Oxford Studies in Philosophy of Religion Volume 9 9780198845492, 0198845499

Table of contents :
Cover
Oxford Studies in Philosophy of Religion: Volume 9
Copyright
Contents
Preface
List of Contributors
1: Privation in the Problem of Evil: Impairment, Health, Well-Being, and a Case of Humans and Betazoids
1. Introduction
2. The Test Case: Betazoids and Humans
3. Definitions: Disability, Impairment, and Health
4. A Solution to the Puzzle
5. Privative Evils and Theodicy
6. A Confusion?
7. Where Are We Now?
Bibliography
2: What If God Makes Hard Choices?
1. Swinburne’s Model
2. A Challenge to Swinburne’s Model
3. A New Model
4. An Advantage
5. A Disadvantage
3: Theism, Pro-Theism, Hasker, and Gratuitous Evil
1. Stage-Setting
2. Restrictions on God’s Permission of Evil
3. Hasker’s Case Against (1)
4. Must Gratuitous Evil Occur, or Must it Merely be Permitted?
5. Criticisms of Hasker’s Argument
5.1 The strategy is self-defeating
5.2 The account of God’s plan for creation in (e) is mistaken
5.3 If God were to prevent all gratuitous evil, we would neither know nor reasonably believe that God had done so
5.4 Our moral motivations would not be undermined
6. Conclusion
Appendix: The Argument From Excessive Gratuitous Evil
References
4: God and Gratuitous Evil: A Response to Klaas Kraay
1. Preliminaries
2. The NGE Defense
3. The Problem of Moral Motivation
4. The Problem of Excessive Gratuitous Evil
5. Conclusion
5: Moral Indulgences: When Offsetting is Wrong
0. Introduction
1. Moral Indulgence
2. The Moral Status of Moral Indulgence
2.1 Two desiderata and what can’t meet them
2.2 Universalizability and the second desideratum
2.3 Application of universalizability and the first desideratum
2.4 Cause-matching
3. Divine Indulgence
3.1 Background assumptions
3.2 Divine indulgence
3.3 Universalizability and divergence on the divine
4. Closing Thoughts
Bibliography
6: Time of Trial
Time of Trial
7: Why Did The One Not Remain Within Itself?
I. Introduction
II. A Framework For Our Inquiry
III. Why Did God Create?
IV. Could God Have Had Any Completely Adequate But Non-Coercive Reason To Create?
V. The Solution to the Aporia of Contingent Creation
VI. The Need for a Further Traditional Resource
VII. Need the Weights of Various Reasons Add for the Problem to Arise?
VIII. The Way Forward
IX. The Transfinite Alternative
X. Cantor on the Transfinite and the Absolute
XI. The Aporia Resolved
XII. The Upshot for Theology
XIII. An Inventory of Kinds of Normative Reasons
XIV. Immediate Elimination of Four Kinds of Normative Reasons
XV. Did God Create out of Reasons of Personal Attachment?
XVI. Creating from Reasons of Love
XVII. Did Generosity Provide God with Reason to Create?
XVIII. There Could Not Be Melioristic Reasons to Create
XIX. Ameliorism
XX. “To Show Forth His Goodness”
XXI. Manifestationism
XXII. How Mistaken We Have Been About What Matters
XXIII. How Corrosive is Ameliorism when it Comes to Other Sorts of Reasons?
XXIV. One Extant Solution to the Corrosive Effect of Ameliorism
8: Organic Unities, Summation, and the Problem of Evil
1. The Principle of Organic Unities and the Thesis of Universality
2. Chisholm on Defeat and the Problem of Evil
3. Is the View That There Are Organic Unities Incoherent?
4. Does the Moorean View Lead to Evaluative Schizophrenia?
5. The Objection from Evaluative Inadequacy
Bibliography
9: Organic Unities and the Problem of Evil: A Reply to Lemos
I
II
III
IV
V
References
10: Unrestricted Actualization and Divine Providence
1. Interpreting Almeida’s Core Ideas about U-actualization
2. The Significance of U-actualizationfor the Theory of Goodness
3. Evaluation of Almeida’s Argument for (3b)
3.2 Should we accept premise (1b)?
References
11: Unrestricted Actualization and Perfect Worlds: A Reply to Langtry
1. Introduction
2. Langtry’s Objections to U-Actualization
2.1 Langtry’s Objection to (2b)
2.2 Langtry’s Objections to (1b)
12: Prediction and Providence: Rejoinder to Almeida’s Reply
1. Almeida’s Reply to my Objection to (2b)
2. Almeida’s Reply to my Objection to (1b)
2.1 The argument from God’s essential comprehensive knowledge
2.2 The argument from divine practical rationality
References
13: A Naturalistic Intrinsic Value Theodicy
1. Introduction
2. Creaturely Freedom
3. The Distribution of Intrinsic Value
4. Organic Unity and the World as a Whole
5. Objections: Value, Existence, Death, Pain
6. Conclusion
References
Index

Citation preview

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Oxford Studies in Philosophy of Religion

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Oxford Studies in Philosophy of Religion Volume 9 Edited by L A R A BU C HA K A N D D E A N  W.  Z I M M E R M A N Co-editor for Volume 9 PHILIP SWENSON

1

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

1 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 2019 The moral rights of the authors have been asserted First Edition published in 2019 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: 2019945436 ISBN 978–0–19–884549–2 Printed and bound in Great Britain by Clays Ltd, Elcograf S.p.A. 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.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Contents Preface List of Contributors

1. Privation in the Problem of Evil: Impairment, Health, Well-Being, and a Case of Humans and Betazoids Alexander R. Pruss and Hilary Yancey

vii ix

1

2. What If God Makes Hard Choices? Paul Draper

18

3. Theism, Pro-Theism, Hasker, and Gratuitous Evil Klaas J. Kraay

31

4. God and Gratuitous Evil: A Response to Klaas Kraay William Hasker

54

5. Moral Indulgences: When Offsetting is Wrong Rebecca Chan and Dustin Crummett

68

6. Time of Trial Frances Howard-Snyder

96

7. Why Did The One Not Remain Within Itself? Mark Johnston

106

8. Organic Unities, Summation, and the Problem of Evil Noah Lemos

165

9. Organic Unities and the Problem of Evil: A Reply to Lemos Michael J. Zimmerman

183

10. Unrestricted Actualization and Divine Providence Bruce Langtry

195

11. Unrestricted Actualization and Perfect Worlds: A Reply to Langtry213 Michael Almeida

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

vi Contents

12. Prediction and Providence: Rejoinder to Almeida’s Reply Bruce Langtry

225

13. A Naturalistic Intrinsic Value Theodicy Scott A. Davison

236

Index

259

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Preface This is the ninth volume of Oxford Studies in Philosophy of Religion. As  with earlier volumes, these essays follow the tradition of providing a non-sectarian and non-partisan snapshot of the subdiscipline of philosophy of religion. This subdiscipline has become an increasingly important one within ­philosophy over the last century, and especially over the past half century, having emerged as an identifiable subfield at the same time as other emerging subfields such as the philosophy of science and the philosophy of language. This volume continues the initial intention behind the series: attracting the best work from philosophers well-known for their contributions to the philosophy of religion, as well as including essays by philosophers working mainly outside this area when their interests intersect with issues in the philosophy of religion. This inclusive approach to the series provides an opportunity to mitigate some of the costs of greater specialization in our discipline, while at the same time inviting wider interest in the work being done in the subfield. Each volume typically includes an essay by the winner of the Sanders Prize in Philosophy of Religion, described within. L. B., P. S., & D. Z. Berkeley, CA; Williamsburg, VA; New Brunswick, NJ

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

List of Contributors Authors Michael Almeida, University of Texas at San Antonio Rebecca Chan, San José State University Dustin Crummett, Ludwig Maximilian University of Munich Scott A. Davison, Morehead State University Paul Draper, Purdue University William Hasker, Huntington University Frances Howard-Snyder, Western Washington University Mark Johnston, Princeton University Klaas J. Kraay, Ryerson University Bruce Langtry, University of Melbourne Noah Lemos, College of William & Mary Alexander R. Pruss, Baylor University Hilary Yancey, Baylor University Michael J. Zimmerman, University of North Carolina at Greensboro

Series Editors Lara Buchak, University of California at Berkeley Dean W. Zimmerman, Rutgers University

Co-Editor for Volume 9 Philip Swenson, College of William & Mary

Managing Editors Christopher Hauser, Rutgers University Isaac Wilhelm, Rutgers University

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

1

Privation in the Problem of Evil Impairment, Health, Well-Being, and a Case of Humans and Betazoids Alexander R. Pruss and Hilary Yancey

1. Introduction Some arguments for the non-existence of God rely on (1) positive evidence of evil and (2) the incompatibility of theism with the evils reported in (1). The kinds of evils often cited in (1) include natural evils, like a hurricane that destroys an island’s infrastructure, and moral evils, like the Holocaust or the Khmer Rouge. Evils can also be categorized in terms of whether they deprive a good or inflict a harm. Privative evils are evils that deprive someone of a good. But this is not merely a good that would benefit the person, as when the admissions committee denies someone entrance into a prestigious university; privative evils involve depriving someone of a good that is due to them. In some substantial sense, that person is owed the good in question. Defenders of theism have reason to be particularly worried about privative evils because cases where someone is owed a particular good present a twofold objection to God’s not providing that good. First, one can object that God should provide that good because of the benefit it would have for the individual (providing it out of charity). But one can also object that God should provide that good because justice demands it. So God would, if God existed, be not only uncharitable but also unjust in failing to provide that good. For theodicists, who aim to give an account of why evil is compatible with God’s existence, privative evils present a steep obstacle. In this chapter we consider a candidate privative evil: impairment. We argue that while impairment does entail the lack of a due good (the good of overall health), this lack does not turn out to be substantial enough to pose a problem for theodicists. The value of overall health, we argue, is easily overcome by the values of other components of well-being. To illustrate our

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

2  Alexander R. Pruss and Hilary Yancey argument, we consider a case of Hannah the human and Bridget the betazoid, each of whom lacks telepathy, where the lack is an absence in one case and an impairment in the other. Our argument proceeds as follows: First, we explain our test case and the problem that it poses. Second, we offer putative definitions of impairment, disability, and health. We argue that Bridget the betazoid’s impairment results in a lack of overall health but does not result in a substantial loss to her overall well-being. Finally, we argue that this is (partial) good news for answering the putative argument from evil based on privative evils. One initial note: One could argue that impairments are not intrinsic bads at all and thus do not intrinsically contribute to the problem of evil (notwithstanding indirect contribution in the form of, say, certain kinds of social disadvantages or discrimination). See Elizabeth Barnes’s (2016) “meredifference” view of disability (Silvers 2003 makes a similar argument with respect to disabilities).1 For the sake of our argument, we assume that an impairment is an intrinsic bad, but then argue that it poses no significant challenge to a theodicy, because the type of intrinsic bad it is results in a very small loss of overall well-being or flourishing.

2.  The Test Case: Betazoids and Humans Star Trek’s betazoids are a species that look human but have telepathic abilities. We will throughout this argument assume that telepathy is highly valuable (if there are worries about privacy, then change telepathy to some other ability that humans don’t have) and (at least partly) constitutive of what it is to be a betazoid—defining the species “betazoid” makes reference to telepathic abilities.2 Now consider two individuals: Hannah and Bridget. Hannah and Bridget live remarkably similar and happy lives of moral and intellectual excellence. They have satisfying engineering jobs with the Federation, they have fulfilling family lives, they have an extensive network of acquaintances and a narrower network of good friends, they are both

1  Barnes is skeptical about the impairment/disability distinction. Her mere-difference view is articulated with respect to disabilities, which are defined in such a way that her view would be applicable to what we call impairments in this chapter. 2  Here the notion would be that the sentence “Betazoids are telepathic” is a generic like “Cats have tails” (see Sarah Jane Leslie and Adam Lerner (2016)). Telepathy is part of the ratio of betazoids, but the sentence is not falsified by the existence of a non-telepathic betazoid, and one does not fail to be a betazoid by lacking telepathy.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Privation in the Problem of Evil  3 deeply committed to furthering interspecies understanding, and they both are champion players of Klingon chess. And neither has any telepathic skills. If you ask Hannah and Bridget whether they would like to have telepathic skills, they both express some worries about inadvertent violations of others’ privacy, but on the whole they see these skills as worth having due to the new forms of flourishing that the skills would make possible. In particular, their work on furthering interspecies understanding could be greatly improved with respect to some recently discovered species that the universal translator is having difficulties with. Each would be better off—that is, the well-being of each would be enhanced overall—if she were telepathic.3,4 Now consider an additional fact: Hannah is human and Bridget is betazoid. We would expect, therefore, Hannah to lack telepathic abilities and Bridget to possess them. After all, it is only characteristic of Bridget’s kind to have such abilities. Furthermore, because it is a lack of a natural power that non-accidentally deprives her of forms of flourishing characteristic of her kind that are naturally supported by that power, Bridget’s failure to have telepathic skills is an evil. And if the lack of telepathy isn’t caused by someone’s wrongdoing, it is a natural evil. But in our story, Bridget doesn’t mind lacking telepathic skills. It’s not something she thinks about—perhaps she doesn’t even know she is a betazoid or perhaps her community is sufficiently warm and accepting that she doesn’t really notice. Is Bridget worse off than Hannah for lacking telepathy? Given their near-identical life circumstances, it is hard to see how either individual could be better off than the other, since neither lacks anything the other has. But the difference in species presents a new problem: the type of lack in each circumstance is not identical. Particularly, Bridget lacks a good characteristic of her kind, while Hannah does not. Does this difference lead to a substantial difference in Hannah and Bridget’s well-being?

