose (CT: 165–6). But some counterexamples challenge (D4)’s adequacy even as an analysis of secular omnipotence (and would remain a challenge to a definition modified to allow for impeccability along the lines Swinburne suggests). My freely raising my arm is a contingent state of affairs that I can bring about; indeed, it feels like a basic action, one that can be undertaken without my trying to do anything else in order to raise my arm. But it is doubtful whether anyone, even God, can decide what someone else should freely do and then be able to make it the case that they do it. I can simply choose that I freely do something, but no one else can. My freely raising my arm is a contingent state of affairs that someone can bring about, but God is the wrong person for the job! Molinists, like Flint and Freddoso, can remain libertarians while denying this; since God knows “conditionals of creaturely freedom,” describing what all the possible creatures would (freely) do in all the circumstances in which they could find themselves, God can “weakly actualize” (to use Plantinga’s term) both those circumstances and a creature’s being free in them, thereby effectively choosing precisely



DEAN ZIMMERMAN

what will be done, freely. Counting the “weakly actualizing” of a state of affairs as a kind of bringing about, Flint and Freddoso sidestep the problem—at the cost of admitting hosts of very important contingent states of affairs that are outside of God’s power: namely, all states of affairs corresponding to the conditionals of creaturely freedom. I shall assume, with Swinburne, that Molinism is false. Other puzzles have nothing to do with our freedom, or God’s essential goodness. If God exists necessarily, or at least is essentially eternal, then God could not successfully issue the decree: “Let the giver of this very decree cease to exist.”

8 An Attractive Solution A standard attempt to modify the account of omnipotence in order to avoid all these difficulties is to say that a being’s omnipotence only requires that it be able to bring about contingent states of affairs that can possibly be brought about by that very being. This allows for a smaller range of things that God must be able to do in order to satisfy the modified definition of omnipotence; “for there are many feats that we can consistently suppose to be performed but cannot consistently suppose to be performed by God” (Geach 1977: 17–18). Causing me to freely sit down, for example, can be excluded on these grounds. It is something someone can bring about easily enough—namely, myself; but, according to libertarians, no one else can do this. If God is essentially impeccable, then God’s doing something immoral, such as lying or punishing the innocent, is not required for omnipotence. And if God is a necessarily existing being, then committing suicide is likewise excluded. Geach himself rejects attempts to define omnipotence along these lines; but it is a popular proposal, and worth examination. Indeed, it is the modification which Swinburne advocates, in his most recent published discussion of the matter (Swinburne 2009). Modifying (D4) in the manner proposed yields: (D5) x is an omnipotent being at t =df for any contingent state of affairs S that could possibly be intentionally brought about by someone, and for any time t, if it is possible, given the actual history of the world up to and including t, that S obtain, and possible that x intentionally bring about S, then x can intentionally bring about S after t.

DEFINING OMNIPOTENCE



The modification eliminates problematic cases like God’s making me freely do something, God’s committing suicide, and God’s doing evil. Is it possible that God brings it about that I freely do something? Not if the libertarians are right about freedom. Is it possible for God to commit suicide? Not if God is a necessary being. Is it possible for God to perform an immoral action? Not if God is essentially impeccable. A glaring problem with (D5) as it stands is the fact that any being that could not possibly be an agent counts as omnipotent; if there is no S that x can intentionally bring about, then x satisfies (D5) trivially. So I had better add that x is capable of intentional action: (D6)

x is an omnipotent being at t = df it is possible for x to intentionally bring about something; and for any contingent state of affairs S that could possibly be intentionally brought about by someone, and for any time t, if it is possible, given the actual history of the world up to and including t, that S obtain, and possible that x intentionally bring about S, then x can intentionally bring about S after t.

With (D6), I have reached something very close to Swinburne’s definition [D] in CT.20 [D]

a person P is omnipotent at a time t if and only if he is able to bring about any logically contingent state of affairs after t, the description of which does not entail that P did not bring it about at t (CT: 156).

Swinburne intends that the term “logically contingent state of affairs after t” be restricted so as to refer only to states of affairs “logically compatible with all that has happened at and before t” (CT: 155); so, although it might not appear to, [D] does in fact incorporate the qualification about time introduced in (D3) and (D4). By his use of the phrase “the description of which,” Swinburne cannot mean just any old true description of the state of affairs in question. Suppose that “the placing of a bet on the horse that Zimmerman does not, in fact, select” correctly describes the state of affairs, Zimmerman’s placing a bet on Beetlebaum. Suppose the race has yet to start, I have money to bet, I am at the racetrack, and so on. The correctness of that description (“a bet

20

It is also pretty close to Wierenga’s (O*) (1989: 25).



DEAN ZIMMERMAN

Zimmerman does not select”) should not be an obstacle to my bringing about this state of affairs. What Swinburne clearly has in mind here is something like a “canonical description” of a state of affairs, one that fully captures its logical structure. The restriction in (D6) of the relevant states of affairs to just ones that it is possible that x bring about finds its functional equivalent in Swinburne’s restriction to states of affairs “the description of which does not entail that P did not bring [them] about.” Are these two restrictions really equivalent? That depends upon whether, for every agent x and state of affairs S, the canonical description of S entails that x does not bring about S if and only if it is not possible for x to bring about S. To have the same consequences as (D6), [D] must be understood in such a way that the canonical description of Dean freely places a bet entails that God did not cause this state of affairs; and God ceases to exist and God performs an immoral action must each entail a contradiction. I shall assume Swinburne’s restriction and mine do, in fact, coincide. [D] is not Swinburne’s final definition of omnipotence in CT. Though he defends [D] against counterexamples (CT: 156–63), he finally rejects it in favor of a definition with an important further qualification: roughly, that an omnipotent being not be required to be able to bring about states of affairs that are believed to be irrational or wrong (CT: 165). This is meant to resolve counterexamples involving states of affairs that a morally perfect being could not bring about; and it places his definition squarely in the tradition of Project 2: the statement of the maximal amount of power that could be had by a being with the other properties ascribed to the deity by a certain theological tradition.21 I shall set aside this modification, and consider problems that arise with or without it. However, assuming that rationality and (hence, by Swinburne’s lights) impeccability are essential to God, it is not at all clear that [D] requires that God be able to bring about overall bad states of affairs—ones that would be wrong for anyone to bring about. The canonical description of the state of affairs that God brings about something that would be wrong for anyone to bring about entails a contradiction if the meaning of ‘God’ includes impeccability; so that would allow 21 More recently, Swinburne has rejected this qualification, and gone back to [D], but [D] modified to take account of the problem of McEar (see Swinburne 2009). The resulting definition will be discussed later.

DEFINING OMNIPOTENCE



the impeccable God to satisfy Swinburne’s [D] and my (D6). The philosophical theologian is still squarely within the secular confines of Project 1 if the modification resulting in these definitions is motivated entirely by the desire to respect libertarianism. If part of the motivation for the modification is to render omnipotence and impeccability compatible, or to allow for a necessarily existing omnipotent being, then, to that extent, the philosophical theologian has drifted into Project 2.

9 The Problem of McEar Definitions in the style of (D6) and [D], promising as they may seem, face a well-known problem. As Swinburne notes, [D] “may however need to be tightened up, in order to take account of the problem of McEar” (2009: 496), a hypothetical creature who was dreamed up by Plantinga (though it has medieval precursors) for the purpose of raising a difficulty for such definitions of omnipotence.22 McEar is a creature limited in its powers as a matter of necessity: there is nothing this sort of being could possibly do other than to scratch its ear (and, as I shall interpret the proposed counterexample, also to perform other bodily motions by means of which it scratches its ear23). So long as McEar can scratch, it counts as omnipotent by (D6)—since no other states of affairs could possibly be brought about by McEar, none of these other states of affairs are relevant to determining whether McEar satisfies (D6). And yet it is a pathetically weak being; so there must be something wrong with the definition. I am assuming that [D]’s use of entailment to identify and set aside impossible states of affairs has the same effect as (D6)’s use of impossibility—so that, if McEar satisfies (D6), it satisfies [D] as well. However silly the example might seem, there is a valid point behind it; for there is nothing silly about the idea that the powers of a particular kind of thing might be limited in certain ways, and limited as a matter of necessity. Electrons have the power to attract protons and to repel other electrons. Would it be possible for a particular electron suddenly to gain 22 McEar makes his first appearance in Plantinga (1967: 168–73), though not under that name (for McEar’s baptism, see LaCroix 1977: 187). Mackie (1962) raised the problem earlier, but more abstractly. Flint and Freddoso (1987: 138, n. 4) find the same kind of counterexample discussed by William of Ockham. 23 Pavel Davydov has pointed out to me that, to be a person, McEar arguably has to have the power to perform mental actions, as well.



DEAN ZIMMERMAN

the power to attract other electrons and repel protons? It is tempting to say: No, it is of the essence of an electron not to do such things. Something similar may well be true with respect to finite agents. Even if the wacky McEar is not a possible creature, there may nevertheless be essential limits to the powers of some beings—and, if so, why not to those of intelligent beings, like us?24 In that case, an essentially weak creature might always be at the height of its puny powers—able to do the full repertoire of things that such a being could possibly do—and thereby qualify as omnipotent (on (D6) or [D]) even though the range of things it can do is very limited. Swinburne proposes to get round McEar “by a clause (loosely) to the effect that P is omnipotent only if he can bring about more logically contingent states than can any other logically possible person” (2009: 496). It is true that McEar “can bring about all that is logically possible for him to bring about, viz. that his left ear is scratched. But other things can bring that about, and much else besides and so they have a greater claim to omnipotence than does McEar.”25 This suggests a definition in two stages, reminiscent of a proposal floated and rejected by Ross: first define what it is for a thing to be as powerful as it can be (Ross talks about a thing’s being able to do everything that is possible for the kind of thing that it is); and then define an omnipotent being as one that is as powerful it can be and . . . here Ross thinks the proposal will break down, because what would have to be added is something like “is a perfect being” or “is a divine being,” which would result in an uninformative characterization (1969: 210). Swinburne’s idea, on the other hand, is to fill in the ellipsis with something about the much greater range of things that the truly omnipotent being can bring about—a strategy that one might even be able to implement while remaining true to the spirit of Project 1, invoking only non-theological criteria for judging the extent of a being’s power. I shall reconstruct Swinburne’s response to McEar in two stages. The first stage is to define what it is for a thing to be as powerful as it can be, 24 For an objection, see Wierenga (1989: 29). For rebuttals, see Wielenberg (2000: 26–47, especially 31), and Pearce and Pruss (2012: 406). 25 Swinburne (2009: 498, fn. 4). Although the point of their article is not primarily to define omnipotence (and they do not claim that their proposal meets all counterexamples), Baillie and Hagen add a similar clause to get round the problem of McEar; see their (2008: 24).

DEFINING OMNIPOTENCE



which I shall do by simplification of (D6) (removing the condition that x is the sort of thing that can intentionally cause something to happen): (D7) x is as powerful as x can be at t =df for any contingent state of affairs S that could possibly be intentionally brought about by someone, and for any time t, if it is possible, given the actual history of the world up to and including t, that S obtain, and possible that x intentionally bring about S, then x could intentionally bring about S after t.26 The second stage makes the move Swinburne suggests, stipulating that the omnipotent being is the one with the power to bring about the greatest range of states of affairs: (D8) x is omnipotent at t =df it is possible for x to intentionally bring about something; x is as powerful as x can be at t; and, for any possible y, if y were to exist at t and be as powerful as y can be at t, the range of things x could intentionally bring about after t is greater than the range of things that y could intentionally bring about after t. The proposal has considerable appeal. However, someone defending it as a definition of a kind of omnipotence applicable to God must answer the question: What is it for one range of states of affairs to be greater than another? The answer is far from obvious when infinitely many states are involved, and when a pair of beings with different degrees of power may have ranges that overlap without either being a proper subset of the other— and those are the circumstances in which (D8) must be implemented. First, I show that someone whose goal is to find a definition of omnipotence that could be ascribed to God, and who shares enough of Swinburne’s assumptions about God, must allow that the set of states of affairs within the scope of an omnipotent being’s powers and the set within the scope of some other possible (patently non-omnipotent) being’s powers may overlap, without either being a subset of the other. Both Divine impeccability and libertarianism about creaturely freedom lead to this result. Then I discuss the sizes of the ranges of states of affairs that must be compared in order to make sense of “greater” in (D8). (D7) is close to Sobel’s notion of an ONSLIP, a being of “only necessarily self-limited power” (2004: 364). 26



DEAN ZIMMERMAN

Assuming God’s impeccability, one must admit that there are bad states of affairs God cannot bring about, but that some limited evil agents could; therefore, these evil agents have some powers God does not. If every evil agent is more limited than God, such an agent will also lack some powers that God has. And, so long as the evil agent can do other benign things that God could also bring about directly, God and the evil agent will share some powers as well. Thus, a philosophical theologian engaged in Project 2, and assuming impeccability, must contrast God’s power with that of a possible agent able to bring about a range of states of affairs that properly overlaps the range God is able to bring about (i.e., they overlap, and neither is a subset of the other); and the latter set of states of affairs must be judged greater than the former, if (D8) is to provide a definition that applies to God. But one need not appeal to explicit theological desiderata in order to reach this conclusion; libertarian assumptions about freedom are enough. If one person’s free choices cannot be brought about by another, even I can bring about states of affairs—e.g., my freely choosing to sit— that God cannot. Again, the range of some of our powers overlap—both God and I can bring it about that I become seated—while others (including all the really impressive ones) are had by God alone. The result is the same: neither of the two ranges of states of affairs—those God can bring about and those I can bring about—properly includes the other. Swinburne accepts both of these assumptions, and so cannot give a certain natural interpretation of what is meant in (D8) by the talk of a “greater range” of states of affairs. He cannot say that one being can bring about a greater range of states of affairs than another if and only if the former can bring about everything the other can, and more besides. What he needs, instead, is a sense of ‘greater’ that allows for comparisons of properly overlapping sets of states of affairs. Furthermore, it cannot be a function of the number of states of affairs involved because, once god-like beings are in the picture, the numbers that need to be compared will be infinite. A general, god-like power to move things, or to create things, will include infinitely many distinct states of affairs that can be brought about by an agent with the power. There are, after all, infinitely many distinct speeds that things could be given and infinitely many distinct possible creatures that could be created. Knowledge of these different speeds and creatures would be required of any decent candidate for an omnipotent being. Without

DEFINING OMNIPOTENCE



finely discriminative cognitive abilities, even a very powerful being would not be able to intentionally bring about the precise states of affairs omnipotence would seem to require. How large should one take the infinities in question to be? The usual mathematical representations of three-dimensional spatial regions— whether Newton’s Euclidean space, or the three-dimensional space-like slices of a relativistic manifold—make use of non-denumerably infinite sets of points, equal in size to the set of real numbers. An omnipotent being ought to be able to select the precise locations of particles in these spaces, or the precise values of fields at points. Such powers will require non-denumerable infinities of options from which to choose, infinities at least as large as the real numbers. (Perhaps even higher infinities of distinct states of affairs must fall within the range of omnipotence, but I should be surprised if the difference between omnipotence and lesser degrees of power were to turn out to be a matter of the cardinality of the ranges of states of affairs being compared.) So long as the set of states of affairs within an omnipotent being’s power is infinitely large, it will have proper subsets of equal size that omit hosts of the states of affairs the omnipotent being can bring about. Presumably a being of such power could make another being with powers limited to the bringing about of states of affairs in at least some of those equally large subsets; but the latter sort of being will clearly be less powerful, despite the fact that the two sets of states of affairs are the same size.27 27 The definition of omnipotence here seems to require quantification over infinities of abstracta, and it will have to dodge puzzles that bedevil attempts to define omniscience in ways that quantify over all propositions. The “Grim result,” that there is no set of all propositions, is taken by Grim to represent an objection to the coherence of any doctrine of omniscience—a point Grim develops in his contribution to (Plantinga and Grim 1993), and elsewhere. Beall (2000) argues that, by denying the existence of more than one necessary truth, one can resist Grim’s proof that there is no set of all propositions. The familiar problem with Beall’s suggestion is that one of the theoretical roles propositions are supposed to play is that of explaining the differences among beliefs and other attitudes toward things that can be true or false; and two people can have beliefs that are necessarily true without believing the same thing—or so it would seem. A metaphysics of states of affairs, however, could more plausibly include the stipulation that there is just one necessary state of affairs. For states of affairs may be conceived of as worldly entities, playing roles that include being causal relata and the background conditions for causation. If states of affairs are thought of as all the actual and merely possible events, and not the objects of propositional attitudes, there is little reason to posit multiple necessary states of affairs. It may, then, be easier to define omnipotence by quantification over all states of affairs than to define omniscience by quantification over all propositions. For a general discussion of the



DEAN ZIMMERMAN

Let us suppose that the infinite cardinality of the range of states of affairs within the power of any potentially omnipotent being is that of the real numbers—though all that matters, for the present argument, is that it be an infinite number. In the interesting cases, then (i.e., cases involving possible beings who are plausible candidates for omnipotence), the differences between greater and lesser ranges cannot be differences in the number of states of affairs that can be brought about. To continue to make sense of greater and lesser, Swinburne must appeal to a measure over these sets of states of affairs, one that allows precise comparison of equally numerous sets, neither of which is a subset of the other. There are bound to be functions from sets of states of affairs to numbers that satisfy the technical, purely mathematical criterion for being a measure—at least, so long as the states of affairs in question are not too numerous to form a set. But what Swinburne needs is a natural way of comparing their sizes—a non-arbitrary measure, something based upon the relative importance of being able to bring about this or that state of affairs. Is it reasonable to hope for objective facts about such matters? Though the problems may be surmountable, there is reason to worry. Compare a being that can create electrons wherever it will, on the one hand, with a being that can create protons wherever it will, on the other; and who otherwise can do little but think: the electron god and the proton god, as it were. Each, if capable of discriminating among the infinitely many perfectly precise ways of adding electrons or protons to the universe, has a non-denumerable infinity of states of affairs it can bring about. Is one field of action larger or smaller than the other? Are they equally large? The most natural thing to say would seem to be: the beings are equally powerful, and the states of affairs in question must therefore be equally large. Here, ‘equally large’ cannot mean equal in number, since pairs of sets like these can have the same number of states of affairs but not be equally large according to the relevant measure; ‘equally large’ must mean “assigned the same size by the measure function appropriate for comparison of powers.” In correspondence, Daniel Rubio has raised interesting puzzles that must be addressed if one is to assign sizes to sets of states of affairs in a many ways Cantorian problems can arise for theories of propositions, and the limited number of ways out of these problems, see Pruss and Rasmussen (2015).

DEFINING OMNIPOTENCE



way that corresponds to greatness of powers. Since the sets of states of affairs include non-denumerably infinite ones, many subsets must be assigned measure zero. Assuming that greater powers correspond to abilities to bring about a wider range of precise states of affairs, the easiest way to arrive at a zero-measure set would be to focus upon extremely precise states of affairs. (Sets of finitely many zero-measure states of affairs will naturally also have measure zero.) Thus, we could end up with the quixotic result that a being (like the infamous McEar) who can only do one thing will be ranked as stronger than a being who can do infinitely many things (just draw the latter unfortunate being’s powers from the measure zero subsets). That’s bad. What’s worse: it will be pretty easy to generate beings who can do strictly more things than other beings (‘strictly’ meaning: I can do everything you can, plus some); and yet they must be deemed equally powerful. All you need to do is to select two beings who have exactly matching powers in the portion of the set with size greater than zero, and then give one of them a bunch of powers from the measure zero portion. (Daniel Rubio, personal correspondence)

These puzzles do raise serious questions about the viability of the whole enterprise. But they are not unlike puzzles elsewhere that arise when measures are defined over sets of infinite cardinality—for example, in assigning volumes to sets of points in a Euclidean space, or to the objects occupying such sets of points. Drawing comparisons between the two applications of measure theory is instructive. In Rubio’s first example, McEar can be compared to a microscopic solid, while the other being is like a disconnected cloud of infinitely many point-sized particles spread throughout the universe. The microscopic solid is almost nowhere, while the cloud may be ubiquitous, with point-sized representatives in every single region of space above a certain size. Still, the measure of the space the cloud occupies is zero; so it is smaller, in volume, than the pathetically small solid. In Rubio’s second example, the two beings could be compared to cubes of the same volume, occupying exactly the same points, save for the fact that one but not the other lacks point-sized parts at the corners. In this case, a pair of things has the same size, but, intuitively, one is ‘larger’ in virtue of having ‘extra’ parts—albeit parts of measure zero. A plausible response to Rubio’s objection begins with the following denial: it is not the case that, for any arbitrary set of states of affairs one chooses, there is a genuinely possible creature with powers limited,



DEAN ZIMMERMAN

essentially, to just the bringing about of the states of affairs in that set. This is, after all, why McEar strikes everyone as so sketchy; no agent could have just the power to scratch without also having all sorts of other powers—e.g., to move one’s finger or head, or at least to wiggle one’s ear. Powers come in packages, and a reasonable constraint on using the space of states of affairs to reflect the greatness of powers is to insist that no pair of sets differing only in states of affairs of measure zero correspond to the maximal powers of two possible beings. The analogy with volume can be extended to illuminate this suggestion, and its motivation. Suppose that analytic geometry proves very useful in describing the behavior of some solids, but that one doubts whether the solids in question really contain any less than threedimensional parts. One thinks, instead, that they are like the subject matter of Tarski’s “Foundations of the Geometry of Solids” (1956): each three-dimensional solid contains further three-dimensional proper parts, ad infinitum, and no parts with zero volume. When using analytic geometry to describe the sizes and locations of such objects, one should abstract away from the zero-volume differences between regions occupied by the solids. Richard Cartwright (1975) provides a very natural way to do so: treat a set of points in the space as genuinely occupiable if and only if it is identical to the interior of its closure. This procedure throws out such oddities as disconnected clouds of points, or cubes with just their corners removed. If the only space-occupiers are Tarskian solids, a pair of regions with only measure-zero differences should not represent two distinct possible locations of such solids. The responses to Rubio’s objections are analogous. To Rubio’s first objection, Swinburne could reply that randomly selected powers to bring about very different but extremely precise states of affairs do not go together to constitute the complete powers of a possible being. To Rubio’s second objection, Swinburne could claim that one cannot take the complete powers of a possible being, eliminate the power to bring about some very specific state of affairs, and thereby arrive at the complete powers of a different kind of possible being. The imposition of this constraint on the correspondence between states of affairs and genuinely possible “power packages” strikes me as well motivated. Still, the defender of judgments of relative size in the domain of states of affairs faces some difficult questions. As noted earlier, the definition of omnipotence must allow for comparison in cases of

DEFINING OMNIPOTENCE



proper overlap. Compare, for example, one being who can create only electrons and neutrons, and another who can create only protons and neutrons; and suppose that, together with suitable rational powers, these are complete power packages for two limited but god-like beings. The axioms of measure theory—which must be respected in order to make sense of larger and smaller sets in this context—impose some requirements on judgments of the sizes of the sets of states of affairs in the range of their powers. For the two gods to be equally great, the sizes of the two largest sets of states of affairs each can bring about will have to be equal; and that would require that the area of non-overlap be equal as well—i.e., that the set of electron-introduction and proton-introduction states of affairs be equally large. On the other hand, for the electron-and-neutron god to be stronger than the proton-and-neutron god, the size of the set of electron-introduction states of affairs would have to be greater than the size of the set of proton-introduction states of affairs—since their neutron-introduction and rational powers are the same. Comparisons of rather similar powers can generate reasonably strong intuitions about comparative importance. Plausibly, the two gods just described are equally powerful. But comparisons of radically different kinds of power raise worries about whether there is an objective, unique way of assigning sizes to all sets of states of affairs that will correspond to their relevance to a being’s power. Here is a stock example, taken from a different domain, of the kind of worry that should arise: It is natural to have doubts about whether one more statue by Michelangelo or one more symphony by Mozart could have altered the relative greatness of their contributions to art; and natural to doubt, more generally, whether the goodness and badness of states of affairs is always commensurable.28 A similar worry is bound to arise here. A god who can make any kind of material thing is greater than our proton god, electron god, and neutron god; its power subsumes all of theirs. Likewise, a god who can make any kind of immaterial thinking being is greater than one who can only make the least impressive sort of angel. Obviously, a god who can make every kind of material thing and every kind of immaterial thinking thing is greater than any of the gods discussed in this section. Although these comparisons seem unproblematic, they require that there be answers to 28

Chang (2002) uses this example in a seminal discussion of the failure of parity with respect to relative goodness and badness.



DEAN ZIMMERMAN

questions about the comparative powers of all the lesser gods as well. And there, our intuitions are likely to give out. Is the one-angel god exactly as strong as the electron-god? If not, precisely how much stronger or weaker is one than the other? Are the any-material-thing and anyimmaterial-thinking-thing gods equal in power? If the any-materialthing god is stronger, is the any-immaterial-thinking-thing god also weaker than the proton-and-neutron god? Stronger? Or are they equally powerful? The imposition of a measure on the space of states of affairs requires that there be answers to these sorts of questions. The situation does not strike me as utterly hopeless, however. One could always hold out for the existence of precise comparisons of the greatness of powers, while acknowledging that we lack the ability to reliably make them in many cases. Or one could accept that there is a plurality of incompatible but equally acceptable ways of judging the relative greatness of powers, while insisting that there is exactly one set of states of affairs that is both (i) the range of the powers of a possible being, and (ii) the largest set according to every acceptable way of measuring the sizes of the sets of states of affairs. Assuming all that, Swinburne’s proposed modification looks promising.

10 The Strongest of the Strongest Swinburne’s way of dealing with McEar depends upon some heavy-duty assumptions about the amenability of powers to measurement in terms of the states of affairs at which they are directed. The questionability of these assumptions makes it worth casting about for a simpler way to get round McEar. As an alternative, I propose identifying the omnipotent being with one that is as powerful as it can be, and that is also able to thwart the will of any other such being. (D9) x is omnipotent at t =df it is possible for x to intentionally bring about something; x is as powerful as x can be at t; and, for any possible y, z, . . . , if y, z, . . . , can intentionally bring things about, and if they were to exist at t and be as powerful as they could be at t, the range of things x could intentionally bring about after t would include preventing y, z, . . . , from intentionally bringing anything about after t. I rather like (D9). It seems to me to be a plausible end-point of Project 1, the search for a conception of omnipotence that is not patently

DEFINING OMNIPOTENCE



theologically motivated; but it also appears to be a kind of power that the Christian, say, could attribute to God, or that the perfect-being theologian could attribute to a greatest possible being. Questions remain about how to interpret it exactly. Much depends upon the essential features of times and other possible agents. If every possible agent could have existed at every possible time, then (D9) is easy to read. But if this modal assumption is false, then one of the conditions in the definition, namely, “if they [i.e., y, z, . . . ] were to exist at t and be as powerful as they could be at t,” would sometimes be an impossibility, and the conditional of which it is a part would be counterpossible. If counterpossibles are all true (in virtue of their necessarily false antecedents), the definition would face counterexamples arising from the possibility of weak creatures existing at times when nothing stronger could possibly exist. I shall assume this problem does not shipwreck the definition—either times are modally forgiving and possible agents are modally flexible, so that any of the latter could exist at any of the former, or a principled distinction can be made between true and false counterpossibles that will save the definition. (D9) is not, however, completely immune to McEar-like problems. Suppose there could be such a creature as Negative Nelly: someone essentially weak, necessarily lacking all but the power to thwart the wills of others by preventing them from being able to exercise their powers. When Negative Nelly is at full strength, she is able to thwart the will of any possible being, and then she counts as omnipotent by (D9). But a definition that is satisfied by a character with such purely negative powers is not adequate to any of the omnipotence projects. Nelly seems to me to be less threatening to (D9) than McEar was to (D6) and to Swinburne’s [D]. Granted, McEar is weird and arguably impossible. But that did not matter, since McEar simply stood in for any possible being that is very limited in its powers but that happens to be operating at its essential maximum. Denying that there could be such a creature is much harder than denying the possibility of McEar; it would entail that any creature with powers, however strong, could be made stronger still—a doubtful proposition.29 So McEar-like considerations seem to me quite telling as objections to (D6) and [D]. Nelly, on the

29

See Wierenga’s defense of this idea (1989: 29), and Wielenberg’s criticism (2000: 31).



DEAN ZIMMERMAN

other hand, must have a very specific and bizarre set of powers: She must have the ability to “freeze” any possible being or group of possible beings, however strong they might otherwise be. But she must also lack sufficiently many other abilities so as to be obviously weak, too weak to conceivably qualify as omnipotent. It is easier to reject this supposed counterexample, declaring it to be impossible. The question whether there could be two omnipotent beings is a venerable one, and there are things to be said on both sides.30 Thinking about this question reveals why it is unlikely that Nelly is genuinely possible. The case for the possibility of two co-existing, essentially omnipotent beings depends crucially upon their wills being modally linked—so that neither could choose to bring about something that the other could choose to prevent. Perhaps two or more essentially perfectly good and perfectly knowledgeable beings could be omnipotent, but the possible co-existence of two or more beings who could will to bring about incompatible things, while remaining omnipotent . . . that should be impossible on any decent definition of omnipotence. And with Negative Nelly, there is little to be said in favor of the idea that more than one possible being could have Negative Nelly’s powers. If there were a possible being, Nelly-two, with the same powers as Negative Nelly, then it could both prevent Nelly from bringing anything about and be prevented by Negative Nelly from exercising this power. For the same reason, there could be no possible being, Nelly-plus, who has Negative Nelly’s powers plus some powers that Nelly is supposed to lack. Granting the possibility of Negative Nelly is, therefore, far from innocuous. Hidden within the admission of her possibility is the denial of the possibility of all sorts of other entities with her thwarting powers plus other abilities. If the realm of possible beings has at most one being with Negative Nelly’s thwarting powers, how plausible is it to suppose that this one lucky being would be otherwise so unlucky—so incredibly, and essentially, weak? Staring down the claim that she is genuinely possible seems much easier than staring down the claim that there could be weak creatures operating at the essential maximum of their powers— and the latter, more difficult feat is what it means to deny the possibility of McEar.

30

For a recent argument against the possibility, see Baillie and Hagen (2008).

DEFINING OMNIPOTENCE



11 A Further Problem and a Major Revision (D8) and (D9) are slightly different notions of maximal power that take Swinburne’s definition [D] as a starting point, and modify it so as to defeat McEar. However, once free actions on the part of creatures are exempted from the control of an omnipotent being, a further difficulty emerges. Suppose I will never, as a matter of fact, freely lie to my spouse—though I will have plenty of opportunities to do so. In counterfactual circumstances in which I do lie to her, God could punish me for (freely) lying to her, or cause me to feel remorse for (freely) lying to her. Since, ex hypothesi, I will not (freely) lie to her, there will never be an opportunity for God to punish me, or to instill remorse in me, for doing so. And since God cannot ensure that I freely lie, God cannot just intentionally bring it about that the opportunity arises. Still, there is nothing about the world up until now, or the nature of the states of affairs involved in my punishment or feeling remorse for lying, that is incompatible with their obtaining in the future; and there is nothing impossible about God’s causing them to obtain. So they seem to be counterexamples to (D8) and to (D9).31 I can see a way around this difficulty, but it requires a more extreme alteration in the approach to defining omnipotence, one that makes use of somewhat controversial assumptions about the nature of action. Happily, they are assumptions Swinburne is quite prepared to make. Since Danto (1963) introduced the terms, it has become a commonplace within action theory to distinguish between “basic acts” and “non-basic Leftow’s definition of omnipotence introduces a notion of “all-things-considered” ability, which “varies over time, and is extrinsic”; one can be “intrinsically such as to bring it about that P” even at times when it is (to take his example) too late to bring it about that P (2009: 186–7). If successful more generally, Leftow’s strategy would take care of my counterexample as well. His view seems to me to involve a greater departure from the range analysis than might at first appear. Leftow’s crucial terminology (“x is intrinsically such as to bring it about that P”) requires elucidation. The phrase cannot mean that, whenever someone is intrinsically such as to bring it about that P, P is going to be happening, barring external interference. An agent need not be constantly trying to bring about P in order for P to be within the range of things she can do. So Leftow’s expression must, rather, mean that the agent has what it takes, intrinsically, to be able to bring about P; and I suppose that “having what it takes, intrinsically, to be able to bring about P”, means something like: “she has enough strength to bring it about that P, even if she cannot bring it about that P for some other reason—i.e., some reason that has nothing to do with lack of strength.” If that is the right way to understand Leftow’s locution, his view is very close to Wielenberg’s pure strength theory (Wielenberg 2000). 31



DEAN ZIMMERMAN

acts.” Very roughly put, the distinction is between things one does by doing something else, and things one does but not by means of doing anything else in order to do them. The precise location of the divide, and how best to mark it, are a matter of contention; but the notion of doing one thing by doing another is firmly embedded in our conception of action, and any theory of action worth its salt must make sense of the idea. Once an account has been given of what it is to do X by doing Y, a theory of the basic–non-basic distinction becomes almost inevitable.32 The distinction is crucial for the modification of the definition of omnipotence I shall propose. The proper way to formulate the basic–non-basic distinction depends upon whether actions are individuated in a coarse-grained or finegrained manner. Whatever their views about the individuation of actions, volition theorists should ultimately recognize a notion of basic action that will serve my purposes. Here, I locate Swinburne’s own views in action theory within the debate about whether actions are coarsegrained or fine-grained, and then I make use of his distinctions in a new definition of omnipotence intended to avoid the counterexamples mentioned earlier in this section. However, I believe my strategy could be implemented within other versions of the volition theory, ones that give different verdicts on the individuation of actions. Some theories of action are fine-grained: raising my arm is one act, and signaling an auction bid (by raising my arm) is a different act; my moving the pool cue is one act, and my striking the cue ball is another. Most metaphysicians who have a place in their ontology for states of affairs take them to be finely individuated, much like propositions. Suppose they are; then a very fine-grained account of actions would result if one also adopted the thesis that, whenever S1 and S2 are different states of affairs, an act of intentionally bringing about S1 is not identical with an act of intentionally bringing about S2—even if performed by the same person at the same time, and even if the person brings about S2 by bringing about S1. Since the states of affairs that my hand goes up and that I place a bid in the auction are distinct, this sort of fine-grained theory treats my bringing each of them about as a separate action—even 32 For examples of the way in which the distinction falls naturally out of different theories of action, see Chisholm (1976: 76–88), Ginet (1990: 11–22), Goldman (1970: 63–72), and McCann (1998: 61–93).