3  We can also add that it’s metaphysically possible for both Hannah and Bridget to come to have telepathic skills (perhaps even more technologically advanced aliens could give them such skills by neurosurgery), so there are no conceptual difficulties involving per impossibile conditionals in saying that they would be better off having them. 4  At this point we might say that overall well-being is enhanced here because of Bridget and Hannah’s subjective preference for telepathic abilities. We do not claim that the addition of telepathy makes an objective difference to either’s well-being such that even in the presence of a positive desire to avoid gaining telepathic ability, Bridget or Hannah would be better off. This will be critical in our discussion of so-called privations relative to a certain natural kind (like betazoids).

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

4  Alexander R. Pruss and Hilary Yancey Let V indicate non-instrumental value and C be the many goods of life that Bridget and Hannah have in common. The intuition that there is no significant difference between Bridget and Hannah’s well-being seems to imply that: V ( C ) + V ( being human ) ≈ V ( C ) + V ( being betazoid ) + V ( being impaired ) .

Intuitively, V(being impaired) is a big negative. On accounts of natural evils wherein impairments constitute significant evils, V(being impaired) should be a big negative. This would mean that being betazoid must have something in it that is of sufficiently great value, as compared to being human, to ensure that the left-hand side of our approximate equality is not much greater than the right. In other words, being betazoid would have to be much better than being human. And there doesn’t seem strong reason to think that being of a different species is inherently of much more value. We thus have a conflict. On the one hand, Bridget and Hannah are roughly on a par in terms of value. On the other hand, Bridget suffers from the negative of being impaired, and does not have a compensating good, since being betazoid is not much better than being human. In other words, we have a tension between the following three claims: (1)   Hannah and Bridget are roughly on a par well-being-wise. (2)   It is not significantly better to be betazoid than to be human. (3)   Being impaired as such is a significant negative. We will argue that the best option is to reject (3), concluding that being impaired as such is not a significant negative. In the next section, we will define our key terms, and then argue for our solution to the puzzle.

3.  Definitions: Disability, Impairment, and Health Following some others in the philosophy of disability literature, we distinguish between impairment and disability. This is not without controversy: notably, Elizabeth Barnes is skeptical about impairment serving as a theoretical groundwork for understanding disability, particularly because we lack a robust theory or explanation of what constitutes an impairment.5 She also worries 5  Barnes (2018, p. 1154). See also Barnes (2016).

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Privation in the Problem of Evil  5 that the impairment/disability distinction risks overly “disembodying” the notion of disability, such that disability becomes solely related to one’s social situation, without adequate account for how those with disabilities have forms of solidarity around their experiences of their bodies.6 As Barnes and others note, impairments have often been (controversially) defined as a “loss of normal function.” Normal function might be (and has been) defined in terms of an Aristotelian-type notion of a species form or a Boorsean-type notion of statistical significance in terms of survival and reproduction. Both of these definitions of normal function have substantial critiques in the literature. However, following Dana Howard and Sean Aas, we think that there is value in retaining some concept of impairment, even if impairment itself might be, as they say, “a socially constructed property of bodies, just a different sort of property—one that is more easily confused with a natural property.”7 It is also possible, we think, to reject naturalizing definitions of disability, even if we are more hesitant about rejecting such definitions of impairment.8 Our purpose in this chapter is to offer putative definitions that help capture some of the conceptualizing done specifically in relationship to certain types of bodily configurations9 and so-called privative evils. Thus we will follow some social models of disability and maintain a conceptual distinction between impairment and disability, because our aim is to focus on bodily configurations rather than the interaction of those configurations with an individual’s environment or context.10 And, as a rough and ready characterization, impairments necessarily involve a relatively long-lasting intrinsic condition of an individual that cannot be easily altered by the individual.11

6  Barnes (2018, p. 1158). See also Barnes (2016). 7  Howard and Aas (2018, p. 1121). 8  See Howard and Aas (2018, p. 1119). 9  Compare Cross (2016, p. 694). Cross observes in a footnote that he describes disabilities and/or impairments in terms of bodily configurations assuming for convenience that mental capacities are functions of bodily configurations (p. 694, n. 2). Like Cross, while we refer here to bodily configurations, we do not mean to assume anything about the mind–body question. It could be that telepathy is not mediated through any physical organ in a betazoid. It might not be that a lack of telepathy is aptly described as a “body configuration,” but we do not think this makes any substantive difference to our case. 10  Cross observes that one objection to social models of disability is the possible collapse of the impairment/disability distinction, depending on how one construes the environment (p.  694). On a sufficiently wide reading of environment—the physical and socio-political environment—what will count as an impairment will turn out to be culturally relative. We recognize this objection, but for some of the reasons discussed later in this section, continue to think there is value in maintaining the distinction. 11  For a survey of some issues concerning impairment and disability, see Wasserman, Asch, Blustein, and Putnam (2011). For recent discussion about the advantages and disadvantages of

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

6  Alexander R. Pruss and Hilary Yancey For our purposes, moreover, we will take impairments to deviate from biotypical12 function in an unfavorable direction: of so-called normal eyesight, below normal eyesight and supernormal eyesight, only the below-normal case would qualify as an impairment (though, interestingly, in some contexts and on some views we do not rule out, the supernormal eyesight might qualify as a disability). On this terminological choice, impairment is a function of the intrinsic features of the individual as well as normative facts about what counts as normal function for the species of which the individual is a member. What exactly determines these normative facts is a matter of substantial philosophical interest and dispute.13 Options include evolutionary history,14 statistical and ecological facts about a species,15 irreducible teleological features, and divine intentions. Disability, in turn, is a function of impairment, environment, and the individual’s choices. For instance, if Alex has chosen to spend the rest of his life alone on a scientific expedition to another star without any ability to communicate with people back home except by email, and he has no interest16 in singing, reciting poetry, praying out loud or cursing malfunctioning machinery, losing the power of speech would be an impairment but not a disability. But given his actual environment and the impairment/disability distinction in models of disability, see Francis (2018), Howard and Aas (2018), and Barnes (2018). 12  Notions of “normal” function are controversial, as Amundson (2000) and others have argued. We aim to get at those bodily configurations that are species atypical, without necessarily implying that such configurations are as a result necessarily suboptimal or harmful. 13 Importantly, as Holmes and Patrick (2018) have observed in their “The Myth of Optimality in Clinical Neuroscience,” theorizing about selection pressures as a winnowing process (p. 242) has led to thinking about psychiatric illness as a dysfunction defined by deviation from a statistical norm. However, they argue, the expression of traits is far more complex and environment/context dependent. They argue, “With the exception of gross pathology, shifts within a selected neurobiological function or behavior in isolation will neither be necessary nor sufficient to generate psychiatric illness . . .While there may be fitness disadvantages at the extremes, population-level variability must be interpreted in terms of cost–benefit tradeoffs that can dynamically fluctuate across environments” (p. 249). Their paper is an important contribution, we think, to the debate about defining and understanding norms of biological function. We flag it here as an area for further exploration and discussion, but one we cannot take up further in the context of this chapter. 14 The classic account is Wright (1973). Modified versions are found in, e.g., Millikan (1984), Bedau (1992), and Koons (2000), though Koons no longer endorses the latter. For criticism, see, e.g., Plantinga (1993). 15  Or something more fine-grained than a species (the Boorsean reference class is usually considered an age group of a sex of a species; see Boorse 1977). Mark Johnston also suggested to one of us that for plants, whether a characteristic is normal may depend on the cultivar. 16  Let’s suppose that Alex has no interest in either the subjective sense of not being interested in, or the objective sense of its not being a contribution to, his well-being.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Privation in the Problem of Evil  7 choices, loss of the power of speech would likely be a disability, at least until he and/or the environment adapted. The final term necessary to our argument is “health” (note that we intentionally do not give a precise definition of well-being, though we will argue that on whatever theory of well-being one adopts, health plausibly makes a net positive contribution to well-being). Health, like impairment and disability, is contentious in its definition; for our purposes, we define bodily health as functional integrity: a person (or other living entity) is healthy to the extent that her body’s systems successfully support homeostasis, reproduction, growth, and the other properties that characterize the biological life of a given species.17 On our definition, health is fairly abstract. But one necessary condition for this functional integrity will be the proper functioning of a dizzying array of concrete systems studied by physiologists and psychologists. The proper functioning of particular systems—cardiovascular, neural, musculoskeletal, etc.—can be further subdivided: a left bicep, a right femur, or a particular neuron can all function properly or not, and independently to varying degrees. Without these parts being healthy, or at least being subsumed in the right way in a larger healthy system,18 it doesn’t seem to be the case that the whole individual can be called healthy. This is borne out by cases where, though the part plays a small (perhaps non-essential) role in the organism’s overall homeostasis and achievement of its ends, intervention is required when it ceases to function properly: appendicitis and tonsillitis are good examples. We take health to be something of a scalar concept, where something’s being less than perfectly healthy does not equate to its being positively unhealthy (as when a left bicep is not maximally strong or flexible). But some minimum threshold of health is necessary among the parts in order for there to be a healthy whole. Now, in our initial story, Hannah is healthy. Is this a good thing? It certainly seems so. Recall how Plato (1974) motivates us to the life of virtue by assuming that we realize that bodily health is a good thing, and arguing that a fortiori moral health—the health of the soul—is worth having. His 17  See Weber (2003). Mental, emotional, or spiritual health will be defined in an analogous way to how we have defined bodily health. These also involve a kind of functional integrity, but both the ends at which that integrity aims and the things integrated/integrating differ. The authors are currently engaged in preparing a paper on definitions of life. 18  There are highly redundant parts, such as single cells, whose demise does not harm the larger system (we are grateful to Philip Swenson for this example). Moreover, the demise of a single part may be a constituent of the healthy self-renewal of the system, as when primary teeth are shed by children.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

8  Alexander R. Pruss and Hilary Yancey argument starts with what he takes to be a commonplace: bodily health is good. In both the Apology and the Charmides, Socrates urges his interlocutors to value the health of the soul above but not in contrast to bodily health. Both are good, Socrates argues, though bodily health depends on the health of the soul. This commonplace is basically compelling: bodily health is valuable. But perhaps we need to look a little more carefully at it.

4.  A Solution to the Puzzle Given our definition of impairment and health, Bridget is not healthy simpliciter, even though she has each of the fine-grained constitutive goods of health that Hannah has. Bridget’s left bicep, right femur, and sight are just as healthy as Hannah’s. Both Bridget and Hannah lack telepathic health, but this makes a difference to one’s overall health only if one is a member of a telepathic species. But it’s not clear that having this good of overall health makes a difference to overall well-being such that Bridget’s well-being can be said to be substantially lower than Hannah’s. One can imagine informing Bridget that because she lacks telepathy, she lacks the good of full health qua betazoid, and Bridget replying that this does not make a material difference to her quality of life. It seems like Bridget’s lacking full health would make a difference to her well-being only if (a) Bridget highly valued full health (perhaps she thought she was human and took great pride in her optimal “human” health) or (b) if upon learning that she is impaired, she came to resent her lack of telepathy as an instance of injustice. But in either case, the mere existence of the impairment (and the corresponding lack of health) does not make (much of) a difference to well-being; rather, the difference to wellbeing arises from some other particular attitude, value, or pursuit of the individual. Let us argue some more for the approximate well-being parity of Hannah and Bridget. Consider the alphazoids. Alphazoid females are telepathic but their males are not. Aleph and Alepha are an alphazoid couple with a genetic anomaly that results in their female offspring being non-telepathic, but otherwise both their female and male offspring would be perfectly fine. Alphazoids conceive boys if they mate during a full moon and girls if they mate during a new moon (they are infertile at other times). Moreover, Aleph and Alepha live on a planet away from other alphazoids, so their nontelepathic girls aren’t going to be teased or made to feel bad. Aleph and

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Privation in the Problem of Evil  9 Alepha now have a choice whether to have a girl impaired by lack of telepathy or a boy who has “overall” or “full” health (because he is not supposed to be telepathic), both of which will have roughly the same goods of life and neither of whom will be telepathic. Again, it does not seem that Aleph and Alepha are acting significantly morally better by mating during a full moon, in order to have a non-impaired child, than by mating during a new moon. And this in turn suggests that the impaired girl is roughly on a par in wellbeing and value with the healthy boy. This line of argument seems to fit with the fact that the good of being “fully healthy” is not particularly motivating for many of us. At least by the time we have reached middle age, most of us are resigned with respect to various chronic infirmities, and hence have no hope in this life of achieving being fully healthy, but this resignation does not lower our motivation to stay healthy in many fine-grained respects. Rather, it seems that our motivation to stay healthy is more closely linked to our particular aims and pursuits, rather than to a desire to achieve “full health” simpliciter. Suppose you have a broken arm, and you believe this is your only health problem. But then you learn that your hearing is significantly below normal and that this cannot be cured. Before you learned this news, you thought that fixing the fracture would restore both the health of the arm and overall health. But after learning the news, you know that fixing the fracture would restore the health of the arm but not overall health. If full health has a sub­ otivation stantial value over and beyond its components, then your level of m for fixing the broken arm should go down. Yet surely it would be irrational for your level of motivation to go down by more than a small amount. Thus, the good of full health as such is at most a minor good. In practice, there are complications. When you learn that your hearing is unfixably deficient, you also learn that you are less likely to become a great violinist: a great violinist at least typically needs to have both healthy arms and good ears.19 So, insofar as the desire to become a great violinist—and who wouldn’t want to be one if they could?—is part of the motivation for fixing the arm, the motivation will go down when you learn that your hearing will still be bad. But importantly the reason for this is not that you’ve lost the good of full health. It’s that your power to play the violin is

19  This may not be universally true; some exceptional musicians are deaf or hard of hearing. But it doesn’t seem unreasonable to think that, in general, having good hearing increases one’s ability to play the violin well and correspondingly, a loss of hearing would decrease the likelihood of becoming a great violinist.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

10  Alexander R. Pruss and Hilary Yancey diminished. We can thus think that full health is not much of a further good, over and beyond the goods of all the particular powers that different aspects of health enable (such as the power to dance well), alone or in combination. And these judgments seem right. One not only wouldn’t be much less motivated to fix the arm, but clearly one shouldn’t be much less motivated. This leads to an interesting result. Bridget is impaired and on this basis suffers a natural evil. But all the goods that Hannah has—with the possible exception of the minor good of overall heath—Bridget has as well. Suffering a natural evil constituted by her impairment, it turns out, has little or no effect on Bridget’s well-being. Not all natural evils—including ones that could be considered “severe”—do real harm. Thus we conclude that being impaired, as such, is not a significant negative. Both Hannah and Bridget lack the goods that telepathy provides. In Hannah’s case, that lack is a mere lack. But goods characteristic of one’s species, such as four-leggedness in the case of sheep and telepathy in the case of betazoids are due goods, and hence their lack is an impairment and an evil.20 But one is no worse off, or not much worse off, for the good that one lacks being a due good. Put differently, the fact that Bridget, as a betazoid, ought to have telepathic abilities doesn’t count for much when evaluating whether her lack makes a difference to her overall well-being.

5.  Privative Evils and Theodicy We have argued that in the case of Bridget the betazoid, the fact that she ought (as a member of a certain species) to be telepathic does not in itself pose a substantial problem for a theodicist. The fact that Bridget is impaired

20  One might wonder whether it is plausible to think that one is “due” all characteristic properties of the species of which one is a member. We could imagine a species where every individual has all but one of the 32 tooth positions in the jaw filled, and it is random which is the unfilled position. Then having position n in the jaw filled is characteristic of the species—31/32 members of the species normally have that feature—but it is not characteristic or normal to have all of the positions filled. However, our use of “characteristic” here can be taken to be not merely statistical, so that having position n filled isn’t characteristic of the species—what is characteristic is having 31/32 spaces filled. Alternately, one might distinguish between characteristic and strongly characteristic features, and say that having position n filled is characteristic, but it is having 31/32 spaces filled that is strongly characteristic. We can then assume that telepathy is a strongly characteristic property of being betazoid, and the lack of a strongly characteristic good is a lack of a due good.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Privation in the Problem of Evil  11 doesn’t make a substantial difference to her well-being, and so it fails to be the kind of evil required to get the problem of evil argument off the ground. Our argument about privations generalizes to evils other than impairments. Suppose that Alice and Bob are in financial need, and each has a different rich uncle that does not give them any gifts. However, ten years ago, Alice’s rich uncle promised Alice that he would give her generous gifts if she should ever have financial needs. Bob’s made no such promises. Alice is not galled, however, by her uncle’s promise-breaking, because she has completely forgotten the promise. The modesty of Bob’s means constitutes Bob as deprived of some goods, but these goods are not due to Bob, and hence the modesty of Bob’s means is not an evil. The modesty of Alice’s means, however, constitutes her as deprived of goods that are due to her from her uncle, and hence she suffers from being deprived of a due good. But even though she suffers a significant evil, she is not much worse off than Bob.21 If the above arguments are right, then when you calculate someone’s level of well-being, and in your accounting you come to a privative evil, while you don’t add the value of the good that they are deprived of, you also don’t decrement, or at least don’t significantly decrement, their well-being to mark the fact that the good that they lack was due to them. As a result, those who are merely lacking a good and those who are lacking it despite its being due have similar well-being. An impairment, then, is significantly deleterious to well-being only when and because it significantly limits one’s abilities or causes some further harm. Bridget may, as a result of her impairment, lack a significant good— she cannot communicate telepathically—while the impairment itself entails only the lack of an insignificant good of “overall” or “full” betazoid health. But even though we do not add the value of this absent telepathic ability to

21  One might wonder whether we are right to argue for the apparent parity between these two individuals. Consider the following possible counterexample, provided by an anonymous referee: Suppose two individuals—alike in all relevant respects—go out for a night on the town. The first is prodigal and spends all of his money recklessly. The second is robbed of his money at gunpoint. Both are now materially identical—broke. Yet is seems both that (a) one is worse off than the other and that (b) a generous third party would have a moral reason to give money to the robbed victim over the prodigal subject. In response, we would argue that vice (like prodigality or recklessness) does provide moral reasons not to give the money to the first individual. A better case would be where someone loses all her money by accident (say she trips and the money falls into the sewer) and the other person is robbed. Then, we think, the generous third party lacks a noticeable moral reason to give the money to the person who was robbed (that is, deprived of a due good) than to the person who lost it accidentally.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

12  Alexander R. Pruss and Hilary Yancey Bridget’s well-being ledger, we shouldn’t go on to subtract a further large chunk for its lack being an impairment. Alice’s and Bob’s rich uncles are in morally significantly different positions with respect to Alice and Bob, but it seems that that isn’t because of differences of value between the situations. Alice’s uncle owes her property. Bob’s doesn’t owe Bob. But it seems that Bob’s uncle’s not giving him money deprives him of approximately the same thing as Alice’s uncle’s not giving her money does. The difference is deontic rather than axiological. To see this point better, imagine a generous third party who can relieve either Alice’s or Bob’s modesty of means. She would be relieving Alice of an evil, a privation of a due good. But in Bob’s case, she would be relieving him of a mere lack. But she doesn’t have a significantly greater reason to give the gifts to Alice, at least as regarding Alice and her goods alone.22 Suppose now you are Bridget the non-telepathic betazoid. We did assume that Bridget doesn’t mind being deprived of telepathy, or at least she doesn’t mind it any more than Hannah minds not having telepathy. But even though as Bridget you may not mind it, you might nonetheless ask whether your impairment provides an argument from evil against the existence of God. But a consideration of the above arguments makes this line of thinking available: “If I were human like Hannah, I wouldn’t be significantly better off than I am now. But God would be perfectly justified in creating another happy human like Hannah. So why shouldn’t he be justified in creating me instead?” In other words, your impairment itself doesn’t result in a substantial loss of well-being (especially when compared with Hannah the human). But if God is perfectly justified in creating Hannah, with her level of well-being, then what difference does it make that God created you, a (technically) impaired betazoid, instead? Granted, our arguments did allow that there might be a small decrement to well-being due to impairment (based on the lack of full health). And perhaps that small decrement could call for a theodicy. But one can’t hang a persuasive argument from evil on small decrements to well-being, because it is so easy to imagine that in a complex enough world such small decrements might serve a proportionate purpose. That’s why nobody seriously runs the argument from evil on the basis of their last stubbed toe. Perhaps one can argue that we have underestimated the deontic considerations. While the cases of Hannah and Bridget are on a par value-wise, 22  But if the generous third party additionally has reason to benefit Alice’s uncle, by relieving him of his burden of unfulfilled duty to Alice, she would have reason to give the gift to Alice.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Privation in the Problem of Evil  13 perhaps the fact that Bridget is deprived of a due good should count for more, as one might argue is true in the case of the two uncles. That Alice’s uncle owes her money might arguably make a difference. Does God have a duty to create beings who fulfill species norms? It may be that there are significant duties of potential parents to try to have children that fulfill the species norm—Robert M. Adams (1972) considers this option. Others (e.g., Kahane and Savulescu  2009; Barnes  2014) consider whether parents would be permitted to select for disability or if there are parental duties to avoid disability where possible. We do not aim to resolve these disputes here. But even if it were the case that parents had some kind of duty to try to have children who embody the species norm, God having an obligation to make Bridget conform to her species norm would depend on God’s being analogous to a parent in this respect. Like Adams, we doubt that God has duties that are analogous to those of human couples with respect to species norms. In any case, it is clear now that given the line of thought we put in Bridget’s mouth, to make an argument from evil on the basis of privative evils we would need some significant work on God’s deontic position.

6.  A Confusion? However, it’s possible that there has been a confusion. Recall the equation: V ( C ) + V ( being human ) ≈ V ( C ) + V ( being betazoid ) + V ( being impaired ) .

The basis for that equation was the thought that Bridget is not much worse off than Hannah. But the equation talks of (non-instrumental) value, rather than about how well off the individuals are. The difference here is between value-of and value-for. We can say that the value for Hannah of C plus being human is approximately the same as the value for Bridget of C plus being betazoid plus being impaired. But to move from this to a claim about the value of these features as such, as opposed to their value for the individuals, appears to be a confusion.23 We are happy to embrace the alleged confusion. Ockham’s razor favors the Aristotelian identification of non-instrumental value-for with noninstrumental value-of. A good flute is a flute that functions well as the kind 23  We are grateful to Larry Temkin for pointing this out.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

14  Alexander R. Pruss and Hilary Yancey of thing it is and it is good for the flute to be uncorroded.24 Functioning well as the kind of thing something is just is flourishing understood broadly (cf. Aristotle 1984, Book 4). Moreover, there is a kind of argument from love for the identification of value-of with value-for. Love has at least three aspects: benevolence, appreciation, and pursuit of union (Pruss 2012). Love isn’t merely a conjunction of these aspects. The aspects are tightly intertwined, with each furthering rather than hampering the others. And the identification of (non-instrumental) value-of with (non-instrumental) value-for gives us a particularly elegant account of part of this intertwining. Appreciation is appreciation of what is valuable. When I appreciate the value of an individual, I seek to preserve and promote that value. Now when I act benevolently for an individual, I seek to preserve and promote what is of value for the individual. If the value-of and value-for are the same, then this appreciation motivates the benevolence and the benevolence is an expression of the appreciation. And a benevolence that is an expression of appreciation is a benevolence that escapes the danger of patronization and condescension, something that is often a danger in benevolent activity for the sake of another. On the other hand, if value-of and value-for were different, then not only would we lack this elegant intertwining, but there could be a real conflict between appreciation and benevolence. For appreciation would naturally lead me to promote the value of the beloved, which could take time away from the benevolent promotion of the value for the beloved, and conversely. The identity of value-of and value-for makes it possible for love to have an intrinsic unity between the appreciative and benevolent aspects. And union can then flow from these, since through benevolence one unites oneself to the beloved in will and through appreciation one unites in intellect (cf. Aquinas 1920, I–II, 28). But all that said, the theodical applications perhaps do not require an identification of value-of and value-for, at least in the Christian tradition. The God of Christianity is primarily a God of benevolent agapê. Theodicy then becomes primarily a task of arguing that the kind of world that we have is a world that a God who loves all his creatures could have created in his benevolence. But here it is the value for creatures that takes center stage rather than the value of creatures, if there is a distinction to be made.