DEFINING OMNIPOTENCE



when I do the second by doing the first. Likewise for moving a cue and striking a cue ball. Coarse-grained theories of action posit one act where a fine-grained approach sees two or more. A typical coarse-grainer identifies the act of raising my hand with the act of bidding, and the act of moving the cue with the act of striking the cue ball. Swinburne opts for a course-grained theory that is similar in many ways to Donald Davidson’s.33 In the case of killing a man by shooting him, Swinburne says, we can think of the agent as simultaneously doing many different actions (pulling the trigger, shooting, killing) or as doing just one action which is described in many different ways. Although we can think in either way, it is, I believe, more natural to think in the latter way and I shall do so (1986: 85–6).

Following Davidson, he prefers to think of the “essence of the action” in a case like this as just “the agent moving his body in a certain way, intending thereby to bring about certain further consequences” (1986: 86). All the intended consequences beyond the movement are not really parts of the action, though they can contribute to its extrinsic description; “the effects of the action (intended or not), other than its basic result, lie outside the action” (89). Swinburne takes the basic results of typical human actions to be things like one’s arm going up, one’s vocal cords producing the sound of the word ‘king,’ or one’s shoelaces becoming tied—states of affairs which, Swinburne thinks, one intends to bring about but not by means of intentionally bringing about something else: “I do not intentionally move certain parts of my body in order to tie my shoe-laces or to say the word ‘king’; for I would not know which parts to move to secure these ends” (1986: 88). These states of affairs are the simplest contents of our volitions, the things we will to bring about in order thereby to bring other things about. According to Swinburne, an action always includes a “basic result”— unless it is a complete failure, in which case it is a mere volition, unaccompanied by any intended results.34 Volition theorists disagree 33

Swinburne (1986: 86, n. 3) explicitly compares his theory to Davidson’s (1980: 57–61). Here Swinburne differs from Davidson, who has no place in his theory of action for a sui generis category of volitions, present whenever one acts. (I expect Davidson would accuse Swinburne of holding “the bad old doctrine that all we ever do is will things to happen” (Davidson 1980: 59).) 34



DEAN ZIMMERMAN

about which results are genuinely basic; but any volition theory that explains how agents manage to do one thing by doing something else (or, on Swinburne’s preferred way of putting it, how agents do something that falls under one description by doing something that falls under a more basic description) will need the series to end somewhere. At that point, one reaches a volition to bring about a state of affairs that the agent believes he or she can bring about merely by willing. Some volitiontheorists, like Ginet (1990: 31–44), take volitions to be a kind of mental state that can only take as its content a state of affairs that can be brought about in a basic way. Swinburne’s view is more complex. His volitions (or “purposings”) seem to be directed, in the first instance, at “ultimate results,” states of affairs corresponding to something like the “highest level” of description under which the agent conceives of her action (1986: 87); but the agent’s mental states must also include attitudes specifying all the “lower level” descriptions under which the agent’s action falls— including the intention to bring about the most basic result (95). I shall somewhat simplify and assume that volitions take both more and less basic results as their objects. On a volition theory, the state of affairs the agent intends to bring about as a basic result serves as the content of something like a mental decree; and when all is well, and the power to do this kind of action is present, and not blocked somehow, the decree is successful. God is supposed to be incorporeal, and not limited to act by means of the motions of a particular body. For such a being, the content of a volition could simply be the basic result that a certain planet stop spinning. For a deity, stopping the planet could be a basic action. If a being is omnipotent, then—corporeal or not—it should be able to bring about such a thing simply by willing that it be so. I shall describe the power so to act as having the basic power to bring it about that such-and-such obtains, where “such-and-such” is the description of a state of affairs or circumstance—e.g., the Red Sea’s parting, or a certain portion of water’s becoming wine. Human beings obviously have much more limited ranges of basic powers than an omnipotent being would have. So far in my attempt to define omnipotence, I have followed Swinburne and others in making use of a very general notion of ability to intentionally bring about. If I can move my finger, then, if the gun is in my hand and loaded and many other crucial conditions are satisfied, I have the ability to intentionally bring it about that someone dies. If

DEFINING OMNIPOTENCE



I can raise my hand, then, so long as I am at an auction and visible to the auctioneer, I have the ability to bring it about that I am the highest bidder. On a theory of action like Swinburne’s—one that distinguishes between basic results of action and more distant results—a very general ability to bring it about that S must be given a disjunctive analysis along something like the following lines: either S is a state of affairs that one is able to bring about as a basic result, or there is a state of affairs S* that one can bring about as a basic result, and one can do so in contexts where S* would obtain as an intended consequence of one’s bringing about S. In other words, the facts about the states of affairs that an agent can intentionally bring about at a time are determined by the states of affairs he or she can bring about directly, as basic results; plus the leverage these abilities give the agent over larger states of affairs in virtue of foreseeable results of the basic action and of conventions that give the basic action added significance. Generally speaking, an agent with greater basic powers is better able to control the holding of the conditions needed for basic results to generate non-basic results. So, generally speaking, an agent with a greater range of basic powers will have a greater range of powers to bring about all kinds of states of affairs. In the limiting case, one should expect that a being with maximal basic powers (assuming there is a maximum) would also be able to determine the widest range of states of affairs of all sorts. In light of this connection, it seems to me that nothing significant is in danger of being lost if one stipulates that the degree of power relevant for defining omnipotence is a function of the degree of an agent’s basic powers alone; in other words, if it is a function of the size of the range of states of affairs an agent is able to bring about by means of volitions that have the states of affairs in question as basic results.35 The two approaches to omnipotence defended so far, encapsulated in (D8) and (D9), could each be refined by restricting attention to the states of affairs that can be brought about as basic results. Modifying either one requires first modifying (D7), as well.

35

Although Wielenberg’s definition of omnipotence (2000: 39) is quite different from those proposed here, his explanation of the source of God’s power supports the idea that omnipotence is a function of just basic powers. He defines ‘willpower’ as “a capacity for making things happen simply by willing them to happen,” and says that God’s omnipotence is due entirely to God’s “tremendous willpower” (39).



DEAN ZIMMERMAN

(D7*)

x is as powerful as x can be at t = df for any contingent state of affairs S that could possibly be intentionally brought about by someone as a basic result, and for any time t, if it is possible, given the actual history of the world up to and including t, that S obtain, and possible that x intentionally bring about S as a basic result, then x could intentionally bring about S after t.

The “greatest range” approach of (D8) could then be restricted to basic results: (D8*)

x is omnipotent at t = df it is possible for x to intentionally bring about something; x is as powerful as x can be at t; and, for any possible y, if y were to exist at t and be as powerful as y can be at t, the range of things x could intentionally bring about as a basic result after t is greater than the range of things that y could intentionally bring about as a basic result after t.

The “unthwartable” approach of (D9) could also be formulated in terms of just basic results: (D9*)

x is omnipotent at t = df it is possible for x to intentionally bring about something; x is as powerful as x can be at t; and, for any possible y, z, . . . , if y, z, . . . , can intentionally bring things about, and if they were to exist at t and be as powerful as they could be at t, the range of things x could intentionally bring about as a basic result after t would include states of affairs that prevent y, z, . . . , from intentionally bringing anything about as a basic result after t.

How does restricting the range of states of affairs to those that can be basic results deal with the counterexamples at the beginning of this section? Note first of all that, whenever another human being punishes me for lying, or causes me to feel remorse for lying, the more basic result by which these outcomes are achieved does not entail that I have actually lied. Punishment for lying, meted out by a friend or spouse, might take the form of stony silence, or leaving me alone to think about what I have done. I might be caused to feel remorse by someone’s reminding me how I felt when someone else lied to me, or simply by seeing the disappointment in my friend’s face. But someone can be silent, leave the room, tell me a story, or look disappointed in me, whether or not I actually lied. In each case, there is a basic result by which the intended effect is achieved,

DEFINING OMNIPOTENCE



and it is independent of my guilt or innocence. Of course God—or any good candidate for omnipotence—would have a much wider array of options for punishing me or causing me to feel remorse, were I to lie to my spouse. What means might God use to punish? Perhaps withering a plant that gives me shade, or sending a mole to dig up my lovely lawn. For generating remorse? Perhaps causing my child to ask me to read the story Pinocchio, or bringing vividly to mind the way I felt when someone lied to me. But again, as in the human case, the basic actions God could take would be possible outside the context of my actually lying to my spouse—although then they would not constitute punishment for lying, nor the generation of remorse for having lied.36 Take any example, S, of a kind of state of affairs that someone can bring about, but that depends, for the possibility of its occurrence, upon the occurrence of another state of affairs, one that could not be under the control of that agent—something that even an omnipotent being would have to allow someone else or something else to cause in an indeterministic manner. A volition theory of action like Swinburne’s requires a basic–non-basic distinction; so the act of bringing about S in its fullness can always be divided up into two parts: a basic result, S*, and all of the more remote or complex results that make it possible for an agent to bring about S by bringing about S*. The strategy of (D8*) and (D9*) is, for any such S, to ignore it and to focus upon the corresponding S*— something that could be brought about as a basic result even if the rest of the world does not cooperate and makes S inaccessible. When comparing ranges of states of affairs in order to determine relative greatness, (D8*) and (D9*) imply that only S* is relevant. Will the restriction to just basic results have unwanted consequences? Not so far as I can see. Of course problems will arise if one thinks that the only results that can be brought about in a basic way are bodily motions. In that case, even if there is a maximum of strength definable by means of (D8*) or (D9*), a being with that degree of strength would be at best a god-like superhero, and would have to rely upon external conditions in order to succeed in doing anything more impressive than moving his or her body (unless the body in question coincided with the entire physical world). But why should one think that the scope of possible basic actions is

36

I thank Michael Bergmann for help clarifying the ideas in this paragraph.



DEAN ZIMMERMAN

so limited? Presumably only because one takes there to be an incoherence in the very idea of an agent capable of making changes in the physical world simply by willing them, and not in virtue of a special connection to a particular body. In other words, the objection would be based on the impossibility of physical effects caused by ghosts, or immaterial deities, or agents with telekinetic powers. Suppose, however, it is so much as possible for a being to bring about, as a basic result, such states of affairs as an atom’s coming into existence from nothing, a human being’s ceasing to exist, a planet’s stopping in its orbit, and so on. Then such worries would be misplaced. The (imagined) critic is trying to cast doubt on whether the definitions apply to anything with truly miraculous powers, as opposed to mere superheroes; but the doubt can only be motivated by the conviction that nothing could have truly miraculous powers. A definition of omnipotence that is held hostage only to the possibility of beings with miraculous powers should be adequate for most purposes.

12 Conclusion I have identified two notions of maximal power, spelled out in terms of the range of states of affairs that can be brought about by an agent. Each seems to me to be promising; there is some chance it will be able to withstand the barrage of counterexamples. But many questions have been raised that deserve a more searching treatment than I have given them here. The two I find most worrisome are: (1) Can the space of states of affairs used in precisification of (D8) or (D8*) be sensibly treated as measureable? (2) Is there a near cousin of Negative Nelly hiding in the woodwork, someone who satisfies (D9) or (D9*) but is more plausibly possible than Nelly? I have also set to one side the familiar question: Could a being omnipotent in one of these senses create a stone too heavy for anyone to lift? So there remains plenty to worry about in these definitions. Still, it seems to me that Geach was overly pessimistic. Swinburne has been Chisholming his way through the underbrush for quite some time. Following his path may ultimately lead to a perfectly acceptable rangetheory of omnipotence.37 I am grateful to the audience at Purdue University’s conference, “Faith and Reason: Themes from Swinburne,” and to my commentator, Meghan Sullivan, for many challenging 37

DEFINING OMNIPOTENCE



References Baillie, James and Jason Hagen. (2008). “There Cannot be Two Omnipotent Beings,” International Journal for the Philosophy of Religion, 64: 21–33. Beall, J. C. (2000). “A Neglected Response to the Grim Result,” Analysis, 60: 38–41. Cartwright, Richard. (1975). “Scattered Objects,” in Keith Lehrer (ed.), Analysis and Metaphysics. Boston: Reidel Publishing Company, 153–71. Chang, Ruth. (2002). “The Possibility of Parity,” Ethics, 112: 659–88. Chisholm, Roderick. (1976). Person and Object. La Salle, IL: Open Court. Chisholm, Roderick. (1979). “Objects and Persons: Revision and Replies,” in Ernest Sosa (ed.), Essays on the Philosophy of Roderick M. Chisholm. Amsterdam: Rodopi, 317–88. Danto, Arthur C. (1963). “What We Can Do,” American Philosophical Quarterly, 2: 141–8. Davidson, Donald. (1980). Essays on Actions and Events. Oxford: Oxford University Press. Fischer, John Martin. (ed.) (1989). God, Foreknowledge, and Freedom. Stanford, CA: Stanford University Press. Flint, Thomas P. and Alfred J. Freddoso. (1987). “Maximal Power,” in Thomas V. Morris (ed.), The Concept of God. Oxford: Oxford University Press, 134–67; originally published in Alfred J. Freddoso (ed.) (1983). The Existence and Nature of God. South Bend, IN: University of Notre Dame Press. Freddoso, Alfred J. (1989). “Accidental Necessity and Logical Determinism,” in J. M. Fischer (1989); originally published in Journal of Philosophy, 80 (1983), 257–78. Geach, Peter. (1977). Providence and Evil. Cambridge: Cambridge University Press. Ginet, Carl. (1990). On Action. Cambridge: Cambridge University Press. Goldman, Alvin. (1970). A Theory of Human Action. Englewood Cliffs, NJ: Prentice-Hall.

queries and useful suggestions. Participants in a reading group at the Rutgers Center for the Philosophy of Religion also contributed a great deal to the chapter’s improvement. In these two contexts, I received valuable feedback from Marilyn Adams, Robert Adams, Michael Bergmann, James Doyle, Chris Hauser, Sam Lebens, Brian Leftow, Howard Robinson, Richard Swinburne, Peter van Inwagen, and Edward Wierenga, among others; I am painfully aware of not having taken full advantage of all the help they offered. Pavel Davydov read more than one draft, and was generous in his comments. Although I was unable to address his deepest worries about the project, he also pointed out plenty of smaller, fixable problems. Correspondence and conversations with Sam Lebens, Andrew Moon, and Justin Mooney enabled me to sharpen the arguments against the possibility of Negative Nelly. My biggest debt is to Daniel Rubio, who helped immensely with the arguments of section 9— the measure-theoretic response to McEar.



DEAN ZIMMERMAN

Hoffmann, Joshua and Gary Rosenkrantz. (2007). “Omnipotence,” in Chad Meister and Paul Copan (eds), The Routledge Companion to Philosophy of Religion. London and New York: Routledge, 271–80. LaCroix, Richard. (1977). “The Impossibility of Defining ‘Omnipotence’,” Philosophical Studies, 32: 181–90. Leftow, Brian. (2009). “Omnipotence,” in Thomas P. Flint and Michael C. Rea (eds), The Oxford Handbook of Philosophical Theology. Oxford: Oxford University Press, 166–98. Leftow, Brian. (2012). God and Necessity. Oxford: Oxford University Press. Mackie, J. L. (1962). “Omnipotence,” Sophia, 1: 13–25. McCann, Hugh. (1998). The Works of Agency. Ithaca, NY: Cornell University Press. Melden, A. I. (1961). Free Action. London: Routledge and Kegan Paul. Pearce, Kenneth L. (2011). “Omnipotence,” in James Fieser and Bradley Dowden (eds), The Internet Encyclopedia of Philosophy, . Pearce, Kenneth L. and Alexander R. Pruss. (2012). “Understanding Omnipotence,” Religious Studies, 48: 403–14. Plantinga, Alvin. (1967). God and Other Minds. Ithaca, NY: Cornell University Press. Plantinga, Alvin. (2003). Essays in the Metaphysics of Modality, Matthew Davidson (ed.). Oxford: Oxford University Press. Plantinga, Alvin and Patrick Grim. (1993). “Truth, Omniscience, and Cantorian Arguments: an Exchange,” Philosophical Studies, 71: 267–306. Pruss, Alexander R. and Joshua Rasmussen. (2015). “Problems with Plurals,” in Karen Bennett and Dean Zimmerman (eds), Oxford Studies in Metaphysics: Vol. 9. Oxford: Oxford University Press, 42–57. Rosenkrantz, Gary and Joshua Hoffmann. (1980). “What an Omnipotent Agent Can Do,” International Journal for Philosophy of Religion, 11: 1–19. Ross, James F. (1969). Philosophical Theology. Indianapolis: Bobbs-Merrill. Ryle, Gilbert. (1949). The Concept of Mind. New York: Barnes and Noble. Sobel, Jordan H. (2004). Logic and Theism. Cambridge: Cambridge University Press. Speaks, Jeff. (2014). “The Method of Perfect Being Theology,” Faith and Philosophy, 31: 256–66. Swinburne, Richard. (1977). The Coherence of Theism. Oxford: Clarendon Press. Swinburne, Richard. (1986). The Evolution of the Soul. Oxford: Clarendon Press. Swinburne, Richard. (1993). The Coherence of Theism, revised edn. Oxford: Clarendon Press. Swinburne, Richard. (2004). The Existence of God, 2nd edn. Oxford: Clarendon Press.

DEFINING OMNIPOTENCE



Swinburne, Richard. (2009). “How the Divine Properties Fit Together: Reply to Gwiazda,” Religious Studies, 45: 495–8. Tarski, Alfred. (1956). “Foundations of the Geometry of Solids,” in Logic, Semantics, Metamathematics. Oxford: Clarendon Press, 24–9. Taylor, Richard. (1966). Action and Purpose. Englewood Cliffs, NJ: Prentice-Hall. Wielenberg, Erik J. (2000). “Omnipotence Again,” Faith and Philosophy, 17: 26–47. Wierenga, Edward R. (1989). The Nature of God. Ithaca, NY: Cornell University Press.

6 Law, Cause, and Occasionalism Alvin Plantinga

Think of natural theology as the activity of coming up with arguments for the existence of God—arguments, roughly, from some feature or other of our universe. Richard Swinburne is certainly the outstanding natural theologian of our day; indeed, his work over the last thirty years or so has resulted in the most powerful, complete, and sophisticated development of natural theology the world has so far seen. One of his arguments starts from the premise that there are natural laws, and in particular, simple natural laws. As Swinburne sees it, the existence of simple natural laws is much more likely given that there is such a person as God, than it is on the proposition that there is no such person as God. Now I am strongly inclined to agree with Swinburne here, even though there are questions. (For example, is it clear that we can come up with anything like a decent probability for the existence of such laws, given the claim that there is no such person as God? And do we have to factor in the antecedent probability of there being such a person as God, and if so, how do we estimate that?) I do not plan to explore these questions, inviting though that prospect is. Furthermore, my project, unlike Swinburne’s, is not apologetic; I do not propose to offer a theistic argument here. Instead, I want to begin by looking at some questions about natural law or laws of nature. Are there good reasons for thinking there are any such things? If so, what sort of things are they? How are they related to determinism? How are natural laws (if there are any) related to God and what are the most promising ways to think of laws of nature from a Christian perspective?

LAW , CAUSE , AND OCCASIONALISM



I will argue that the three most promising ways are (1) the idea that laws of nature reflect the causal powers of the creatures God has made, (2) the idea that natural laws are divine ordinances, part of God’s way of directing and ordering creation, and (3) the idea, due to Del Ratzsch, that laws of nature are counterfactuals of divine freedom. The first, so I will argue, fits best with the thought that there are secondary causes as well as divine causation; the second and third fit best with occasionalism, the thought that all causal activity is divine causal activity. I will conclude by giving qualified support to occasionalism.

1 Natural Law and its Nature First, the question of the nature of natural laws—what sorts of animal would a natural law be, if indeed there are such things? Newton’s Gravitational Law and his three Laws of Motion would be putative examples, as would be the laws of conservation of momentum, energy, and angular momentum. Laws of nature, typically, are universal generalizations, although perhaps there are also some probabilistic natural laws. But of course not just any universal statement is a law: All the books on my desk belong to me and All the birds in Sam’s backyard are sparrows are universal in form but hardly laws. We might think the problem here is that these propositions make essential reference to a particular time or place or person. But this is not the real problem. Here are a couple of historically important examples (van Fraassen 1989: 27): (1) All solid spheres of enriched uranium have a diameter of less than one mile and (2) All solid spheres of gold have a diameter of less than one mile. Neither makes reference to any particular time or place or person; still, one is inclined to doubt that (2) is a law of nature, but much more likely to afford that status to (1).1 Why? What makes the difference? Something along the following lines: One wants to say that (2), if true, is just 1

Strictly speaking, (1) is a consequence of natural law, not itself a law of nature.



ALVIN PLANTINGA

an accident—there certainly could have been a solid sphere of gold more than a mile in diameter, and if one were discovered—say, on the moon— all the civilized nations would fight over it. But there could not have been a solid sphere of enriched uranium a mile in diameter; the critical mass for enriched uranium is only about 110 pounds. The point is that laws of nature seem to be, in some sense, necessary. The thought is that it is necessary that material objects attract each other with a certain force; it is necessary that momentum is conserved in an isolated system; it is not possible that energy increases in a closed system, or that a material object attains a velocity greater than the speed of light. So another and crucially important characteristic of laws of nature, if there are any, is that they are necessary.

2 The Necessity of Law 2.1 Absolute necessity This necessity, however, while it may be their glory, is also, so to speak, their Achilles’ heel. First, what kind of necessity are we talking about here? Some philosophers, for example Sydney Shoemaker (1980) and more recently, Chris Swoyer (1982: 203 ff.), Evan Fales (1990), and Alexander Bird (2005: 353 ff.), argue that the laws of nature are absolutely necessary,2 or strictly necessary, or necessary in the broadly logical sense—necessary in the same way as bachelors are unmarried or that red is a color or that 7 + 5 = 12. Clearly the laws of nature are not among the truths of mathematics or the truths of logic. Of course there are many absolutely necessary propositions that are neither truths of logic nor truths of mathematics: for example, whatever is red is colored, and no human beings are prime numbers. But the laws of nature do not seem to be of this kind either. The fact is, so we are told, that any two objects attract each other with a force directly proportional to the product of their masses and inversely proportional to the square of the distance between them; this is a law of nature. But it does not seem to be absolutely necessary. It certainly seems 2 We cannot really give a definition of absolute necessity here—or if we can, it will be in terms of other notions just as much (or little) in need of explanation (possibility and impossibility, for example). But we also do not really need a definition: we can get the idea from examples.

LAW , CAUSE , AND OCCASIONALISM



that this attractive force could have been inversely proportional to some other power of the distance between them; if it had been, no doubt things would have been different, but that is not to say that it is absolutely impossible. That nothing can travel faster than light is thought to be a law of nature;3 but it certainly seems possible, in the strict sense, that elementary particles of some sort (or, for that matter, a spaceship) should do that. And even if we humans could not make a spaceship capable of that feat, could God not do so? True, there may be (and perhaps are) absolutely necessary propositions whose necessity we cannot detect just by thinking about them. But our best guide to necessity is intuition; we see that some propositions are necessary, and we learn that others are by way of seeing that they follow from those of the first sort. So if the laws of nature seem to be contingent (i.e., not absolutely necessary), we should suppose they are contingent, in the absence of powerful argument for their (absolute) necessity. And as far as I know there is no decent, let alone powerful, argument for their absolute necessity.

2.2 Contingent necessity? D. M. Armstrong suggests that the laws of nature are necessary, but not absolutely necessary: Suppose it to be a law that Fs are Gs. F-ness and G-ness are taken to be universals. A certain relation, a relation of non-logical or contingent necessitation, holds between F-ness and G-ness. This state of affairs may be symbolized as ‘N(F, G) (Armstrong 1983: 85).4

This does seem to square pretty well with our intuitions about laws of nature. There does seem to be necessity of some kind associated with them, but they do not seem to be absolutely necessary. But what is this relation of non-logical or contingent necessitation? So far, Armstrong has told us nothing at all about this relation, except that it is non-logical and contingent. Of course he has given it a name: he says it is a relation of non-logical and contingent necessitation. As David Lewis says, however, 3

Strictly speaking, nothing can accelerate from a velocity less than that of light to a velocity greater than that of light; perhaps there are tachyons that are always moving faster than light. In what follows I shall ignore this qualification. 4 Others who adopt this same approach (i.e., take it that laws are “contingently necessary”) are Dretske (1977: 248 ff.) and Tooley (1977: 667 ff.).



ALVIN PLANTINGA

there has to be more to this relation than just that name: just bearing the name cannot equip it to explain laws of nature: it cannot do that just by bearing a name, “anymore than one can have mighty biceps just by being called ‘Armstrong’ ” (1983: 366). But what is this more? Laws of nature, therefore, if there are any such things, are necessary in some sense, but not necessary in the broadly logical sense. And this is the problem: what is that sense in which laws of nature are necessary? How are we supposed to understand that? Strict necessity we know and love; but what is this nonstrict necessity?5 How are we supposed to construe it? Armstrong does not tell us.

2.3 Humean conceptions of law If it is so hard to say what kind of necessity is enjoyed by natural law, perhaps we should follow Davids Hume and Lewis in rejecting the whole idea of necessity for laws of nature. Thus, the late David Lewis took it that laws of nature are simply the theorems common to the true axiomatic systems that enjoy a best combination of strength and simplicity (e.g., Lewis 1973 and 1983). Some axiomatic systems enjoy great simplicity: for example, one in which the only axiom is 2 + 1 = 3. Others enjoy great strength: for example, a system in which every true proposition is an axiom. The former system, however, lacks strength, and the latter lacks simplicity. Strength and simplicity clearly compete; a maximally simple system—e.g., one whose only axiom is 2 + 1 = 3—will be weak; a maximally strong system—one such that every true proposition is a theorem—will not be simple. We do not have much of a grasp on what Lewis’s best systems will be like,6 but presumably they will display a great deal of strength, but also a great deal of simplicity. There is little reason to think there is just one such system—perhaps there are a multitude of systems each unsurpassed by any other. There is more than one problem with this view of the laws, but I wish to point out a particularly interesting consequence of this way of looking at the matter: it is that determinism (at least as widely understood) and freedom are compatible. Here I am thinking of freedom in the classical libertarian sense. A person is free, with respect to a given act, and at a 5

Van Fraassen’s (1989) discussion remains perhaps the best discussion of this problem. We do not even know that strength and simplicity are equally important: perhaps strength counts for more than simplicity, or simplicity for more than strength. 6

LAW , CAUSE , AND OCCASIONALISM



particular time, if and only if at that time he has the ability (it is within his power) to perform that act, and also, at that time, the ability to refrain from performing it. If I am now free with respect to the action of raising my left hand, then right now I can raise it, and also right now I can refrain from raising it. And even if I do raise it, I could have refrained from doing so. Now the most common definition of determinism goes as follows. Let Ut and Ut* be the complete states of the universe at times t and t*, respectively; and let L be the conjunction of the laws of nature. Then determinism holds just if the conjunction of L with Ut entails Ut* (e.g., van Inwagen 1983: 16). To put it more colloquially, determinism holds just if the state of the universe at any one time together with the laws of nature entails the state of the universe at any other time. But now suppose that the laws of nature are partly dependent upon what, as a matter of particular fact, does happen: past, present, or future. And suppose on a given occasion I raised my left hand (maybe to volunteer an objection to Lewis’s view). Given determinism, the laws together with the state of the universe, say, 1,000 years ago (call it Uminus 1000), entail that I raise my hand then. But it does not follow that it was not within my power, then, to refrain from raising my hand. What the laws are depends upon, among other things, what I do. I did raise my hand on that occasion, and the axioms of some favored system together with Uminus 1000 entail that I did so; but that fact does not entail that it was not within my power to refrain from raising it then. Of course if I had refrained on that occasion, then (given Uminus 1000) some proposition that was a law of nature would have been false, and hence would not have been a law of nature. So the important point here is that on these broadly Humean conceptions of laws of nature are at bottom merely descriptive of what actually happens, the laws of nature are (or can be) within my power in the sense that I have the ability to act in such a way that their conjunction would have been false. We ordinarily think of the laws of nature as outside our power, but on these Humean conceptions this is not true. If indeed we do have libertarian freedom, it will be within our power to break the laws, i.e., the laws that actually hold are such that it is within my power to go contrary to them. This gives us a good reason, I think, to reject these Humean conceptions of the laws. If there are any laws of nature, it is not within my



ALVIN PLANTINGA

power, or even yours, so to act that what is a law would not have been. This is of course connected with the apparent necessity of these laws; the Humean accounts leave out that necessity and hence, so I say, they cannot be thought of as correct accounts of the laws. Indeed, these Humean accounts are really versions of anti-realism with respect to the laws. One way for an anti-realist to proceed would be to say forthrightly: “there are no laws of nature; there are only non-necessary universal generalizations (of one sort or another).” But another way to be an anti-realist with respect to laws is to say: there are laws of nature, and what they are, are non-necessary universal generalizations of a certain sort.7

2.4 “Creaturely inviolability” Natural laws, therefore, are universal generalizations that enjoy a certain kind of necessity. But what kind? Not broadly logical necessity, but what other kind is there? Perhaps we can approach this question along the following lines. The Apollo 11 was launched on July 16, 1969 from Kennedy Space Centre and landed on the moon on July 19. It took a total of 3 days, 3 hours, and 49 minutes to fly to the moon. Later, the NASA Pluto probe New Horizons, at a speed of 58,000 km/hr, took only 8 hours and 35 minutes to get to the moon from Earth. Perhaps future technological innovations will be able to reduce that time to 1 hour, or, indeed, even 1 minute. No matter how good our technology gets, however, we will not be able to reduce that time to 1 second. That is because nothing can travel faster than the speed of light; but the speed of light is about 186,000 miles per second and the moon is 238,900 miles from earth. The speed of light is a sort of universal speed limit; nothing can exceed it. No matter how good our technology gets, we will not be able to build a spaceship that travels faster than light. That nothing can travel faster than the velocity of light is a law of nature; and we might say that it is creaturely inviolable. No creature can act in such a way as to violate this law (or bring it about that it was not a law). But the same holds for the other laws of nature: no matter how

7

Compare nominalism:

(a) there are no universals; what there are instead are nomina, names; (b) There are universals, and they are nomina, names.

LAW , CAUSE , AND OCCASIONALISM



good our technology gets, we will not be able to build a machine that violates a law of nature or in some other way acts so as to falsify any of them. So what sort of necessity do the laws have? A first approximation would be this: laws of nature are necessary in the sense that they enjoy creaturely inviolability, but no creatures can develop a technology whereby they can act in such a way as to violate a law of nature. No doubt God can act in such a way; but we cannot. So the necessity of law is a matter of their being creaturely inviolable—or at any rate humanly inviolable.

3 God and the Laws So much for a preliminary and general description of the laws of nature. But how are they related to God? There seem to be three possibilities, three ways in which God could be related to the laws of nature, again assuming that there are some. For the sake of concreteness, let us consider a particular law: Newton’s Law of Gravitation (and here I will not be concerned with special or general relativity or quantum mechanics). How is this law related to God? One possibility is secondary causalism: that laws of nature reflect the causal powers of the creatures God has made. God creates material objects with a certain nature or certain powers, and in such a way that they have the property P of being such that any two of them attract each other in the way specified by Newton’s Law; that is, they exercise a certain kind of force with respect to each other. Newton’s Law specifies the degree and nature of this force. A second possibility would be decretalism. Newton’s Law is or represents a divine decree, a decree God issues which specifies how material objects will move under various conditions. A third possibility is given by Del Ratzsch’s very interesting suggestion that natural laws are counterfactuals of divine freedom: Newton’s Law specifies how God acts and would act, how he would treat the stuff he has made, under various different conditions (Ratzsch 1987). Let us take these one at a time and in order.

3.1 Secondary causalism God creates concrete objects with causal powers, the power to cause behavior of certain kinds. According to this way of looking at things, there are two kinds of causality: primary causality, exerted by God alone; and secondary causality, exerted by some of God’s creatures. God creates



ALVIN PLANTINGA

all the concrete objects;8 God also sustains them in existence. But they are so created that they too can get involved in causation—not just as effects, but also as causes. Here we must make two distinctions. First, there is strong secondary causalism, the notion that a secondary cause, given God’s sustaining it and its powers, can cause something to happen without any further divine action or aid. But there is also concurrentism: the more common idea that any causal transaction involving secondary causation must also involve God’s concurrence—his, so to speak, ratifying that particular exertion of causal power. Concurrence, of course, is not simply the absence of objection or countervailing activity; it is instead a positive activity on the part of God. (Peter van Inwagen thinks concurrentism merely pays God empty metaphysical compliments.) That’s the first distinction: the second is this. We might think concrete objects have these causal powers by nature, so that it is an essential property of a material object to exert the forces it does exert. No physical object could have existed without exerting that force; not even God could have created a physical object that did not exert this force. On the other hand, we might think that material objects do indeed exert those forces, but their doing so is not essential to them, and God could have created material objects that lacked this property. Secondary causalism is perhaps the common-sense way of thinking of the matter. We ordinarily take it for granted that many created objects can cause changes in other created objects. I can cause a row of dominoes to fall by exerting a small force on the first domino; then each domino causes the succeeding domino to fall by hitting it. I can cause the billiard ball to roll by striking it with a cue stick; that billiard ball can strike another ball, thereby causing that other ball to roll away. (Does that cue ball cause the other ball to roll away? Or is it rather the event of the cue ball’s striking the other ball that causes the event consisting in the other ball’s rolling away? The first would be a matter of agent causation and the second of event causation.) This way of thinking also has impressive historical pedigree; it is fully developed by Thomas Aquinas, who, as is his wont, follows Aristotle. Nonetheless, it has problems and difficulties.

8 Except, of course, for himself. There are also abstract objects such as numbers, propositions, properties, and the like; I take these to be necessary beings but dependent upon God. Here I will not be concerned with abstract objects.

LAW , CAUSE , AND OCCASIONALISM



Perhaps the main difficulty here, is that the very idea of creaturely causality is obscure. Of course we can use other terminology: we can speak of forces, or powers, or bringing it about that, or . . . But do we really understand any of these locutions when we are speaking of creatures? Is there a reasonably clear and coherent concept or idea associated with these terms? It pains me to agree with Hume, but is he not right here? We see the first billiard ball roll up to and strike the second, and we see the second roll away. We do not, of course, see or experience anything like a causal connection between the first ball and the second, or the motion of the first ball and the motion of the second, or the event consisting in the first ball’s striking the second and the second’s moving away. We just see the first ball roll up to the second and become juxtaposed with it, we hear a click, and then we see the second ball roll off. Furthermore, we do not seem to have a coherent idea of a necessity linking the two events. What is this idea of creaturely causation? Of course we can just “take it as primitive,” refusing to offer an explanation for it. But again, of course, that does not really help. If we really do not grasp this idea of creaturely causation, it will not do a lot of good to take it as primitive. Alternatively, we might try to follow Immanuel Kant, who, wakened by Hume from his dogmatic slumbers, argued that causation must be a sort of prism or lens thorough which we look at the world, a sort of idea that we impose on the world, one that the world as it is in itself does not display. But again, this does not really help. If we do not have a good grasp of the notion, it will not help to declare that it is a contribution from our side—we still do not have a grasp of it.