24  Of course, the case of the artifact may only be analogous to real flourishing. After all, on Aristotle’s own metaphysics, it does not appear that artifacts are substances.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Privation in the Problem of Evil  15

7.  Where Are We Now? If the above line of thought is right, then in the case of privative evils like impairments, the fact that the good involved is a due good does not contribute significantly negatively to the subject’s well-being. What seems lost by an impairment—full health—turns out to be of much less consequence than it might have initially seemed, and certainly needn’t make a huge difference to well-being. Thus, the privation component of an impairment alone does not give it much weight in the problem of evil. Given that no theodicy is required in the case of God’s creating beings which lack such a good when that good is not due, it follows that the proponent of an argument from privative evil needs to give us reason to think that when that good is due, God should provide the good. In other words, the arguer from privative evil needs to show us that the absent due good is due from God. (Recall the case of the two uncles, where the wealth was due to Alice from the uncle but not from the generous third party.) This would require careful work on the particulars of the God–creature relationship.25 Moreover, the results of this work may depend on the particulars of the privative evil in question. We do not know how things will look once this work is done, so this chapter is in significant part an encouragement to engage in such work. In the case of privative evils, then, the arguer from evil needs to work harder. But neither can we say that it is clear that there is no problem of privative evil. Thus Augustine’s theory that all evil is privative evil (a) would advance the theodicist’s cause, but (b) would not guarantee an automatic universal theodicy for all evils. Both (a) and (b) should be welcome to theists who like Augustine’s theory. It’s obvious why (a) is welcome to them. And the reason (b) is welcome is that we should be suspicious of a theory that would give a theodicy of all imaginable evils. A world with billions of innocent people all of whom suffer dreadfully for eternity would be a world for which we should not be able to give a theodicy (and hence the classical theist who thinks God is a necessary being should conclude that such a world is impossible26), and if Augustine’s theory would give a theodicy for it, there is something fishy about it. Our arguments here only concern the privative aspects of impairment. If there is positive evil involved—for instance, if there is pain and pain is a 25  For some work that would suggest that such goods are not due from God, see Murphy (2017). 26  But see Gulesarian (1983).

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

16  Alexander R. Pruss and Hilary Yancey non-privative evil—then such positive evils will require separate treatment. But some progress has been made. And each time we make progress on what was previously a troubling form of the problem of evil, we get some inductive support for the claim that the problem of evil in general has a solution.27

Bibliography Adams, Robert M. 1972. “Must God Create the Best?” Philosophical Review 81: 313–32. Amundson, Ron. 2000. “Against Normal Function.” Studies in History and Philosophy of Science Part C: Studies in History and Philosophy of Biological and Biomedical Sciences 31: 33–53. Aquinas, Thomas. 1920. Summa Theologica. Translated by the Fathers of the English Dominican Province. London: Burns, Oates and Washbourne. Aristotle. 1984. Nicomachean Ethics. In The Complete Works of Aristotle, vol. 2, edited by J. Barnes, 1729–1867. Princeton, NJ: Princeton University Press. Barnes, Elizabeth. 2014. “Valuing Disability, Causing Disability.” Ethics 125: 88–113. Barnes, Elizabeth. 2016. The Minority Body. Oxford/New York: Oxford University Press. Barnes, Elizabeth. 2018. “Against Impairment: Replies to Aas, Howard and Francis.” Philosophical Studies 175: 1151–62. Bedau, M. 1992. “Where is the Good in Teleology?” Philosophy and Phenomenological Research 52: 781–801. Boorse, Christopher. 1977. “Health as a Theoretical Concept.” Philosophy of Science 44: 542–73. Cross, Richard. 2016. “Impairment, Normalcy, and a Social Theory of Disability.” Res Philosophica 93: 693–714. Francis, Leslie. 2018. “Understanding Disability Civil Rights Non-Categorically: The Minority Body and the Americans with Disabilities Act.” Philosophical Studies 175: 1135–49.

27  We are very grateful to the participants of the Rutgers Value and Religion Workshop for many helpful comments, to the John Templeton Foundation for sponsoring that workshop, and to the comments of anonymous referees as well as of Christopher Hauser and Philip Swenson. The paper is much better for these comments.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Privation in the Problem of Evil  17 Gulesarian, Theodore. 1983. “God and Possible Worlds: The Modal Problem of Evil.” Noûs 17: 221–38. Holmes, Avram J. and Lauren Patrick. 2018. “The Myth of Optimality in Clinical Neuroscience.” Trends in Cognitive Sciences 22: 241–57. Howard, Dana and Sean Aas. 2018. “On Valuing Impairment.” Philosophical Studies 175: 1113–33. Kahane, Guy and Julian Savulescu. 2009. “The Moral Obligation to Create Children with the Best Chance of the Best Life.” Bioethics 23(5): 274–90. Koons, Robert  C. 2000. Realism Regained: An Exact Theory of Causation, Teleology, and the Mind. Oxford/New York: Oxford University Press. Leslie, Sarah-Jane and Adam Lerner. 2016. “Generic Generalizations.” Stanford Encyclopedia of Philosophy (Winter 2016 Edition), ed. Edward  N.  Zalta. https://plato.stanford.edu/archives/win2016/entries/generics/ (retrieved April 10, 2019). Millikan, Ruth Garrett. 1984. Language, Truth and Other Biological Categories. Cambridge, MA: MIT Press. Murphy, Mark  C. 2017. God’s Own Ethics: Norms of Divine Agency and the Argument from Evil. Oxford: Oxford University Press. Plantinga, Alvin. 1993. Warrant and Proper Function. Oxford: Oxford University Press. Plato. 1974. Plato’s Republic. Translated by J.  M.  A.  Grube. Indianapolis, IN: Hackett. Pruss, Alexander R. 2012. One Body: An Essay in Christian Sexual Ethics. Notre Dame, IN: Notre Dame University Press. Silvers, Anita. 2003. “On the Possibility and Desirability of Constructing a Neutral Conception of Disability.” Theoretical Medicine and Bioethics 24: 471–87. Wasserman, David, Adrienne Asch, Jeffrey Blustein, and David Putnam. 2011. “Disability: Definitions, Models, Experience.” Stanford Encyclopedia of Philosophy (Summer 2016 Edition), ed. Edward N. Zalta. https://plato.stanford.edu/entries/disability/ (retrieved February 3, 2016). Weber, Bruce. 2003. “Life.” Stanford Encyclopedia of Philosophy (Spring 2015 Edition), ed. Edward  N.  Zalta. https://plato.stanford.edu/archives/spr2015/ entries/life/ (retrieved April 20, 2018). Wright, Larry. 1973. “Functions.” The Philosophical Review 82(2): 139–68.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

2

What If God Makes Hard Choices? Paul Draper

By a “hard choice” between actions A and B, I mean one in which the chooser believes that A is better than B in some respects and worse in others, but has no inclination to believe that A or B is better overall than the other or that A and B are equally valuable overall.1 It is not surprising that human beings face hard choices in this sense. After all, there is much we do not know about value. Could, however, an omniscient God face hard choices? At first glance, it seems not, because an omniscient being would know either that A and B are equally valuable overall if in fact they are equally valuable overall or which one is better than the other if they are not equally valuable. One problem, however, with this claim about what an omniscient being would know is that it assumes the truth of what Ruth Chang calls the “trichotomy thesis,”2 which states that “better than,” “worse than,” and “equal in value to” exhaust the relevant value relations that one action can bear to another.3 As Chang points out, this thesis is tempting because we mistakenly

1  An action can be valuable either because of its own properties (e.g., it might be graceful and for that reason possess non-instrumental aesthetic value, in which case there would be an aesthetic reason to perform it) or because of the properties of its effects (e.g., it might bring about a distribution of goods that is just or promote someone’s welfare and for that reason possess instrumental moral value, in which case there would be a moral reason to perform it). The issue of whether aesthetic reasons are also moral reasons will be addressed (though not settled) later in the chapter. 2  “The Possibility of Parity,” Ethics 112 (July 2002): 659–88. The trichotomy thesis is defined on p. 680. 3  Another problem is that it ignores the real possibility that open theists are right about the nature of perfect knowledge and hence that God, like humans, makes decisions under uncertainty. Presumably a God would, in deciding what to do, consider the long-term consequences of the various actions She might perform. (I assume here that plausible forms of nonconsequentialism do not go so far as to assert the complete moral irrelevance of consequences.) If, however, the value of the various actions open to God depends on their consequences and in particular on the future free choices of human beings or other persons, and it is impossible to know beforehand what those choices will be, then even if the trichotomy thesis is true, God might be forced to choose between two or more actions none of which is known or even believed by Her to be more valuable than or equal in value to any of the others.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

What if God Makes Hard Choices?  19 assume that values are like the properties that interest scientists—that they can be represented by real numbers. There is no good reason to think, however, that the structure of properties in the realm of value should mirror those in the physical sciences in this way. Thus, the trichotomy thesis is at best speculative. It ignores at least one other real possibility, namely, that in some cases two alternative actions are “on a par” (in Ruth Chang’s technical sense of that term): They are comparable with respect to their value—they are in some sense in the same neighborhood when it comes to value—but they are not equally good overall and neither is better overall than the other.4 If this possibility is actual, then even an omniscient God, if there were such a being, might face hard choices. While such a God might not literally struggle to make these choices, She could not base them solely on a comparison of the strength of competing reasons. The goal of this chapter is to explore a new model of God’s motivational structure based partly on the idea that, if there were a God, then some of God’s choices would be hard, and partly on the idea that moral goods are different in kind from, for example, aesthetic goods. In constructing a model of how a theistic God makes choices, the goal of philosophers, regardless of their personal religious beliefs or lack thereof, should be to make theism as plausible as possible. This is certainly one of Richard Swinburne’s goals, and he pursues it with clarity and rigor. So I will start by describing his model. Then I will identify three meta-ethical assumptions made by that model and begin to motivate my model by challenging one of those assumptions. After explaining my model and identifying an important advantage that it has over Swinburne’s, I will close by discussing its most important disadvantage. Throughout the chapter, except when context screams otherwise, I will employ a classical theistic conception of God. Accordingly, I assume that, in order to qualify for the title “God” a being must be the sovereign ruler of all things, perfect in power, knowledge, moral goodness, practical rationality, and freedom (from non-rational desires), and maximally worthy of love, praise, gratitude, and obedience. For reasons that will soon be clear, I do not assume that a classical theistic God must possess all of Her intrinsic properties necessarily.

4  Another possibility is that the value of one action is incomparable to the value of another, but the choices I will be considering in this chapter don’t seem to be like that. Comparisons can be made, just not one of the usual three.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

20  Paul Draper

1.  Swinburne’s Model Swinburne claims that on his understanding of what a theistic God would be like, God is the simplest sort of person possible. A significant part of his justification for that claim is his further claim that God’s perfect moral goodness follows necessarily from Her being omniscient and perfectly free. His justification for this further claim has, at least implicitly, two steps. The first step attempts to show that no being could be omniscient and perfectly free without also being perfectly rational. The second step attempts to show that perfect rationality entails moral perfection. Concerning the first step, to say that a person is “perfectly free” is to say that only reasons, which Swinburne equates with beliefs about the value of various actions open to that person, influence that person’s choices. Further, Swinburne holds that, in the absence of causal factors like bad desires, believing that an action has value entails having a reason to act, and such reasons, when stronger than competing reasons, are necessarily motivating. Swinburne concludes that perfectly free persons who are also omniscient are perfectly (objectively) rational: they always (i) do the overall best action open to them when there is such an action, (ii) an equal overall best action when more than one action open to them is tied for overall best, or (iii) an overall good action when at least one action open to them is overall good and no action is overall best (for example, because every action open to them is overall worse than another action open to them).5 The second step in the argument is in one sense very short because Swinburne takes an action to be morally good or morally best just in case it is overall good or overall best to do. He adds that uniquely best actions will in some cases be moral obligations and so morally wrong not to do and in other cases supererogatory. This means that a perfectly rational person is guaranteed never to do wrong and always to exceed moral duty whenever possible and to the maximum degree possible (whenever there is a maximum degree possible). In short, although Swinburne doesn’t put it this way, his position is that an omniscient and perfectly free person must be perfectly morally good because such a person must be perfectly rational

5  This is not quite right because (i) Swinburne believes (correctly I think) that even an omniscient being would lack knowledge of causally undetermined future events, and (ii) the value of an action might depend on whether or not such events occur. I owe this point to Luke Wilson.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

What if God Makes Hard Choices?  21 and “perfect objective rationality” and “perfect moral goodness” refer to one and the same property.6 This argument depends for its success on at least three controversial meta-ethical assumptions. The first is minimal moral realism: moral judgments are true or false (cognitivism) and some moral judgments are true (that is, error theory is false). The second is a weak form of motivational intellectualism: in the absence of non-rational causal influences, beliefs about what is overall the best thing to do are necessarily motivating. The third is that a moral reason is not one sort of reason among others but instead an all-things-considered reason—a meta-level reason that exists in a particular case only when all object-level reasons relevant to a particular case can (if they conflict) be weighed against each other on some objective scale.

2.  A Challenge to Swinburne’s Model I am inclined to accept the first two of Swinburne’s three meta-ethical assumptions, but I have doubts about the third. I am not convinced that what is morally good is in all cases identical to what is good “overall,” and for this reason I am skeptical about Swinburne’s identification of perfect rationality with perfect moral goodness. An alternative position that I find no less plausible is that certain goods are specifically moral (for example, charity and fairness) while others (for example, majesty and elegance) are not. Call this position “value partitionism.” I will not try to prove that value partitionism is true. Instead, I will assume its truth for the purpose of exploring its theological implications. If value partitionism is true, then there appears to be significant tension between perfect rationality and moral perfection. For moral perfection would seem to demand always choosing moral goods over non-moral ones while perfect rationality would demand choosing non-moral goods whenever they are more valuable than competing moral goods. Of course, if moral goods were always more valuable than non-moral goods, then no tension would exist, but such a view is not plausible. A more promising solution to this problem relies on the fact that perfect moral goodness is optimal moral goodness, and the optimal degree of a perfection need not always be the maximal degree. A comparison to perfect freedom may make 6  Richard Swinburne, The Existence of God, 2nd edition (Oxford: Clarendon Press, 2004), chapter 5. See especially pp. 99–106.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

22  Paul Draper this idea more palatable. God is thought to be perfectly free, but not maximally free. Instead, it is assumed that to possess perfect or optimal freedom is to possess as much freedom as possible compatible with being perfectly rational and thus being constrained by reasons. Similarly, a theist who accepts value partitionism might want to claim that to possess moral goodness to the optimal degree is to possess as much moral goodness as possible compatible with being perfectly rational.7 If this solution works (and I will assume it does), then the next task is to consider cases in which the value of an action promoting one or more moral goods is on a par with the value of an alternative action promoting one or  more non-moral goods. I submit that in such cases, perfect morality, even if it is compatible with perfect rationality, demands more than perfect rationality. Consider, for example, the story of Mary of Bethany as told in the following mashup of verses from the Gospels of Mark and John: Six days before the Passover, Jesus came to Bethany, where Lazarus lived, whom Jesus had raised from the dead. Here a dinner was given in Jesus’ honor. Martha served, while Lazarus was among those reclining at the table with him. Then Mary took about a pint of pure nard, an expensive perfume; she poured it on Jesus’ feet and wiped his feet with her hair. And the house was filled with the fragrance of the perfume.8 Some of those present were saying indignantly to one another, “Why this waste of perfume? It could have been sold for more than a year’s wages and the money given to the poor.” And they rebuked her harshly. “Leave her alone,” said Jesus. “Why are you bothering her? She has done a beautiful thing to me.”9

Now consider the problem of comparing the value of Mary’s beautiful act to the alternative charitable act of selling the perfume and giving the money to the poor. To the extent that Mary’s action involves caring for or honoring Jesus or just giving him pleasure, it has moral dimensions. But here I want to downplay those dimensions and instead focus on the aesthetic aspects of the action—its tactile and olfactory beauty, as well as its gracefulness and sensuality (at least as I imagine it). It seems naïve to assume that there is  some objective scale of value according to which it could be judged 7  Mark Murphy defends the even more radical claim that an Anselmian God is not morally perfect because moral perfection implies being constrained by more than reasons and that is incompatible with perfect freedom understood in the way I just described. See his God’s Own Ethics: Norms of Divine Agency and the Argument from Evil (Oxford: Oxford University Press, 2017). 8  John 12:1–3. 9  Mark 14:4–6.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

What if God Makes Hard Choices?  23 correctly that one of these two alternative actions is better overall than the other. The values in question are too different from each other. It is at least plausible to claim instead that the two actions are on a par with respect to their goodness. They are comparable, but neither is better than the other and they are not equal in value. To see that they are not equal in value, suppose that we know exactly how expensive the perfume was and that Mary also owned a slightly more expensive perfume. Selling the more expensive perfume and giving the money to the poor would be overall better than selling the less expensive perfume and giving the money to the poor. Thus, if the two actions originally being compared were equally valuable, then it would follow that selling the slightly more expensive perfume and giving the money to the poor would be overall better than what Mary did. But surely that does not follow.10 Suppose I am right that this is a counterexample to the trichotomy thesis. Then, if Swinburne were right that moral reasons are generated only by judgments of overall value, it would follow that selling the perfume and giving the money to the poor was not morally better than what she did. Clearly, though, it was. If Mary were morally perfect, that (or something similar) is what she would have done. Mary’s reasons for doing what she did were (we are supposing) primarily aesthetic. In fact, so far as we can tell, what she did is compatible with her being aesthetically perfect, at least to the extent of making the promotion of aesthetic values her first priority whenever doing so is compatible with her being perfectly rational. Thus, the alternative act suggested by “some of those present” (Judas is named as Mary’s critic in the book of John) is better morally and worse aesthetically than what Mary chose to do, but it is not overall better and so is not more rational than what Mary did. Thus, if the trichotomy thesis is false because of cases like Mary of Bethany and if value partitionism is true, then Swinburne’s assumption that what is morally better is what is overall better is false. And this in turn implies that, even if moral perfection is compatible with perfect rationality, it is not entailed by it: perfect rationality and moral perfection are distinct properties.

3.  A New Model This suggests a new model of God’s motivational structure, one based on the epistemic possibility that, if God exists, then She makes hard choices of 10  This is called the “small improvement argument.” Ruth Chang defends it in section 1 (pp. 667–73) of “The Possibility of Parity.”

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

24  Paul Draper a certain special sort, namely, choices in which two (or more) options are on a par with respect to their value, with one option being morally better than another but not better overall. In such cases, both options are compatible with perfect rationality but only one is compatible with moral perfection, which means that God could be essentially perfectly rational and yet only accidentally morally perfect. Surprisingly, or maybe not surprisingly, some parts of the Bible seem to portray God as prioritizing non-moral values over moral ones. Consider, for example, the Book of Job, whose protagonist, a righteous man who suffers horrifically, accuses God of lacking sufficient commitment to the moral value of justice. The vast majority of commentators agree that God does not directly respond to Job’s charge.11 Instead, speaking out of the whirlwind, He describes His design of the cosmos and of the animal kingdom in a way clearly intended to emphasize His power and the grandeur of His creation. Were it not for theological worries about God’s moral perfection, the most natural interpretation of this part of the story would be either that God agrees with Job’s charge that He is unjust or that God denies that Job can sensibly apply terms like “just” and “unjust” to Him because He and Job are not members of any shared moral community.12 This is why Job’s first response (before capitulating in his second response) is just to refuse to repeat his (unanswered) accusation. To paraphrase Jesus’s rebuke of Mary’s accusers, God appears to be saying to Job, “Leave me alone. Why are you bothering me?13 What I have done is majestic and sublime.” On this interpretation, Job’s God does not do what is morally best. Thus, He is not maximally morally good, though He might still be morally perfect (i.e., he might still possess the optimal degree of moral goodness) if what is morally best in this case is not compatible with perfect rationality. One might object that there is an important disanalogy between the case of Mary of Bethany and the case of Job’s God: Mary was forced to choose between charity and beauty, while Job’s God, as suggested by the epilogue of the book of Job, was capable of both designing a magnificent universe and attending to the well-being of His creatures in a manner compatible with 11 One notable exception is Eleonore Stump. See her Wandering in Darkness (Oxford: Clarendon Press, 2010), chapter 9. 12  For a defense of the position that the God of the book of Job is not a member of Job’s moral community, see Wes Morriston, “Protest and Enlightenment in the Book of Job,” in Paul Draper and J.  L.  Schellenberg (eds.), Renewing Philosophy of Religion: Exploratory Essays (Oxford: Oxford University Press, 2017), chapter 13. 13  Actually, Job’s God asks, “Who is this that darkens my counsel?” (Job 38:2), which brings to mind Aquinas’s view that light or clarity is a necessary condition of beauty.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

What if God Makes Hard Choices?  25 other moral values like justice. The existence of such a disanalogy, however, is at best impossible to prove. Consider, for example, the aesthetic value of drama (or that of a good story, whether fictional or not). Drama requires of necessity an intense conflict of forces, and nothing is more dramatic than a struggle between good and evil. A world with no conflict, no pain, no immorality, and no tragedy would, quite possibly of necessity, have dramatically less drama and so dramatically less aesthetic value than our world. Defenders of the natural order theodicy also lend support, though unintentionally, to the idea that a God might be forced to make hard choices between aesthetic value and moral value. For example, it may be that producing a universe governed by a few laws expressible as elegant mathematical equations is an impressive accomplishment, not just because of the wisdom and power required for such a task, but also because of the aesthetic value of such a universe. That value may very well depend, however, on God’s choosing not to intervene regularly in nature to protect His creatures from harm. Much of the aesthetic value of the animal kingdom may also depend on its being the result of a long evolutionary process driven by mechanisms like natural selection. As Darwin famously said in the last lines of On the Origin of Species by Means of Natural Selection, There is grandeur in this view of life, with its several powers, having been originally breathed into a few forms or into one; and that, whilst this planet has gone cycling on according to the fixed law of gravity, from so simple a beginning endless forms most beautiful and wonderful have been, and are being evolved.14

Unfortunately, such a process, if it is to produce sentient life, may also entail much suffering and countless early deaths. One questionable assumption of some natural order theodicists is that such connections between aesthetic goods and suffering provide a moral justification for God’s allowing horrific suffering. It is no less plausible that in such a scenario the moral value of preventing horrendous suffering is on a par with (or cannot be compared to) the aesthetic value of regularity, sublimity, and drama. If so, then a morally perfect God would not trade the former for the latter though a perfectly rational one might or might not. Suppose, then, that a God would sometimes be forced to choose between an action that promotes moral value and an action that promotes aesthetic 14  The Origin of Species: A Facsimile of the First Edition (Cambridge, MA: Harvard University Press, 1954 [originally published in 1859]), p. 490.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

26  Paul Draper value or some other sort of value. (Who knows how many dimensions of value an omniscient being would have to consider?) Suppose further that in some of these cases the two actions would be on a par with respect to value. Then a God would face hard choices between options both of which are compatible with Her perfect rationality, but only one of which is compatible with Her being morally perfect. Notice that this model has significant implications for God’s agency. If, for example, God faced multiple hard choices between moral and aesthetic goods, then She could choose to be morally perfect or choose to be aesthetically perfect or choose to fall anywhere on a spectrum between those two extremes. Such a God would have a great deal more autonomy than, for example, a God who is constrained by reasons to perform only a single act of will, namely, actualizing the possible world that She knows to be better overall than any other possible world that She can actualize. Such a God would also have greater autonomy than a God who performs multiple distinct acts of will, but only because She sometimes makes arbitrary choices between options that are equally good overall. On my model of God’s motivational structure, God’s sovereignty would extend, not just to the external world, but also to the sort of person She is and to the values She prioritizes. The reasons on which She acts and even the reasons She has to act would in part depend on Her. Like the various Gods depicted in the Bible, such a God would not be “mechanical”—a mere slave to reasons—but would have genuine individuality and perhaps even a robust personality or character.15 None of this shows, however, that my model is superior to Swinburne’s. For one thing, it is not obvious that a God with greater autonomy is a good thing, theologically speaking. In addition, Swinburne’s model is compatible with God facing hard choices.16 Given Swinburne’s rejection of value partitionism, however, such choices would always involve alternative actions that are morally on a par, so nothing as significant as God’s moral perfection could turn on them. Once again, though, it’s not clear to me that this feature of his theory is disadvantageous. In the next section, however, 15  The argument in this paragraph owes much to (but shouldn’t be blamed on) Ruth Chang’s wonderful 2014 TED talk, “How to Make Hard Choices.” https://www.ted.com/talks/ ruth_chang_how_to_make_hard_choices?language=en. 16  Swinburne doesn’t specifically consider the possibility of actions being on a par with respect to value, but he does believe that “Frequently, perhaps normally, there is no objective scale in which competing reasons can be weighed” (The Existence of God, p. 101). Presumably, this is why he later says that “so often, perhaps almost always, reason provides no unique best action for God to do” (p. 105).

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

What if God Makes Hard Choices?  27 I will argue that a model of divine decision-making that not only allows for a God who makes hard choices, but also affirms value partitionism, does have one very important theological advantage over Swinburne’s model.