3.2 Decretalism A second possibility is decretalism. Perhaps the relation between God and the laws is that God just decrees that objects—material objects, say— shall behave in accordance with the laws. Material objects do not, in fact, exert force on each other and they do not, in fact, display causal efficacy; rather, God issues a decree. He says: let it be that material objects behave as if there is an attractive force between any pair of them, a force that varies directly with the product of their masses and inversely according to the square of the distance between them. On this alternative, there is not really any force between them—i.e., they do not exert forces on each



ALVIN PLANTINGA

other—but they behave as if there were. These objects do not really have any causal powers; they do not in fact exert any forces on each other. They simply behave in accord with the divine decree.

3.3 Counterfactuals of divine freedom On this third possibility, natural laws are or represent counterfactuals of divine freedom: they specify how God would (freely) treat the stuff he has made under various different conditions. They are of the general form under conditions C, God would cause state of affairs S. On this suggestion as on the second, objects do not have causal powers; they do not exert forces on each other or in other way act as causes. And on this suggestion, as on the second, all causal activity is divine causal activity. The second and third possibilities are related in an interesting way. On each of them, the only causal activity is divine activity. But on decretalism, it is as if God, in issuing the decree in one causal act, the issuing of that decree causes whatever happens at any time. On the other suggestion, the suggestion that laws are counterfactuals of divine freedom, God’s causal activity consists in many different actions spread out over time. But again, on either suggestion, all causal activity and all causal power is divine. And this means that both of these suggestions are variants on what has traditionally been called occasionalism.

4 Occasionalism This is the view that the only causal power is divine causal power. God causes every change that occurs. God is the only real cause. Sometimes, however, there is a correlation between certain events and God’s causing some other event; for example, there is a more or less constant correlation between my willing to raise my arm and my arm’s rising. That is because God ordinarily takes my willing to raise my arm as the occasion for causing my arm to rise. Occasionalism may go back to Nicholas of Autrecourt (1300–d. after 1350). Sadly enough, his ideas met with less than overwhelming approval: his works were burned and Nicholas himself was prohibited from lecturing. Another Nicholas, Nicholas Malebranche, is the best-known occasionalist in our tradition, and he put it like this: “There is only one true cause because there is only one true God; . . . all natural causes are not true

LAW , CAUSE , AND OCCASIONALISM



causes but only occasional causes.”9 Elsewhere he says, “But natural causes are not true causes: they are only occasional causes that act only through the force and efficacy of the will of God” (trans. Lennon and Oscamp 1997: 449). On occasionalism, therefore, there is no creaturely causation; creatures do not have the power to cause events or changes or anything else. All causal power is divine causal power. Now one advantage of occasionalism is just the other side of the main problem with secondary causalism. We do not have a clear conception of creaturely causality, but that problem does not arise with divine causality. Divine causality, as we may suppose, just goes by way of divine fiat: God says, “Let there be light” and there is light. God wills that there be light, or that there be light at a particular time and place, and there is light then and there. And the connection between God’s willing that there be light and there being light is necessary in the broadly logical sense: it is necessary in that sense that if God wills that p, p occurs. Insofar as we have a grasp of necessity (and we do have a grasp of necessity), we also have a grasp of causality when it is divine causality that is at issue. I take it this is a point in favor of occasionalism, and in fact it constitutes a very powerful advantage of occasionalism.

5 Which is Best? Should we therefore award the palm to occasionalism, taking it to be proved, or, since proofs rarely occur in philosophy, in better shape than its rivals? Hardly. That is because occasionalism might have serious difficulties of its own. After all, one cannot establish that the set of non-self-membered sets is not a member of itself by showing that it could not be a member of itself. Are there serious difficulties for occasionalism? As far as I know, it is not a presently popular doctrine: what do people see as problems with it? Well, for one thing the idea that created substances often cause events and changes in the world seems to be no more than part of common sense: for example, I go over to the fridge and open its door. This certainly seems a pretty clear case of my causing something to happen, namely the refrigerator’s door opening. Now perhaps an idea’s just being common sense gives it an initial 9

The Search after Truth and Elucidations of the Search after Truth, trans. Lennon and Oscamp (1997: 448), cited in Lee (2014).



ALVIN PLANTINGA

advantage or an initial claim on our credence. But here this is not much of an advantage: if the idea of creaturely causation really is wholly obscure, the fact that it is apparently endorsed by common sense will not help a lot.

5.1 Strong occasionalism Still, there may be something in the neighborhood that really is a serious point against occasionalism—at any rate one version of it. I can easily see how it could be that when at t I undertake to raise my arm, God does the actual raising, my undertaking being the occasion for his so acting. But what about my willing this in the first place? Here we have a change, a change in myself, i.e., my self. Before t I have not undertaken to raise my hand; after t I have. Does God cause that change? As far as I can see, the problem does not depend on any particular position on the relation between mind or self and body; just for definiteness, however, and because this is a volume on Swinburne, I will think of the problem in terms of substance dualism. I myself am a substance; I am not a body or material object, but am closely and uniquely related to a certain particular physical object, namely, my body. And of course changes occur in me or perhaps to me. My finger gets hit with a hammer; I am in pain. (Perhaps I then also say something to myself, wisely keeping it to myself.) You ask me what I had for lunch; I think for a moment, and then form the belief that what I had for lunch was a hotdog. I try to remember the name of the person talking with me, and suddenly it comes to me. As I sit at my computer, I try to figure out how best to put the next point I am trying to make. So there are changes in me, in my self. Presumably these changes have causes. But what is the cause of these changes? According to strong occasionalism, all causation is divine causation; God, therefore, is the cause of these changes. I idly decide to think about the Exum Ridge route on the Grand Teton and then immediately do so; it is God who causes that decision, and causes my thinking of that route. Can that be right? We think of making decisions along the following lines: I marshal the reasons for one course as opposed to another, think about the matter, and then opt for one course. On strong occasionalism, when I marshal the reasons for the decision, it is really God who does the marshaling; it is God who causes me to think of a given circumstance as a reason for a

LAW , CAUSE , AND OCCASIONALISM



particular line of action, God who causes that circumstance to come to my mind, and God who causes me to make the decision I do indeed take. But then does it make sense to say that it is I who takes that decision? How can it be that I take that decision, when it is God who causes every circumstance in the whole process? Is there any room left for agency, for me to be an agent? It looks like the answer is no. Nothing I do here exemplifies anything that could be thought of as my agency; it is all God’s agency. How can I be thought of as an agent under these conditions? Now materialists and others sometimes appear to be willing to give up the notion of human agency. But from a Christian point of view this is not a real possibility. If God causes me to do whatever I do, then, when, for example, I make a wrong decision, deciding to act in a selfaggrandizing way, it is not I who am responsible for or who causes that decision; it is God who does so. Indeed, I do not really do anything that could sensibly be called “making” that decision: it is rather that God just causes a particular mental state to occur at that time. So consider an occasion on which I fall into sin—I make an unpleasant and unflattering remark about someone mainly because I resent that person’s academic success. What happens on this occasion? God causes me to resent this person’s success; God also causes me to make that remark, and it is not the case that my resentment causes the remark. The only connection between my resentment and my remark is that God causes each of them, causes them to occur in the order in which they do occur, and causes them for reasons of his own. An important aspect of Christian belief is that we human beings are proper subjects for moral evaluation, for praise and blame. The central truth of Christian belief is that we human beings have sinned, to which God’s magnificent response is Incarnation and Atonement. But if it is true that God causes my every thought, my every “decision,” as well as my every action, how can I possibly be blamed, or even thought responsible, for that remark? I had no agency either in its occurrence or in the resentment out of which it arose. But then my actions and mental states are not the proper subject of moral evaluation. More exactly, I am not the proper subject of moral evaluation. It is not merely that I could not have avoided taking this action or being in this mental state; it is that I did not really do anything here at all. God did whatever was done. Still another obvious difficulty for strong occasionalism: God is perfectly loving and perfectly good. But if God causes whatever happens,



ALVIN PLANTINGA

causes it to happen by willing that it happen, then is God not the cause of evil? Indeed, is he not, then, the cause of all the evil that occurs? Christian belief, therefore, precludes strong occasionalism. According to Christian belief, I am a proper subject of moral evaluation; but if so, I am responsible for my envious condition, and I am responsible for that snide remark. God is not the cause of the evil the world contains; it is creatures that cause evil.

5.2 Me and my undertakings But do we now not have a problem? The attraction of occasionalism is just that divine causality, involving as it does just logical necessity, is clear and understandable; but causation on the part of creatures is not. If we reject strong occasionalism, however, taking it that human beings and perhaps other persons cause changes—decisions, for example, or undertakings—in themselves, in the way in which I cause the event consisting in my taking a certain decision, then are we not back to obscurity of the notion of creaturely causation? Well, perhaps we are confronted with an obscure notion here, but it is a different obscure notion. The relation between me and one of my decisions or undertakings, even if it is properly thought of as causal, is very different from a causal relation between me and my hand’s going up. Suppose I undertake to raise my hand; the relation between me and that undertaking is very different from a causal relation (if there is one) between me and my hand’s going up. Perhaps we can see this by turning to the divine case. I said that we can understand divine causality, just because it is a matter of broadly logical necessity: necessarily, if God wills that so and so, then so and so happens. But there is not anything like that in the case of the relation between God and that willing itself. God causes there to be light by willing that there be light; he does not cause it to be the case that he wills that there be light by willing that he wills that there be light. This question of the relation between a person and certain of his or her mental states such as decisions and undertakings is quite a different question from the question about (possible) causal relations between a person and such events as her arm’s rising. As for the first, we know that people do in fact make decisions, they do in fact will that certain states of affairs be the case, they do in fact undertake certain actions, even if we cannot say a lot about how they do it, or what is involved in such cases, or whether it is causation that is involved in these cases.

LAW , CAUSE , AND OCCASIONALISM



Objection: you said earlier that the obscurity of creaturely causation was a strong point in favor of occasionalism; here you say that we must just accept this relation between persons and certain of their mental states, even if we cannot really say what is going on there. Why should we not say the same about creaturely causation? Reply: because there is an alternative in the case of creaturely causation: we can instead opt for occasionalism. But there is no alternative in the case of that relation, whatever precisely it is, between me and my decisions and undertakings. In particular, we cannot sensibly say that they are caused by God.

5.3 A problem for weak occasionalism? Strong occasionalism, therefore, is too strong. Let us distinguish strong occasionalism from weak occasionalism. I will not take the time to try to give a serious definition of weak occasionalism: let us just say, for present purposes, that it is the thought that the only creaturely causation is of the sort involved in my causing my decisions, volitions, and undertakings (if indeed that relation is one of causation). According to weak occasionalism, I am the cause, in that peculiar sense, of my decisions and undertakings, but when I will to do something—raise my hand, for example—it is God who causes my hand to rise. But is weak occasionalism not really subject to one of the same difficulties that beset strong occasionalism? One objection to strong occasionalism is that it makes God the cause of evil in the world; God is the sole cause of my evil decisions and volitions. On weak occasionalism, that is not true, but God still causes whatever evil effects in the world are associated with my decisions and volitions. I decided to rob a bank; in the course of the robbery I shot someone. God does not cause me to decide to rob the bank or to shoot the guard, but he does cause all of the motions of my limbs, including my shooting the guard, and he also causes the guard’s injury or death. Does this not make God the cause of evil? This is not an easy question. First, however, we must make a distinction. Evil, I take it, is a matter of some person’s performing a wrong action—in the context of weak occasionalism,10 performing an evil act of will or an evil undertaking. God does not cause any such acts of will or 10

Henceforth I will drop the ‘weak’ of ‘weak occasionalism.’



ALVIN PLANTINGA

undertakings. However, in addition to evil, there are bad situations or states of affairs—someone’s suffering, or being treated unjustly, for example. Let us refer to these situations or states of affairs, inelegantly, as “the bad.” Is not God, on occasionalism, responsible for the bad? God is not the cause of evil, but he is the cause of the bad, and is that not bad enough? Perhaps we can approach this question by asking how the other main position, secondary causalism, fares with respect to this matter. Does secondary causalism do better with respect to the question of God’s causing evil? Here we must distinguish two cases: the scenario in which only immaterial personal agents have causal powers, and the scenario in which both personal agents and material substances, material objects, have such powers. First, what about material objects as causes? It is obvious, I take it, that material objects as such are not personal agents and do not act freely. Therefore, whatever they do is by way of chance or by way of determination by prior cause. But that material objects do what they do by way of chance is implausible. How could something occur just by chance, given the existence of God? We might think that God could issue a decree and say, “Let it be that A or B, and I don’t care which.” But being omniscient, God would know which of A and B would occur if he issued that decree. And how would that differ, in any significant respect, from his just decreeing A or decreeing B? So what about the case of material objects acting as they do by virtue of their being caused to cause whatever they do? Take any particular event E: it could be that there is an unbroken chain of causality ending in E and going all the way back to creation. If so, God’s relation to E would be as follows: God created an initial set of these objects, with those causal powers, sustained them and their successors in existence, knowing that E would eventually occur as a result of his creation and sustenance. God indirectly causes E, i.e., sets in motion and sustains a train of events that issues in E. Suppose E is a bad event: on occasionalism, God directly causes E, but in the current scenario, he does so indirectly. Is there any reason to think that in the first case, God bears more responsibility for the bad than in the second case? I do not think so. So far, secondary causalism fares no better than occasionalism on this point. Turning to the other alternative, suppose that a created personal agent freely causes some event in the causal ancestry of E—for example, suppose I undertake to stab someone. On secondary causalism, my

LAW , CAUSE , AND OCCASIONALISM



undertaking (presumably) causes certain events in my brain, which in turn set in motion a causal chain of events, the last member of which is the stabbing. On the (weak) occasionalist reading, God takes my undertaking as the occasion for his causing the events in my brain as well as the event issuing in the stabbing. In each case my undertaking results in the stabbing’s occurring. Presumably God endorses and underwrites this arrangement in order to confer on us significant freedom; my undertakings can result in morally significant events, including events that are part of the bad. On both occasionalism and secondary causalism, God permits me to undertake something bad. On the occasionalist reading, God then directly causes the events that are posterior to my undertaking, and that culminate in the stabbing; on the secondary causalism reading, my willing causes the events in my brain, and God establishes the causal relations that hold between the brain event and the subsequent members of the chain culminating in the stabbing, thereby indirectly causing those events, including the stabbing. On occasionalism, God directly causes the brain events; on secondary causalism, I not God cause those events (although of course God conserves me in existence and concurs with my causal activity). But this difference does not seem relevant to the question whether God causes the bad in a way incompatible with his being wholly good. On occasionalism, God directly causes the events subsequent to my undertaking, including the stabbing; on secondary causalism, God indirectly causes these events, including the stabbing. In either case God directly or indirectly causes the bad. As I have argued, however, it is hard to see how God is more responsible for causing the bad if he causes it directly than if he causes it indirectly. It is hard to see how it could be that his causing the bad directly is incompatible with his being wholly good, while his causing the bad indirectly is not. Still, I have to admit that there remains a sort of intuitive pull toward the thought that God’s directly causing the bad involves him more intimately with the bad than does his indirectly causing it.

6 Conclusion Let us take stock. The problem with secondary causalism is that we have no clear conception of causation as accomplished by creatures; we



ALVIN PLANTINGA

understand divine causation, but creaturely causation is at best dubious. On occasionalism, of course, there is no creaturely causation, so that on this head occasionalism enjoys a clear advantage. On occasionalism, however, there is that intuitive idea that God’s direct causation of the bad fits less well with his being wholly good than his indirectly causing the bad. These two considerations are not really commensurable; but it does seem to me that the problem with occasionalism is a smaller problem than the problem with creaturely causation. I therefore suggest that the best compromise is weak occasionalism.

References Armstrong, David. (1983). What is a Law of Nature? Cambridge: Cambridge University Press. Bird, Alexander. (2005). “The Dispositionalist Conception of Law,” Foundations of Science, 10 (4): 353–70. Dretske, Fred. (1977). “Laws of Nature,” Philosophy of Science, 44: 248–68. Fales, Evan. (1990). Causation and Universals. London: Routledge. Lee, Sukjae. (2014). “Occasionalism,” The Stanford Encyclopedia of Philosophy (Spring 2014 edn), ed. Edward N. Zalta, . Lennon, Thomas M. and Paul J. Oscamp. (trans.) (1997). The Search for Truth and Elucidations of the Search for Truth. Cambridge: Cambridge University Press. Lewis, David. (1973). Counterfactuals. Cambridge: Harvard University Press. Lewis, David. (1983). “New Work for a Theory of Universals,” Australasian Journal of Philosophy, 64 (4): 343–77. Ratzsch, Del. (1987). “Nomo(theo)logical Necessity,” Faith and Philosophy, 4 (4): 383–402. Shoemaker, Sydney. (1980). “Causality and Properties,” in Peter van Inwagen (ed.), Time and Cause. Dordrecht: D. Reidel. Swoyer, Chris. (1982). “The Nature of Natural Laws,” Australian Journal of Philosophy, 60 (3): 203–23. Tooley, Michael. (1977). “The Nature of Laws,” Canadian Journal of Philosophy, 7: 667–98. van Fraassen, Bas. (1989). Laws and Symmetry. Oxford: Clarendon Press. van Inwagen, Peter. (1983). Essay on Free Will. Oxford: Oxford University Press.

II

Philosophical Theology

Atonement and Liturgy

7 Love and Forgiveness Swinburne on Atonement Eleonore Stump

1 Introduction Over the course of his career, in addition to his many books and articles on various subjects in metaphysics and philosophy of religion, Richard Swinburne has undertaken to produce a systematic philosophical defense of Christianity. While there are other philosophers who have worked at defending one or another piece of traditional Christian philosophical theology, Swinburne is the only one to have engaged in the mammoth task of a systematic exposition and defense of all the most central doctrines of Christianity. He began with a trilogy on God’s existence and the nature of faith: The Coherence of Theism (1977), The Existence of God (1979), and Faith and Reason (1981), and he followed this trilogy with a tetralogy on topics in Christian philosophical theology. The first volume was Responsibility and Atonement (1989); the next was Revelation (1991); this was followed by a volume on trinity and incarnation, The Christian God (1994); and then there was the final volume of the tetralogy, Providence and the Problem of Evil (1998). The labor of a large project of this sort is clearly revealed in an endearing but utterly implausible claim Swinburne makes several times in Responsibility and Atonement: “The strongest desire of all, which influences us to fail to fulfill our obligations . . . is the desire to rest” (1989: 112). Both those who are inclined to attack Christianity and those who mean to defend it can be grateful for Swinburne’s unresting efforts, which have yielded a lucid exposition of Christian doctrine and a

LOVE AND FORGIVENESS



philosophically powerful defense of it. I myself have learned a great deal from Swinburne’s work over the years, and I am impressed not only by the depth and breadth of his work, but also by the personal courage which it exemplifies. In this chapter, I want to look at the account of the atonement1 that Swinburne gives in the first volume of his tetralogy. It is widely supposed by both Christians and non-Christians alike that the doctrine of the atonement is the distinctive doctrine in Christianity; and Christians often speak of the value of the atonement itself as infinite or at least so great as to be incommensurate with all other created goods. But, on the Christian doctrine of the atonement, what is in fact so valuable about the atonement? Although the answer to this question might seem elementary, on reflection the question turns out to be remarkably difficult to deal with. Any attempt to answer it depends on finding an answer to another, even more difficult question: On Christian doctrine, what exactly is the atonement? That is, what is it that is accomplished by the passion and death of Christ (or the life, passion, death, and resurrection of Christ)? An answer to this question, and thus also to the question about the value of the atonement, is the project of Swinburne’s book on the subject. Here it needs to be said that the doctrine of the atonement differs from other major Christian doctrines in having no formula specifying its interpretation. Although creedal or conciliar statements rule out some interpretations as unorthodox, for the doctrine of the atonement there is nonetheless no analogue to the Chalcedonian formula for the incarnation, for example. For this reason, it is possible for there to be highly divergent interpretations, all of which count as orthodox. Within the boundaries of orthodoxy, constraints on acceptable interpretations have to come from questions about fit with relevant biblical texts and 1 Different religious or scholarly communities, operating at different times and places, have understood the word ‘atonement’ in highly various ways. In this chapter, I am using the word ‘atonement’ in what I hope is its broadest and most general sense to mean whatever is accomplished by the passion and death of Christ, or by the life, passion, death, and resurrection of Christ. (The point of the disjunction is to avoid adjudicating among complicated theological disputes not relevant to the discussion, at least not relevant at the outset.) It is necessary to have some word to designate the doctrine in question, and, of the various terms available from the long history of the discussion of the doctrine, ‘atonement’ seems to me to be the least tendentious, carrying the least theological baggage. As far as I can see, this usage does not put me at odds with Swinburne’s use of the term.



ELEONORE STUMP

coherence with moral intuitions and theories. That is no doubt why the entire first half of Swinburne’s book on the atonement is devoted to a discussion of moral concepts such as guilt and forgiveness. It is helpful here to be clear about the basic theological claims that are part of the doctrine. There is little if any controversy in theological circles, now or in the history of Christian thought, over the claim that the atonement is the solution to a problem. For Augustine and the orthodox Christian tradition after him, the problem is the human proneness to sin, that is, the tendency to will contrary to the will of a perfectly good and loving God. As Augustine famously called to everyone’s attention, one hallmark of this defect is the will’s intractability to itself, its proneness to moral wrong even against its own desires for the good. This proneness to sin, this fragmentation in the will, is the heart or the source of the problem to which the atonement is the solution. So much is fairly uncontroversial. The difficulty lies in trying to explain what the passion and death of Christ, or the life, passion, death, and resurrection of Christ, have to do with this problem. What is it about Christ’s life and death that could constitute a remedy for human sinfulness? Many theologians and philosophers in the history of Christian thought have tried to answer this question. The importance of Swinburne’s book lies in its effort to do better than they have done. Although his account is Anselmian in its general outlines, it differs significantly from Anselm’s theory of atonement in numerous ways, especially in the details of its moral theory.

2 Swinburne on Atonement On Swinburne’s view, there is an ordinary human atonement that is required for the removal of guilt consequent on wrongdoing to other human beings. Swinburne takes this atonement to consist at least in repentance and apology, and usually also in reparation and penance (1989: 81). As he uses the term, penance is something which a wrongdoer gives to a person he has wronged and which is in addition to the reparation by which the wrongdoer tries to undo the damage he has done. Forgiveness of the sort that removes guilt is not possible without atonement, according to Swinburne, although in some limited circumstances it is possible for the victim of wrongdoing to forgive the

LOVE AND FORGIVENESS



wrongdoer without requiring reparation and penance in addition to repentance and apology (1989: 85). For Swinburne, it is “within the victim’s power . . . to determine, within limits, just how much atonement is necessary before he is prepared to give the forgiveness which will eliminate guilt” (1989: 86). And so, as Swinburne says, “An agent’s guilt is removed when his repentance, reparation, apology, and penance find their response in the victim’s forgiveness” (1989: 85). In the case of God and God’s forgiveness of human wrongdoing against God, Swinburne holds that God undoubtedly could forgive human sin without requiring reparation and penance. But, Swinburne says, “it is good that if we do wrong, we should take proper steps to cancel our actions . . . as far as logically can be done” (1989: 149). And so God chooses not to forgive sin without reparation and penance, too. For God’s forgiveness, full atonement is needed. The problem, then, becomes this: what possible penance could human beings give to God? That is, in addition to trying to repair whatever damage one’s wrongdoing has done, one needs to try to offer God something valuable in addition, as an extra gift. But what could this extra gift be? On Swinburne’s view, human beings are not in a position to offer God anything that would count as penance for their sins or even for the sins of others.2 In these circumstances, God himself makes the needed penance available to human beings in the life and death of Christ. On Swinburne’s account, what human beings can offer to God as an extra gift, and so as penance for their past moral wrongdoing, is the life and death of Christ. Swinburne is concerned to distinguish his view from the view of those Reformers who accepted a penal substitution theory of atonement. For Swinburne, the best way to understand Christ’s role in providing reparation and penance is to take it as a sacrifice. (In this regard, of course, Swinburne can adduce substantial biblical support.) On Swinburne’s view, the idea of a sacrifice to God is that:

2 As Swinburne puts it, “[w]e ordinary men, even if we had not sinned ourselves, would owe so much to God anyway . . . that the little extra we could do for [ourselves or for] our brethren in this way would not amount to very much. Only when I owe you nothing can I give you something” (1989: 157).



ELEONORE STUMP

God takes something valuable as a gift of reconciliation whose benefits he will often share with worshippers—like, to use a humble modern analogy, the box of chocolates which one gives to one’s host, who then offers one in return a choice from the box. The sacrifice of Christ is then Christ giving the most valuable thing he has—his life; both a lived life of obedience to God, and a laid-down life on the Cross—as a present to God, whose benefits will flow to others (1989: 152).3

And, Swinburne adds, “if, as some Reformers have claimed, God forgives men before men seek him, God would not be taking men seriously. . . . Rather, the sinner has to use Christ’s death to get forgiveness” (1989: 153). We use Christ’s death to get forgiveness when we plead it before God in this way: ‘Our life is a failure’, we may now say. ‘We have made a mess of the life which you gave us, we have made no reparation of our own for our sins. . . . But we have been given a perfect life, not owed to you, O God. We offer you this life instead of the life we should have led. . . . Take its perfection instead of our imperfection. We are serious enough about our sins to repent and apologize and to offer you back an offering of this value as our reparation and penance.’ (1989: 154–5)4

So, by doing something that counts as offering Christ’s perfect life to God, and adding her own repentance, apology, and whatever reparation she can make, a human being makes atonement to God. In consequence, God forgives her, and her guilt is removed. This, then, is Swinburne’s theory of the role that Christ’s life, suffering, and death have in solving the problem of human sinfulness. There has by now been a lot of discussion about this theory of Swinburne’s, and some of the criticisms raised with respect to it seem to me challenging for Swinburne. For example, it is not so easy to see, on 3 As Swinburne recognizes, this picture is somewhat complicated by the doctrine that Christ is also fully God. He says, “in so far as Christ is God himself, he cannot offer a sacrifice to himself. The sacrifice model has then to be somewhat transformed. God makes available the sacrifice (of himself), but it is we who have to offer it. . . . Any man who is humble and serious enough about his sin to recognize what is the proper reparation and penance for it may use the costly gift which another has made available for him to offer as his sacrifice. On this model Christ’s death has no efficacy until men choose to plead it in atonement for their sins” (1989: 153). 4 At this point, we might wonder why the needed reparation and penance have to be made by God’s becoming incarnate. In response, Swinburne says, “[T]he point of reparation is to restore the status quo as nearly as possible. Likewise, the best penance is that which more than makes it up to you in the respect in which I harmed you. . . . Since what needs atonement to God is human sin, . . . appropriate reparation and penance would be made by a perfect human life, given away through being lived perfectly” (1989: 156–7).

LOVE AND FORGIVENESS



Swinburne’s theory, why Christ’s suffering and death should be part of the penance given to God. On Swinburne’s favorite metaphor for penance, it operates as a gift that helps to make amends for wrong done. But it is not easy to see why Christ’s suffering and death should be anything that God finds a present, something he likes to have. It seems that Christ’s living a sinless life ought to be the present. Suffering and death, especially at the hands of sinful human beings bent on torture, seem an unlikely candidate for a present. And I myself would add, as another concern, that on Swinburne’s theory, Christ’s life, suffering, and death are a remedy only for past sins, for things that human beings have already done. It is not clear, on Swinburne’s theory, how Christ’s life, suffering, and death have any impact on the defect in the human will, its tendency to internal fragmentation as regards the good, that makes humans prone to sin in the first place. No doubt, Swinburne has good and useful things to say about these problems and others as well. In this short chapter, I do not want to dwell on familiar or obvious points of controversy as regards his theory. Rather, I want to look at some issues in moral psychology that tell against the theory, in my view.

3 The Nature of Love I want to begin with a Thomist account of love, because it is the best account of love I know.5 On Aquinas’s account of the nature of love, love requires two interconnected, mutually governing desires (ST II–II, q. 25, a. 3):6 (1) the desire for the good of the beloved,7 and (2) the desire for union with the beloved.8 5

I have discussed Aquinas’s account of love in detail in my (2010, ch. 5). References to Aquinas’s Summa Theologiae (ST) follow the standard division of the text into parts (I, II, III), questions (qq), articles (aa), objections (objs), and replies (ad 1, 2, etc.). 7 Cf. ST I–II, q. 26, a. 4, where Aquinas says that to love is to will good to someone. Cf. also ST I–II, q. 28, a. 4, where Aquinas explains the zeal or intensity of love in terms of the strength of a lover’s desire for the good of the beloved. 8 Cf., e.g., ST I–II, q. 26, a. 2, ad 2, and q. 28, a. 1, sed contra, where Aquinas quotes approvingly Dionysius’s line that love is the unitive force. Cf. also ST I–II, q. 66, a. 6, where 6



ELEONORE STUMP

The good of the beloved should be understood as that which truly is in the interest of the beloved and which truly does conduce to the beloved’s flourishing. What exactly this is depends, of course, on the standard of value adopted. Aquinas shares with other thinkers in the Christian tradition the conviction that personal relationship is the genus within which the greatest goods for human beings fall. For Aquinas, there is one mind and one will in God; and so God is a person, in our sense of the word ‘person.’ A union of love with God is thus a personal relationship, too; and, on Aquinas’s views, it is the greatest of personal relationships. The greatest good for human beings is to be in union of love with God. Because of the strong connection Aquinas sees between God and goodness,9 on his view, anything that contributes to the objective good for a person also brings her closer to God. The beloved’s closeness to God and her flourishing as the best person she can be will therefore be covariant, for Aquinas. So to desire the good of the beloved is to desire for the beloved those things that in fact contribute to the beloved’s flourishing, and these will also increase the beloved’s closeness to God.10 On traditional Christian doctrine, God is perfectly loving to all human beings. And so, on Aquinas’s account of love, God desires the good of each human person and union with her (e.g., ST Ia, q. 20, a. 2). That is, Aquinas’s general account of love applies to God’s love, too; but there are some things worthy of note as regards God’s love.11 In the first place, for God, the two desires of love converge. That is because the ultimate good of a human person, her ultimate flourishing, just is union with God. So in desiring the good for a human person and

Aquinas explains the superiority of charity to the other virtues by saying that every lover is drawn by desire to union with the beloved, and ST I–II, q. 70, a. 3, where Aquinas explains the connection between joy and love by saying that every lover rejoices at being united to the beloved. Also, see ST Ia, q. 20, a. 1, ad 3, where Aquinas explains that an act of love is aimed at two things: the good that one wills, and the person for whom one wills it. 9

For further discussion of this claim, see the chapters on goodness and simplicity in my (2003). 10 Since union with God is the ultimate good for each and every human person, union with God is a shareable good. A person who is in union with God is also in union with the other persons in union with God. 11 In addition to the special case of love and human persons, there are also other special cases, such as self-love, for example, that need an additional word of explanation. For further discussion of special cases, see my (2010, ch. 5).

LOVE AND FORGIVENESS



union with a human person, God is desiring the same thing, at least where the ultimate good is concerned. Second, in the case of love between human beings, when one person desires union with another, she needs to be responsive to him. Suppose, for example, that he is incurably musically illiterate (because of a righthemisphere stroke, for example), but that she is a composer. Then, if she understands his condition, it would be cruel of her, not loving, to insist on trying to share her musical creativity with him. But in God’s case things are the other way around. When God loves a human person, God’s love is not responsive to goodness in her. Instead, God’s love is the source of whatever goodness or excellence there is in her. On Aquinas’s view and on all resolutely anti-Pelagian views, goodness in creatures is derived from God directly or indirectly. Like many others in this tradition, Aquinas supposes that there is nothing good in any creature, including human persons, which is not in one way or another a gift of God, whereby God enables creatures to imitate something in God’s nature.12 On this view, in loving a human being and desiring the good of that human being, God is offering goodness to her. If she does not resist God’s love, then God’s love is productive of goodness in her, not responsive to the goodness she has already produced in herself by herself.13 For these reasons, anything good in a human person and any closeness to God on her part is brought about in her entirely by God, but it is also true that a human person is ultimately responsible for whether or not she is in union with God. Even on Aquinas’s resolutely anti-Pelagian views, because the beloved can always resist the love of God, God’s bringing about goodness in the beloved is responsive to something in the beloved.14 12

If Pelagianism is understood as the position that there can be something good in a human person which is not a gift of God’s, then Aquinas is resolutely, vehemently antiPelagian. For further discussion of these elements of Aquinas’s philosophical theology, see the chapters on faith in my (2003). 13 On Aquinas’s views, God gives grace to every person who does not reject grace, but God does not give grace to a person who rejects grace, since doing so would violate that person’s will. God’s giving of grace to a person is therefore responsive to whether that person rejects grace or not. But it is always open to a human being to reject God’s grace or to cease rejecting God’s grace. With regard to God’s giving of grace, then, alternative possibilities are always available to a human person. 14 I have argued elsewhere that God can bring about goodness in a person without thereby violating her free will, where free will is to be understood in a libertarian sense. Nothing about God’s love of a human person or God’s giving goodness to her takes away



ELEONORE STUMP

So, mutatis mutandis, Aquinas’s account of love applies also to God’s love for human beings. Although God’s love for human beings is in some respects a special case, nonetheless Aquinas’s account of love as consisting in two mutually governing desires, for the good of the beloved and for union with the beloved, holds with regard to God’s love, too.