4.  An Advantage The advantage in question concerns the following argument against ­classical theism. (1)  No being that is perfectly free can be significantly free. (2)  Any being that is maximally worthy of praise and gratitude must be significantly free. (3)  Any being qualifying for the title “God” (as classical theists use that title) must be both perfectly free and maximally worthy of praise and gratitude. So, (4) No being qualifies for the title “God” (as classical theists use that title). My model, unlike Swinburne’s, has the advantage of implying that this valid argument for the nonexistence of a classical theistic God is unsound. Premise (1) is the key. That premise sounds odd. How could a being that is perfectly free not also be significantly free? The answer is that two different sorts of freedom are involved here. The term “perfectly free” is used in Swinburne’s sense. It refers to a sort of freedom from certain influences or constraints as opposed to a freedom to do something. More precisely, a person is perfectly free when their choices are influenced only by reasons (understood as beliefs about value) instead of by (other) causal factors like felt desires or compulsions. The term “significantly free” combines a moral sense of “significantly” with a libertarian sense of “free.” Accordingly, a person is significantly free if they are free in the libertarian sense to choose between two or more alternative actions even though they believe that one of those actions is morally worse overall than another. On Swinburne’s model, perfectly free beings cannot be significantly free in this sense because, according to that model, (i) all-things-considered reasons are beliefs about overall value, (ii) beliefs about what is morally worse overall are beliefs about what is worse overall, and (iii) in the absence of non-rational desires, beliefs about overall value are necessarily motivating. On my model, even if what is overall best is always morally best, moral

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

28  Paul Draper value is not identical to overall value because what is morally best is not always overall best. An action that is morally best (like, we may suppose, Mary’s selling the perfume and giving the money to the poor) will not be overall best in those cases in which the moral value of that action is on a par with the sort of value some alternative action possesses (like the aesthetic value of Mary’s anointing Jesus with perfume). Therefore, on my model, a being that is perfectly free might still have morally significant libertarian free will, which implies that God could be essentially omniscient and essentially perfectly free and yet be only accidentally morally perfect. One might object that this difference with respect to premise (1) makes no difference because, while premise (3) is true by definition, premise (2) is false, so the argument is unsound no matter what. In defense of the claim that premise (2) is false, the objector might argue that what is required for praiseworthiness is responsibility, and a sufficient condition for responsibility— even assuming libertarianism—is that the agent be the uncaused cause of their actions—the power to choose or do otherwise is unnecessary.17 This argument might work just fine were it not for the fact that classical theism demands not just praiseworthiness but maximal praiseworthiness. For the classical theist, it simply will not do to praise God while thinking correctly that She could have been more worthy of that praise. Maximal praiseworthiness, however, requires the ability to choose otherwise. To appreciate the plausibility of this position, consider this case. Suppose I discover that a powerful supernatural agent named Wilma made a benevolent choice to benefit me in some way. Suppose further that Wilma is the uncaused cause of her choice even though her nature (perhaps she is essentially morally perfect) entails that she could not have chosen otherwise. Then it may be correct for me to hold Wilma responsible for benefitting me and so to praise her for it. Imagine, however, that I also learn that a powerful supernatural agent named Betty knowingly benefitted me. Imagine further that, unlike Wilma, Betty is free to choose and do otherwise and yet she deliberately (as opposed to arbitrarily or randomly) chooses to benefit me. Plausibly, if Wilma and Betty benefitted me equally and all other factors are held equal, then Betty would have greater responsibility and so be more deserving of my praise than Wilma precisely because she could have done otherwise and yet still deliberately chose to benefit me. Thus, even if

17 For a defense of this position, see Michael Bergmann and J.  A.  Cover, “Divine Responsibility without Divine Freedom,” Faith and Philosophy 23.4 (2006): 381–408.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

What if God Makes Hard Choices?  29 praiseworthiness does not require the ability to do otherwise, it is at least plausible that maximal praiseworthiness does. One might object that if Wilma’s choice flows naturally and inevitably from her character, then that is itself praiseworthy and thus it is far from clear that Betty is overall more praiseworthy than Wilma. This objection, however, conflates two senses of “praiseworthy.” An action or quality can be worthy of praise simply because it has positive value that is appropriately recognized. (Think of praising someone for good looks that they were born with.) That is different, however, from being worthy of praise in the sense of deserving it. Betty is more praiseworthy in this second sense and the God of classical theism must be maximally praiseworthy in both senses. Further, even if I am wrong about this and the objection succeeds, the argument can be reformulated so that it appeals to thank-worthiness instead of praiseworthiness. Clearly Betty is more worthy of my gratitude than Wilma, even if Wilma is no less or even more worthy of praise. A plausible justification for my intuitions about this thought experiment is that there are multiple sorts of control that an agent can have over how the world turns out, one of which involves being the ultimate source of some state of affairs and the other of which involves the power to settle which of a number of alternative possibilities is actual. Each of these generates a distinctive kind of responsibility and thus maximal praiseworthiness (in the appropriate sense), or at least maximal thank-worthiness, requires both. Further, while it is true that a being could have the power to settle which of a number of alternative possibilities is actual without having significant freedom, possessing significant freedom in addition to freedom that is not significant increases the second sort of responsibility and thus increases praiseworthiness and thank-worthiness. If this is right or at least plausible, then the fact that my model, unlike Swinburne’s, allows for God to possess morally significant freedom of the will is a significant theological advantage.

5.  A Disadvantage The main disadvantage of my model (from the perspective of classical theists) is that it exacerbates the problem of evil and, more generally, makes the project of natural theology even more difficult than it already is. On Swinburne’s model, one can plausibly claim (as Swinburne does) that theism, because of its simplicity, is the most plausible form of supernaturalism,

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

30  Paul Draper arguably (although this is controversial) leaving naturalism as the only serious competitor to theism. My model implies that there are a whole host of supernaturalist alternatives to theism that are just as simple as theism is. These alternatives involve deities that have all the properties God has except they are not morally perfect: when they face hard choices they do not always do what is morally best. One such alternative is aesthetic deism, which posits a deity just like the theistic one except that it is aesthetically perfect instead of morally perfect. This alternative is particularly troubling, because, like naturalism, it explains the pattern of good and evil in the world much better than theism does (as I pointed out earlier, a mixture of good and evil may be required for certain sorts of beauty and drama to exist), but, unlike naturalism, aesthetic deism accounts for data like complexity, cosmic fine-tuning, consciousness, and the extraordinary beauty we perceive in nature just as well or better than theism does. Ignoring the inevitable appeal by connoisseurs of the incredible to private evidence, this all by itself shows that theism is probably false. More generally, my model greatly increases the number of serious competitors to theism that are no less simple and no less capable of accounting for the relevant data. Together, these competitors leave very little probability space for classical theism.18

18  I am grateful to Robert  M.  Adams, Scott Davison, Wes Morriston, John Schellenberg, Larry Temkin, Luke Wilson, and an anonymous OSPR referee for helpful comments on earlier drafts of this chapter, and also to Ruth Chang, whose brilliant philosophical work inspired it. I am also grateful to Philip Swenson and Dean Zimmerman for inviting me to participate in a workshop (sponsored by the John Templeton Foundation) on value theory and the philosophy of religion. Were it not for that invitation, this paper would not have been written (so blame them).

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

3

Theism, Pro-Theism, Hasker, and Gratuitous Evil Klaas J. Kraay

1. Stage-Setting On the vertical axis of the table below are three familiar judgments concerning the claim that God exists: theism affirms it, atheism denies it, and agnosticism, of course, suspends judgment about it. On the horizontal axis are four judgments about the difference that God’s existence would—or does—make to the value of the world and its inhabitants.1 Pro-theism holds that God’s existence would make the world better than it would otherwise be.2 Anti-theism, on the other hand, claims that God’s existence would make the world worse than it would otherwise be. Indifferentism is the position that God’s existence would make the world neither better nor worse, and agnosticism suspends judgment about this axiological matter.3 Each cell in the table represents a unique combination of existential and axiological judgments. Some of these (such as theistic pro-theism or atheistic anti-­ theism, for example) are fairly common combinations. Others are perhaps less familiar, but all of them are, I think, coherent possibilities, whether or not they have (m)any adherents.

1  I say “would or does” in order to be neutral concerning whether theism is true. Hereafter, I will omit the second disjunct for ease of exposition. 2  The terms “pro-theism” and “anti-theism” are due to Guy Kahane  (2011), although my definition here is slightly weaker than his, due to some pressure from Moser (2013). Kahane’s paper sparked the current literature on this topic, which now includes Mawson (2012), Luck and Ellerby (2012), Kraay and Dragos (2013), Kraay (2013), Moser (2013), Davis (2014), Davis and Franks (2015), Penner (2015), Penner and Lougheed (2015), McLean (2015), Dumsday (2016), Mugg (2016), Lougheed (2017), the papers collected in Kraay (2018), Azadegan (forthcoming), Linford and Megill (forthcoming), Lougheed (forthcoming a), Lougheed (forthcoming b), and Penner (forthcoming). 3  These positions are best construed as sharing a commitment to axiological realism—the view that there are objective facts about such axiological matters.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

32  Klaas J. Kraay Axiological judgments

Existential judgments

Pro-Theism Anti-Theism Indifferentism Agnosticism Theism Atheism Agnosticism

The simplicity of this table masks the formidable complexity of the underlying issues. In order to fully assess the positions on the vertical axis, for example, one must be clear on what is (or should be) meant by ‘God,’ and what it means (or should mean) for God to exist. These are notoriously controversial matters. These requirements pertain equally to the horizontal axis. Moreover, in order to have a fruitful debate about the positions on the horizontal axis, plausible accounts must be given of terms like ‘world,’ ‘better,’ and ‘worse,’ and of the relevant counterfactual (or counterpossible4) judgments. These, too, will be highly contentious. Still further, various versions of these axiological judgments can be identified. Kahane (2011), for example, distinguishes between wide and narrow, and between personal and impersonal forms of anti-theism, and these distinctions can also be applied to the other axiological positions.5 So a complete assessment of the axiological issues will also require careful attention to these distinctions.6 In this chapter, I won’t try to provide a comprehensive account of the terrain, nor will I try to defend any particular combination of existential and axiological positions. Instead, I will simply try to make a little bit of headway by considering one particular claim: (1)  If God exists, no gratuitous evil occurs. This claim enjoys widespread assent in contemporary analytic philosophy of religion.7 It is also an important claim for both the axiological and the 4  Among contemporary analytic philosophers of religion, theism is typically thought to be either necessarily true or necessarily false. If this is so, then these axiological positions involve counterpossible judgments. For more on this, see Kahane (2011), Kraay and Dragos (2013), Davis and Franks (2015), and Mugg (2016). 5  For more discussion of this, see Kraay and Dragos (2013). 6  For a more detailed overview of the relevant issues, see my “Invitation to the Axiology of Theism” in Kraay (2018). 7  David O’Connor calls it the “Establishment Position” (1998, 72, 74), and Jeff Jordan dubs it the “Standard Claim” (2003, 236). William Rowe has said that (1) “accords with basic moral principles . . . shared by both theists and nontheists” (1979, 337). Stephen Wykstra, putting the

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Theism, Pro-Theism, Hasker, and Gratuitous Evil  33 existential issues displayed in the table above. For one thing, it’s plausible that (1) could be harnessed into an argument for pro-theism: it certainly looks like a reason for thinking that God’s existence would make the world better than it would otherwise be, at least if there is an appropriate causal connection between the antecedent and the consequent. For another, (1) is also the first premise of a widely discussed argument for atheism that continues as follows: (2)  Gratuitous evil occurs. Therefore, (3)  God does not exist. In recent decades, much of the debate about this argument (and probabilistic variants of it) has concerned the claim expressed in premise (2). Critics of this claim have defended accounts of our epistemic circumstances and capacities according to which this premise cannot reasonably be asserted. This controversial position has become known as skeptical theism, and it has generated a large and very technical literature.8 A few authors, however, have attempted to resist (1). One such strategy is developed in a series of important publications by William Hasker (1992, 2004b,  2008).9 If Hasker’s argument were to succeed, this would be an important result for the overall debate about whether God exists, since it would count against a prominent kind of argument for atheism. While Hasker does not explicitly consider the question of the axiological consequences of God’s existence, his argument, if sound, would also be an important result for this debate, since it would count against one line of apparent support for pro-theism.10

point more strongly, has said that (1) is “a basic conceptual truth deserving assent by theists and nontheists alike (1984, 76). In more recent papers, Rowe has even deemed it a necessary truth (1996, 284), and has said that to deny it is “radical, if not revolutionary” (1991, 79). (Note, however, that there are some differences in how these authors understand the term ‘gratuitous evil.’) 8  For recent surveys of this vast terrain, see McBrayer (2010), Dougherty (2011), and especially Dougherty (2014). 9  Hasker also defends his view against critics in three other papers (Hasker  1995,  1997, and 2010a). 10  Indeed, as we will see, Hasker thinks that things would be rather worse in certain important respects (and perhaps even overall) if God were to prevent all gratuitous evil. So Hasker would presumably think that the apparent support (1) offers for pro-theism is, on careful reflection, illusory.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

34  Klaas J. Kraay In Section 2, I discuss certain restrictions on God’s permission of evil in order to illuminate claim (1), and in Section 3, I set out Hasker’s case against it. In Section  4, I clarify an important point about Hasker’s argument: I show that it does not require Hasker to maintain that God’s plan for creation requires the actual occurrence of gratuitous evil. In Section 5, I set out and evaluate four criticisms of Hasker’s argument. Finally, in an Appendix, I  consider the merits of a successor argument for atheism—one that is compatible with Hasker’s view.