4 Forgiveness This account of love has implications regarding the nature of forgiveness.15 Since love emerges from the interaction of two desires, for the good of the beloved and for union with her, the absence of either desire is sufficient to undermine love. To the extent to which love is implicated in forgiveness,16 the absence of either desire undermines forgiveness, too. So, for example, a resentful desire for revenge is incompatible with love of the person against whom the desire for vengeance is directed. A resentful or vengeful person does not love his enemy in virtue of the fact that he desires the bad, rather than the good, for his opponent. For that same reason, such a person does not forgive his enemy either. On Aquinas’s account of love, however, this is not the only way to fail to forgive. To forgive a person, it is not enough to have only one of the desires of love: the desire for the good of the perpetrator. It is also necessary to have a desire for union with him—if nothing else, then at least the generic desire for union that is an element of the general love for her free will. It is therefore possible both that anything good in a human person, on which closeness to God is based, is brought about in her entirely by God and also that a human person is ultimately responsible for whether or not she is in union with God. Since this is so, the love of God has to be responsive to something in the beloved, even on Aquinas’s resolutely anti-Pelagian views. See, for example, the chapter on grace and free will in my (2003). 15 For more detailed discussion of forgiveness in connection with Aquinas’s account of love, see my (2010, ch. 5). I am indebted to Jeffrie Murphy for helping me to think through issues related to forgiveness, especially issues involving retributive justice (see footnotes 23 and 25). 16 In his oral comments on this chapter, Mike Rea characterized my position as reducing forgiveness to love, but this is a mistaken characterization, in my view, at least on some common understandings of the nature of reduction. On my view, love is necessary and sufficient for forgiveness. But it may or may not be the case that forgiveness is nothing but love. Analogously, being risible is necessary and sufficient for being human—anything that is risible is human and nothing that is not risible is human—but being human is not reducible to being risible. Risibility picks out human beings by an accident which is had by all and only human beings, but the nature of human beings is not nothing but risibility.

LOVE AND FORGIVENESS



humanity. A person who desires the good of someone who has injured her but who rejects totally any desire for union with him fails to love him, in virtue of having the opposite of a desire for union with him. For that reason, she also fails to forgive him. So failure to forgive can find expression not only in resentment or vengefulness, but also in a rejection of union with the wrongdoer. One person forgives another person who has injured her only if she desires the good for him and union with him even in the face of his injustice against her. On this account of love, it is possible for a person to forgive unilaterally, without repentance on the wrongdoer’s part, because it is up to her alone whether she desires the good for him and union with him. So forgiveness lies entirely within the power of the wronged party. This is why Christ can say about those who are crucifying him, “Father, forgive them,” and this is why the father in the parable of the prodigal son exemplifies love by forgiving before he hears his son’s speech of repentance. Of course, forgiveness is to be distinguished from reconciliation in this regard. Reconciliation requires a certain mutuality; that is because the way in which the desires of love are fulfilled, or whether they are fulfilled at all, will depend crucially on the condition of the wrongdoer being forgiven. A person’s desire for the good for someone who has wronged her cannot be fulfilled if, in self-destructive impulses, he refuses the good offered him. And her desire for union with him cannot result in any kind of union with him as long as his state of character and current condition effectively keep her from being close to him. So, for reconciliation between a wrongdoer and the person he has wronged, certain states of mind and will are required in the wrongdoer, in addition to the forgiveness of the person wronged. But for forgiveness it is enough that the person wronged loves the wrongdoer. Furthermore, on traditional Christian ethical views, love is the heart and source of all morality, and it is obligatory. For any person, the absence of love is morally blameworthy, and the presence of love is necessary for moral good or excellence.17 Given the connection between love and forgiveness, it follows that forgiveness is also obligatory in the

17

For more discussion of this claim, see my (2011).



ELEONORE STUMP

same way and to the same extent. It does not follow that any given wrongdoer has a right to forgiveness from any person whom he has harmed or against whom he has committed an injustice. Rights and obligations need not be correlative in every case.18 One person can have an obligation with regard to another person even if that person does not have a correlative right with regard to her. So even though a wrongdoer has no right to forgiveness from the person he has wronged, the wronged person would be subject to appropriate moral censure if she refused to forgive him. In refusing to forgive him, she would be unloving toward him; and in being unloving, she would be worthy of moral disapprobation. Since God is perfectly loving, these conclusions about forgiveness apply to God as well. Like human persons, God can forgive a wrongdoer unilaterally, in the sense that, even without any repentance on the wrongdoer’s part, God can still desire the good for her and union with her. Since God is perfectly loving, God always desires the good for a person, which is union with her. Since failure to forgive involves failure to desire the good for a wrongdoer or failure to desire union with her, failure to forgive is also a failure to love. Since God loves every person, then God also forgives every person, whether or not she accepts either God’s love or God’s forgiveness.

5 Hatred At this point, someone might object to this conclusion because of a concern with biblical texts that seem to be about God’s hatred. It is true that the book of Wisdom claims that God loves everything and hates nothing that he has made.19 But there is a line in Malachi in which God says that he hated Esau (cf. ST Ia, q. 23, a. 3, sed contra). Furthermore, there is the Gospel text in which Christ says that on Judgment Day, he will say to some people, “I never knew you. Depart from me, you who do 18

I have argued for this claim at length in my (1992). To see the point here, consider, by way of analogy, that if Paula were a very rich tourist traveling in a very poor country, she would be obligated to give some of her money away for charitable purposes in that country if she were solicited to do so; if she refused all such solicitations, she would be subject to appropriate moral censure. But it would not be the case that any particular recipient of her donations would have a right to her money. 19 See, for example, ST Ia, q. 20, a. 2, sed contra, where Aquinas quotes with approval the line from Wisdom that God loves all things and hates nothing that he has made.

LOVE AND FORGIVENESS



evil” (Matt. 7:23; cf. Lk. 13:27). Telling some people to go away does seem the opposite of desiring union with them. And so it seems as if Christ does not love these people he sends away but rather hates them. Contrary to what I have claimed, then, it can seem on the basis of these and other biblical passages that God does not love and forgive everyone. Rather, God’s love is replaced by hatred at least sometimes, with regard to some people. But here it is worth noticing that, on Aquinas’s account of love, there are two different kinds of hatred: one of which is the opposite of love and one which is actually a species of love.20 (In this respect, hatred is like anger. Anger also has two forms: it can take the form of righteous indignation, which is compatible with love, or the form of wrath and vindictiveness, which is not compatible with love.) Hatred of both kinds will be the same in desiring what is not good (in some sense) for the beloved, or in desiring not to be united (in some sense) with the beloved, or both. The two kinds of hatred will differ, however, in their ultimate desires for the person who is the recipient of these attitudes. To see this distinction in kinds of hatred, it will help to have in mind a particular case of very serious moral wrongdoing: so consider the case of Otto Moll. Moll was at the head of the work detail managing the exterminations at Auschwitz; he participated in the gassing and cremation of very many people, and he himself shot people too weak to be herded into the gas chambers.21 A contemporary of Moll’s who desired that Moll lose his position at Auschwitz would have been desiring something which was in some sense bad for Moll. That is because it is bad for a person when he loses what he himself very much wants, and Moll wanted very much not to lose his position at the camp. On the other hand, if what a person has his heart 20 Cf., e.g., ST II–II, q. 34, a. 5. In that question, Aquinas explains that there are two kinds of hatred: one kind arises when something is hated insofar as it is incompatible with what is naturally good; the other kind arises when something is hated which is a real good but is taken to be an evil by someone who has become corrupt. This latter kind of hatred is vicious, Aquinas says, but the former kind is not. See also ST I–II, q. 29, a. 2, ad 2, where Aquinas maintains that love and hatred are contraries if considered in respect of the same thing; but if they are taken in respect of contraries, they are not themselves contrary, but rather are consequent to one another. As he puts it, it amounts to the same that one love a certain thing or that one hate its contrary. 21 For some discussion of Moll’s case, see Overy (2001: 392–3).



ELEONORE STUMP

set on is itself very bad, then it is good, all things considered, that he suffer the bad of not getting what he wants. It would have been better for Moll if he had lost his position at Auschwitz; that is, it would have been better for Moll himself if he had died as someone other than the moral monster he was enabled to become in consequence of the position he held at Auschwitz. And so, in this sort of case, to want what is in some sense bad for a person can be to want what is ultimately and objectively the good for that person. Something analogous can be said as regards union. Someone contemporary with Moll who wanted to be united with Moll during Moll’s time in power in Auschwitz, who wanted to join Moll in his evil acts and ambitions, would have been wanting what is in effect not good for Moll. Moll would have been much better off being opposed than being joined, given that his actions and desires were so evil. For all these reasons, desiring the bad for someone and desiring not to be united with that person can be a species of the desires of love provided that the person in question is bad enough that the badness of losing what he wants and the badness of the alienation of others is the best thing for him in the circumstances, and one wants these things as the best for him in the hope of ultimate union with him.22 On the other hand, there is also a kind of hatred that is the opposite of love. In the grip of this kind of hatred, one will desire what is bad in some sense for Moll and will desire to be at a distance from Moll; but one will desire these things as what is ultimately bad for Moll. That is, a person in the grip of this kind of hatred will be glad if Moll becomes worse, increasingly distant from his own flourishing and from shared union with God. This is hatred, too, but it is the kind of hatred that is incompatible with love. So one can have a kind of hatred for a person such as Moll that is one of the species of love. In the grip of that kind of hatred, one will desire 22 Cf., e.g., ST II–II, q. 34, a. 3. As Aquinas explains here, love is due to one’s neighbor on account of the nature and grace he has from God, but not on account of the sin or lack of justice he has from himself. Although one cannot hate the nature and grace in one’s neighbor without sin, it is lawful to hate the sin in one’s neighbor. In fact, it is part of love for one’s neighbor that one hate the lack of good in him, since the desire for another’s good is equivalent to hatred of his evil. In the objections in this question, Aquinas considers cases in which God is said to hate some human persons or in which Christ seems to be commanding hatred of others; and in the replies to these objections Aquinas explains these cases as ones in which hatred is in fact a kind of love.

LOVE AND FORGIVENESS



what is bad (in some sense) for Moll and distance from Moll, but one will also desire a change of heart in Moll. Moll’s ultimate flourishing will also be something one wants. For these reasons, one will want Moll ultimately to have as much real good as he can, and one will desire as much shared union with Moll as Moll’s condition allows. In the kind of hatred at issue here, one’s hatred of Moll is a kind of love of him. But, on the other hand, one can also have a hatred of Moll that makes one desire, ultimately, the very opposite of the good for Moll and union with Moll. This hatred is incompatible with love. Forgiveness requires one’s maintaining toward a wrongdoer the desire for his good and the desire for union with him, but both these desires of love have to be responsive to things in the beloved. If the beloved is intransigent in his moral wrongdoing, then the desires of love might be restricted to the desires that constitute a kind of hatred, provided that they are encompassed within a desire for the beloved’s reform and for union with him. Encompassed in this way, the desires of this kind of hatred are a species of love. Without such an encompassing desire, the hatred in question is opposed to love. The two biblical lines about God, that he loves everyone and that he hates Esau, are therefore compatible if the hatred at issue in the statement about Esau is understood as the kind that is a species of love. As for the Gospel text, when Christ expresses a desire that on Judgment Day some people go away, the reason he gives for this desire is implied in his characterization of them: “you who do evil.” Those who do evil are those who have persisted in rejecting the divine love that would have produced in them the good missing when they do evil. Because they finally reject God’s love, the hatred that is one of the species of love is the only kind of love with which God can love them. It is crucial to see, however, that this result is a function only of the rejection of love on the part of human beings, and that even it does not result in a loss of love or forgiveness on God’s part. It results only in God’s having a desire for union that cannot be fulfilled. If God tells such people to go away, it is an expression of God’s recognition of the impossible. It is impossible even for God to unite himself in love to someone who rejects him. So the impossibility of a fulfillment of the desires of love on God’s part for some human persons co-exists in God with the desire for union with those persons. God loves and forgives those he tells to go away. If they were to want not to go



ELEONORE STUMP

away from God, his love and forgiveness would bring them to union with him. And so the preceding conclusion stands: God always loves and forgives every person, even those who reject God’s love wholly. A fortiori, repentance, apology, reparation, and penance are not required to get God’s forgiveness. Consequently, whatever Christ’s life and death bring about, it is not the case that they are needed as a component of a process requisite for forgiveness on the part of God.

6 Retributive Punishment Given the importance of forgiveness in accounts of atonement, it is necessary to say something briefly about retributive punishment in order to round out the discussion of forgiveness. The heart of the notion of justice in the Western tradition influenced by Aristotle is the idea that justice consists in giving to each person what is his due, that is, what is owed to him or what counts in some way as his own. To a person in favor of retributive punishment in the interests of justice, it may seem as if Aquinas’s account privileges love over justice in a morally inappropriate way, because it mandates forgiveness. On the view of such an objector, the wrong a person does might make it morally necessary to maintain toward that person the attitudes of anger or hatred that are not part of love, because what the wrongdoer deserves, what is his due, is only retribution. On this view, it is wrong to suppose that forgiveness is always obligatory, even if the wrongdoer has no right to it. On the contrary, it can be obligatory to fail to forgive, in the interest of justice. For the putative objector, Aquinas’s views of love and forgiveness do not take moral wrongdoing seriously enough. If Jerome has committed a serious enough moral wrong against Paula, then Paula should neither forgive him nor love him. Instead, in the interest of justice, she should seek only retributive punishment for him. To do anything else is to make the world a worse place than it might otherwise be in virtue of in effect condoning, by inappropriately forgiving, moral wrongs.23

23

For a well-argued statement of such a case, see, e.g., Murphy and Hampton (1988) and Murphy (2003).

LOVE AND FORGIVENESS



To evaluate these claims as an objection to Aquinas’s account of love and forgiveness, it is crucial to understand whether retributive punishment is a good for the person punished. That is because if the putative objector believes that retributive punishment would be a good for the person punished, then he has no quarrel with Aquinas’s views of love and forgiveness. On Aquinas’s views, as I have explained, a person could count as loving and forgiving of someone even if he also desired punishment, even retributive punishment, for that person, provided that he desires the punishment as a good for the other. To constitute an objection to Aquinas’s account of love and forgiveness, then, the objection must include the supposition that retributive punishment is not a good for the person punished. So we need to include this supposition in the objection. But then we also need to examine the supposition. For this purpose, it helps to return to the case of Moll. On the supposition the putative objector accepts, retributive punishment is not a good for the person punished. That is, for the objector, it would be better for Moll if he escaped retributive punishment; Moll would be better off if he were not judged and imprisoned or otherwise punished for the acts that he did. In the view of the objector, it would be better for Moll if he were to live out the remainder of his life without punishment. But this view seems to me false. To see that it is false, consider the period between the time of Moll’s wrongdoing and the time when retributive punishment might have been imposed. There are three possibilities for Moll in this period: Moll could be repentant; or, if he were unrepentant, he could be capable of repentance or incapable of it. So let us take these possibilities in turn and look at the objector’s supposition in connection with each of them. If Moll were repentant during the period in question, then presumably he himself would welcome punishment as a way in which he could do something, no matter how limited, to make amends for his evil. Albert Speer presented himself as a repentant Nazi at the Nuremberg trials, and it was part of the repentance he expressed that he willingly accepted the punishment assigned him at his trial.24 And, in general, it does seem true

24 I put the point in this awkward way to withhold judgment about whether Speer was in fact really repentant. Although many people believe he was, there is strong evidence to support a claim that his repentance was either insincere or double-minded. See, e.g., Van der Vat (1997).



ELEONORE STUMP

that a person genuinely and deeply repentant for serious moral wrongdoing will see his punishment as appropriate, as a good he himself wants, and so as a good for himself. He will be more pained over his wrongdoing and more tormented by his guilt if there is nothing he does to make up for it. And this attitude seems not only understandable, but also in fact reasonable and right. Consequently, for cases of repentant wrongdoers, the putative objector’s supposition that retributive punishment is not good for the punished person is false, and seeking retributive punishment for a repentant Moll is compatible with the first desire of love. Suppose, then, that during the period in question, Moll were unrepentant but capable of repentance. As Moll’s own testimony during his interrogations makes clear, engaging in morally horrific acts such as those Moll did takes a terrible psychic toll on a person. As Moll described himself, he suffered a nervous breakdown after his work at Auschwitz and continued to endure an afflicting mental illness afterwards in consequence of his Auschwitz experiences. If there is any possibility of peace and psychic healing for a person such as Moll, it seems much more likely to come if Moll is repentant than if he is not. But it is also plausible that such a person is much more likely to find repentance if he is compelled through judgment and the imposition of punishment to accept the verdict of the rest of the world on him. The problem for Moll in his unrepentant state is that he is still in denial about the evil he did and still determined to validate his morally monstrous actions as morally acceptable. As the story of the Nuremberg trials shows, being forced to go through that trial and face its verdict and punishment can bludgeon the denial of wrongdoing on the part of even such hardened perpetrators as the Nazi elite. And so, understood in this way, the judgment and punishment of a wrongdoer such as Moll are a good for him, even if he lacks repentance, provided only that he is capable of it. So in this case too, the objector’s supposition is false, and seeking punishment for an unrepentant Moll is compatible with the first desire of love. So consider the third and final case. Suppose that Moll is unrepentant and, for the sake of this argument, stipulate that Moll is incapable of repentance, for one reason or another. Even so, it remains the case that Moll can become morally worse. Increase in moral depravity is itself an evil for Moll, and it will take a further toll on Moll’s psyche. Insofar as

LOVE AND FORGIVENESS



punishment restricts Moll and reins him in, it helps to prevent in him the additional moral evil attendant on his holding the view that he can do horrific acts with impunity. If Germany had not lost the war, if Moll had not been captured and brought to judgment, who does not suppose it likely that Moll would have sunk even deeper into evil and concomitant psychic pathology? And so, even if judgment and retributive punishment do not make Moll better, they can keep him from getting worse. In this case also, then, retributive punishment of Moll can be seen as a good for Moll and so as compatible with the first desire of love. For these reasons, the supposition of the putative objector that retributive punishment is not a good for a person such as Moll is mistaken.25 But since this supposition is mistaken and retributive punishment can be a good for the wrongdoer, then the objection to Aquinas’s views of love and forgiveness is mistaken as well. It is possible to take justice very seriously, it is possible even to hold that imposing retributive punishment on a wrongdoer is sometimes obligatory, and still to love and forgive even wholly unrepentant wrongdoers, on Aquinas’s account of love.

7 God’s Justice At this point, an objector might suppose that, even if considerations of retributive punishment do not militate against the connection between love and forgiveness I am arguing for, considerations of God’s justice do. For the putative objector, God’s justice requires that God not love some people, the unrepentant evildoers such as Otto Moll, for example. On the putative objector’s view, God’s justice requires that God hate such people with the hatred that is the opposite of love. There is, of course, some question about how to understand God’s justice toward a human person, if justice is a matter of giving every person his due but everything good in a human person is already a product of grace given him by God. Aquinas recognizes, however, that God’s justice can also be a matter of something’s being owed to God. And so the objector’s position can be formulated this way: God owes it to 25

In this connection, see also Jeffrie Murphy’s reasons for softening his original stern stand about retributive punishment (Murphy 2012).



ELEONORE STUMP

himself not to love the unrepentant perpetrators of serious wrongdoing, but to hate them with the hatred that is the opposite of love. It is reasonable to suppose that the justice of God requires that God have what is owed to God. But, Aquinas thinks, what is owed to God is that God should have what his perfectly good will wants. As I have been at pains to show, however, God’s perfectly good will is the will with which God loves everything that he has made, unrepentant wrongdoers included. Since love includes a desire for the good of the beloved, what is owed God’s justice is therefore that God should have what his love desires. And this is that God’s goodness should be given to his creation to the maximal extent possible. God’s justice is therefore not a response to the merits in a human person that she has acquired for herself. On all anti-Pelagian views, there are no such merits. On the contrary, God’s justice is God’s having what is owed to God; and that is God’s having God’s will, which is the will of love. And so God’s justice is encompassed in God’s love; it is not in any way opposed to it. In fact, Aquinas maintains, God’s justice is just one part of God’s goodness and love; God’s mercy is another. Like God’s justice, God’s mercy is a matter of giving goodness to human persons. God’s mercy perfects God’s justice by expelling from a human person defects that keep that person at a distance from God if only the human person does not reject that grace of God. And so God’s justice is not opposed to God’s mercy either. God’s justice and mercy, like God’s love generally, always seek the good of every human person. And that good is union with God (ST Ia, q. 21, a. 1, ad 3 and ad 4, and q. 21, a. 30).

8 An Objection: Forgiveness and Reconciliation At this point, someone might object that the difference between the view of forgiveness I have argued for and Swinburne’s view is just terminological. On Swinburne’s account, a wronged person’s ability to forgive is dependent on the decisions of the wrongdoer. If the wrongdoer decides not to make atonement, his victim cannot forgive him. On the position I have defended, contrary to Swinburne, forgiveness is entirely within the power of the wronged person. We can forgive people who are still not

LOVE AND FORGIVENESS



speaking to us, we can forgive our parents long after they are dead, and we can be commanded to forgive our enemies (as distinct from being commanded to forgive those people who were our enemies but have now repented, apologized, made reparation, and brought a present). But, the putative objector will claim, we speak of forgiveness ambiguously. Sometimes—especially when we think about forgiveness from the point of view of the person wronged—we mean by ‘forgiveness’ the wronged person’s putting away all resentment or alienation with respect to the wrongdoer. In that sense, it is perfectly possible for a victim to forgive the person who has wronged her no matter what he decides to do. But, on the other hand, sometimes—especially when we think about forgiveness from the point of view of the person committing the wrong— we mean by ‘forgiveness’ the restoration and healing of a broken relationship. In this sense of ‘forgiveness,’ forgiveness is not entirely within the power of the wronged person; it does depend on the decisions and actions of the wrongdoer as well, at least on his repentance. So, the objector will conclude, we can distinguish forgiveness alone from forgiveness-plus-reconciliation. If we take Swinburne to be talking about forgiveness-plus-reconciliation, then it seems that he is right in thinking that something is needed on the part of the wrongdoer. Forgiveness-plus-reconciliation cannot be achieved unilaterally by one person, not even God, because reconciliation requires a kind of mutuality between two persons. In response, I would say that the putative objector’s conclusion is no doubt right, but it does not lend support to Swinburne’s theory of atonement. On the understanding of forgiveness as forgiveness-plus-reconciliation, nothing is requisite for God’s share of the process, the forgiveness alone, so if something is needed for forgiveness-plus-reconciliation, it is needed for the reconciliation part. Consequently, if we are taking forgiveness as forgiveness-plus-reconciliation, Christ’s life and death should not be understood as a present offered to God in order to help bring about God’s forgiveness. On the understanding of forgiveness as forgiveness-plus-reconciliation, what is left for Christ’s life and death to influence is the process by which human beings come to repentance (and the other parts of reconciliation). For these reasons, a theory of the atonement focused on forgiveness-plus-reconciliation rather than on God’s forgiveness alone would need to explain the influence of Christ’s life and death in changing human wills. However such a theory might be



ELEONORE STUMP

spelled out, it would be a very different theory from the one Swinburne has argued for.

9 Conclusion On Aquinas’s account of love, which is the best I know, the central claims of Swinburne’s theory of atonement are ruled out. God’s love embraces every human being, and so God always desires union in love with every human being. Because failure to forgive is unloving, God also forgives each human being. Consequently, God can and does forgive unilaterally, just as he can and does love unrequitedly. It is true that, like human beings, God cannot have what he wants in love or forgiveness just by wanting it. In order for there to be union of love between God and a human person, she has to be willing to accept God’s love and forgiveness, too. But these are good things in her, and since they are good, they are the effects of God’s grace working in her, a grace that God, in love, will give to each person who does not reject it.26 To suppose that repentance, apology, reparation, and penance are a prerequisite to God’s forgiveness is to suppose that until a wrongdoer has accomplished these things, God does not forgive her. But if God does not forgive her, God does not love her either. To suppose, then, that God does not forgive until full atonement has been made to him by a wrongdoer is to suppose that, until it has been made, God is not God. Furthermore, on anti-Pelagian views, any good—in repentance and apology or the payment of reparation and penance—is itself a gift of God’s love. For Aquinas, the good of anything that makes amends for moral wrongdoing therefore comes in consequence of God’s love and forgiveness. It cannot be a condition for them. So although there is

26 The rejection of Pelagianism does not imply that Paula is just a puppet in God’s hands. It is open to Paula to reject grace or to cease rejecting grace; and God’s giving of grace to Paula is responsive to which of these states Paula is in. Since this is so, Paula is ultimately responsible for her state and condition even though God is responsible for all the good in Paula. By the same token, if there is little good in Paula, the person ultimately responsible for this state of Paula’s is only Paula. The good will in Paula that would bring Paula into union with God in love would have been produced in Paula by God if Paula had not rejected it. So God’s love, like human love, has to be responsive to something in the loved person, even though, unlike human love, God’s love is responsible for all the good that there is in the beloved human person.

LOVE AND FORGIVENESS



undoubtedly a point to repentance, reparation, and penance, it cannot be the removing of an obstacle in God to forgiveness of the wrongdoer. It is true, even on Aquinas’s account of love, that God will punish some perpetrators of moral wrongdoing and that some perpetrators are even punished with the worst thing for human beings: endless exclusion from union with God. There are some human persons whom God hates. But this hatred is a species of love and not opposed to it, and those punished and hated in this way are not those to whom God is unwilling to give forgiveness. They are those to whom forgiveness is continually offered by the infinite and unstoppable love of God, which never fails to embrace even those punished and hated. Those who are hated and rejected are so not because they are not loved and forgiven, but rather because they themselves will not cease rejecting the love and forgiveness of God, which are always there for them.27 And so, although we can be grateful to Swinburne for his careful and reflective construction of his theory of atonement, it seems to me that we still lack a workable explanation of Christ’s role in solving the problem of human sinfulness.28

References Murphy, Jeffrie. (2003). Getting Even—Forgiveness and Its Limits. Oxford: Oxford University Press. Murphy, Jeffrie. (2012). Punishment and the Moral Emotions. Essays in Law, Morality, and Religion. Oxford: Oxford University Press. Murphy, Jeffrie and Hampton, Jean. (1988). Forgiveness and Mercy. Cambridge: Cambridge University Press. Overy, Richard (ed.). (2001). Interrogations. The Nazi Elite in Allied Hands, 1945. New York: Viking Penguin. Stump, Eleonore. (1992). “God’s Obligations,” in James Tomberlin (ed.), Philosophical Perspectives 6. Atascadero, CA: Ridgeview Publishing, 475–92. Stump, Eleonore. (2003). Aquinas. London: Routledge. Stump, Eleonore. (2010). Wandering in Darkness. Narrative and the Problem of Suffering. Oxford: Oxford University Press. 27 The only sin God does not forgive, Aquinas says, is final impenitence, that is, a refusal of love that never ends. See, e.g., De Malo, q. 3, a. 15. 28 I am grateful to Michael Rea for his helpful comments on an earlier draft of this chapter.



ELEONORE STUMP

Stump, Eleonore. (2011). “The Non-Aristotelian Character of Aquinas’s Ethics: Aquinas on the Passions,” Faith and Philosophy, 28 (1): 29–43. Swinburne, Richard. (1977). The Coherence of Theism. Oxford: Oxford University Press. Swinburne, Richard. (1979). The Existence of God. Oxford: Oxford University Press. Swinburne, Richard. (1981). Faith and Reason. Oxford: Oxford University Press. Swinburne, Richard. (1989). Responsibility and Atonement. Oxford: Oxford University Press. Swinburne, Richard. (1991). Revelation. Oxford: Oxford University Press. Swinburne, Richard. (1994). The Christian God. Oxford: Oxford University Press. Swinburne, Richard. (1998). Providence and the Problem of Evil. Oxford: Oxford University Press. Van der Vat, Dan. (1997). The Good Nazi: The Life and Lies of Albert Speer. New York: Houghton Mifflin.

8 The Liturgical Present Tense Nicholas Wolterstorff

Contemporary analytic philosophy of religion has concentrated on four topics: the epistemology of religious belief, the nature of God, the problem of evil, and the nature of religious experience. If someone who knew nothing about the religious life formed conclusions about such a mode of life from reading this literature, she would come to a very distorted view. She would come to the view that, apart from the mystical experiences that some people have, the religious life is all about belief. She would have no inkling of the fact that communal worship is central in the lives of most adherents of almost all religions. Analytic philosophers of religion have paid almost no attention to communal worship and liturgy.1 On this occasion I will refrain from addressing the question, why is it that what analytic philosophers of religion have found important to discuss is so out of accord with what religious people typically find important in the practice of their religion? Instead, I will break out of the mold and reflect philosophically on a feature of the Christian liturgy that I have long found both fascinating and baffling, namely, the use on high holy days of the present tense to sing and speak about events in the past. Let me call this usage, the liturgical present tense. Let us have a few examples.

1 Examples of the Liturgical Present Tense The first verse of Charles Wesley’s Christmas hymn, “Hark, the Herald Angels Sing,” goes like this: 1

The main exception to this generalization is that there have been a few essays discussing the metaphysics of the Eucharist.



NICHOLAS WOLTERSTORFF

Hark, the herald angels sing, “Glory to the new-born king. Peace on earth and mercy mild, God and sinners reconciled.” Joyful all ye nations rise, Join the triumph of the skies, with the angelic hosts proclaim, “Christ is born in Bethlehem.” Hark, the herald angels sing, “Glory to the new-born king.”

Here is the first verse of a Christmas hymn by Paul Gerhardt, in English translation: All my heart this night rejoices as I hear, far and near, sweetest angel voices: “Christ is born,” their choirs are singing, ‘til the air everywhere now with joy is ringing.

The first verse of a hymn for Holy/Good Friday by Bernard of Clairvaux goes as follows, in English translation: O sacred head, now wounded, With grief and shame weighed down, Now scornfully surrounded, With thorns thine only crown. O sacred head, what glory, What bliss till now was thine! Yet, though despised and gory, I joy to call thee mine.

The first verse of an Easter hymn by Charles Wesley goes like this: “Christ the Lord is risen today,” Alleluia! Sons of men and angels say; Alleluia. Raise your joys and triumphs high; Alleluia. Sing, ye heavens, and earth reply. Alleluia!

It is likely that, when Wesley composed those lines, the old Latin hymn, Surrexit Christus Hodie, was ringing in his ears. Its first verse, in English translation, is this: Jesus Christ is risen today, Alleluia! Our triumphant holy day, Alleluia!

THE LITURGICAL PRESENT TENSE



Who did once, upon the cross, Alleluia! Suffer to redeem our loss. Alleluia!

It is characteristic of hymns in the liturgical present tense to insert such indexicals as “now,” “today,” “this night,” “this day,” and “this happy morn,” as if to make doubly sure that we do not miss the point. What is the point? How are we to understand the use of the liturgical present tense? If the use of the present tense in these hymns is to be taken literally, then those who sang the Christmas hymns on Christmas Day 2013 were declaring that Christ was born earlier that very day, and those who sang the Easter hymns on Easter 2014 were declaring that Christ arose earlier that day. If that is what they were declaring, did they also believe it? Did they believe that Christ was born on Christmas Day 2013 and arose on Easter 2014? Those who sing these hymns believe that Christ was born, died, and rose well in the past. They are reminded of this when Scripture is read. That is what they declare they believe when they recite the Apostles’ or Nicene Creed. They use the past tense: “He was conceived by the Holy Spirit, born of the Virgin Mary”; “He came down from heaven and was incarnate by the Holy Spirit of the Virgin Mary.” So if they also believe that Christ was born on Christmas Day 2013, then they believe that an event in the distant past became present on that day—that it happened again, that it was reactualized. What could possibly bring about something so extraordinary as that? Before I address the question of how the liturgical present tense should be understood, let us take a look at the Orthodox liturgy for Vespers on Holy Friday; here we find a density in the use of the tense that goes beyond anything that we find in the hymns commonly sung in Catholic or Protestant churches.2 To forestall ambiguity in references, let us suppose that we are discussing the enactment of the liturgy that took place on Holy Friday 2014. Early in the liturgy the choir sings, “Today the most pure virgin saw thee hanging on the cross . . . ; and with a mother’s love she wept and bitterly her heart was wounded. She groaned in anguish from the depth 2 The liturgy can be found in The Lenten Triodion, trans. in Ware and Mary (1978). Page numbers in the text refer to this edition. The liturgy can be found in The Lenten Triodion, trans. by Mother Mary and Archimandrite Kallistos Ware (London: Faber and Faber, 1978, p. 611 ff.). I have removed many of the capitals that are found in the text.



NICHOLAS WOLTERSTORFF

of her soul, and in her grief she struck her face and tore her hair” (612). Here the crucifixion is sung about as if it had happened earlier on Holy Friday 2014. Just a bit later the crucifixion and the events immediately preceding it are sung about as happening simultaneously with the enactment of the liturgy. Today the Master of Creation stands before Pilate; today the maker of all things is given up to the cross, and of his own will he is led as a lamb to the slaughter. He who sent manna in the wilderness is transfixed with nails; his side is pierced, and a sponge with vinegar touches his lips. The deliverer of the world is struck on the face, and the creator of all is mocked by his own servants (612).

The tense then switches abruptly to the past tense: “He prayed to his father, saying: ‘Forgive them this sin’ ” (612). Then, shortly, back to the present tense: “A dread and marvellous mystery we see come to pass this day. He whom none may touch is seized; he who looses Adam from the curse is bound. He who tries the hearts and inner thoughts of man is unjustly brought to trial” (613). Then, a bit later, back to the past tense: “Down from the tree Joseph of Arimathaea took thee dead, who art the life of all, and he wrapped thee, O Christ, in a linen cloth with spices. Moved in his heart by love, he kissed thy most pure body with his lips” (614). This is, to say the least, extraordinary—extraordinary in the employment of paradox, extraordinary also in the use of the present tense. And extraordinary in the seamless weaving back and forth between the present and the past tense; such weaving back and forth is not typical of the Good Friday hymns sung in Catholic and Protestant churches. Any theory as to the significance of the use of the present tense in liturgical contexts will have to enable us to explain why the Orthodox liturgy would weave back and forth between the present tense and the past. We can safely assume that it is not because the Orthodox are careless or insecure in their use of tenses! The passages just quoted are followed by passages in which words in the present tense are put into the mouths of Joseph of Arimathaea and of Mary, mother of Jesus. The passages are extraordinary in the intensity of the emotions expressed. Let me quote the passage in which Mary is speaking: The pure virgin mother wept as she took him on her knees; her tears flowed down upon him, and with bitter cries of grief she kissed him. “My son, my lord and

THE LITURGICAL PRESENT TENSE



God, thou wast the only hope of thine handmaiden, my life and the light of mine eyes; and now, alas, I have lost thee, my sweet and most beloved child. . . . I see thee, my beloved child, stripped, broken, anointed for burial, a corpse. . . . In my hands I hold thee as a corpse, O loving lord, who has brought the dead to life; grievously is my heart wounded and I long to die with thee. . . . I reflect, O master, how never again shall I hear thy voice; never again shall thy handmaiden see the beauty of thy face as in the past; for thou, my son, hast sunk down before mine eyes. . . . Release me from my agony, and take me with thee, O my son and God. . . . Leave me not to live alone” (618–20).