2.  Restrictions on God’s Permission of Evil Evidently, (1) intends to express a moral restriction on God’s permission of evil. Before setting out and evaluating Hasker’s argument, it will be helpful to clarify just what sort of restriction Hasker has in mind, and to distinguish it from other restrictions that have been, or might be, proposed. There are many definitions of ‘gratuitous evil’ in the literature. In what follows, I will work with Hasker’s most recent definition, which says that a token or type of evil is gratuitous if and only if God, if he exists, antecedently knows it to be certain or extremely probable that he could prevent it in a way that would make the world overall better (2010a, 308).11 It is important to see that, given this definition, (1) expresses a very general constraint on the divine permission of evil. For contrast, here is a more specific constraint: (1*)  If God exists, then for any human S, God permits S to suffer evil e only if S ultimately benefits from suffering e, and this benefit defeats12 the harm involved in S’s suffering e. The requirements expressed by (1) and (1*) are logically independent: God might satisfy one but not the other. In one respect, (1) expresses a stricter requirement than (1*), since it ranges over animal suffering as well as human suffering. In another respect, (1*) is stricter than (1), since it requires that the human sufferer receive an evil-defeating benefit. Conjoining the

11  I have replaced “could antecedently know” in Hasker’s formulation with “antecedently knows,” since if God could know p, God knows p. Hasker here means to refine a definition of gratuitous evil offered by Rhoda (2010, 287–9), which Rhoda, in turn, takes to be an improvement over William Rowe’s (1979) conception. Other critics of Rowe’s account of gratuitousness include William Alston (1991, 33–4) and van Inwagen (1991, 164, n. 11). Rhoda’s definition appears to be inspired by certain remarks by van Inwagen (2001, 69; 2006, 97). 12  In the sense explained by Chisholm (1968).

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Theism, Pro-Theism, Hasker, and Gratuitous Evil  35 requirements expressed by (1) and (1*) would, of course, yield a stricter requirement than either one expresses individually. Restrictions on God’s permission of evil that require the sufferer to ­benefit in some way from the evil are sometimes called patient-centered restrictions, and the view that some such restriction is needed to secure God’s goodness has been called theodical individualism.13 Many patient-centered restrictions, varying widely in strength, have been proposed and defended by ­philosophers. In a few places, Hasker explicitly criticizes certain patientcentered restrictions.14 But he also clearly means to target the more general, non-patient-centered requirement expressed by (1).15 In what follows, I will confine myself to assessing Hasker’s argument against the latter.16 Before turning to Hasker’s argument, here is one more observation concerning restrictions on God’s permission of evil. It seems that the existential and axiological debates are connected in the following interesting way: The stricter the requirements on God’s permission of evil are held to be, the more plausible pro-theism becomes, but, equally, the more theism becomes vulnerable to arguments from evil that assert that these requirements are (probably) not satisfied. Consider, for example, this very strict requirement: (1**)  If God exists, then (i) there is no gratuitous evil; and (ii) for any human S, God permits S to suffer evil e only if S ultimately benefits from e in a way that defeats the harm involved in S’s suffering e; and (iii) this benefit renders S’s life significantly overall better than it would otherwise be; and (iv) this benefit accrues to S within ten minutes of e’s occurrence; and (v) and all those who witness e, including S, are conscious of this ­benefit and of exactly how it results from e. Clearly, the truth of (1**) would support pro-theism (given an appropriate causal connection between the antecedent and consequent), but, equally clearly, this claim could also be harnessed into an argument for atheism, since it is obvious that the consequent is false.17 13  This term is due to Jeff Jordan (2004). While I’m mentioning terminology, I should note that the requirement expressed by (1) is termed meticulous providence by Peterson (1982). 14  See, for example, Hasker (1992, 27–9; 2008, 89–191). 15  This is clear in the following places: Hasker (1992, 29–30; 2004b; and 2008, 191). 16  Jordan (2004) offers an argument very similar to Hasker’s, but focused on theodical individualism. Other papers that discuss theodical individualism include Maitzen (2009,  2010), Gellman (2010), Mawson (2011), and Crummett (2017). 17  If this is right, then the theistic pro-theist needs to strike a careful balance: she needs to defend divine restrictions on the permission of evil that are strong enough to support pro-theism, while not being so strong as to support atheism. For more on this relationship between the existential and axiological debates, see Penner and Arbour (2018).

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

36  Klaas J. Kraay

3.  Hasker’s Case Against (1) Hasker does not exactly argue that (1) is false. Instead, he maintains that given a certain account of God’s nature and plan for humanity, (1) can reasonably be resisted.18 In this section, I set out and assess Hasker’s innovative and important argument. Hasker believes that if God were to prevent all gratuitous evil, we would come to know this. He offers two related reasons. First, he claims that one of God’s chief purposes in creating rational beings is to bring them to knowledge of his nature. This knowledge, Hasker supposes, would surely include the proposition that God, due to his essential moral goodness, prevents all gratuitous evil (1992, 39; 1997, 391; 2004b, 92; 2008, 197). Second, Hasker claims that if God were to prevent humans from gaining this knowledge, this would amount to a “pervasive policy of deception” (1992, 37) or a “massive disinformation campaign” (1992, 39). Hasker takes this to be morally objectionable, and hence unworthy of God. Hasker’s argument here turns on what we would come to know about God’s nature and policies, given theism, but surely it could also be expressed—more modestly—in terms what we would come to reasonably believe.19 Hasker then imagines what would happen if we were to come to know (or, let’s hereafter add, reasonably believe) that God prevents all gratuitous evil. In the vast literature on the problem of evil, a distinction is standardly

18  Hasker says that premise (1) “ . . . should be rejected by theists, since it comes into conflict with other, better-entrenched elements of the theistic worldview” (2004b, 81; and see 2008, 189). It’s worth adding that non-theists could also resist (1) on Haskerian grounds. Such individuals would hold that if theism were true, then these “better-entrenched elements of the theistic worldview” would be more plausible than (1), in which case, if (1) conflicts with them, (1) should be resisted. 19  Dustin Crummett has pointed out to me that on the more modest formulation, it will be more difficult for Hasker to argue that moral motivation is undermined if God prevents all gratuitous evil. I agree, but I suspect that Hasker would still think it abundantly clear that moral motivation would be significantly undermined, even if is not quite as clear as on the more ambitious formulation. Moreover, the advantage of the more modest claim, presumably, is that it is more plausible. A related issue, pressed by an anonymous referee, concerns the scope of “we” in the claim that if God prevents all gratuitous evil, we would come to know (or reasonably believe) this. Hasker does not say exactly how many people would gain this knowledge or reasonable belief. I take it that Hasker needn’t hold that all people would—and this, in any case, would be highly implausible. Nor does Hasker need to show that some specific proportion of people would gain this knowledge or reasonable belief. He does, however, need to hold that enough people would, such that the consequences for moral motivation would be sufficiently deleterious for God’s plan for creation. I return to this issue in Section 5.4, below.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Theism, Pro-Theism, Hasker, and Gratuitous Evil  37 drawn between moral evil and natural evil. Roughly, the former is ­wrongdoing perpetrated by moral agents, and the latter is pain and suffering not due to moral agents.20 Hasker considers each in turn. If God were known or reasonably believed to prevent all gratuitous moral evil (hereafter GME), he argues, moral motivation would be undermined. This is because, according to Hasker, “an important part of what leads human beings to attribute great significance to morality is the perception that pointless harm and suffering very often result from morally objectionable behaviour” (2004b, 82). Absent that ‘perception,’ human beings would be rather less likely to deem morality significant, and as a result, their motivation to act morally would be seriously impaired. After all, why refrain from performing or permitting moral evil when you are confident that God will ensure that any moral evil that occurs is non-gratuitous? Hasker next imagines what would happen if God were known or reasonably believed to prevent all gratuitous natural evil (hereafter GNE). Hasker thinks that our motivation to prevent or minimize natural evil would be compromised. Specifically, he claims that our motivation to respond to natural evil by acquiring or developing goods such as “knowledge, prudence, courage, foresight, cooperation, and compassion” would be reduced or eliminated (1992, 38–9; see also 2004b, 88 and 2008, 193). This is because we would be confident that any natural evils that occur would also be non-gratuitous—so why bother trying to minimize or prevent them?21 Hasker believes that God deems it very important for human beings to place a high priority on fulfilling moral obligations, and, in particular, for them to assume major responsibility for the welfare of their fellow human beings (2004b, 82; 2008, 191). Hasker also thinks that God deems it very important that human beings respond to natural evil by acquiring and developing the various goods mentioned above. Let’s say that God’s ‘plan for creation’ involves these priorities. Hasker thinks that God’s preventing all gratuitous evil would compromise his plan for creation, and, accordingly,

20  These definitions may be a bit too broad: Perhaps it’s not the case that all moral wrongdoing counts as evil, and perhaps it’s not the case that all pain and suffering due to natural processes counts as evil. There’s no need to hash this out here, though, since nothing turns on this for Hasker’s purposes, or for mine. 21 So, while it may seem natural to suppose (as I mentioned earlier) that (1) supports ­pro-theism, Hasker evidently disagrees. In fact, he seems to think that, if God were to prevent all gratuitous evil, things would be rather worse than they would otherwise be in certain important respects.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

38  Klaas J. Kraay that God will not prevent all gratuitous evil. Pulling all this together, Hasker’s challenge to (1) can be expressed as follows: (a)  If God prevents all gratuitous evil, this fact is known or reasonably believed by us. (b)  If we know or reasonably believe that God prevents all GME, then our motivation to behave morally is undermined. (c)  If we know or reasonably believe that God prevents all GNE, then our motivation to acquire or develop various goods (g1 . . . gn) in response to natural evil is undermined. Therefore, (d)  If God prevents all gratuitous evil, then then the motivations mentioned in (b) and (c) are undermined.22 (e)  God’s plan for creation includes ensuring that the motivations mentioned in (b) and (c) are not undermined. Therefore, (f)  It’s not the case that God prevents all gratuitous evil.23 After clarifying an important point about this argument in Section  4, I evaluate it in Section 5.

4.  Must Gratuitous Evil Occur, or Must it Merely be Permitted? Hasker’s argument might be thought to suggest that the actual occurrence of gratuitous evil is required for God to bring about his plan for creation. And this, in turn, might seem theologically problematic. After all, someone might say, if God needs gratuitous evil to occur in order to execute his plan for creation, he should just come up with a different and better plan— perhaps even one in which he doesn’t create free moral agents in the first place! Hasker, however, appears to deny that God’s plan requires the actual 22  Presumably Hasker thinks that premises (b) and (c) only apply to creatures: it’s not the case that if God knows that he prevents all gratuitous evil, God’s moral motivations are undermined! 23  In his 1992 paper, Hasker explicitly argues for the weaker conclusion that “It is not the case that God is morally required to prevent gratuitous evil” (30). But this conclusion is too weak for his purposes, since it does not, by itself, count against (1). This is because even if God is not morally required to prevent gratuitous evil, he might do so anyway.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Theism, Pro-Theism, Hasker, and Gratuitous Evil  39 occurrence of gratuitous evil. He says: “Clearly, God might be willing to permit evils, gratuitous or otherwise, without those evils actually occurring” (1992, 41, n. 10; 1995, 423–4).24 In order to clarify his overall argument, and to forestall misconceptions, it will be useful to explore this claim. Let’s begin with GME, and, for the moment, let’s grant premises (a) and (b). If these claims are true, then God’s preventing all GME suffices for the undesirable consequence identified in the consequent of (b): the undermining of moral motivation. So, in order to bring about his plan for creation, God should not prevent all GME. The contradictory of ‘prevent’ is ‘permit,’ so God should permit GME. But God’s permitting GME to occur does not entail that it actually occurs: libertarian-free creatures, for example, might be permitted by God to perform GME, but might nevertheless refrain from so doing. So: God’s preventing all GME suffices to undermine moral motivation, given (a) and (b), but this does not entail that the actual occurrence of GME is necessary for moral motivation not to be undermined. But now suppose there are libertarian-free creatures who in fact manage to refrain from performing any GME. In this case it is they, not God, who prevent all GME. If the prevention of all GME suffices to undermine moral motivation, then it seems we have reached a paradoxical result: these creatures’ moral rectitude undermines moral motivation! That seems very strange, to say the least, and it is presumably a consequence that Hasker should resist. Luckily for Hasker, he has the resources to do so. Premise (a) says that if God prevents all GME, then this fact is known or reasonably believed by us. But Hasker could say that our preventing all GME has no such consequence. Here is the difference between the two cases. As we have seen, Hasker defends (a) by saying that God cannot deceive, and that God wants to bring us to knowledge of his nature, which, by hypothesis, involves preventing all GME. So there is an iron-clad divine guarantee that we would learn about God’s prevention of GME. But, of course, there is no such guarantee for creaturely prevention of GME. If libertarian-free creatures prevent all GME, it needn’t be the case that this fact is known or reasonably believed by them. Why? Well, for example, it’s highly plausible to think that there is at least one world (and probably very many) in which libertarian-free creatures prevent all GME, but in which they fail to learn this fact about themselves.25 24  Hasker also says that the title of his 1992 paper (“The Necessity of Gratuitous Evil”) was “deliberately paradoxical and provocative” (1995, 423–4). He goes on to clarify that “the strictly correct title of this paper would be, ‘The Necessity of the Possibility of Gratuitous Evil’ ” (1992, 41, n. 10, emphasis added). 25  More modestly: If there are any worlds in which libertarian-free creatures prevent all GME, it is likely that in very many of those worlds, most creatures neither know nor reasonably