Though the present tense is prominent in this passage, its use here is not an example of what I have in mind by “the liturgical present tense,” but rather occurs within what are presented as quotations of what was said long ago, these quotations being introduced by words in the past tense.

2 The Reactualization Interpretation A view that was commonly espoused by liturgical scholars in the twentieth century is that, by the enactment of the liturgy, some of the central events narrated in Scripture are reactualized. Christ’s birth was reactualized on Christmas Day 2013, as it has been on every other Christmas Day; his resurrection was reactualized on Easter 2014, as it has been on every other Easter. Let me call this the reactualization interpretation of the liturgy. I am not aware of any defender of the reactualization interpretation who explicitly presents his view as an explanation of the liturgical present tense, but I assume that, in part at least, they all see themselves as articulating the understanding implicit in the liturgy and its accompanying hymnody when these employ the liturgical present tense. So let us begin our inquiry by getting a grip on this interpretation and then assessing its tenability. Philosophers are familiar with the distinction between event-types and event-tokens: event-types are universals; event-tokens are particular instances of event-types. Reactualization theorists are not claiming that the event-type, Christ being born, is reinstantiated every Christmas Day. That claim, taken literally, would be remarkable. What they are claiming is not that, but, rather, that the event-token that was Christ’s birth is reactualized every Christmas Day, made present again. They are claiming that by enacting the liturgy we reactualize certain event-tokens from the past. This, too, is a remarkable claim. As with remarkable claims in



NICHOLAS WOLTERSTORFF

general, one wants to be doubly sure that we are not misinterpreting the words of those who seem to be making the claim. So rather than just stating the view in my own words, let me allow a few proponents of the view to state it in their words. The most influential exponent of the reactualization interpretation was probably the Catholic liturgical theologian Dom Odo Casel, who called it the “mystery” conception of the liturgy. His mystery conception of the Christian liturgy was apparently suggested to him by his prior study of archaic rituals. About those rituals he says this: for the ancients . . . the play is a sacred thing. Men carry out its action in a way all can see; but the real actors are the gods who dwell at the feast. They are the ones through whom men fulfill the actions which they do. For this reason what the play represents is made real in the deepest fashion (Casel 1962: 159).

Casel holds that the relation between events in the biblical narratives and enactments of the Christian liturgy is to be interpreted along somewhat similar lines. Here are a few passages in which he presents his interpretation. His description of certain events in the biblical narratives as “primeval” is presumably meant to suggest a similarity between the biblical narratives and archaic myths. When the church on Septuagesima Sunday begins her reading of the book of Genesis, she brings the primaeval fact to life: creation, the Fall and all the rest. . . . With the reading of the Scripture we return to the first age; we place ourselves into the primaeval act which is made present (Casel 1962: 124). Still more strongly do we perceive the power of the primaeval saving act made present, in the Eucharist; by the transformation and the consumption of the bread and wine man is filled with the power of Christ. He returns to that primaeval force with which God gave life to the world in the death and resurrection of Christ. Man’s action in the rite is made one with God’s action (Casel 1962: 125). In the view of some theologians Christ placed the grace which he had earned by his sacrifice in the sacraments; when the sacrament is performed those who perform it receive the grace it contains. The sacrament is the cause of grace. In the mystery conception it is the primaeval saving act which is made present. . . . The myth is lived out in worship; the rite is living myth (Casel 1962: 124–5). Worship is the means which brings back the Origin; in it the new Beginning is made present for Christians. What he experiences in his worship is not only an aftereffect of the saving act; the saving act itself takes on presence (Casel 1962: 128). The mystery is no mere recalling of Christ and his saving deed; it is a memorial in worship. The church does what the Lord did, and thereby makes his act present (Casel 1962: 141).

THE LITURGICAL PRESENT TENSE



Five times over Casel speaks of some primeval saving act as “made present” when the liturgy is enacted. In one passage he speaks of a primeval saving act as “taking on presence.” I discern no clue indicating that we are to interpret him as speaking figuratively. Whereas Casel moved from a study of archaic rituals to liturgical theology, the Reformed liturgical scholar J.-J. von Allmen confined his writing to liturgical theology. In his Worship: Its Theology and Practice (1965) he says that liturgy recapitulates the biblical narratives and, more generally, recapitulates the history of salvation. He explains that what he means by saying that the liturgy “recapitulates” the history of salvation is that it “sums up and confirms ever afresh the process of saving history which has reached its culminating point in the intervention of Christ in human history” (von Allmen 1965: 33). But then, in the course of developing this summing-up idea, he employs the language of reactualization rather than the language of recapitulation. He describes an enactment of the liturgy as an anamnesis, a memorial. And he explains as follows what he understands an anamnesis to be: “It is a restoration of the past so that it becomes present . . . ” (von Allmen 1965: 34). It is “a real actualization of the past in the present . . . ” (von Allmen 1965: 34). “The past is reenacted and becomes present” (von Allmen 1965: 34). I judge that von Allmen’s use of the language of reactualization is the more significant for having been almost inadvertent; this way of speaking and thinking about the liturgy was in the air at the time. Mircea Eliade was an anthropologist of religion, not a liturgical scholar. On occasion, however, he too turned his attention to the Christian liturgy, and when he did, he, like Casel, interpreted what he saw through the lens of his interpretation of archaic rituals. Eliade emphasized that archaic persons did not regard their myths as mere stories; the myths relate what really happened. “Myth tells how, through the deeds of Supernatural Beings, a reality came into existence, be it the whole of reality, the Cosmos, or only a fragment of reality—an island, a species of plant, a particular kind of human behavior, an institution” (Eliade 1963: 5). Though archaic persons regarded their myths as narrating what really happened, they thought of those events as occurring in another time, in mythic time rather than historical time. To that time, our time bears no relation; the events that occur in mythic time bear no temporal relations to events that occur in historical time. The thought of trying to find out when, with respect to historical



NICHOLAS WOLTERSTORFF

events, the events in mythic time occurred would have made no sense to archaic persons. Eliade’s central thesis concerning the understanding by archaic persons of their rituals is that they regarded themselves, when performing the rituals, as entering into the mythic time of origins. They regarded themselves as doing the things narrated in the myth and doing them in mythic primordial time. The participants in the festival become contemporaries of the mythical event. In other words, they emerge from their historical time—that is, from the time constituted by the sum total of profane personal and intrapersonal events—and recover primordial time, which is always the same, which belongs to eternity. Religious man periodically finds his way into mythical and sacred time, re-enters the time of origin, the time that “floweth not” because it does not participate in profane temporal duration, because it is composed of an eternal present, which is indefinitely recoverable (Eliade 1959: 88).

In ritual, the events narrated in the myths are reactualized.3 The view that Eliade attributes to archaic persons concerning the relation between myth and reality could not possibly be true. On one interpretation of the view, a human being, when participating in the ritual, becomes a participant in what took place in the mythic time of origins. For that to happen, that human being has to be, for a time, one of the personae of the myth. But in the myths there are no human beings, only gods: “The actors in myths are supernatural beings” (Eliade 1963: 6). On another interpretation, a human being, when participating in the ritual, loses his identity as a human being for a time and is one of the gods in that mythic time of origins. But we who are human beings cannot lose our identity as human beings; it is essential to us.4 The fact that the view could not possibly be true is not a reason for concluding that archaic people did not hold it. It is, however, a reason for requiring strong evidence that they held it. Is Eliade perhaps interpreting too literally what archaic people said about their rituals and what they said when performing them? He seems to allow them no figures 3 Cf. Eliade (1959: 105–6): “The religious festival is the reactualization of a primordial event, of a sacred history in which the actors are the gods or semidivine beings. But sacred history is recounted in the myths. Hence, the participants in the festival become contemporaries of the gods and the semidivine beings. They live in the primordial time that is sanctified by the presence and activities of the gods.” 4 In my (1990: 125–6), I take note of some additional impossibilities and puzzles.

THE LITURGICAL PRESENT TENSE



of speech. Since I am not a scholar of archaic religions, I can do no more than register my doubts. After discussing the archaic understanding of mythic time, Eliade observes that “interest in the ‘irreversible’ and the ‘new’ in history is a recent discovery in the life of humanity.” “Archaic humanity . . . defended itself, to the utmost of its powers, against all the novelty and irreversibility which history entails” (Eliade 1974: 48). In the rituals of the archaic person we see his “refusal to grant value to memory and hence to the unusual events (i.e., events without an archetypal model) that in fact constitute history. In the last analysis, what we discover in all these rites and all these attitudes is the will to devalue time . . . ” (Eliade 1974: 85–6). Eliade argues that the discovery of the new and irreversible in history is to be credited to the interpretation and valorization of history as theophany that we find in ancient Israel. In Judaism, he says, Yahweh no longer manifests himself in cosmic time (like the gods of other religions) but in a historical time, which is irreversible. Each new manifestation of Yahweh in history is no longer reducible to an earlier manifestation. . . . Hence the historical event acquires a new dimension; it becomes a theophany (Eliade 1959: 110–11).

When Eliade looks at the Christian liturgy, he thinks he sees some of the same phenomena that he identified in archaic rituals; he interprets the presence of those phenomena in the Christian liturgy as remnants from archaic times. Narratives are recited; the participants leave behind “profane” time and enter the “sacred” time of the narrated events; the narrated events are reactualized; etc. The “Christian liturgical year,” says Eliade, “is based upon a periodic and real repetition of the Nativity, Passion, death, and Resurrection of Jesus, with all that this mystical drama implies for a Christian; that is, personal and cosmic regeneration through reactualization in concreto of the birth, death, and resurrection of the Savior” (Eliade 1974: 130). There is a big difference, however: “when a Christian of our day participates in liturgical time, he recovers the illud tempus in which Christ lived, suffered, and rose again.” So far, a similarity to archaic rituals. But the illud tempus “is no longer a mythical time, it is the time when Pontius Pilate governed Judaea” (Eliade 1959: 111). That time is historical time. Instead of speaking of events from the past as being “made present,” as Casel does, Eliade speaks of them as being reactualized; he says that there



NICHOLAS WOLTERSTORFF

is a “real repetition” of those events. (Von Allmen says both that they become present and that they are reactualized.) The ideas are slightly different: made present versus made actual again. But if an event from the past is made present, then it is reactualized, and if it is reactualized, then it is made present. So the two views are equivalent. Eliade does appear to part ways from Casel when he speaks of the participants in the Christian liturgy as leaving behind profane time and entering sacred time; Casel does not speak of leaving behind one time and entering another. But surely Eliade’s use of the metaphors of leaving and entering has to be judged as misleading; he himself emphasizes that Christians regard the time of the events that they hold sacred as historical time. His thought is not that participants in the liturgy see themselves as entering some non-historical time in which the events they hold sacred occurred; his thought is that certain sacred events that occurred in the historical past are regarded as becoming actual again. I mentioned that the reactualization interpretation was common among liturgical scholars in the twentieth century; it lives on into the twenty-first century. In the Introduction to her translation of six “festal orations” by Gregory of Nazianzus (Harrison 2008),5 the church historian and liturgical scholar Nonna Verna Harrison quotes three passages from Gregory’s sermons that she interprets as expressing the reactualization view; from her comments it is clear that that is also her own view. Here are the three passages from Gregory: Christ is born; give glory; Christ is from the heavens, go to meet him; Christ is on earth; be lifted up. “Sing to the Lord, all the earth,” and, to say both together, “Let the heavens be glad and let the earth rejoice,” for the heavenly one is now earthly. Christ is illumined, let us flash like lightning with him. Christ is baptized, let us go down with him [into the water], that we may also come up again. Today salvation has come to the world, to things visible and to things invisible. Christ is risen from the dead; rise with him. Christ has returned to himself; return. Christ is freed from the tomb; be freed from the bonds of sin. The gates of hades are opened and death is destroyed.6

Harrison interprets Gregory as thinking of the enactment of the liturgy as an anamnesis. Her interpretation of anamnesis is essentially the same as von Allmen’s: “Anamnesis means re-presentation of God’s saving 5 6

I thank Terence Cuneo for calling this book to my attention. The passages from Gregory are in Harrison (2008: 25).

THE LITURGICAL PRESENT TENSE



works so that the worshipers can participate in these events as present realities. . . . [I]t is important to note that the saving events are made present in their liturgical celebration, not only the persons who once participated in those events and are now glorified in heaven. Since God’s saving actions transcend the limitations of temporal sequence, the historical events in which God has acted can be present now and in the future” (Harrison 2008: 24–6). This is the language of the reactualization interpretation: “The saving events are made present,” “the historical events in which God has acted [are] present now.” Be it noted, however, that whereas Gregory does use the liturgical present tense, he does not use the telltale language of the reactualization interpretation; he does not speak of events from the past as becoming present, as being reactualized. Harrison is over-interpreting the passages from Gregory, interpreting them in the light of the twentieth century reactualization theory concerning the relation between biblical narrative and Christian liturgy. Eliade interprets archaic persons as believing in a certain sort of time travel: participants in the ritual exit historical time and enter the mythic time of origins, and then, a bit later, exit mythic time and reenter historical time. Neither he nor anyone else who embraced the reactualization interpretation of the Christian liturgy attributed to Christians the counterpart belief that, when participating in the liturgy, one travels back in time to when Christ lived. It is noteworthy, for example, that Eliade does not say that the participants enter the illud tempus in which Christ lived, suffered, and rose again; he says that they recover the illud tempus. The idea is not that participants in the liturgy travel back to when those events occurred, but rather that those events become actual again, become present. It is not we who travel about within time; it is events that travel about within time, from past to present. The very same eventtoken that happened in the past is happening again now, in the present— apparently while continuing to be an event that happened in the past, though on this point the writers I have read are unclear.

3 Is Reactualization Possible? Is this possible? Is it possible for an event-token that was actual in the past—began, lasted for a while, and ended—to become actual again in the present? Is it possible for it to become actual again next week, yet



NICHOLAS WOLTERSTORFF

again the week following, etc.? I judge that it is not. An event-type can occur again; the event-type, eating the bread of the Eucharist, has recurred billions of times. But that event-token of my eating the bread of the Eucharist which happened last Sunday can never happen again; it cannot become actual again, cannot become present again. That event-tokens cannot happen again, seems to me obviously true. Evidently those liturgical theologians whom I have quoted do not find it so; I have learned that some philosophers also do not find it obviously true. So the claim calls for a defense. Unfortunately, the space available to me in this chapter does not permit such a defense. I shall have to content myself with arguing that, even if certain events can be reactualized, those that liturgical theologians claim to be reactualized by enactments of the liturgy have never been reactualized.7 What would it take for the event of Christ’s birth to be reactualized? Presumably that event had certain essential properties and relations, among those, that of Mary standing to Christ in the relation of delivering him as an infant from her womb, that relation being one that can only be instantiated in space. So for Christ’s birth to be reactualized, the original actor, Mary, and the original “patient,” the infant Christ, would once again have to show up in space-time, and Mary would once again have to perform the action of giving birth to that infant. The reactualization theory of the liturgy implies that each enactment of the liturgy on Christmas Day brings that about. Since there have been millions and millions of such enactments, the theory implies that it has been brought about millions of times. It can confidently be affirmed that nothing has ever brought about the reappearance of Mary and the infant Jesus and the recurrence of her giving birth to him. Enactments of the Christmas liturgy have not brought this about, nor has anything else. The reactualization theory of what happens when the Christian liturgy is enacted is simply not true. We have to look elsewhere for an interpretation of the significance of the liturgical present tense. Harrison says of “God’s saving actions” that they “transcend the limitations of historical sequence” so that “the historical events in which God has acted can be present now and [again] in the future”

7

The point that follows was suggested to me by Trenton Merricks.

THE LITURGICAL PRESENT TENSE



(Harrison 2008: 24–6, emphasis added). She gives no reason for holding that this remarkable view is true. Charity requires that one give serious consideration to the possibility that she and all the other writers who have used the language of reactualization were not speaking literally. But I discern no clue in the four writers I have quoted that they were not speaking literally. They do not say, for example, that when we participate in the liturgy we speak as if events from the past are happening presently. They say that events from the past become present.

4 On Speaking As If Just now I used the phrase, “speak as if.” I suggest that when those who participate in an enactment of a Christian liturgy use the liturgical present tense, they are speaking as if certain event-tokens from the distant past are happening now. In so doing, they are employing a distinct rhetorical trope. They know full well that the events they are singing and speaking about happened in the past. Nonetheless, they sing and speak about them as if they were happening now. Evidently Christians find it important, at least on the high holy days, to sing and speak about certain events in the distant past as if they were happening now. We saw that in an enactment of the Orthodox liturgy for Vespers on Holy Friday, events in the distant past are at certain points sung and spoken about as if they had happened earlier on that Holy Friday, before the liturgy was enacted, rather than as if they were happening at the time of the enactment. In such cases, the past tense is used. Obviously the issues raised by the use of the present tense to speak or sing about events in the distant past are also raised by the use of the past tense on a certain day to speak or sing about events in the distant past as if they had happened earlier that day. It would be cumbersome, however, always to keep both of these uses in play. So I will speak only of the use of the present tense to speak or sing about events in the distant past, while urging the reader to keep in mind that everything I say by way of interpreting such speech and song applies also to the use of the past tense when speaking about events in the distant past as if they had happened earlier that day. I said that when we use the liturgical present tense, we are employing a distinct rhetorical trope. It is not metaphor, it is not simile, it is not personification, it is not synecdoche. It is a sui generis trope. Though its



NICHOLAS WOLTERSTORFF

use is common, encompassing much more than just the liturgical present, I am not aware that the rhetorical tradition has ever singled it out, identified it, and given it a name. So let me give it a name. Let me call it the as-if trope. Before I set out to elucidate this trope, I should explain why I reject the answer to our question as to the significance of the liturgical present tense that probably some readers regard as the obvious answer, namely, that the liturgical present tense is simply a particular use of what grammarians call “the historical present.” The historical present is the use, in English and a number of other languages, of the present tense to describe events that happened in the past. Those of us who write about thinkers from the past regularly use the historical present. I did so when stating the views of the reactualization theorists. What seems to me a decisive reason for not interpreting the liturgical present tense as a particular case of the historical present is the use, in hymns cast in the present tense, of such temporal indexicals as ‘today,’ ‘this night,’ ‘this day,’ ‘this happy morn,’ and so forth. These are used to refer to the day on which the hymn is sung, not to that day in the distant past when the event that is sung about took place. The first line of Charles Wesley’s Easter hymn, “Christ the Lord is risen today,” is a good example of the point. Christ is declared to have arisen earlier on whatever be the day on which the hymn is sung.8 Not all hymns that employ the liturgical present tense include such temporal indexicals as those I have mentioned; Wesley’s Christmas hymn, “Hark, the herald angels sing,” does not. But I submit that such indexicals can always be inserted without any change of sense. “Hark, the herald angels now sing . . . ”9 The use of the as-if trope pervades ordinary life. We all employ it, and we all know how to identify and interpret it. So far as I am aware, the use of the trope to speak about an event from the distant past as if it were happening now is not common outside enactments of the liturgy. But other uses of the trope are common. So let us reflect for a while on the trope as such, with the aim of discerning how it works and why we use it. 8 By itself, the first line of the hymn might be interpreted to mean that Christ today retains the status of being risen. But when the opening line is interpreted in the context of the hymn as a whole, it becomes clear that that is not what it means; it means that the event of Christ’s resurrection took place earlier today. 9 And of course the angels have to be singing when we are singing if we are to hearken to them.

THE LITURGICAL PRESENT TENSE



Then, in the expectation that what we have learned about the as-if trope in general will illuminate special uses of it, let us look at the liturgical present tense. We use the as-if trope both when speaking to someone or something and when talking about someone or something. Let me begin with an example of the former. A dog, so it is said, is a boy’s best friend. Boys sometimes talk to their dogs as if the dog were their best personal friend. The language they use is a selection from the language they would use when talking to their best personal friend: personal best-friend language. The boy does not believe that his dog is a person; he knows full well that it is not. Nonetheless, he uses personal best-friend language in talking to the dog. Just listening to him talk, one would infer that he was talking to a personal friend—a friend who, for some reason, makes no response. Why does the boy talk to his dog as if the dog were his personal friend, all the while being fully aware that the dog is not a person at all? Well, the language that we use in talking to personal friends has a distinct resonance, as I shall call it. That is to say, it is associated in one’s mind with distinct feelings, thoughts, and images. When the boy talks to his dog as if the dog were a personal friend, he projects a good deal of that resonance onto the dog. That he does so is not arbitrary. He will already have an attitudinal take on his dog: certain feelings about his dog, certain thoughts, certain images. His attitudinal take on his dog coincides with, or comes close to coinciding with, the resonance of a good deal of personal friendship language. So it feels right to him to talk this way. Speaking this way serves to express his attitudinal take on his dog. At the same time it reinforces those feelings, those thoughts, those images about his dog. The boy might be able to state in literal language some of the attitudinal take on his dog that he expresses when he talks to the dog as if the dog were a personal friend. But the literal description will never be the equivalent. The as-if trope is open-ended in the same way that metaphors are open-ended. By virtue of their open-endedness, both metaphors and instances of the as-if trope have an evocative power that literal speech lacks. Everything I have said about using the as-if trope when talking to someone or something applies as well to using it when speaking about someone or something. Let me introduce an example of the point that will, at the same time, make another important point about the use of the



NICHOLAS WOLTERSTORFF

as-if trope. My example of the boy and his dog may leave the impression that the employment of the as-if trope is interesting, quaint, amusing, and the like, but that it carries no moral freight. To the contrary, it often carries profound moral freight. Often it is used to demean certain of one’s fellow human beings. The trope is often employed as an instrument of injustice. Slavery as depicted in Harriet Beecher Stowe’s novel Uncle Tom’s Cabin was substantially true to slavery as it existed in the American South in the mid-nineteenth century. Stowe depicted slave owners as often talking about male slaves as if they were work animals. Owners discussed the physical strength of their slaves, their endurance, their health, their age, whether or not they were of a compliant nature, how much food they required, etc. At least some of those who talked this way knew that their slaves were human beings, not mere work animals. Nonetheless, they talked about them as if they were work animals; they used work-animal language in talking about them. If one did not know that they were talking about human beings, one would infer from their speech that they were talking about work animals. Their talking about them as if they were work animals was part and parcel of their treating them as if they were work animals: they put them on the block for sale, inspected their teeth and muscles before purchasing them, etc. Augustine St. Clare, one of the slave owners in Stowe’s novel, describes as follows what it was like to buy a slave: “buying a man up, like a horse,—looking at his teeth, cracking his joints, and trying his paces, and then paying down for him” (Stowe 2005: 195). Upon St. Clare’s death, his wife put Tom up for sale. The narrator describes as follows the scene in the slave warehouse in New Orleans where Tom and others were auctioned: “Various spectators, intending to purchase, or not intending, as the case might be, gathered around the group, handling, examining, and commenting on their various points and faces with the same freedom that a set of jockeys discuss the merits of a horse” (Stowe 2005: 282). Shortly before the auction was to begin Simon Legree arrived on the scene; his attention was drawn to Tom: “He seized Tom by the jaw, and pulled open his mouth to inspect his teeth; made him strip up his sleeve, to show his muscle; turned him round, made him jump and spring, to show his paces” (Stowe 2005: 283). The language we use when talking about work animals has a certain resonance; in speaking of their male slaves as if they were work animals

THE LITURGICAL PRESENT TENSE



the slave owners projected a good deal of that resonance onto their slaves. Their doing so was not arbitrary; they already had an attitudinal take on these human beings that coincided with the resonance of a good deal of work-animal language. So it felt right to them to talk this way; talking about them in this way expressed their attitudinal take on these creatures: their feelings, their thoughts, their images. At the same time it reinforced those feelings, thoughts, and images. And one person’s speaking about them as if they were work animals encouraged others to do so as well, thus encouraging others to feel and think about them, and to image them, in accord with the resonance of work-animal language. Apparently, some slave owners believed that African-Americans really were non-human animals; they believed that these creatures lacked whatever is necessary for being a human being. So of course they talked about them as if they were animals. In so speaking, were they using what I have identified as the as-if trope? They were not. A condition of using the sui generis rhetorical trope that I am calling the as-if trope, when speaking about some entity that is not ø as if it were ø, is not believing that it is ø. If one is under the misapprehension that it is ø and speaks about it accordingly, one is speaking literally; one is not using the as-if trope. The slave owners who believed that their slaves were not human beings but rather mere animals and spoke accordingly were speaking literally. They were speaking as if they were mere animals when in fact they were human beings; but they were not using what I have identified as the as-if trope. Either way, whether those who spoke about these human beings as if they were work animals were or were not using the as-if trope, they wronged them deeply. They profoundly demeaned them. Such speech has to be counted among the deep injustices of slavery. Were those who used the as-if trope when speaking about their slaves as if they were work animals speaking falsely? Was their demeaning speech a case of false speech? This question did not arise when we were reflecting on the boy speaking to his dog as if the dog were a personal friend; when one speaks to some person or animal, one’s speech is in the mode of address rather than assertion. So, too, the question does not arise for slave owners speaking to their slaves as if they were work animals. But when slave owners, using the as-if trope, spoke about slaves as if they were work animals, their speech was in the declarative mood. So the question does arise: was their demeaning speech a case of speaking falsely?



NICHOLAS WOLTERSTORFF

If they had asserted about them that they were mere animals, they would have spoken falsely. If what they said carried the implication that they were animals, that implication would have been false. But in speaking about them as if they were animals, knowing that they were not, they were neither saying nor implying that they were animals. Nor were they saying that they were like animals; that would have been to use the term ‘animal’ as a simile. They were speaking of them as if they were animals while knowing full well that they were not. In so doing, were they speaking falsely? When we employ the as-if trope in speaking about someone or something, are we making true–false assertions? I think we are not. What we are doing instead, so I have suggested, is projecting the resonance of some component of our language onto a person, an animal, a thing, and thereby expressing an attitudinal “take” on that entity which more or less coincides with that resonance. Is the use of the as-if trope a species of pretending? Not in general. It is not clear to me whether or not the boy who uses the language of personal friendship when speaking to his dog is thereby pretending that his dog is a person who is a friend of his; perhaps he is. But it seems clear that those slave owners who used the as-if trope of work-animal language in speaking about their slaves were not, in so doing, pretending that their slaves were mere animals. Nor, I think, were they imagining that they were mere work animals. Though the use of the as-if trope is closely related to pretending and imagining something to be what one knows it is not, it is not, in general, a species of either of those.10 It is important that we do not confuse using the as-if trope, to speak about something that is not ø as if it were ø, with saying about some event which is not ø, literally speaking, that it is as if ø. An example of the latter is this: “It was as if all the air had been sucked out of the room.” If one used this sentence to describe the atmosphere in some room, one

10

Michael Bergmann has called my attention to the fact that imagining something to be what one knows it is not is rather common in certain forms of therapy and in certain forms of meditative practice. Therapists sometimes invite clients to relive a traumatic experience, to imagine it as happening now, perhaps with some loving figure present who was not in fact present. Those spiritual directors who employ the strategies of Ignatian spirituality invite their directees to imagine themselves as present at, say, the crucifixion. (Recall the first line of the spiritual, “Were you there when they crucified my Lord?”.) Others invite the directee to imagine Christ as bodily present in the room and speaking to them now.

THE LITURGICAL PRESENT TENSE



would be using a simile rather than the as-if trope. The sentence is the equivalent of, “It was like all the air being sucked out of the room.” The as-if trope is a sui generis form of non-literal speech, not to be identified with either simile or metaphor. My sense is that it is often much more powerful than either simile or metaphor. It is extremely demeaning for one person to say to another, “You are nothing but an animal.” My sense is that it is even more demeaning for one person to talk about another as if he were nothing but a work animal.

5 Interpreting the Liturgical Present Tense With these general reflections on the as-if trope in hand, let us turn to the liturgical present tense. Those who participate in the Christian liturgy on Christmas Day know full well that Christ was born roughly 2,000 years ago. And, apart from liturgical theologians, few believe what is ontologically impossible, that that event-token is being reactualized on that Christmas Day and made present again. Nonetheless, they sing, “Christ is born in Bethlehem.” Not was born, but is born. What they are doing, so I suggest, is employing the as-if trope. They are singing about Christ’s birth as if it were happening at the time of their singing, or earlier that Christmas Day. Those who participate in the Christian liturgy on Easter likewise employ the liturgical present tense. They sing, “Christ the Lord is risen today.” They know full well that Christ’s resurrection took place thousands of years ago. Nonetheless, they sing about his resurrection as if it were happening at the time of their singing, or earlier on that Easter day. They are employing the as-if trope. The resonance of the language that we typically use when talking about something that is happening now, or that happened earlier today, is very different from the resonance of the language that we use when talking about something that happened in the distant past. “Hark, the herald angels sing,” present tense, has a very different resonance from “The herald angels sang,” past tense. “Christ the Lord is risen today,” present tense, has a very different resonance from “Christ the Lord arose,” past tense. When Christians celebrate Christ’s birth on Christmas Day, they find that the resonance of the present tense fits their attitudinal take on Christ’s birth far better than does the resonance of the past tense. So, too, when they celebrate Christ’s resurrection on Easter they find that



NICHOLAS WOLTERSTORFF

the resonance of the present tense fits their attitudinal take on Christ’s resurrection far better than does the resonance of the past tense. Of course, their having that attitudinal take on those events is due, in part, to their often having sung about those long-ago events with hymns in the present tense. Is it possible to put into words some of the resonance of present-tense language as opposed to past-tense language? Perhaps. Start with pasttense language. When we think and talk about events that occurred 2,000 years ago, spatial images of length are evoked: the events are long past, far in the past, long ago. They are distant from us. The opposite is true when we talk about events that are happening now or that happened earlier today: there is no distance, no gap. When we talk about events in the distant past as if they were happening now we project onto them this resonance of present-tense language. One might say that thereby we make them present to us. We do not make them present again; that is impossible. We make them present to us. My guess is that it was this phenomenon that led liturgical theologians to develop their theory of reactualization. Christians feel intuitively that for hymning Christ’s birth and resurrection, the right language is not language whose resonance is that of distance, but rather language whose resonance is that of immediacy. Christ’s birth and resurrection are as immediately relevant to us as they were to those who were living at the time. The language of distance that we use for speaking of what happened long ago is literally accurate but theologically and emotionally all wrong. The language we need is the language we use for speaking of what is presently happening. The Eastern Orthodox liturgical scholar Alexander Schmemann makes a closely similar point in the following passage: To be sure, the Virgin does not give birth today, no one “factually” stands before Pilate, and as facts these events belong to the past. But today we can remember these facts and the Church is primarily the gift and the power of that remembrance which transforms facts of the past into eternally meaningful events. Liturgical celebration is thus a re-entrance of the Church into the event, and this means not merely its “idea,” but its joy or sadness, its living and concrete reality. . . . We were not there in Bethany at the grave [of Lazarus] with the crying sisters. From the Gospel we only know about it.

THE LITURGICAL PRESENT TENSE



But it is in the Church’s celebration today that an historical fact becomes an event for us, for me, a power in my life, a memory, a joy (Schmemann 1969: 82–3).11

6 The Liturgical Present Tense in Memorial Celebration The examples of the use of the liturgical present tense that I gave at the beginning of this chapter were all taken from Christian hymns and liturgies; a full discussion would bring into the picture hymns and liturgies from other religions as well, especially Judaism. And they were all taken from hymns and liturgies that Christians use on high holy days in the church calendar: Christmas, Holy Friday, and Easter. The liturgy on those days is enacted as a memorial of Christ’s birth, of Christ’s death, and of Christ’s resurrection. To the best of my knowledge, all Christian uses of the liturgical present tense occur in hymns and liturgies that are used on those days in the church’s calendar when the enactment of the liturgy is a memorial celebration of special events in the history of salvation. The passages from Gregory of Nazianzus quoted earlier, in which he employed the liturgical present tense, all came from “festal orations” by Gregory, that is, from sermons that Gregory preached on one or another of the church’s high holy days. The Eucharist is also a memorial celebration of a special event in the history of salvation. Given that Christians naturally employ the liturgical present tense on those high holy days when enactment of the liturgy is a memorial celebration of a special salvific event, one would expect them also to employ the liturgical present tense weekly within the Eucharistic memorial. To the best of my knowledge, that is not the case. A good many contemporary Eucharistic liturgies include the acclamation, spoken by the people, “Christ has died, Christ is risen, Christ will come again.” And a good many include the words, spoken by the celebrant just before he or she distributes the bread and wine, “Christ our Passover is sacrificed for us.” The verbs ‘is risen’ and ‘is sacrificed’ are in the present tense. The question is whether it is the liturgical present tense. 11

I thank Terence Cuneo for calling this passage to my attention.



NICHOLAS WOLTERSTORFF

I think it is not, but that it is, instead, the continuous present tense. Both describe Christ’s present status: Christ has the status of being risen and the status of being sacrificed. A test for whether some use of the present tense is an example of the liturgical present tense is whether, without changing the sense, one can insert the indexicals ‘today’ or ‘this morning.’ I judge that these two sentences fail this test. The congregants are not saying that Christ arose this morning; the celebrant is not saying that Christ is sacrificed today. Is it significant that most if not all uses of the liturgical present tense in Christian hymns and liturgies occur within the context of liturgical enactments that are memorial celebrations of some special event in the history of salvation? Is there something about doing something as a memorial of some important event in the past that makes it natural, within the context of the memorial, to speak about that past event as if it were happening now? Is the celebration of the Eucharist exceptional in that it does not employ the liturgical present tense? In earlier days, July 4 each year was the occasion for Americans to participate in enacting memorial celebrations of the signing of the Declaration of Independence on July 4, 1776. Today such memorial celebrations on July 4 have all but disappeared; for most Americans, the day is now nothing more than an early summer holiday. Back in the days when July 4 was still the occasion for enacting memorial celebrations, those celebrations almost invariably included a patriotic speech by some public figure. It would be interesting to know whether those orators sometimes slid into what I have been calling “the liturgical present tense” when speaking about the Declaration of Independence from Britain that was adopted on July 4, 1776. Is there some general human tendency to employ the liturgical present tense when commemorating important events from the past, or is this a peculiarity of liturgical memorials? I do not know the answer to this question.

References Casel, Odo. (1962). The Mystery of Christian Worship. Westminster, MD: Newman Press. Eliade, Mircea. (1959). The Sacred and the Profane. New York: Harcourt Brace Jovanocich. Eliade, Mircea. (1963). Myth and Reality. New York: Harper and Row.

THE LITURGICAL PRESENT TENSE



Eliade, Mircea. (1974). Myth of the Eternal Return. Princeton: Princeton University Press. Harrison, Nonna Verna. (trans.) (2008). Festal Orations: St Gregory of Naziansus. Crestwood, NY: St Vladimir’s Press. Schmemann, Alexander. (1969). Great Lent: Journey to Pascha. Crestwood, NY: St Vladimir’s Press. Stowe, Harriet Beecher. (2005). Uncle Tom’s Cabin. Mineola, NY: Dover Publications, Inc. von Allmen, J. J. (1965). Worship: Its Theology and Practice. London: Lutterworth Press. Ware, Archimandrite Kallistos and Mother Mary. (trans.) (1978). The Lenten Triodion. London: Faber and Faber. Wolterstorff, Nicholas. (1990). “The Remembrance of Things (Not) Past,” in Thomas P. Flint (ed.), Christian Philosophy. Notre Dame: University of Notre Dame Press, 118–61.

Immortality, Body, and Soul

9 The Rev’d Mr Bayes and the Life Everlasting Peter van Inwagen

Most philosophers will have encountered the “Doomsday Argument,” a Bayesian argument for the conclusion that our species will come to an end sooner than we might have expected—that the probability that we should assign to the hypothesis that the end of the human species will come within the next few centuries is significantly higher than the probability we should assign to that hypothesis if we considered only causal factors like thermonuclear or environmental or epidemiological catastrophe—or extraterrestrial disaster scenarios: “big rock hits the earth,” “nearby supernova,” and the like. The question has recently been raised whether, if this argument indeed establishes its conclusion, a closely parallel argument might show that it is highly unlikely that there is any such thing as an afterlife—or at any rate a very lengthy one.1 This question is my subject. I begin with a review of the Doomsday Argument.

1 In order to eliminate some of the messy issues that a satisfactory analysis of the Doomsday Argument as it applies to the real world would involve (such as increasing and decreasing population levels), I’m going to show 1 See Page (2010: 399–401) and Leslie (2008). In this chapter I attempt to think that question through from first principles and do not directly discuss Page’s and Leslie’s arguments.

THE REV ’ D MR BAYES AND THE LIFE EVERLASTING



how the argument would work in application not to the end of our species, but to the end of a universe—and not our universe, but a very simple universe: the Toy Universe, I’ll call it. Imagine that I am an inhabitant of the Toy Universe. At a certain point in my life, when our story opens, I know very little about the temporal features of the Toy Universe (which, within the fiction that I am one of its inhabitants, I shall refer to as simply the universe). Among the few things I do know is this: The universe had a beginning in time: there was no first moment at which it existed, but the set of moments at which it has existed has a greatest lower bound. I do not, however, know anything about how long it has existed, other than this: The universe has existed for less than 1 million years. (This implies that I know nothing about how long I have existed: for example, for all I know, I have existed for only a few seconds. Make this an intelligible assumption in any way you like—one possibility would be to suppose that for all I know, I was created, along with the rest of the universe, a few seconds ago, complete with memories of “a wholly unreal past.”) I know that the universe has existed for less than 1 million years because I know that: The total span of the existence of the universe will be 1 million years or less. But I would like to give a rather more precise description of my “initial” knowledge of the temporal features of the universe than this. In order to do this, I will introduce a precise system of dates. Let our unit of temporal measure be the terrestrial sidereal year. Let any real number greater than or equal to 0 be a “date.” Then the first year of the existence of the universe comprises those dates greater than 0 and less than or equal to 1, the second year of the existence of the universe comprises those dates greater than 1 and less than or equal to 2, and so on. And the date 23.48592 denotes a period of about 5 minutes that ends approximately 23 years and 15,334,736 seconds after 0. One thing I know is this: The universe will come to an end at some point in the interval (0, 1M].



PETER VAN INWAGEN

(‘1M’ being my abbreviation for ‘1,000,000.’) And I also know this: The date of the End is the outcome of some random process—of some ontologically random process that “chose” a single number x that satisfied the condition 0 < x
1M (and might “just as well” have chosen any other number in that interval as the one it did choose). More exactly, the objective probability of the proposition that the universe would end in any given year (not greater than 1M) was (at the moment of creation, so to speak) the same as the objective probability of the proposition that it would end in any other given year: for each year, that probability was literally one in a million. Similarly, the objective probability of the proposition that the universe would end in any given 100-second subinterval of (0, 1M] was the same as the objective probability of the proposition that it would end in any other 100-second subinterval of that interval. Generalizing: For any four dates, m, n, i, and j (all less than or equal to 1M). If jmnj = jijj, then the objective probability of the proposition that the End would come between m and n was equal to the objective probability of the proposition that the End would come between i and j: jmnj/1M or jijj/1M. Now consider the following hypothesis, which I will call ‘K’ (for kiloyear). The End will fall within (0, 1,000]. What subjective probability should I assign to this hypothesis, given what I know “as our story opens?” Neglecting the “nuisance fact” that, if I consider this question, I must know that, since I have had time to consider it, at least a few seconds of (0, 1,000] must already have passed, the answer is evident: the same probability that I should assign to the hypothesis that it will end in any other millennium: one in one thousand, or 0.001. Now suppose that at some point I acquire the following piece of information (which I will call ‘H’ for hundred or hectoyear): The present date falls within the interval (0, 100]. (That is to say, this moment is a moment that falls during the first century of the existence of the universe.) It is important to emphasize that I acquire no more specific information about the present date than

THE REV ’ D MR BAYES AND THE LIFE EVERLASTING



that. Imagine, if you like, that the whole episode of my acquiring this new piece of knowledge consisted in God’s handing me a slip of paper on which this alone was written: “The present date falls within the interval (0, 100].” So, for all I know, the present year could be 1 or 23 or 87 or 100 or any other year that is not greater than 100. The Doomsday Argument (applied to the present simple case) is essentially a Bayesian argument for the conclusion that my knowledge that the present moment falls within (0, 100] should lead me to raise the probability I assign to any hypothesis to the effect that the End will come “fairly soon.”2 And not only raise it, but to raise it significantly. It should, for example, lead me to raise—significantly—the probability I assign to K. But why—and by how much? It is possible to reason one’s way to an answer to these questions, but the required piece of reasoning is rather complicated. At any rate, it is complicated if one carries it out from scratch and does not simply begin in medias res—by saying something like, “We can calculate the probability I should assign to K, given my new piece of knowledge, H, using Bayes’s Theorem, which, as we all know, is . . . ” I propose to start from scratch. I will present the complex piece of reasoning I have alluded to in the form of a commentary on some figures—a sort of visual aid—that I will call probabilistic Venn diagrams (PVDs). A PVD is much like an ordinary or “logical” Venn diagram except that: (i) A PVD contains an outer circle, L, which represents the whole of “logical space;” all the other circles in a PVD are inside this outer circle. (ii) In an ordinary Venn diagram, each circle represents a class of things. A circle in a PVD, however, represents a proposition or hypothesis (or, if you like, a class: the class of possible worlds in which that proposition is true). In fact, any subregion of L, however shaped or gerrymandered, represents a proposition, whether that subregion has been “outlined” or not. For example, the region common to two or more circles represents a proposition, the conjunction of the propositions represented by those 2 And, in fact, for any positive number a, a < 1M, to raise the probability that I assign to the hypothesis that the End will come in the interval (0, a]. As a approaches 1M, however, the prior probability of that hypothesis and the revised probability become closer and closer to each other.



PETER VAN INWAGEN

circles. (L, the outer circle, represents the proposition that something is the case.) (iii) In a PVD, unlike an ordinary Venn diagram, the size (that is, the area) of a region within the outer circle—its area in proportion to the area of the outer circle, the proportion of the outer circle it occupies—is significant. (We will use the area of the outer circle as our—dimensionless—unit of area measure: its area is 1. If, therefore, a region inside the outer circle has an area that is 1/17th the area of the outer circle, its area is 1/17 or about 0.059.) The area of a region represents the probability of the hypothesis the region represents. For example, the region within the outer circle mentioned in the parenthesis represents an hypothesis that has one chance in 17 of being right—or has a probability of 0.059. In the sequel, we will, to simplify our exposition, speak as if a circle in a PVD and the hypothesis it represents were one and the same thing. Now consider two overlapping—that is, consistent—hypotheses: A and B. We abbreviate “the proportion of B that is A” as “A//B”. If, for example, 10 percent of the region B coincides with a part of A (if the area of A \ B is one-tenth the area of B), A//B = 0.1. If a point in B is chosen at random—by throwing a dart, perhaps—, there is one chance in ten that that point will also be a point in A. We may therefore think of A//B as the probability of A, given that B—or the conditional probability of A on B. Now let us draw a PVD that displays our two hypotheses: K and H. (These two hypotheses are obviously consistent and, obviously, neither entails the other. We therefore represent them by circles that overlap, but only partly. And each obviously has a probability of less than unity, so we draw them both as “subcircles” of the outer circle.) Having done this, let us proceed to “normalize” the region of logical space that comprises all the propositions about the existence of the Toy Universe and its temporal features that I knew before I discovered that the present date fell within (0, 100]. That is, let us agree to ignore the region of logical space in which the conjunction of those propositions is false and assign a new measure of 1 (formerly the measure of the whole of logical space) to the region in which it is true. Call this region T and represent it by a circle—our “new” outer circle. (The circles H and K of course fall within T.) Doing this will have the effect of increasing the

THE REV ’ D MR BAYES AND THE LIFE EVERLASTING



probability measure of every subregion of T by the same proportion as that by which the measure of T was increased. Suppose, for example, that T occupies one-seventeenth of logical space. Then, when the measure of T is normalized to 1, the “new” measure of every subregion of T must be seventeen times its “old” measure—since 1 is seventeen times oneseventeenth. If, for example, A is some subregion of T whose old measure was 0.02, the new measure of A is 0.34 (= 17  0.02). More generally, suppose that a region of logical space A is normalized and that B is a subregion of A; if Aold is the old area of A and if Bold is the old area of B, then Bnew, the new area of B, is equal to (1/Aold)  Bold. We may now draw Figure 9.1, a PVD, on the left below. Now the way I have drawn Figure 9.1 displays more than I have stipulated about the relations between T and H and H \ K. Figure 9.1 represents H as occupying a certain determinate proportion of T, and it represents H \ K as occupying a certain determinate proportion of H and a certain determinate proportion of K. (K does occupy a certain determinate proportion of T: it occupies one one-thousandth of T— obviously the figure is not drawn to scale—, but nothing that has been said has any implications as regards the size of H or of H \ K.) In displaying such “extraneous information,” the diagram resembles a diagram illustrating a proof of a theorem concerning, say, the interior angles of a triangle in a plane-geometry textbook; all but certain features of the “sample” triangle shown in the diagram are to be ignored: all of them that do not actually figure in the proof of the theorem.

H∩K

H H

K

Figure 9.1

T

Figure 9.2



PETER VAN INWAGEN

In what follows, I will abandon the first person: we shall try to see how to solve the problem that faces the fictional “me,” the inhabitant of the Toy Universe: How should my coming to know the truth of H affect the probability I assign to K? When we look at Figure 9.1, therefore, we are to attend only to certain of its features: that both circles lie inside T, that they partly overlap, and that the (new or normalized) area of K is 0.001—although this last “feature” cannot be said to be perspicuously displayed in the figure. (A triangle in a geometry-text diagram has to be drawn with some determinate interior angles; the spatial regions that represent H and H \ K in a PVD have to have some sizes.) Now here is what we want to know: what is the value of K//H? (That is, what probability should we assign to the hypothesis that the End will come in the first thousand years of the existence of the Toy Universe, given that we know that the present falls within the first hundred years of its existence?) We can determine this value (given a certain assumption that I will come to presently), and the reasoning by which we obtain it will, as I said, be presented in the form of a commentary on a PVD—much as the reasoning that leads to “The sum of the interior angles of a triangle is equal to the sum of two right angles” can be presented in the form of a commentary on a diagram that displays, among other things, a triangle and its interior angles. We know that H is true. Let us therefore normalize the measure of H to 1. The result of this operation is shown in Figure 9.2, given earlier. The shaded region, of course, is H \ K, or the conjunction of H and K. As I have said, to normalize a region of logical space A is to multiply the area of all its subregions by the reciprocal of the area of A—that is, the area of A before we “redefined” the measure of that area to be 1. If, for example, the “old” area of A was 0.2, the “new” area of A is 0.2  1/0.2, which is, of course, 1. And if B was a subregion of A with area 0.013, the new area of B is 0.013  1/0.2 or 0.065. Now consider the shaded region in Figure 9.2, H \ K. Now that we are in effect treating H as the whole of logical space—now that its area has been normalized to 1—, the new area of H \ K is simply the proportion of H that is K. (If, for example, the new area of H \ K is 0.3, then threetenths of H is occupied by a part of K.) That is to say, now, after normalization, the area of H \ K = K//H. (Note that the values of

THE REV ’ D MR BAYES AND THE LIFE EVERLASTING



proportions like K//H and H//K do not change when we normalize the area of H to 1, since the areas of H and K increase proportionately.) And we can say something about the present area of H \ K (now equal to K//H) in terms of H//K (the proportion of K that is H) and the old areas of H and K and H \ K. The old area of H \ K was equal to H//K times the old area of K. (If a plane region is one-tenth red, the area of its red part is one-tenth of its total area.) As we did on an earlier occasion, let us use subscripts to keep all these areas straight: ‘Hold’ will denote the old area of H, and similarly for ‘Hnew,’ ‘Kold,’ ‘Knew,’ ‘(H \ K)old,’ and ‘(H \ K)new.’ We have: ðH \ KÞold ¼ H==K  Kold And we know that (H \ K)new is equal both to K//H and (H \ K)old  1/Hold. We therefore have K==H ¼ ðH \ KÞnew ¼ H==K  Kold 1=Hold ; or K==H ¼ H==K  ðKold =Hold Þ; an expression some of you may recognize despite my non-standard notation.3 And a way of calculating K//H—that is, the probability I should assign to the hypothesis that the End of the Toy Universe will come within its first thousand years (K), given my knowledge that the present moment lies within the first century (H), is just what I promised to deliver. Of course, to make that calculation, I still need to determine three numbers: the prior probability that the Toy Universe will end in its first thousand years (the probability I assigned to that hypothesis prior to my discovery that the present moment lay in the first century), the prior probability of the present moment’s falling within the first century, and the probability of the present moment’s falling in the first century given that the Toy Universe will end at some point in its first millennium. The prior probability that the Toy Universe will end at some point in its first thousand years is easy to come by: it is, as we have seen, 0.001 or one one-thousandth. The other two numbers are more difficult. Let us

3

In standard notation: P(KjH) = P(HjK)  (P(K)/P(H)).



PETER VAN INWAGEN

begin with the third, H//K or the proportion of K that is H or the probability that the present falls in the first century, given that the End falls in the first millennium. I can deliver this number if you allow me a certain assumption. The Crucial Assumption I may regard the present moment as a moment chosen at random from the interval (0, the End]. This, I believe, is the assumption on which the Doomsday Argument turns. I will use it in the calculation that follows, and in one other calculation: the calculation of the prior probability of H. Let P be the present moment and let E be the End. Let us assume that E falls within the first millennium and see what this assumption can tell us about the probability we should assign to H. Notice that if E fell within the first century, the probability of H would be unity, and that if E fell in the interval (100, 1,000] it would be less than unity. If the date of E were exactly 200, the probability of H would be 1/2—owing to the fact that in that case P is a moment chosen at random from the interval (0, 200] and hence has a 50–50 chance of falling within the interval (0, 100]. (Note that the Crucial Assumption was a premise of that little argument.) And, similarly, if the date of E were exactly 300, the probability of H would be 1/3—and so on. Suppose we knew that the date of E was one of the “century” dates (100, 200, 300, . . . , 900, 1,000)—and knew the probability of the date of E’s being any one of these was equal to the probability of its date being any one of the others— and knew nothing else about its date. Then, intuitively, we should have to assign to H a probability that was the average value of the numbers 1, 1/2, 1/3, 1/4, 1/5, 1/6, 1/7, 1/8, 1/9, and 1/10—or roughly 0.2929. This intuitive calculation suggests that, if we measure time in centuries rather than years, the probability of H on K should be the average value on the interval (0, 10]—that is, the interval that runs from 0 centuries to 10 centuries—of a function f defined on that interval such that f ðxÞ ¼ 1 if 0 < x
1 f ðxÞ ¼ 1=x if 1 < x
10: The arguments of f are dates greater than 0 centuries and less than or equal to 10 centuries. The values of f are probabilities. For example,

THE REV ’ D MR BAYES AND THE LIFE EVERLASTING



f (5.34) is the probability that P falls within the first century given that the date of E is the final moment of the year 534—about 0.187. The average value of a function on an interval is given by the Mean Value Theorem, which is highly intuitive, at least in the first quadrant. Since the interval we are considering ((0, 10]) has 0 as its greatest lower bound, we need consider only such intervals. The following diagram (Figure 9.3) illustrates the Mean Value Theorem for this special case. In this diagram, g is the function whose average value on some interval (0, a] we wish to determine. That interval is shown in the diagram as the interval between the g(x)-axis and the vertical line AB—an interval whose length is a. The line CD represents the average value of g on that interval—that is, its “height” above the x-axis (‘h’ in the diagram) is the average value of g on that interval. The Mean Value Theorem tells us that the area of the rectangle with base a and height h is equal to the area of the shaded region—the area of the region under the graph of g whose width is the same as the rectangle with base a. And that’s

g (x)

A g

C

D

h

x

O a

B

Figure 9.3



PETER VAN INWAGEN

so intuitive an idea that in some developments of real analysis it’s the basis of the definition of the average value of a function on an interval. Incidentally, since we’re in that part of mathematics, real analysis, we treat both the areas of regions and the lengths of intervals as simple numbers (the measures of the regions and the intervals)— that is, we needn’t worry about units and square units as we do when we make calculations about distances and areas in the physical world. So armed, let us return to our function f and calculate its average value v on the interval (0,10]. The hard part of this problem is to determine the area A under f in that interval. If we know what A is, then, by the Mean Value Theorem, A = 10v—that is, A is equal to the area of a rectangle whose base is 10 and whose height is v. And since A = 10v, v = A/10. Here is a diagram (Figure 9.4) that represents the relations between A and v. Figure 9.4 represents a consequence of the Mean Value Theorem: that the average value of f on the interval (0, 10] is equal to the height of the rectangle drawn in the figure, the rectangle whose area is equal to the area (A) of the shaded region. And what is A? What is the area under f on the interval (0, 10]?—or since the area is the same whether the interval is closed or open at either

f (x)

1 f 10x

0

Figure 9.4

THE REV ’ D MR BAYES AND THE LIFE EVERLASTING



end, we can simply ask what the area under f is in the region that starts at 0 and ends at 10. The function f is defined partwise, so the question divides. The area of the region (a square) under the “flat” part of the function is 1  1 or 1. The area under the curved part of the function, the “1/x” part, is Z 10 1 dx x 1 And that definite integral, if you remember your first-year calculus, is equal to the natural logarithm of 10 minus the natural logarithm of 1. Any sort of logarithm of 1 is equal to 0, so the area under the curved part is equal to the natural logarithm of 10, and the area under the flat part plus the curved part is 1 plus the natural logarithm of 10 (usually written “1 + ln10”). And, since, as we have seen, the average value of f between 0 and 10 is equal to one-tenth the area under its graph on that interval, the average value of f between 0 and 10 is one-tenth the sum of 1 and the natural logarithm of 10, or (1+ ln10)/10. You can look up the values of logarithms in tables, or, these days, on your smart phone (which will also do the arithmetic for you), and, having consulted my iPhone, I’m in a position to tell you that the result of our calculations is roughly 0.3302— not too different from the figure (0.2929) we got by averaging the ten fractions 1/1, 1/2, 1/3, . . . 1/9, 1/10. And that—0.3302—is the probability that P lies in the interval (0, 1], given that E lies in the interval (0, 10]. (Or, measuring temporal intervals in years rather than in centuries, the probability that P lies in the interval (0, 100], given that E lies in the interval (0, 1,000].) Now, what is the prior probability of H—the hypothesis that P falls in the first century? We can answer this question very simply by the method we have just used if we note that it is equivalent to the question: What is the probability of the hypothesis that P falls in (0,1] given that E falls somewhere, at some randomly chosen point, in (0, 10,000]? (One million years is ten thousand centuries.) Our previous answer was (1 + ln10)/10  0.3302. To get the number we now require, we need only substitute 10,000 for 10: the prior probability of H is (1 + ln10,000)/10,000 or (the heavy-duty calculator in my iPhone tells me) approximately 0.00102. And now, finally, we have the numbers we need to use the equation K==H ¼ H==K  ðKold =Hold Þ



PETER VAN INWAGEN

The probability that the Toy Universe will end at some point in its first thousand years, given that the present moment falls in the first century of the existence of the Toy Universe (and given the Crucial Assumption— that I may treat the present moment as a moment drawn at random from all the moments at which the Toy Universe exists4), is, therefore, 0:3302  ð0:001=0:00102Þ You don’t need a heavy-duty calculator for that one: any old calculator will do. And the answer is . . . 0:3234 Before I realized that the present moment was in the first century—when I had no idea what the present date was other than that it was not greater than 1,000,000—, I assigned a probability of 0.001 to the hypothesis that the Toy Universe would end in the first millennium, the same probability I assigned to the hypothesis that it would end at some point in the 576th millennium or at some point in any other given millennium among the first 1,000 millennia. Now, it seems, having gained that piece of knowledge, I must assign that hypothesis a probability more than 300 times greater than 0.001—certainly a case of a significant upward revision of a prior probability! And, of course, if I revise some probabilities upward, I must revise others downward. The prior probability of the hypothesis that the Toy Universe will end at some point in the interval (1,000, 1M] was 0.999. The revised probability is 1  0.3234 or 0.6733: a near certainty has become a probability of about 2/3. The prior probability of the hypothesis that the Toy Universe would last at least half a million years was 0.5. The revised probability is close to 0.07 or 7 percent: an even chance has become a very bad bet. I will not, in this chapter, examine the Crucial Assumption or question the cogency of the Doomsday Argument in any other way. I want instead 4 Given the Crucial Assumption, the problem of determining K//H becomes essentially a problem of pure mathematics. Stated very abstractly, it is this:

Let a number B be chosen at random from the interval (0, 1M]. Once that choice is made, let a number A be chosen at random from the interval (0, B]. What is the probability that B falls within the interval (0, 1,000], given that A falls within the interval (0, 100]? The solution is, of course, 0.3234.

THE REV ’ D MR BAYES AND THE LIFE EVERLASTING



to raise the following question: Assuming that the kind of reasoning illustrated here is correct, can parallel or analogous reasoning be used by each of us human beings to show that he or she should assign a significantly lower probability to the hypothesis that he or she will exist 1,000 years from now than the probability that that person would otherwise have assigned to that hypothesis? You may wish to remind me that many people already assign a probability of 0 (or a probability so close to 0 that the difference is of no more than abstract, theoretical interest) to this hypothesis. And that is certainly true. Still, not everyone does. Its probability, conditional on the truth of the teachings of more than one religion, is 1, and many people who practice a religion assign (in effect, perhaps) a probability of 1 to the hypothesis (if Kierkegaard will forgive my use of that word in this context) that the teachings of their religion are true—or, if not 1, then certainly some probability significantly higher than 0.5. My question could be put this way: Can Bayesian reasoning parallel to or analogous to or in some way suggested by the “Doomsday” reasoning I have presented be used to show that these probability assignments are wrong or irrational or in some other respect objectionable?

2 Let us construct an argument parallel to the Toy Universe argument, an argument that turns on the question how long I shall exist. Let us suppose that God has revealed to me that my existence will end at some randomly chosen point within 1 million years of my conception (or whenever it was that I began to exist). Let us suppose I know that I am now 100 years old or less, and that I know nothing more specific about my age: for all I know I am 1 year old or 23 years old or 99 years old. Or, again refining our system of chronology, for all I know I am 0.783 years old (to the nearest thousandth of a year) or I am 78.662 years old or . . . What probability should I assign to the hypothesis (K) that I shall exist for 1,000 years or less if I know (H) that my present age is 100 years or less? I can answer this question if—and only if—I am granted a certain assumption: the “personal analogue” of the Crucial Assumption of the Doomsday Argument.



PETER VAN INWAGEN

The Crucial Assumption (Personal) I may regard the present moment as a moment chosen at random from the interval (0, the final moment of my existence5]. (Here, of course, ‘0’ represents the greatest lower bound of the set of my “possible ages” and my “possible durations”: whatever my present age may be, it is greater than 0 years; whatever the length of the total span of my existence may be, it is greater than 0 years.) Given this assumption, we have already, in effect, calculated this probability: it is 0.3234. That is, my knowledge that I am less than 100 years old raises the probability of K from 1 in a thousand to about 1/3. And, therefore, it lowers the probability that I shall exist for more than a thousand years from 0.999 to about 2/3—and lowers the probability that I shall exist for at least 500,000 years from 1/2 to about 7 percent. It seems, therefore, that Doomsday-style reasoning can be adapted to show that even if I can expect to continue to exist after my death, I cannot reasonably expect to exist for very long—not as the religions that preach a post-mortem existence measure “very long,” at any rate. Not, for example, “so long as the sun and the moon endure.” If we say that the sun will endure till it leaves the main sequence (it will destroy the moon when it does, if the moon is still around), that period is about 5 billion years. Suppose God has told me that I myself shall “endure” for some randomly chosen number of years between now and 10 billion years from now and that I know nothing more specific about the duration of my existence than that. Then the prior probability that I shall endure for as long as the sun and the moon is 0.5. But the probability conditional on my being less than 100 years old (given the Personal version of the Crucial Assumption) is 0.0361—less than 4 percent.6 5 I assume for the sake of simplicity that there will be a moment that is the last moment of my existence—rather than assuming only that there will be a moment that is either the last moment of my existence or the first moment of my non-existence. Incidentally, I doubt whether such mathematical niceties have any application to the real world. If I am one day destined to cease to exist, I doubt whether there will be even a final nanosecond of my existence, never mind a mathematical point in time that is the least upper bound of the set of points in time at which I exist. 6 “The upper bounds in your two calculations—a million years, ten billion years—are entirely arbitrary. It would be more interesting if you told us what the probability was that you would exist for (say) five billion years if you knew that you were less than 100 years old and knew that you would cease to exist at some moment chosen at random from the whole

THE REV ’ D MR BAYES AND THE LIFE EVERLASTING



These calculations, however, ignore the fact that the afterlife that the religions of the world promise, those that do promise an afterlife, is supposed to be everlasting, to go on for ever. No religion I know of teaches that a supernatural authority has decreed that we human beings are to exist post-mortem, but only for some finite period. Suppose, then, that we address the following question: Where F is the hypothesis that I shall exist for ever, what is F//H? According to the Rev’d Mr Bayes, the answer is given by the equation F==H ¼ H==F  ðFold =Hold Þ (Or so we are often given to understand. I have never read Bayes’s “Essay towards Solving a Problem in the Doctrine of Chances,” but I have often been told that it does not actually contain “Bayes’s Theorem.”) Let us see whether we can indeed use Bayes’s Theorem to calculate F//H. We begin by considering H//F. We must first ask how H//F is to be defined. We can’t define it this way (again we shall be measuring intervals of time in centuries): Chose a date (that is, a real number) at random from the set of all dates (from the set of all real numbers greater than 0). H//F is the probability that that randomly chosen date is less than or equal to 1. We cannot define H//F this way because there is no such thing as choosing a real number at random from that set—or, indeed, choosing at random a number from any set of numbers that has a lower bound and no upper bound (cf. footnote 6). I would suggest a “limit” definition. We have in effect seen that if the Crucial Assumption (Personal) is true, then for any date x greater than 1, if x is the final moment of my existence, the probability that the present moment lies in (0,1] is equal to 1/x. I propose, therefore, that we define H//F as the limit of the reciprocal function as its argument increases without bound. On this definition, and no other definition suggests itself, the value of H//F is 0, since the function f(x) = 1/x is asymptotic to the x-axis.

infinite future.” I will be happy to tell the Interlocutor what this probability is—but first she must tell me what it means to choose a moment at random from the whole infinite future (or to choose a number at random from the set of all real numbers greater than 0, or from any set of numbers with a lower bound but no upper bound).



PETER VAN INWAGEN

Given this value for H//F, it is tempting to reason as follows: F//H is 0, since it is equal to the product of 0 and the ratio or fraction Fold /Hold. But this reasoning is correct7 only if the ratio Fold/Hold exists. And it does not exist if Hold is equal to 0. Is it? Before we can answer this question, we must ask what we mean by ‘Hold.’ What is the prior probability of H, the probability I should assign to H before I consider the “effect” on that probability of my knowledge that I am 100 years old or less, and when I know nothing about the total span of my existence—when, for all I know, I shall (a) exist for some unknown finite period of time, or (b) exist for ever. I can think of only one way to define this ‘Hold.’ We have in effect seen that if the Crucial Assumption (Personal) is true, then for any date x, the probability that the present moment falls in (0,1], given that the final moment of my existence is some date in (0, x], is equal to (1 + ln x)/x. I propose, therefore, that we define Hold as the limit of f(x) = (1 + ln x)/x as x increases without bound. That limit exists, and it is equal to 0.8 This thesis is hardly unintuitive. The natural logarithm of x is the power to which the number e must be raised to obtain x, e being an irrational number approximately equal to 2.718. To “obtain” 10,000, for example, we must raise e to a power very close to 9.21. (2.7189.21 is about 9,987.) That is to say, the natural logarithm of 10,000 is approximately 9.21. It should be obvious that although 1 + ln x increases as x increases, x increases “much faster” than 1 + ln x—with the consequence that as x increases, (1 + ln x)/x is going to get very small very fast. And it is intuitively evident that for every real number y greater than 0, a sufficiently large choice of x will result in (1 + ln x)/x having a value greater than 0 and smaller than y.

7

Correct or not, it ought to arouse our suspicions immediately, because it implies that the probability of the hypothesis that I shall exist forever conditional on the hypothesis that I have not “already” existed for an infinite period of time is 0. 8 Here is the sketch of a proof: (1 + ln x)/x = (1/x) + (ln x)/x. So, by a well-known fact about limits, the limit of f(x) = (1 + ln x)/x as x increases without bound is equal to the sum of the limits of f(x) = 1/x and f(x) = (ln x)/x as x increases without bound. We have seen that the limit of f(x) = 1/x as x increases without bound is 0. That the limit of f(x) = (ln x)/x as x increases without bound is 0 can be shown by an application of L’Hôpital’s rule.

THE REV ’ D MR BAYES AND THE LIFE EVERLASTING



If, therefore, this definition of Hold is adopted, the ratio Fold/Hold does not exist and the product of Fold/Hold and H//F does not exist. (For no x does 0  (x/0) exist.) And, therefore, Bayesian reasoning cannot be used to calculate the value of F//H.9

3 In closing, I will say something more about “vast but finite afterlife” cases. These are interesting for their own sake, but considering them will also afford an opportunity to examine and evaluate the Crucial Assumption (Personal). (If the position defended by Lara Buchak—see footnote 9—is correct, this examination will have important ramifications for the “infinite afterlife” case.) Imagine a species I will call the Centurions, since, without exception, they all live for exactly 100 years. The Centurions inhabit our own planet a million or so years in the future. Some Centurions are followers of a religion called Eonism, whose principal teaching is that when Centurions die, they continue to exist—but not eternally: the total span of their existence (including their century of pre-mortem existence) will be 5 billion years. Although they will not exist forever, they will exist as long as the sun and the moon endure. (Or till a’ the seas gang dry and the rocks melt wi’ the sun. The two periods are more or less equal, since the sun’s departure from the main sequence will do for the sun in its present form and the moon and the seas—if the moon and the seas still exist—in one fell swoop. And, as for the rocks, they will quite literally melt wi’ the sun.)

9 When this chapter was read at a conference at Purdue University in honor of Richard Swinburne’s 80th birthday, the commentator was Lara Buchak. In her brilliant comments on the chapter, Buchak presented an alternative way of calculating the value of H//F. If the value of H//F is calculated in that way, it is equal to 0. While I have questions and reservations about Buchak’s way of calculating H//F, the issues involved in evaluating it are very subtle, and I have to concede that she may be right. (She is certainly a much better mathematician than I am.) If I were to attempt to evaluate her proposal, however, this chapter would have to be much longer than it is. I will content myself with pointing out that her way of calculating H//F depends—as does mine—on a sort of “limit analogue” of the Crucial Assumption (Personal), and that in the sequel I shall present reasons for rejecting the Crucial Assumption (Personal).



PETER VAN INWAGEN

Can Bayesian reasoning be used to prove to the Eonists (during their earthly lives) that it is highly improbable that this central teaching of their religion is correct? We have already seen that (always granting the Crucial Assumption (Personal)) if one knows that one’s existence will end at some randomly chosen point in the next 10 billion years, the discovery that one has existed for at most 100 years lowers the probability of the hypothesis that one will exist for at least 5 billion years from 0.5 to about 0.036—lowers it to about one-fourteenth of its original value. On this pass, however, I’ll consider a different sort of Bayesian Argument, one that contains hardly any mathematics (no doubt to the reader’s considerable relief). Among the Centurions who are not Eonists, there are, so to call them, Eonic agnostics. Eonic agnostics hold that either Centurions enjoy no post-mortem existence or that the Eonists are right and that all Centurions exist for 5 billion years—and Eonic agnostics accept neither of the disjuncts of that proposition. Eonic agnostics, moreover, do not assign a negligible probability to either disjunct. One of them may assign a lowish probability to one of the disjuncts (1 percent, say), but none of them assigns to either disjunct a probability on the order of the probability that I assign to the proposition that I shall die at 11:00 a.m. on June 17, 2019 as the result of having been gored by a water buffalo while crossing Times Square. Can Bayesian reasoning provide them with grounds for lowering the probability that each of them assigns to the 5-billion-year hypothesis to some fraction, perhaps some minute fraction, of the probability he or she had hitherto assigned to it? Let us consider a particular Eonic Agnostic, Agnes. Every day, Agnes devotes some serious thought to the question: Which is true, “No postmortem existence” (NPE) or “Five billion years” (FBY)? One day, while she is thinking about this question, the Crucial Assumption (Personal) forces itself upon her mind as true (to borrow a phrase of Gödel’s). And that assumption, we remember, is: I may regard the present moment as a moment chosen at random from the interval (0, the final moment of my existence]. “Let’s see,” Agnes muses. “I’m 43 years old. I may therefore regard the number 43 as a number drawn at random from the number of years from

THE REV ’ D MR BAYES AND THE LIFE EVERLASTING



the beginning to the end of my existence.” Reflection on this fact—she supposes it to be a fact—leads her to construct a carefully formulated analogy: Suppose that there are two enormous urns, A and B, standing side by side. Urn A contains, as an urn should, only empty space within its walls. Urn B, however, could almost be described as a fake urn or a mock urn. It contains only a little empty space at the top; if you looked down into it, you’d see a sort of floor not very far below you: beneath that level surface lies only solid urn-stuff. The empty space in each urn is at some point filled with tennis balls. A then contains 5 billion balls and B only a hundred. A demon or some such agency proceeds to number the balls in each urn consecutively, starting with 1, and writing its number on each ball. The balls in A are, of course, numbered 1 to 5 billion, and the balls in B 1 to 100. The positions of the balls in each urn are then thoroughly scrambled or randomized, by some process I’ll leave to your imagination. Ernest, a Centurion who knows all this stuff but does not know which of the two urns before him is A and which is B, chooses one of the two by tossing a coin, reaches into the chosen urn, and withdraws a ball. (The distance between the opening at the top of B and its solid “floor” is considerably greater than the length of Ernest’s arm.) He examines the ball and discovers that it is numbered 43. He concludes that the urn he drew the ball from is almost certainly B—since he drew the ball from one or the other, and the probability of his having drawn it from B is much, much higher on the evidence he has (to wit, that the ball he drew was numbered 43) than is the probability of his having drawn it from A on that evidence.10 Agnes, after careful consideration of her own case and this imaginary case, reasons as follows: “My case and the case I have imagined are exactly parallel. Just as Ernest rightly concluded that the urn from which he drew the ball was almost certainly the urn that contained

10

Exercise: where A is the hypothesis that Ernest drew the ball from A, B the hypothesis that he drew the ball from B, and H the proposition that the ball he drew bore a number less than or equal to 100, calculate A//H and B//H. Assume that the prior probability of both A and B is 0.5. (After all, Ernest chose the urn he drew a ball from by a coin-toss.)



PETER VAN INWAGEN

only 100 balls, so I may conclude that the span of my existence is almost certainly 100 years—and that NPE is therefore almost certainly true and FBY almost certainly false.” Has Agnes reasoned rightly? Well, she was certainly right when she said that Ernest’s reasoning was correct. And if the Crucial Assumption (Personal) is true, she was certainly right when she said that her case was exactly parallel to his. But is the Crucial Assumption (Personal) true? I think that this thesis or principle or assumption, or whatever you want to call it, is at best extremely doubtful. I will explain why someone like you and me, a present-day analytical philosopher, should find it extremely doubtful “in his or her own case”: (remember that it is a first-person principle). (What I will say about why I should find it extremely doubtful and why you should find it extremely doubtful could easily enough be adapted to Agnes’s case.) Let us call a piece of reasoning that turns on or essentially involves the Crucial Assumption (Personal) a CAPA—for “Crucial Assumption (Personal) Argument.” Suppose that one is considering a CAPA at a certain moment t, and that this occasion is one of the first few occasions on which one has considered a CAPA. Can one regard t as a moment chosen at random from the whole span of one’s existence (including one’s postmortem existence, if such there be)? Or, more cautiously, can one regard t as a moment chosen at random from the whole span of one’s existence exclusive of the initial segment of one’s existence during which one was not yet intellectually capable of considering a CAPA? It seems to me that one cannot—not, at least, if one is an analytical philosopher or a cosmologist who is alive at the present time (a category in which you and I and John Leslie and Don Page all fall). It seems to me quite likely that one will encounter a CAPA during the first century of one’s existence (as you in fact have and I in fact have and Leslie and Page in fact have), however long the whole span of one’s existence may be. It was quite likely that Page would devise a CAPA during the first 100 years of his existence (given his interests, his cognitive capacities, his education, and the ideas available to him in the intellectual environment he had inhabited during his life up to the point at which he devised his argument). And, moreover, it was quite likely, given the way ideas or “memes” are transmitted in our culture, that this CAPA would find its way to most of us during the first century of our existences.

THE REV ’ D MR BAYES AND THE LIFE EVERLASTING



Now, if I knew that God had chosen a moment t0 at random from the set of moments at which I shall exist (a supposition that is possible only on the assumption that I shall exist only for a finite period of time) and had so arranged matters—by whatever means Providence arranges matters—as to ensure that I should first encounter a CAPA at t0 , then I should indeed conclude that the fact that t0 falls within the first century of my existence significantly lowers the probability that I shall exist for a number of years that is a large multiple of 100. But nothing suggests that this is so—or suggests that any other factor has ensured that the moment at which I first consider a CAPA (or any other moment at which I consider a CAPA) can be regarded as a moment chosen at random from the whole span of my existence. As we have seen, moreover, there are reasons for regarding it as doubtful whether the moment at which I first consider a CAPA is a moment chosen at random from that span. And, of course, when I am considering a CAPA, that moment is a moment at which I am considering a CAPA. There is, therefore, no reason to suppose that when I am considering a CAPA, I have any right to regard that moment—the moment the CAPA refers to as “the present moment”—as a moment chosen at random from the set of moments at which I exist. I will supplement these arguments with an “intuition pump.” Consider a person I’ll call Alice because her real name is unpronounceable by human beings. Alice was a member of a long-extinct species that flourished about 7 billion years ago. Nevertheless, she was a colleague of ours, a philosopher of our sort, and like most of us present-day human philosophers, she eased into that condition over some rather fuzzy stretch of time during her twenties. She worked at philosophy for pretty much the rest of her life till she died at the age of 89. But that was not the end of her: like all members of her species, she had a post-mortem existence of 5 billion years (less 89 years in her case). The obliging God of the Philosophers has chosen for us at random a moment t from within the 5 billion years of Alice’s existence. At some point during that 5 billion years, moreover, Alice first encountered a CAPA. Now, which do you think probably came first: t or the moment at which she first encountered a CAPA? My conclusion is that when you consider an argument for the conclusion that you are almost certainly not going to have a long but finite post-mortem existence, an argument that depends on the principle that



PETER VAN INWAGEN

you may regard the present moment as a moment chosen at random from the total span of your existence, the moment you then call “the present moment” cannot be regarded as a moment chosen at random from that span: that moment was in effect chosen by the fact that it is a moment at which you are considering a CAPA. You know, moreover, that this is one of the first few times that you have considered a CAPA. And you have no reason to regard that moment, the moment you call the present moment or this moment when you are considering a CAPA for the first time or one of the first few times, as a moment chosen at random from the set of moments at which you will exist. And you have at least some reason to suppose that that moment was likely to do what it has in fact done: to fall within the first century of your existence, and to have done that even if you are going to exist for a thousand years or for 5 billion years—or for ever. I will end with a parable. Christian set out one day to make his journey to the Celestial City. He passed through the Gate which is called Baptism and he began to walk along the Road to the Celestial City. That Road is as straight as any arrow, but no one who has not traveled its whole length can know that length—which for all the traveler knows when he commences his journey may be only a few short miles or it may be many thousands of miles. Christian had walked but a little way, not so much as three miles, when he came to a stone column on which some words were inscribed. “I will read these words,” said Christian. And the words he read were these: Reason, O traveler, as follows: “I may regard my present position on the Road as a point chosen at random from its whole extent. I have traveled only a few miles. It is therefore likely that I have only a few more miles to travel.” As Christian was pondering these words, he looked up and saw on the Road an old man leaning on a staff, and the man’s name was Right Reason. Right Reason said to Christian, “I see, O Christian, that thou hast read the Words of False Comfort.” “Certes they be words of comfort,” Christian replied, “but why dost thou call their comfort false?” And Right Reason answered, “Dost thou not see that the place which the Words entreat thee to call ‘my present position on the Road’—and in that they do not deceive thee, for it is indeed thy present position on the Road—is no more than the place at which thou hast encountered them? Art thou so blind that thou canst not see that the column on which those Words are written hath been put in such a place as to stand very near that Gate whereat thou beganest thy journey— and that this is the place at which thou must encounter them, howsoever long the Road may be? Didst thou know that whatever agency had set the column in this place had chosen this place at random from the whole extent of the Road, then

THE REV ’ D MR BAYES AND THE LIFE EVERLASTING



thou shouldest indeed reason as the Words entreat thee to reason. But think on this, Christian: thou dost not know this—nor knowest thou of any reason which argues that it is so.”11

References Leslie, John. (2008). “Infinitely Long Afterlives and the Doomsday Argument,” Philosophy, 83: 519–24. Page, Don N. (2010). “Scientific and Philosophical Challenges to Theism,” in Melville Y. Stewart (ed.), Science and Religion in Dialogue, Vol. I. Oxford: Wiley-Blackwell, 396–410. (An online version of this paper is available at .) van Inwagen, Peter. (2005). “Human Destiny,” in William E. Mann (ed.), The Blackwell Guide to the Philosophy of Religion. Oxford: Blackwell, 245–65.

11 Here are two important questions I have not considered: (1) Can the arguments I have given to show that the Crucial Assumption (Personal) is doubtful be adapted to show that the original Crucial Assumption is doubtful? (I had thought not, but conversation with Brian Cutter has convinced me that my reasons for thinking this were mistaken.) (2) If the Crucial Assumption does stand, does the conclusion of the original Doomsday Argument, which depends on the Crucial Assumption (full stop) but not on the Crucial Assumption (Personal), imply that there is no personal immortality?—or no post-mortem existence that is very, very long in comparison with threescore years and ten? I have very briefly addressed this question in van Inwagen (2005: 261–2).

10 What about Hylomorphism? Some Medieval and Recent Ruminations on Swinburne’s Dualism Marilyn McCord Adams

1 Swinburne’s Dualism In the second edition of his book The Evolution of the Soul, Swinburne announces his intention to defend soul–body dualism (Swinburne 2007: 8–9). Lurking in the background is his notion that a full story of the world must be comprehensive in the sense of including all the kinds of events and substances that there are. Front and center is his insistence on the unmarginalizability of the mental: for Swinburne, mental events (sensations, thoughts, purposes, desires, and beliefs, not to mention moral awareness and free choice) are data not to be dealt with the way secondary qualities are in Cartesian and contemporary physics. On the contrary, mental events feature among the primary explananda that must be given pride of place in any adequate theory of the way things are (Swinburne 2007: 3). In The Evolution of the Soul and—all the more so—in his more recent book Mind, Brain, and Free Will (2013), Swinburne’s targeted opponents are materialists, who assume that the physical sciences (physics and chemistry) as we know them constitute a theory of everything, and defend this conviction against alien incursions by holding that the mental is somehow reducible to or at bottom constituted by the physical (2007: xiii–xiv). Swinburne is out to “escape” the mental from their theoretical domains into a room of its own.

WHAT ABOUT HYLOMORPHISM ?



Swinburne’s strategy is to move from event or property dualism which recognizes that mental events and properties are irreducibly different in kind from physical events and properties, to substance dualism, which concludes that mental and physical substances are fundamentally different in kind. Nevertheless, in The Evolution of the Soul, Swinburne goes out of his way to give the physical sciences their due. A full story of the world will have to acknowledge their achievements and incorporate their results. Accordingly, Swinburne declares that the findings of neuro-psychology demonstrate de facto cradle-to-grave dependence of mind on the brain (2007: 145–99, esp. 175, 280; 2013: 117). Further, he makes it criterial: my body is the one through which my mind is relevantly causally interactive and from which it takes its perspectival point of view (2007: 146, 333; 1977: 102–3). From the strength and pervasiveness of mind– brain correlations, Swinburne concludes that even though the soul is a distinct substance from the body, empirical evidence would not lead us to expect that the soul and its mental functions would survive separation from the body (2007: 176–9, 298–311). Brain transplant thought experiments convince Swinburne that it is, however, logically possible for the soul to exist without the body (2007: 147–52, 307, 322–4; 2013: 163–5).1 Christian Scriptures and creeds inform him: this possibility will be actualized (Swinburne 2007: 311–12). Yet, if a full story of the world includes the mental as well as and as distinct from the physical, Swinburne despairs of discovering some superscience that would embrace and give a unified treatment of both. No matter how many mind–brain correlations accumulate, Swinburne thinks there is no hope of our arriving at scientific natural-law-governed explanations of why mental events are correlated with physical events at all, much less of why these mental events are correlated with those brain states. Our full story of the world should aim for descriptive completeness. But a unified and comprehensive scientific explanation of everything lies beyond our reach (Swinburne 2007: xi–xiii, 3–4, 185–96, 198–9, 307).

2 Remarkable Omission? Swinburne’s polemical context explains his initial estimate of the theoretical options: “hard” and “soft” substance dualisms, “hard” 1

See also Swinburne (1994: 19–20, 22) and Swinburne (2004).



MARILYN M C CORD ADAMS

(i.e., reductive) materialisms, and “soft” materialisms (which make physical substances the subjects of both physical and mental properties) (2007: 8–9). My own medieval preoccupations prompt me to ask: what about hylomorphism, that family of theories which see things here below as shaped-up stuff, as constituted by matter and substantial form? What about the view shared by Aquinas, Scotus, and Ockham (albeit with significant variations), that the human soul is the form of the body—not a distinct substance from the body, but the formal principle which structures the human being into a single substance with distinctive functional integrity? Does Swinburne leave hylomorphisms off his opening list because they did not strike him or his opponents as live options? Or did Swinburne suspect that his substance dualism might not be so different from hylomorphisms after all? Surprisingly, the answer is not obvious.

2.1 Surface similarities For starters, in The Evolution of the Soul, Swinburne forwards what medievals would find a “thin” view of substance. He maintains that a substance is just a particular subject of properties that is a continuant and causally interactive (Swinburne 2007: 5). In Mind, Brain, and Free Will, he adds that substances are not processes, but exist all-at-once (Swinburne 2013: 4). Aquinas, Scotus, and Ockham agree that the intellectual soul is a concrete particular, a causally interactive continuant, and the proximate subject of properties of which bodies cannot be proximate subjects. Because they reify qualities, they also hold that qualities are concrete particulars, causally interactive continuants, and subjects of properties. So they could agree that all of these are Swinburnean substances. Yet, for Aquinas, Scotus, and Ockham, neither human intellectual souls nor reified qualities count as complete substances in their sense. Swinburne emphasizes that he is not a “hard” Cartesian dualist who believes that the soul is by nature simple and immortal. Swinburne thinks no empirical evidence puts us in a position to predict post-mortem survival (2007: 10). Likewise, early Scotus2 and 2 The Opera Omnia of Scotus are found in two editions: the 1639 Lyon edition by Lucas Wadding in twelve volumes (hereafter Wadding, followed by volume number and page number); and the as yet incomplete 1950–2008 Vatican edition under the direction of Carl Balic

WHAT ABOUT HYLOMORPHISM ?



Ockham3 find Aquinas’s argument that the intellectual soul is capable of independent existence to fall short of the standard of evident proof.4 One way or another, whether by nature or by Divine power, Swinburne and my trio of hylomorphists affirm: the intellectual soul does survive. Aquinas, Scotus, and Ockham also agree with Swinburne about cradleto-grave functional dependence of the mental on the physical, while insisting that—one way or another—independent mental functioning will go on between death and resurrection.5 Thus, when it comes to the human intellectual soul, Swinburne shares considerable common ground with these hylomorphists. They concur that the intellectual soul is the proximate subject of thought and choice, that it is actually but possibly not functionally dependent on the body, that it is not demonstrably but possibly—by Divine power, likely—the survivor of death.

2.2 Muddied waters My interest is in how Swinburne’s dualism relates to medieval theories that the intellectual soul is the (dominant) substantial form of the human (hereafter Vaticana, followed by volume number and page number). Allan B. Wolter has also issued several volumes of Latin texts with English translations, sometimes basing his text on independent consultations of the manuscripts. I will later refer to his edition from the manuscripts and translation of a Christology question in Franciscan Christology, ed. Damian McElrath (St. Bonaventure, NY: Franciscan Institute Publications, 1980) (hereafter McElrath 1980, followed by page numbers). Note further that Scotus commented on Peter Lombard’s Sentences at least three times. In the Vatican edition, the earliest set is entitled Lectura, while the next set given at Oxford is entitled the Ordinatio. In the earlier Wadding edition, what the Vatican edition entitles Ordinatio is called Opus Oxoniense. The present reference is to Scotus, Opus Oxoniense IV, d. 43, q. 2, nn. 13 and 16, 29–33; Wadding XI.2, 836–8. See also Quodlibeta I, q. 9; Wadding XII, 227. 3 Ockham’s Opera Theologica are edited by the Franciscan Institute in nine volumes (hereafter OTh, followed by the volume number and page number); his Opera Philosophica by the Franciscan Institute in seven volumes. The present reference is to Ockham, Quodlibeta I, q. 10; OTh IX, 63–5. 4 Aquinas, Summa Theologica I, q. 76, a. 1, c. Scotus, Opus Oxoniense IV, d. 43, nn. 13 and 17; Wadding XI.2, 836–7. Ockham, Quodlibeta I, q. 10; OTh IX, 63–5. 5 Aquinas thinks it is essential to human cognitive psychology that there is no abstract thought without phantasms, and so thinks intellectualizing between death and resurrection will involve a miraculously enabled mode of cognition (Summa Theologica I, q. 89, a. 1, c and ad 3; a. 2, c and ad 2; a. 3, c). Scotus (Ordinatio, Prologus I, q. un, nn. 29 and 61–2; Vaticana I, 17–18, 37–8; Ordinatio I, d. 3, p. 1, q. 3, n. 187; Vaticana III, 113–14; Opus Oxoniense IV, d. 45, q. 1, n. 7; Wadding X, 163; q. 2, nn. 12 and 16; Wadding X, 182–3) maintains that—because being is the proper object of the intellect—human souls will be able to intellectualize independent of bodily sensation after death.



MARILYN M C CORD ADAMS

body. Toward the end of both books, where he is treating issues about synchronic and diachronic identity, and in his earlier work, The Christian God, Swinburne not only gives hylomorphists explicit attention, but he also flirts with appropriating hylomorphic conceptuality. First, he declares that physical substances have a hylomorphic composition: the matter is the concrete particular subject, while the form is their properties which are universal, apparently abstract objects (1994: 11, 35, 46). Certainly, such usage would not encourage Swinburne to say that the soul is the form of the body, because he (and all medieval parties to the dispute) understand human souls to be concrete particulars. Aquinas, Scotus, and Ockham all insist that abstract objects as such have no extra-mental existence; properties are completely universal only as objects of thought. Insofar as properties constitute or characterize substance particulars, they also are individual. When Aquinas, Scotus, and Ockham say that the soul is the form of the body, they understand that form to be a concrete particular. Swinburne’s speculative focus is somewhat different: on how souls might be individuated. If matter individuates physical substances, maybe matter is the individuator for souls as well. Swinburne cites Bonaventure’s commitment to the hylomorphic composition of souls (2007: 327–32; 1994: 46–9).6 Caution is in order, however. For matter plays many theoretical roles in medieval hylomorphisms: first and foremost, it is the substratum of substantial change; second, it is the ultimate subject of the properties; third, it plays a role in individuation. To these, Aquinas adds a fourth: that matter is potency limiting act (e.g., contracting bovine nature into Beulah the cow). Bonaventure adds yet another: that matter is the principle of contingency that explains the metaphysical possibility of non-existence. Aquinas and Bonaventure agree that the profound difference between a necessarily extant God and contingent creatures ought to be reflected in their metaphysical constitution. Bonaventure’s proposal is that all creatables—angels and human souls included—are matter–form composites, while Godhead alone is utterly immaterial.7 Aquinas fears that because matter is

6

But see Swinburne (2013: 170–1) for Swinburne’s change of mind. Bonaventure’s Opera Omnia are edited in ten volumes and published by the College of St. Bonaventure at Quaracchi (hereafter Quaracchi, followed by volume number and page 7

WHAT ABOUT HYLOMORPHISM ?



already pressed into service as an individuator, hylomorphic composition of intelligences would make it impossible for them to exercise their natural functions. Given his cognitive psychology, the reason is straightforward. Cognition is literally in-formative. Intellectual cognition is pre-eminently of universals, but forms received into material subjects are individuated thereby.8 Aquinas’s counterproposal is that all creatables are compositions of essence and esse, while essence is identical with esse in God alone. Aquinas argues that—because intellectualizing is the human soul’s best function—the human intellectual soul is not itself composed of matter and form. Rather it is the form of the human body. But— Swinburne joins commentators down the centuries in demanding to know—how could it be individuated in numerically distinct human beings? Aquinas usually maintains that individuation is due to matter “signed by quantity.” Swinburne sees Aquinas as caught in a dilemma. Either human souls are individuated by quantified matter, in which case they will not retain their individuation when separated from the body; or they are individuated independently of body, in which case, how is Aquinas’s position so different from the substance dualism Swinburne espouses?9 Here again, caution is in order. Instructed by Stump, Swinburne recognizes that Aquinas’s view is slightly more complicated: not that the soul is individuated by actual inherence, but rather by “its relation to matter” “on the occasion”10 of its actual inherence—by its aptitude to combine with this hunk rather than that (Swinburne 2013: 171–2). Instructed by Scotus, Swinburne responds that this will not do, because being this very individual thing is intrinsic to an individual. Nothing is individuated by its relation to something extrinsic to it (Swinburne 2013: 171–3).

number). The present reference is to Bonaventure, Commentaria in Quatuor Libros Sententiarum Magistri Petri Lombardi II, d. 3, p. 1, q. 1, qq. 1–3; Quaracchi II, 88–102. 8 Aquinas, Summa Contra Gentiles II, chs. 50 [2], 59 [2], 69 [12]; Summa Theologica I, q. 75, a. 5 c and ad 1um and ad 4um; q. 76, a. 2 ad 3um; Com. De Anima, Lect III, lectio vii, secs. 684–5. 9 Aquinas, De ente et essentia, c. 2, Roland-Gosselin ed., 11–12; Summa Theologica I, q. 75, a. 7, c; q. 76, a. 4 c; q. 76, a. 6 ad 2um; q. 77, a. 2 c. 10 Aquinas, De ente et essentia, c. 5; Roland-Gosselin ed., 39–40.



MARILYN M C CORD ADAMS

Yet, even if we agree with Scotus and Swinburne that souls are individuated by thisnesses (primitive unsharables), the waters are still muddy. For the claim that the soul is the form of the body can be and was combined with a variety of theories of individuation. Aquinas, Scotus, and Ockham all insist that the soul is the form of the body, but they disagree about individuation as much as any philosophers can.

3 Striking Contrasts Since the waters are already muddy, let me stir them up some more. In The Evolution of the Soul, Swinburne begins with mental–physical property dualism and infers soul–body substance dualism. These moves are multiply vexed from a medieval point of view.

3.1 Peculiar property divisions First, medievals would be startled by Swinburne’s epistemic “publicversus-private” rubric for distinguishing the physical from the mental. Thus, in The Evolution of the Soul, he declares that mental properties are those “to which one subject has privileged access, which he is necessarily in a better position to know about than anyone else,” while a physical property is “one that no subject is better placed to know that it is instantiated than is any other subject” (2007: 6). In Mind, Brain, and Free Will, he defines “a mental property as one to whose instantiation in it a substance necessarily has privileged access on all occasions of its instantiation, and a physical property as one to whose instantiation in it a substance necessarily has no privileged access on any occasion of its instantiation” (2013: 67–8). In this more recent work, Swinburne assures, it is not that others cannot know about the mental properties at all, but that the subject “has a further way (by experiencing it) which it is not logically possible that others can use” (2013: 67–70). Of course, if the extra way I have of knowing my mental state is my experiencing it, it is trivially true that no one else can do that. But privileged access was supposed to be a non-trivial doctrine, and nontrivial privileged access was not thought to be a marker of the mental in medieval times. Scotus and Ockham declare: intellects and their acts are knowable in principle. If being is the proper object of the intellect, only contingent obstacles prevent one created mind from seeing into

WHAT ABOUT HYLOMORPHISM ?



another—in their language, from having an intuitive acquaintance with other minds and their mental acts. Neo-platonist pictures suggest that higher-up intellects have cognitive access to whatever is known by lowerdowns. The well-known eleventh-century collect for purity addresses “Almighty God, unto whom all hearts are open, all desires known, and from whom no secrets are hid.” The Divine persons do have privileged access to the Divine mind in that they alone have intuitive knowledge of it by natural necessity. Others have cognitive access to it only insofar as it is freely and contingently granted by the Divine will. The mental life of God enjoys privacy, because creatures have no natural powers to pry. Yet, happily, because it is metaphysically possible, Godhead is free to share Itself with others! In The Evolution of the Soul, Swinburne—having identified physical properties as the ones that are publicly accessible—only belatedly forwards the criterion that every body is extended. In Mind, Brain, and Free Will, he comes closer to medieval notions when he says that physical substances are essentially divisible. For medievals themselves, what is in the genus of body has parts that can be and naturally are “outside”—at a distance from—other parts. Some bodies may have some constituents with parts “outside” parts and other constituents that lack them. For example, Aquinas holds that the human soul combines with prime matter to constitute the human body, but the intellectual soul itself is simple, neither divided nor divisible into integral parts. Moreover, the contrast between what cannot versus what can be extended does not coincide with the distinction between souls and everything else. Plant and non-human animal souls—for Ockham, also the human sensory soul—are extended. According to Aquinas, prime matter is not of itself divisible, but rather capable of being rendered divisible and divided by the inherence of quantity. For Scotus, prime matter is essentially divisible; for Ockham, it is essentially divided into parts which may or may not be at a distance from one another. Neither is the division between material and immaterial merely a function of extension. True, what is immaterial has no parts, while the material is divided or at least divisible into parts. But there is the further question of whether or not a form is the form of some matter, which it will be only if it is or includes formal functional principles for organic functioning; and whether it is or includes formal functional principles for



MARILYN M C CORD ADAMS

inorganic functioning. Angelic forms structure and empower angels for understanding and willing, neither of which is an organic function. So angelic forms are immaterial thrice-over: they are simple and indivisible, not the form of any matter, and include formal functional principles for the inorganic functions only. Cow and human souls are the form of some matter, because both include formal functional principles for organic functioning (respiration, digestion, reproduction, sensation). But the human soul is immaterial in the double sense of being without parts and being or including formal functional principles for understanding and willing.

3.2 Improper property-to-substance inferences Swinburne reckons that (pure) mental properties belong to mental substances, while (pure) physical properties belong to physical substances. Mental properties belong to human beings by virtue of belonging to their souls; physical properties, by virtue of belonging to their bodies (Swinburne 2007: 145–6). By contrast, medieval hylomorphists would not recognize any straightforward inference from property-dualism to substance-dualism. True, they agree that there are properties of which neither the inanimate nor animate body can be the proximate subject, e.g., thinking universal concepts or syllogistic schema. Likewise, there are properties of which immaterial things cannot be the proximate subject, e.g., being red, being six feet tall. But where Swinburne treats sensation as a mental property belonging to the soul, Aquinas11 and Scotus12 reckon that the proximate subject of sensation is not the soul, but rather the living sense organ itself. Still more remarkable is the Aristotelian idea that inanimate bodies can be subjects of intentional contents. For Aristotelian physics denies action at a distance. For the eye to see the red wall, “same form inherence” cognitive psychology requires the form or species of redness to be propagated through the medium of the air into the aqueous humor of the living eye. Following Arab commentators, Aquinas distinguishes two ways in which a form may be received: naturally, in such a way that the subject is characterized by the form; or spiritually or intentionally, in such a way that it is not characterized. Since neither the medium nor the 11 12

Aquinas, Summa Theologica I, q. 75, a. 3, c; q. 76, a. 5, c; q. 89, a. 1, c. Scotus, Opus Oxoniense IV, d. 44, q. 2, nn. 5–6; Wadding X, 129–30.

WHAT ABOUT HYLOMORPHISM ?



eye jelly becomes red, it must be that they receive the form not naturally, but rather spiritually or intentionally!13 Descartes’s interest in dualism is driven by his desire to secure res extensa as the domain of mechanistic physics. Swinburne’s motivation is the other way around: a desire to snatch the mind and the mental out of the physicists’ hands. But Swinburne is happy to “divide and conquer,” handing plant and animal bodies over to physical sciences that treat besouled and soul-less bodies as made up of the same fundamental stuff and governed by the same natural laws. In Swinburne’s terminology, all bodies are of the same mini-essential kind. For medieval Aristotelians, however, body is a high-level genus immediately under the category of substance, while the natures of things are the infima species, fully differentiated kinds. Cow bodies, oak tree bodies, and rocks are definitely not of the same mini-essential kind. This means that from the sameness of physical properties such as the color, flavor, odor, temperature, and texture, one cannot infer that their subjects all belong to the same mini-essential kind.14 Scotus makes the same point about what Swinburne would regard as mental properties. Angels and humans not only differ in species, but they also differ in genus: angels under the genus incorporeal substance; humans under the genus body. Nevertheless, their acts of understanding and willing are of the same species. So, once again, we cannot move from same-kind of property to same-kind of subject.15 Aquinas is clear: animal bodies and souls each may be proximate subjects of various properties, but they are not the proper subjects. He agrees with Swinburne that the proper subject is a concrete particular substance, but disagrees about what concrete particular substances there are. Important for present purposes, neither animal bodies nor animal

13

Aquinas, Summa Theologica I, q. 14, a. 1, c; q. 78, a. 3 c and ad 2um and ad 3um. The fact that, for medievals, body is a genus and not an infima species, also keeps hylomorphists from numbering among Swinburne’s “soft” materialists. Mutatis mutandis for the differences between contemporary and Cartesian physical sciences, Swinburne’s soft materialists start with res extensa, but then conclude that they have to stick some mental properties in it after all. When hylomorphists say that Ferdinand the bull is black and sees red, they are predicating ‘black’ and ‘sees red’ of an animal, of what is essentially a living bovine substance. 15 Scotus, Opus Oxoniense III, d. 13, q. 4, nn. 1, 18; Wadding VII.1, 261, 275; III, d. 14, q. 2, n. 17; Wadding VII.1, 297; III, d. 14, q. 4, n. 1; Wadding VII.1, 309. See also Ockham, Quodlibeta III, q. 6; OTh IX, 225–8. 14



MARILYN M C CORD ADAMS

souls qualify as Aristotelian substances. The soul is the formal functional principle that is the root-structuring principle of the substance. The body is formed out of matter to aid and abet the animal’s natural functions that the soul empowers. The Aristotelian substance is the animal whose besouled body it is. Since neither animal soul nor animal body is a fullfledged Aristotelian substance, a fortiori soul and body are not distinct Aristotelian substances. Because Aquinas believes there can be only one substantial form per thing,16 his conclusion is even more radical: an animal body is not even really distinct from an animal soul, because the animal soul is a metaphysical constituent of the body, along with prime matter that has no actuality of its own.17

4 Theoretical Fundamentals Such a jumble of incongruities between Swinburne and medieval hylomorphists signals deep-structure differences.

4.1 A hierarchy of sciences Swinburne’s “hard” materialists see contemporary physical sciences as bound to reach their ideal of a unified science. Their strategy for winning through to a naturalistic theory of everything is to reduce everything that does not look physical to something that obviously is. Swinburne’s strategy for spoiling their party is—once again—“divide and conquer”: res extensa is theirs to explain, but res cogitans is freed for personalistic rational explanation. Aristotelians approach the explanatory project with a different philosophy of science.18 The goal of inquiry is not a single, but a whole hierarchy of sciences. Sciences are defined by their proper object or subject matter. The most general sciences get at fundamentals one way, but the special sciences are needed to tell a full explanatory story of the world. Thus, while metaphysics studies what pertains to plants and animals insofar as they are beings, and physics analyzes natures 16

Here Aquinas’s position contrasts with that of Scotus and Ockham, who recognize a plurality of substantial forms in living things. See my (1987: 633–69). 17 Aquinas, Summa Theologica I, q. 76, a. 4 c and ad 1um. 18 For a discussion of medieval treatments of Aristotle’s hierarchy of the sciences, see Livesey (1982).

WHAT ABOUT HYLOMORPHISM ?



insofar as they are sources of change, both are too general to explain the distinctive vital functions of biological organisms. Moreover, Aristotelians recognize a certain autonomy among the sciences, with the specialized sciences proceeding from their own first principles by methods adapted to their particular subject matters. Metaphysics can tell us that plants and animals are substances, and physics can predict their free fall from a cliff. But neither suffices to explain why broad-leaf plants lose their leaves, why cows have necks and multiple stomachs, or why animals have the sense organs they do and why they are arranged as they are for the duration that they have. For answers to these questions, we need to look to the specialized sciences of biology and psychology.

4.2 Natural kinds? I have already remarked on Swinburne’s “thin” conception of substance as a continuant and causally interactive subject of properties. In many passages of his books, Swinburne appears also to have a metaphysically light understanding of natural kinds. When he speaks of what it would take to tell the history of the world, he repeatedly insists that accurate accounts could be achieved by alternative conceptual schemes (1994: 12–15, 26–7, 44; 2007: 150, 298; 2013: 142–3). The abandon with which he suggests that it might not matter whether we divided animals into water-dwellers and flyers instead of fish, birds, and mammals (2013: 22–3), not to mention talk of humans shape-shifting into gorillas (1994: 31–2), would send Aristotelian biologists reeling. Swinburne seems to imply that choice among alternative conceptual schemes is purely pragmatic: conceptual carvings of the world into kinds are meant to be useful, to serve an epistemic function as markers for prediction. Even when Swinburne talks about essences and allows that kinds (especially mini-essential kinds) might be essential to individuals (1994: 33), these remarks are so woven into a tapestry of observations about what we would say and what usage (loose or tightened) might permit, that I come away with an anti-realist feeling. My reaction intensifies with Swinburne’s ready lumping of artifacts with plants and animals, and his sometime suggestions that what counts as a miniessential kind (motor vehicle versus car or boat) is up to us (1977: 242–4, 246–8; 2013: 28–39). This makes it sound as if it is a function of our conceptual choices and linguistic practices whether X is essentially F, and so a matter of our conceptual and linguistic practices whether



MARILYN M C CORD ADAMS

X could exist without being F. Or, is Swinburne like Wittgensteinians who in effect tell us that it is a matter of our linguistic practices (the language games that we play) that it is not a matter of our linguistic practices whether X could exist without being F? Wriggling around such maneuvers, I want to protest: “Forget about our telling the story of the world! What kinds of things would be mentioned in God’s narrative? Is God a realist about natural kinds? If so, which?” Perhaps Swinburne would join Wittgensteinians in refusing the questions. My hunch is that if he accepted them, he would give the answer that the category “dualism” should lead us to expect: res extensa and res cogitans. For medieval Aristotelians, substances are not merely causally interactive continuant subjects of properties; rather, they are marked by functional integrity that is metaphysically underwritten by their being essentially instances of a natural kind. This understanding fits with Aristotelian scientific method. Aristotelians agree that the history of the world involves always or for the most part regularities and determinate sequences. They assume that what is always or for the most part cannot “just happen,” but instead requires to be explained. Functional powers are posited in functioning things to explain functional regularities. Most substances engage in a variety of functions and so have many functional powers. Thus, there is a further regularity: always or for the most part those powers travel together and always or for the most part are exercised in a coordinated fashion. Substantial form is posited in the functioning thing to explain these regularities. Substantial form is what makes an individual be the thing it is and so places it in its infima species. Aristotelians are metaphysical realists about natural kinds, which would be as they are whether or not any humans theorized about them. It is, however, the job of science to investigate them. Aristotelian science moves from a determination of the facts to an identification of the explainers and explananda, and from thence to the formulation of a real definition of natural kinds in terms of genus and differentia. The results of scientific inquiry are displayed in syllogisms in which the real definition is the middle term, and explains why the predicate is true of the subject in the conclusion.

4.3 Unexplained correlations? Swinburne thinks that there is not a ghost of a chance that human beings will come up with a super-science that will explain why the mental

WHAT ABOUT HYLOMORPHISM ?



should be correlated with the physical, why this sort of mental life with this body, why these mental events with those bodily states rather than others.

4.3.1

NATURAL TELEOLOGY

Aristotle agrees that the properties of the elements could not explain why plant and animal bodies are always-or-for-the-most-part shaped as they are, why they remain so for an always-or-for-the-most-part determinate life span, instead of the elements out of which they are constructed breaking loose and heading off in different directions toward their natural places (earth to the center, fire to the outer perimeter, water and air in between). Aristotle insists that physics and chemistry are general sciences whose principles do not address themselves to questions of biological structure, much less biological and psychological dynamics. Nevertheless, Aristotle believes in the unmarginalizability of the biological. For Aristotle, it can be no accident that vital functions and organic bodies go together. It is these data that most of all demand explanation. To achieve this, the special methods of biology and psychology are needed. And in these sciences, teleological explanation takes center stage. Characteristic functions are the goal or telos. Functional powers and static structures in the functioner are conditions of the possibility of such functioning. They are conditionally necessary for reaching the functional goal. Aristotle asks: why do animals with lungs have necks? He answers: because in animals with lungs, lungs serve for respiration. Windpipes are needed to deliver the air, and necks are required to house the windpipes. Bodily structure cannot be explained by the lower-level regularities of elemental and mixed bodies, any more than vital functioning can be an efficient causal resultant of their chance convergence. Organic bodies exist for the sake of the vital functions they make possible.19 Aquinas develops this idea when he posits an explanatory structure within the substances themselves. The substantial form is the root that gives rise to (he says, “emanates”) and coordinates the functional powers and structures matter into a body with tissues and organs apt to receive 19

For an excellent analysis of Aristotle’s scientific method in biology, see Lennox (2001).



MARILYN M C CORD ADAMS

the powers:20 e.g., multiple stomachs for digesting grass; long jaws with flat teeth for chewing cud. It is no mystery why this sort of body is associated with this type of soul or why this soul acts and experiences through this body. Aquinas insists that the body exists and has the structure it does for the sake of the soul. And the living organ is the proximate subject of such organic functions.21

4.3.2

KNOWLEDGE BY ASSIMILATION ?

Swinburne joins reductive materialists in the judgment that the mental and the physical are so different from one another that it is difficult to see how there could be a natural explanation of what and why they have to do with each other. Interestingly, some medievals including Aquinas did not share this worry, at least where cognition is concerned. Mixing Neo-Platonism with Aristotle, they hold that the world is intelligible; that the same intelligible content can be received into subjects of different sorts; and that what is received, is received after the manner of the receiver. Thus, Aquinas takes a page from Avicenna when he comments that in creation, God pours the forms of natural kinds at once into matter and into angelic intellects.22 In each case, the role of the form is to explain structure: specifically the same form shapes up matter into cows and horses and informs the intellect with thoughts of what it is to be a cow or a horse. Aristotle’s own cognitive psychology wrestled with Plato’s puzzle: how can you think of something without having it in mind. Aristotle’s answer was that one can come close to having it in mind by receiving its form without its matter (if any). Human cognition begins with sensation. Once again, psychological functioning dictates bodily structure. Only certain sorts of matter are apt to receive the sensible form. The eye as the organ of vision has to be filled with aqueous humor because only transparent media are apt to receive color forms intentionally. Aquinas joins Averroes in assuming that only immaterial things can be the proximate subjects of understanding. But he also agrees that it is 20

Aquinas, Summa Theologica I, q. 77, a. 6, c and ad 2um; a. 7, c; a. 8, c. Aquinas, Summa Theologica I, q. 75, a. 3, c; 76, a. 5, c; q. 89, a. 1, c. Pace Ockham, Quaestiones in Librum Quartum Sententiarum, q. 9; OTh VII, 162–4; Quodlibeta II, q.10; OTh IX, 158–9. 22 Aquinas, Summa Contra Gentiles II, c. 50 [5]; c. 51 [3 and 4]. 21

WHAT ABOUT HYLOMORPHISM ?



essential to human intellectual functioning—given that we are the dimmest wits in the universe—that we get our intellectual contents only by abstraction from sense images. So the human body must have not only sense organs for the external senses, but also a fancy phantasmsorter to prepare the way for abstraction.23 If the human intellect is bottom of the line, its phantasm-sorter must be top of the line, and because the senses and imagination are organic powers, the human body must be structured accordingly.24 Here again, knowledge is by assimilation. Intelligence is a different sort of subject from matter, but the same species of form shapes intelligence into thoughts and stuff into things. Given ancient and medieval cognitive psychology, there is no mystery here. Functional teleology explains why bodily organs are structured the way they are.25 Even though the medievals inherited a biology that did not assign the brain a major place in mental life, the same teleological explanatory strategy could be deployed to the mind–brain correlations to which Swinburne gives so much weight. Looking across the animal kingdom, a contemporary Aristotelian could say that big brains or brains with a certain structure exist for the sake of doing philosophy!

4.4 Summary contrasts It is time to bring together the contrasts uncovered so far. Swinburne posits substance-dualism. There are two fundamentally distinct kinds of substance: souls that are subjects of mental properties, and bodies that are subjects of physical properties. Probably, for Swinburne, res cogitans and res extensa are real essences, but human being and other plant and animal kinds are not. A human being has two parts: one of which is essential—namely, the soul—and the other of which is non-essential— namely, the body (2007: 333; 2013: 170). My body (if any) is the one from which my soul takes its point of view on the world and the one through which it interacts with the world. Swinburne is a “soft” dualist in that he thinks the vast correlations between mind and brain leave us with no reason to believe that the soul could function without the body. At the 23

Aquinas, Summa Contra Gentiles II, c. 60 [1]; c. 67 passim. Aquinas, Summa Theologica I, q. 79, aa. 2–4; q. 85, a. 1, c and ad 2um and ad 3um and ad 4um. 25 Aquinas, Summa Theologica I, q. 85, a. 2, c and ad 1um and ad 2um. 24



MARILYN M C CORD ADAMS

same time, neither the physical sciences nor psychology give us a clue as to why these correlations hold. And the differences between the mental and the physical make the prospects of any super-scientific natural law explanation dim. Thus, for Swinburne, the mental is unmarginalized, the physical sciences are appreciated, but hopes for a naturalistic theory of everything are dashed. For medieval hylomorphists, a human being is a functional integrity metaphysically underwritten by its membership in a natural kind. Human being is hylomorphic: substantial form roots its functional powers and structures matter into the organic body needed for those functions. Function defines the species. For Aquinas, teleology explains why this body has these organs and tissues rather than some other.26 Neither the human soul nor the human body is a complete substance. Animal body is not, because—once again—a substance is a functional integrity. Nothing about the form of mixed body or the forms of the elements from which the body is mixed, would explain its structure. The functional powers to which elemental and mixed-body forms give rise could not account either for the structure or the maintenance of the animal body. Substances are functional integrities with normal life spans. Cow body sans cow soul has no such functional integrity. Likewise, even if the human soul has functions (understanding and willing) that the body does not share, it is a formal functional principle that roots many organic powers and activities. Indeed, given Aquinas’s cognitive psychology, the separate soul cannot continue to carry out its intellectual functions in the natural way. With miraculous assistance, it can be subject to functioning. But it does not have functional integrity apart from the hylomorphic composite.27 Thus, for medieval hylomorphists, the vital and the mental are unmarginalized, the functional integrity of plants and animals is preserved, and the connection between the mental 26 Not all medieval hylomorphisms embraced natural teleology. Scotus and Ockham insist that substances are functional integrities underwritten by natural kinds. They agree that every species of plant or animal has characteristic functions and is perfected by their exercise. But they maintain that ends can be causes as opposed to mere resultants of the coordinated exercise of efficient causal powers, only if they can explain the action of those efficient causal powers. Scotus and Ockham argue that only in the case of rational agents does the end explain the action of the efficient cause. Consequently, they see only material and efficient causal connections within non-rational substances and so no natural teleology. See my (2000: 153–83) and (1998: 1–46). 27 Aquinas, Summa Theologica I, q. 89, a. 1, c; a. 3, c.

WHAT ABOUT HYLOMORPHISM ?



and the physical is explained. The dangling question is whether its understandings of biology and psychology could survive the replacement of Aristotelian physics and four-element stuff chemistry with contemporary physical sciences.

5 Cosmic Perspectives 5.1 Nagel’s challenge In his recent book Mind & Cosmos (Nagel 2012), Thomas Nagel insists on the unmarginalizability of the mental: mind, meaning, and value are—pace Bertrand Russell (1957: 44–54)—data too insistent to be written off as accidental side effects (Nagel 2012: 7–8, 36). On the contrary, they are a key to the nature of the cosmos. Nagel concludes that—despite the fact that the mind is the product of a partly physical process (2012: 8, 36)—reductive materialism is almost certainly false (2012: 8–9, 81). To be sure, Nagel acknowledges the vast accomplishments of the physical sciences as we know them, and the need to allow theoretical space for them (2012: 8). Indeed, scientific inquiry presumes what Nagel insists upon: that the world is intelligible, and that its rational structure is—at least with effort and eventually—accessible to us (2012: 15–20, 73–6, 82–5). Nagel concludes—pace Swinburne and intellectual humility notwithstanding—that it is appropriate for us to hope for a unified theory of everything (2007: 7–8, 85–6, 128). Where the unmarginalizable mental is concerned, regularities and functional organization cry out for explanations that show why it is that they occur. Here Nagel distinguishes two questions: a constitution question and a historical question. The constitution question is, for any type of organism, why it did/does or did/does not have conscious life, and why it did/does have the kind of conscious life it had. The historical question is why and how conscious organisms arose, which—Nagel insists—must be answered by a theory that does not make the mental an accidental and unexplained concomitant of something else (2012: 47, 51–9, 85–6). In Nagel’s estimation, dualism despairs of this (2012: 37), and reductive materialism fails at this because forwarding the physical sciences as we know them as the true theory of everything marginalizes the mental. Scanning for alternative theoretical directions, Nagel recognizes several



MARILYN M C CORD ADAMS

other possibilities. Theism, which he does not favor, seeks an explanation of both questions outside the natural order in the intentions and purposes of a transcendent being. In Nagel’s estimation, theism has the theoretical disadvantage of stopping the explanations prematurely. He wants to ask, what made Divine choices reasonable or attractive? Answers, he thinks, would be little more than a projection of human points of view (2012: 12, 21–6). Nagel would much prefer an explanation internal to the natural order, and leans toward a neutral monism about the basic constituents of the universe, according to which these constituents have properties that explain not only its physical but also its mental character (2012: 22, 29–33, 56), a theory according to which propensity for the development of a subjective point of view existed in the fundamental particles from the beginning (2012: 32–3, 62). Such a theory would require a radical revision in our metaphysical conception of what the physical sciences study (2012: 26). No longer would their domain be a distinct kind of substance; rather, their assignment would be to investigate one aspect or (to use Spinoza’s language) attribute of fundamentally bivalent stuff. Nagel speculates that fundamental constituents with such bivalent tendencies would be governed by two kinds of laws: the nonteleological and timeless laws of physics that govern the fundamental elements of the universe, whatever they are; and the teleological laws of nature that are both universal (not restricted to individual cases) and “biased toward the marvelous”—presumably, the unmarginalizable mental that embraces mind, meaning, and value. The laws of physics predict what—given ideal conditions—will happen immediately, but teleology governs a temporally extended developmental process. Teleology does not explain the end-point in terms of what preceded; instead, it explains the whole long evolutionary process in terms of the marvelous unmarginalizable mental that is its goal (2012: 66–7, 91–3).

5.2 Theism subsuming dualism Swinburne also takes the marvelous unmarginalizable mental to have cosmic implications. To theorize it, Swinburne takes the polar opposite option, the theism that Nagel rejects. Seeing no possibility of a unifying natural-law-governed explanation, Swinburne sees the marvelous unmarginalizable mental as contributing to his cumulative case for theism

WHAT ABOUT HYLOMORPHISM ?



(2007: 199).28 What there is, is res extensa and res cogitans. But the ultimate explainer is a high-powered soul, indeed a trinity of highpowered souls (Swinburne 1994: 170–91). For Swinburne, teleology does have a role in explaining what there is and how things are arranged. But he does not recognize teleology in nature. He locates teleology where it is easiest to understand: in the purposings and deliberations of rational agency. Nagel’s two questions will find their answers in God’s creative choices. But Nagel will still wonder what reasons God had not only for making a world with souls and bodies, but also for making souls that are causally interactive with bodies as much as ours are, for associating souls of this type with bodies of that type, for correlating this kind of conscious experience with that sort of brain state in particular, for using a long evolutionary process to evolve complicated enough bodies, and for beginning the soul–body association only after eons of geological time. Many of these, Swinburne does not try to answer. Presumably, he thinks that to understand what makes these soul–body associations and those mind–brain correlations so apt, we would have to be in possession of the super-science that—he is convinced—eludes us. Nevertheless, Swinburne does attend to an aspect of the cosmic problem on which Nagel does not focus: the problem of evil. If human beings are essentially souls and only contingently connected with bodies, we may ask why God would deliberately create souls embodied in a world like this, when God would have known that such a high level of causal interaction with bodies would cause souls great suffering and—in millions of cases— bring on prima facie personal ruin. Why not skip the sufferings of this present life in the world as we know it, in favor of careers of disembodied bliss? Swinburne’s soul-making theodicy in Providence and the Problem of Evil addresses itself to these concerns (Swinburne 1998).29

5.3 Hylomorphisms and the cosmos Nagel hopes to solve the constitution problem by looking to the fundamental constituents of the universe and reconceptualizing their

28 29

See also his earlier treatment in Swinburne (1979: 152–79). See also his earlier treatment in (1979: 180–99).



MARILYN M C CORD ADAMS

nature. This is precisely what Aristotelians do not do. When accounting for life, consciousness, rational thought, and choice here below, Aristotelians do not seek it in natural necessities governing the four elements. Aristotelians located explanatory teleologies, not in stuff that is common to everything, but rather in the infima species, in each natural kind itself. On medieval readings, even the four elements had ends, each its natural place (earth, the center; fire, the perimeter; water and air in between), and natural tendencies to seek it. By themselves, these could not explain the regular stable structures of plant and animal bodies as opposed to their flying apart at the seams. That means that the teleological explanations of Aristotelian biology and psychology are not universal but “local” to a species or at most a relatively near genus (not body, but mammal or hawk). As to Nagel’s historical question, medieval Aristotelians do not consider it as such, because they hold to the omnitemporality of the species, that whatever species are instantiated at any time in the history of the world, are instantiated at every time in that history. Nevertheless, they do address themselves to an analogous cosmic question: why are these natural kinds instantiated in the universe at all? All appeal for their answer to a transcendent ultimate explainer. For all of them, the cosmic puzzle is not why the marvelous mental, but rather why does the physical world exist at all? Their answers went in a variety of directions, of which I will mention three. Avicenna’s cosmology posits the first cause as the only nature that explains its own necessity. Other natures really exist (in the case of things here below, are really instantiated) only because the first cause directly or indirectly emanates them by natural necessity. Avicenna envisions a great chain of emanation: the first cause directly emanates the first intelligence, the first intelligence directly emanates the second intelligence and one of the heavenly spheres, and so on down to the agent intellect that necessarily emanates prime matter and floods it with forms of natural kinds. Godhead is the nature that includes the efficient causal tendencies, directly and indirectly, to produce animate and inanimate beings, and Godhead and the subordinate emanators all act in emanating by natural necessity to the limit of their powers. On Avicenna’s picture, value comes into it, because cosmic emanation results in a great chain of cosmic excellence. Material things exist because the chain would be incomplete without them. The whole sublunary world is populated by

WHAT ABOUT HYLOMORPHISM ?



the lower-grade natures that are subject to generation and corruption, and matter is the substratum of substantial change.30 Aquinas’s God acts by intellect and free choice in the production of the universe. The goal of creation is to share the goodness and to produce an excellent cosmic Godlikeness. The infinite gap between Divine and creatable natures has two consequences that guide Divine choice: first, there is no best among congeries of creatable things since for every one, there is a better one; second, because infinite excellence is multi-faceted, a world containing only intelligent natures will produce a less excellent Godlikeness than a world containing a wide variety of creatable natures, including material natures, even though they are in themselves less excellent than purely immaterial natures. Cosmic excellence explains why angels and humans, cows and stones were included.31 For Scotus’s God, the primary goal in creation is not cosmic excellence, but rather widening the Trinitarian friendship circle to include the soul of Christ and—because friendship love is not jealous or possessive— other created co-lovers, angels, as well as more human souls. Scotus’s God focuses first on upgrades (hypostatic union for the soul of Christ, grace and infused virtues for all of them) that would fit human souls and angels for such high society. Scotus’s God decides to make the material world, only because human souls are incomplete substances and God wants them to exist and function as metaphysically complete.32 Avicenna, Aquinas, and Scotus all appeal to a transcendent explainer to account for why our world here below includes both the mental and the material. For Avicenna, the great chain of natural excellence is produced by natural necessity. For Aquinas, cosmic excellence is among the goals at which God aims. For both Aquinas and Scotus, our world’s containing the mental and the material is a product of Divine free choice. When Nagel suggests that theism would leave cosmic questions unanswered, he may be thinking that theological explanations would make the natures of things products of an arbitrary Divine will. But medievals did

30

For an excellent account of Avicenna’s position, see Davidson (1992: 74–94). Aquinas, Summa Theologica I, q. 15, a. 2, c; a. 3, c; I, q. 25, a. 6, c and ad 2um and ad 3um; I, q. 47, a. 1, c; a. 2, c. 32 Scotus, Ordinatio III (suppl.), d. 32 (Assisi, com. 137, fol. 174ra–va), ed. and trans. Allan B. Wolter in McElrath (1980: 154–7). 31



MARILYN M C CORD ADAMS

not anticipate Descartes on this score. For Avicenna, Aquinas, and Scotus, the constitution of creatable natures is independent of the Divine will. For Avicenna and Scotus, the constitution of creatable natures is primitive, that human is rational animal pertains to human nature of itself. For Aquinas, creatable natures are ways of imperfectly imitating Godhead. Bovinity is grounded in the Divine essence, but the fact that rational animal does and goat-stag does not constitute a relevant Divine imitation lies outside the scope of God’s will. Moreover, all three thinkers seek an explanation of why the mental and the material get included in the natures of things. Avicenna’s God emanates by natural necessity, and that creatable natures are included within its scope is a function of their constitution. So also, whether and how things of a given nature can contribute to cosmic excellence is a function of their natural kind. Finally, what prompts Scotus’s God to decide on a material world is Divine reflection on the prior fact of what it is to be a human being. For all of them, God transcends the sublunary world in which we live and move and have our being. But because they believed in real natural kinds, whose constitution was metaphysically necessary and independent of anyone’s will or linguistic practices, these natures could stand as something independent of the Divine will that could furnish reasons for God’s choice. Nagel’s suspicion—that theological explanations would invariably sidestep the natures of things and in consequence be “external”—should be extinguished by what Avicenna, Aquinas, and Scotus actually say.

5.4 Evil and the marvelous mental In his books The Evolution of the Soul and Mind, Brain, and Free Will, Swinburne concentrates polemical attention on other positions in the philosophy of mind. In Mind & Cosmos, Nagel recalls us to systematic issues of cosmic significance. Unsurprisingly, the systematic desideratum that drives my own approach to both is solving the problem of horrendous evils.33 I agree with Swinburne and Nagel that the mental is and ought to be unmarginalizable: that any adequate account of the cosmos ought to “front and center” mind, meaning, and value. But the neutral monism to which Nagel inclines leaves me cold, because the world as we 33

These preoccupations are expanded upon in my (1999) and (2006).

WHAT ABOUT HYLOMORPHISM ?



know it exhibits not only a bias toward the marvelous, but also tendencies to wreck and ruin the marvelous. If millions have been lucky enough to lead lives filled with positive meaning, billions more have experiences so miserable that they reasonably wish never to have lived at all. History is a slaughter-bench. Moreover, the physical sciences as we know them, whose results—Nagel thinks—must be mutatis mutandis incorporated, predict end-time entropy and so offer no prospect of eschatological resolution. My chief hesitations about dualism arise not from romantic visions of the causal closure of the physical,34 but rather from the voiced fear that it makes the problem of evil more difficult. If human being is not a natural kind, if embodiment is not essential to us, and if embodiment in this world exposes us to horrors, why would God create us as personal animals in a world such as this? Swinburne’s exceptionally thorough development of the soul-making theodicy and his celebration of the value of being of use, do not—for reasons I have no time to repeat here—lay my worries to rest. While I appreciate that God’s ways are higher than our ways and agree with medievals and reformers that God does not “owe” us, I have maintained that God wants to be good to us anyway and to guarantee to each of us lives that we will be able to value as great goods to us on the whole and in the end. In this connection, I am wondering whether realism about natural kinds and hylomorphism about human nature would be a help. Why would God create us personal animals in this world? Scotus suggests one answer: “because personal animals is what we are essentially, because this world is our natural home.” These are motives that have to do with us, reasons that express an interest in us as we are. I am thinking, they are reasons that we might be able eventually to accept, not as morally sufficient or God-justifying, but rather as not unfriendly, when the plot has resolved in a happy ending.

References Adams, Marilyn McCord. (1987). William Ockham. Notre Dame: University of Notre Dame Press. 34

Swinburne argues at length that the causal closure of the physical is unprovable in (2013: 117–23).



MARILYN M C CORD ADAMS

Adams, Marilyn McCord. (1998). “Ockham on Final Causality: Muddying the Waters,” Franciscan Studies, 56 (Festschrift for Girard Etzkorn): 1–46. Adams, Marilyn McCord. (1999). Horrendous Evils and the Goodness of God. Ithaca and London: Cornell University Press. Adams, Marilyn McCord. (2000). “Final Causality and Explanation in Scotus’ De Primo Principio,” in Chumaru Koyama (ed.), Nature in Medieval European Thought. Leiden: Brill, 153–83. Adams, Marilyn McCord. (2006). Christ and Horrors: The Coherence of Christology. Cambridge: Cambridge University Press. Davidson, Herbert A. (1992). Alfarabi, Avicenna, & Averroes on Intellect: Their Cosmologies, Theories of Active Intellect & Theories of Human Intellect. New York and Oxford: Oxford University Press. Lennox, James G. (2001). Aristotle’s Philosophy of Biology: Studies in the Origin of Life Science. Cambridge: Cambridge University Press. Livesey, Stephen John. (1982). Metabasis: The Interrelationship of the Sciences in Antiquity and the Middle Ages. UCLA PhD dissertation. McElrath, Damian. (ed.) (1980). Franciscan Christology. St. Bonaventure, NY: The Franciscan Institute Press. Nagel, Thomas. (2012). Mind & Cosmos: Why the Materialist Neo-Darwinian Conception Nature is Almost Certainly False. Oxford: Oxford University Press. Russell, Bertrand. (1957). “A Free Man’s Worship,” in Mysticism and Logic. Garden City, NY: Doubleday-Anchor, 44–54. Swinburne, Richard. (1977). The Coherence of Theism. Oxford: Clarendon Press. Swinburne, Richard. (1979). The Existence of God. Oxford: Clarendon Press. Swinburne, Richard. (1994). The Christian God. Oxford: Clarendon Press. Swinburne, Richard. (1998). Providence and the Problem of Evil. Oxford: Clarendon Press. Swinburne, Richard. (2004). “The Possibility of Life after Death,” in Peter Cave and Larvor (eds), Thinking about Death. London: British Humanist Association, 38–42. Swinburne, Richard. (2007). The Evolution of the Soul, rev. edn. Oxford: Oxford University Press. Swinburne, Richard. (2013). Mind, Brain, and Free Will. Oxford: Oxford University Press.

Index a priori 49–52, 54, 59–62, 87 n. 4 Abraham 25, 32–9 Adams, Marilyn McCord 122–3 n. 37, 230 n. 16, 236 n. 26, 242 n. 33 Adams, Robert 91 n. 9, 122–3 n. 37 aesthetics 64, 70–83 Alston, William 4, 5–6, 19–20, 38–9 Anselm of Canterbury 88 n. 5, 150 anti-Pelagianism, see Pelagianism Aquinas, see Thomas Aquinas Aristotle 134, 162, 230 n. 18, 233–4 Armstrong, David 129–30 as-if trope 183–9 atonement 139, 148–70 Audi, Robert 5, 17–18, 27 n. 1, 34, 38–9 Augustine of Hippo 88 n. 5, 150 Baier, Annette C. 8, 19 Baillie, James 104 n. 25, 114 n. 30 Bayes 27, 65, 72, 196–219 Beall, J. C. 107 n. 27 beauty 11, 50, 64, 70–83, 174–5 belief 1, 5–8, 13, 15–17, 22–3, 26–45, 59, 92, 107 n. 27, 138–40, 171, 181, 220 Bellah, Robert N 5 Bergmann, Michael 44 n. 9, 63 n. 5, 68, 81 n. 1, 121 n. 36, 122–3 n. 37, 188 n. 10 Bird, Alexander 128 Bishop, John 42 n. 8 Brower, Jeffrey 44 n. 9, 63 n. 5, 81 n. 1 Buchak, Lara 5, 213 Cartwright, Richard 110 Casel, Odo 176–7, 179–80 cause 36, 91, 93, 95, 99, 102, 104–5, 155, 120–2, 126–44, 176, 236 n. 26, 240 Chisholm, Roderick 84, 91 n. 9, 92–3, 98, 116 n. 32, 122 Christ, see Jesus Clegg, Jerry 5 cognitive 4–25, 28, 38–44, 55, 106–7, 216, 223 n. 5, 225–8, 234–6

coherence 48, 52–63, 107 n. 27, 149–50 Coleman, James 8 Collins, Robin 65–6 Copernicus 56–8 counterfactuals 115, 127, 133, 136 credence 34, 37–8, 138, creeds 7, 27, 29–35, 42, 149, 173, 221 Apostles’ 173 Danto, Arthur C. 115–16 Davidson, Donald 117–18 Davidson, Herbert 241 n. 30 decretalism 133, 135–6 determinism 121, 126, 130–1 Doomsday Argument 196–210 Draper, Paul 37 n. 6, 44 n. 9 Dretske, Fred 129 n. 4 dualism 138, 220–44 Eliade, Mircea 177–81 ethics 21 n. 6, 42–4 explanation 9, 18, 23, 50–3, 60–2, 135, 169, 175, 220–1, 230–43 causal 48–9, 52 teleological 233–5 explanatory power 49–53 Fales, Evan 128 fine-tuning 64–83 Fischer, John Martin 97 n. 17 Flint, Thomas P. 85 n. 1, 87, 89–91, 98–100, 103 n. 22 forgiveness 148–70 Freddoso, Alfred J. 85 n. 1, 87, 89–91, 98–100, 103 n. 22 freedom 66, 86, 99–101, 105–7, 127, 130–1, 133, 136, 143 Geach, Peter 89, 99, 100, 122 Genesis 14, 176 Ginet, Carl 92, 116 n. 32, 118 Goethe 72–3, 77 Goldman, Alvin 116 n. 32



INDEX

Hagen, Jason 104 n. 25, 114 n. 30 Hardin, Russell 8, 18 Harrison, Noona Verna 180–3 hatred 158–69 Hoffmann, Joshua 90, 98 n. 19 Holton, Richard 8 hope 4–5, 12–13, 15, 21–2, 78, 81, 108, 160 Howard–Snyder, Daniel 5, 38–40, 44 n. 9, 68, 81 n. 1 Hudson, Hud 67–8, 76 hylomorphism 220–44 immortality 195–219, 222 Jesus 7, 22, 34–6, 149–53, 157, 158–62, 167, 169, 172–82, 188–92, 241 John Duns Scotus 88, 222–30, 236 n. 26, 241–3 Jones, Karen 8 Kepler, Johannes 56–8 Kvanvig, Jonathan 5, 11 n. 2, 32 n. 4 LaCroix, Richard 103 n. 22 law 50, 56–9, 62, 126–44, 160 n. 22, 221, 229, 236, 238 Lee, Sukjae 137 n. 9 Leftow, Brian 86 n. 3, 88 n. 5, 89 n. 6, 90, 95 n. 16, 115 n. 31, 122–3 n. 37 Lennon, Thomas M. 137 Lennox, James G. 233 n. 19 Leslie, John 196 n. 1, 216 Lewis, David 129–31 liturgical present tense 171–93 liturgy 171–93 Livesey, Stephen John 230 n. 18 Locke, John 5 love 4, 6, 8, 37, 41–2, 148–70, 173–5, 241 McCann, Hugh 92, 116 n. 32 McEar 84–5, 90 n. 8, 102 n. 22, 103–15, 123 n. 37 McElrath, Damian 223, 241 n. 32 McKaughan, Daniel 6 Mackie, J. L. 103 n. 22 Manson, Neil 65, 67, 81 n. 1 materialism 142, 220–30, 234–7 Melden, A. I. 92 n. 10

Moll, Otto 159–65 morality 8, 42–4, 64, 67–72, 86, 99–102, 130–40, 143, 150–1, 157–65, 168–9, 220, 243 moral blame 139–40, 157 moral goodness 44, 157–65 moral praise 43, 139–40 Murphy, Jeffrie 156 n. 15, 162 n. 23, 165 n. 25 Nagel, Thomas 237–43 natural kinds 220, 229, 231–2, 234–6, 240–3 natural theology 48–50, 54, 60–3, 126 necessity 52–4, 69, 77, 103, 128–33, 135, 137, 140, 227, 240–2 noncognitive, see cognitive occasionalism 126–44 Ockham, see William Ockham omnipotence 48, 61, 67, 69, 74, 77, 84–125 ontology 116 Oscamp, Paul J. 137 n. 9 Overy, Richard 159 n. 21 Page, Don 196 n. 1, 216 parsimony 49–51, 56 qualitative 50, 56 quantitative 50, 56 Pearce, Kenneth 89–90, 104 n. 24 Pelagianism 155–6, 166, 168 n. 26 perfection 48, 67, 69–72, 74, 77–9, 88, 99, 102, 104, 108, 113–14, 139, 150, 152, 154, 158, 166 Plantinga, Alvin 5, 84–5, 93, 99–100, 103, 107 n. 27, 126 Pojman, Louis 5 post-mortem existence, see immortality Probabilistic Venn Diagram (PVD) 199–202 probability 26–30, 32–3, 35, 37, 42, 48–63, 65–7, 72–4, 78–9, 81, 126–7, 196–219 problem of evil 35, 67–9, 171, 239, 243 Pruss, Alexander 104 n. 24, 107–8 n. 27 Ptolemy 56–8 punishment 94, 100, 115, 120–1, 162–5, 169

INDEX

purposes 11–12, 27, 33, 35–6, 38, 43–4, 220, 238 Rasmussen, Joshua 107–8 n. 27 Ratzsch, Del 127, 133 reactualization theory 175–84, 190 reconciliation 152, 157, 166–8 religion 5, 26, 28, 31–3, 42–4, 85–6, 171, 191, 209–11, 213–14 Rosenkrantz, Gary 90, 98 n. 19 Ross, James F. 89–90, 104 Rubio, Daniel 108–10, 122–3 n. 37 Russell, Bertrand 84, 237 Ryle, Gilbert 92 n. 10



Swinburne, Richard 19–20, 26–45, 48–63, 64–83, 84–125, 126, 138, 148–53, 166–70, 213, 220–44 Swoyer, Chris 128 Tarski, Alfred 110 Taylor, Richard 92 n. 10 Tennant, Frederick R. 5 Thomas Aquinas 6, 88, 134, 153–69, 222–30, 233–6, 241–2 Tooley, Michael 129 n. 4 trust 4–25, 39–40 uniformity 52–62

salvation 7, 28–9, 35–6, 177, 191–2 Schellenberg, J. L. 5, 31, 38, 63 n. 5 Schmemann, Alexander 190–1 Scotus, see John Duns Scotus Sessions, William Lad 5 Shoemaker, Sydney 128 simplicity 48–63, 65, 130, 154 n. 9 Simpson, Thomas W. 8–10, 19 skeptical theism 67–70, 74, 77–81 Skyrms, Brian 55 Sobel, Jordan H. 85 n. 1, 87, 105 n. 26 soul 174, 220–30, 234–43 Speaks, Jeff 86 n. 3 state of affairs 43, 68, 74–5, 78–81, 84–5, 89–123, 129, 136, 140 Stowe, Harriet Beecher 186 Stump, Eleonore 15–16, 21, 153 n. 5, 154 n. 11, 155–8, 225

value 9–10, 17–18, 26–45, 69–81, 149, 152, 154, 179, 202–5, 212–14, 237–44 Van der Vat, Dan 163 n. 24 van Fraasen, Bas 127, 130 n. 5 van Inwagen, Peter 122–3 n. 37, 131, 134, 219 n. 11 von Allmen, J. J. 179–80 White, Roger 67, 81 n. 1 Wielenberg, Erik J. 89 n. 6, 104 n. 24, 113 n. 29, 115 n. 31, 119 n. 35 Wierenga, Edward R. 90, 101 n. 20, 104 n. 24, 113 n. 29, 122–3 n. 37 William Ockham 88, 103 n. 22, 222–30, 234 n. 21, 236 n. 26 Williams, Bernard 10 Wolterstorff, Nicholas 178 n. 4