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

40  Klaas J. Kraay To know or reasonably believe that they have prevented all GME in a given world, its denizens would have to possess vast quantities of data about the moral status of all salient actions and omissions throughout the total history of their world, including enormously many facts about the relevant consequences, motivations, obligations, virtues, vices, and much else besides.26 It’s hard to believe that any individual creature could acquire all the relevant information, let alone that most or all would, in many or all of the relevant worlds. So, if it’s the case that there is at least one world (and probably very many) in which libertarian-free creatures prevent all GME, but fail to learn this fact about themselves, then Hasker can consistently maintain that God’s prevention of GME would undermine moral motivation, but that our prevention of GME would not do so. Let’s now turn to GNE, and grant premises (a) and (c). Given these claims, God’s preventing all GNE suffices for the undesirable consequence identified in the consequent of (c): the reduction or elimination of our motivation to acquire and develop goods (g1 . . . gn). So, in order to bring about his plan for creation, God should not prevent all GNE. The contradictory of ‘prevent’ is ‘permit,’ so God should permit GNE. In the case of GME, as we just saw, Hasker can claim that God’s permission of GME does not entail its occurrence. But it might seem that there is no comparable way to open up space between the divine permission of GNE and its occurrence, in which case if God permits GNE, GNE occurs. Here are two ways Hasker might reply. First, he could subsume natural evil under moral evil, perhaps by following Alvin Plantinga, who famously speculated that natural disasters and the like are really due to the misuse of libertarian freedom by Satan and his cohorts.27 Second, Hasker could follow

believe this fact about their world. Incidentally, there being no gratuitous moral evil in such a world does not entail that all its denizens are morally impeccable: they may well perform lots of morally evil actions, just not gratuitous ones. So my claim here needn’t bother friends of the thesis that Plantinga (1974) terms trans-world depravity. 26  Different moral theories will have different views about what exactly needs to be known or reasonably believed. But, presumably, to know or reasonably believe that their world lacks GME, its denizens would have to be pretty confident about which moral theory is correct. It’s easy to imagine a world—indeed, lots of worlds—in which there is no widespread confidence about which moral theory is correct. And this provides further reason for thinking that there is at least one world (and probably very many) in which moral agents bring about no GME, but fail to learn this fact about themselves and their world. 27 Plantinga thinks it possible that “ . . . Satan rebelled against God and has since been wreaking whatever havoc he can. The result is natural evil. So the natural evil we find is due to free actions of non-human spirits” (1974, 192).

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Theism, Pro-Theism, Hasker, and Gratuitous Evil  41 Peter van Inwagen in holding that God would allow chance to affect the relevant parts of creation such that GNE might (but needn’t) ensue.28 Presumably, these suggestions need to be more than mere logical possibilities for Hasker to employ them. Let’s grant that one of these stories is indeed plausible enough for Hasker to adopt it without compromising his overall story.29 If so, then God can be said to permit GNE without guaranteeing its occurrence.30 The upshot of this section is that Hasker can reasonably insist that God’s plan for creation does not require the actual occurrence of either GME or GNE, and so Hasker can thereby avoid whatever theological drawbacks there might be to denying this. Now, whether or not the actual occurrence of gratuitous evil is part of God’s plan, Hasker is entirely convinced that it really does occur (1992, 2004a, 2010b). And, as we saw in Section 3, Hasker aims to show that this fact in no way counts against theism. I now turn to some criticisms of Hasker’s argument.

5.  Criticisms of Hasker’s Argument 5.1  The strategy is self-defeating Hasker’s strategy can appear paradoxical, even self-defeating. If God must permit gratuitous evil in order to prevent various motivations from being undermined, then it seems as though these evils are no longer gratuitous: they are permitted precisely for the sake of securing these outweighing goods!31 In reply to this objection, Hasker offers the following analogy (2004b, 89; 2008, 195). Suppose that a musical performance is rewarded with enthusiastic and sustained applause from the audience. Any individual audience member’s clapping could surely be prevented without compromising the 28  Van Inwagen identifies two relevant sources of chance in the created order. The first is “natural indeterminism” and the second is “the initial state of things” (1988, 54–60). 29  In order to be sufficiently plausible for Hasker’s purposes, it would have to be reasonable to think (in the case of the first story) that God could be justified in permitting Satan to misuse his free will, and (in the case of the second story) that God could be justified in permitting chance to play such a role in creation. 30  One might object that God would foreknow that Satan and his cohorts would misuse their freedom, so as to bring about natural evil, and, equally, that God would foreknow that chancy or indeterministic processes would issue in natural evil. Hasker, of course, needn’t be troubled by this objection, given that he is a prominent defender of Open Theism, a view which denies that God has just this sort of foreknowledge (see, for example, Hasker 1989). 31  Rowe (1991), Chrzan (1994), and Howard-Snyder and Howard-Snyder (1999) all press this criticism against Hasker.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

42  Klaas J. Kraay phenomenon of applause. But if someone were to prevent every member of the audience from clapping, there would be no applause at all. Equally, Hasker says, while God could surely prevent any individual instance of gratuitous evil without compromising his plan for creation, God could not prevent every such instance without compromising his plan. So each instance of such evil is, after all, gratuitous—but the class of permitted gratuitous evils serves an important function in God’s plan for creation, and so it is not gratuitous.32 This reply seems decisive, and so I turn to another objection.

5.2  The account of God’s plan for creation in (e) is mistaken It is important to see that the theist who wishes to wield Hasker’s strategy against arguments from gratuitous evil needn’t claim to know that (e) is true: justified belief will do. A theist who justifiably believes (e), along with the other premises in Hasker’s argument, would be justified in believing its conclusion, and hence would be justified in resisting (1). Now, something like (e) is in fact generally accepted by theists, and it seems rather churlish to suppose that no theist can ever be justified in believing it. Moreover, nontheists can surely also be reasonable in believing that, given the truth of theism, (e) is plausible. All else equal, such individuals could then be justified in using Hasker’s strategy (mutatis mutandis) for resisting (1). In the absence of a compelling reason to believe that nobody can be justified in asserting (e), I set this objection aside.33

5.3  If God were to prevent all gratuitous evil, we would neither know nor reasonably believe that God had done so This objection attacks premise (a). Here is William Rowe’s expression of it: If God really is in the paradoxical corner Hasker thinks He is in, then clearly the best course is not to make his presence and policy so decisively known that his very purposes for human life are undermined, if not 32  This way of putting the point is due to Rowe (1988), and Hasker quotes it approvingly in Hasker (1992, 33; 2004b, 89; and 2008, 195). 33  That said, this objection could be pressed in a specific way against a particular individual. If, for example, S wants to use Hasker’s strategy against some argument from gratuitous evil, and if someone can show that S is herself not justified in asserting (e), then this would presumably preclude her from reasonably employing Hasker’s strategy on that occasion.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Theism, Pro-Theism, Hasker, and Gratuitous Evil  43 defeated . . . Surely, other things being equal, God will judge that our lack of decisive knowledge of His presence and policy is better than letting horrendous, gratuitous evils abound in the world.  (1991, 85)

Hasker might reply by flatly denying that it is metaphysically possible for God to deceive us or to prevent us from coming to know that he prevents all gratuitous evil. But this seems implausible. Perhaps, as many philosophers have held, it is metaphysically impossible for God to do moral wrong, but it is not clear that God’s deception concerning (or shrouding of) this particular policy would be morally wrong—particularly if God does so precisely to secure the greater good of preventing the dire consequences mentioned in (b) and (c).34 If this is so, then Hasker needs some other reason to think that it is metaphysically impossible for God to deceive us or shroud his policy from us. None seems on offer. Alternatively, Hasker might reply by contesting Rowe’s axiological intuitions, perhaps—ambitiously—by claiming that divine deception or shrouding really would be worse than God’s preventing all gratuitous evil, or—modestly—by claiming that Rowe has not shown otherwise. It is difficult to know how to assess either reply. To do so properly, we would have to hold before our minds all the relevant states of affairs in which God prevents all gratuitous evil but ensures that we never discover this, thereby blocking us from knowledge of an important aspect of his nature, and we would then have to, somehow, compare these with all the relevant states of affairs in which God permits some gratuitous evil. This is a daunting prospect, to say the least. Luckily for Hasker, there is a simpler reply he can make—one that Rowe himself anticipates (1991, 86). Hasker can point out that if his critic means to endorse the claim that, if God exists, God would ensure that we never discover his policy of preventing all gratuitous evil, this critic can no longer reasonably assert (1). To defend (1) by asserting something inconsistent with reasonably believing (1) is hardly sound philosophy: it’s dialectical suicide. And so I set this objection aside as well.

5.4  Our moral motivations would not be undermined This objection targets Hasker’s (b), or his (c), or both. Let’s suppose that God prevents all gratuitous evil, and let’s further suppose, with Hasker, that 34  Gelinas (2009, 573, n. 17) argues in a similar vein.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

44  Klaas J. Kraay God would ensure that we come to learn this important fact about his nature. One way to press this objection is to hold that God could—and would—simply block us from making whatever inferences are subsequently needed to undermine the motivations mentioned in (b) and (c).35 By doing this, God would ensure that his plan for creation is not compromised. How might Hasker respond? First, he might insist that God’s interference in our ability to draw the relevant inferences constitutes morally objectionable mental meddling. But if, as Hasker supposes, it is a very important part of God’s plan for creation that the relevant motivations not be undermined, his critic could surely retort that God’s interference is here justified for the sake of executing that plan. Second, Hasker could say that God’s interference would compromise some other very important divine goals, perhaps ones involving the unfettered (or at least non-supernaturally-fettered!) use of our intellectual capacities. To make this work, Hasker would have to offer a plausible account of these goals, and of the ways in which they would be compromised by God’s interference, and he would have to claim that it would be overall better for God to achieve these goals than for him to intervene in this way to keep our moral motivations intact. This would be a highly complex response involving rather controversial large-scale value judgments. Here is another way that Hasker might respond. In several places, Hasker points out that, lamentably, some theists have indeed allowed their belief that God prevents all gratuitous evil to undermine their moral motivation, and that this has led to very deleterious results (1992, 39; 1997, 392; 2004b, 87–9; 2008, 193–5). I think that Hasker is quite right that there have been, and are, such individuals. There being some such people does not, of course, establish (b) or (c), but can it defeat the version of the objection to these claims presently under consideration? Unfortunately, no. The existence of such people counts against the idea at issue (namely, that God would block us from forming the motivation-undermining inferences) only on the assumption that theism is true—and this assumption is, of course, illicit in the context of a debate about an argument for atheism.36

35  Interestingly, when considering how it can be that many theists believe that God prevents all GME without having their moral motivation undermined, Hasker speculates that “Perhaps the Holy Spirit is actively at work in preventing God’s people from carrying out in practice the implications of their mistaken beliefs” (1997, 392). The objection I am considering is importantly different from what Hasker suggests in this passage. Instead of blocking people from performing morally wrong actions, God would be blocking people from making certain inferences. 36  I thank Philip Swenson and Dean Zimmerman for helping me to see this.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Theism, Pro-Theism, Hasker, and Gratuitous Evil  45 Here is another way to object to (b) or (c). Several authors have held that the motivation to follow the dictates of a deontological moral system would not be compromised by the knowledge or reasonable belief that God prevents all gratuitous evil.37 This challenge can be made vivid by appeal to actual theists, many of whom hold both that God prevents all gratuitous evil and that they are indeed subject to a deontological moral system. In response, Hasker has granted that there are such theists, and has speculated that they simply fail to see the tension between these parts of their worldview (1997, 392). This response is unsatisfactory. Hasker’s model predicts that theists who reasonably believe that God prevents all gratuitous evil will find their moral motivation undermined. I have granted that some actual theists exemplify this. But when Hasker’s critics point to counterexamples, Hasker in effect dismisses them by rendering an a priori diagnosis of subconscious cognitive dissonance. But whether such individuals’ worldview is incoherent is simply beside the point. The existence of theists who simultaneously believe that God really prevents all gratuitous evil and that they are really subject to a deontological moral system (and whose moral motivation is really not undermined) does count against Hasker’s (b) or (c), whether or not their worldview is consistent. Here is a different response that Hasker might offer. He might say that the consequents of (b) and (c) should not be understood to assert that all or even most people would find their relevant motivations undermined.38 If Hasker can show, more modestly, that enough people would find their moral motivation undermined (to a sufficient degree, perhaps), such as to compromise God’s plan for creation, this would inoculate his argument against these counterexamples. At this point, of course, the burden of proof shifts back to Hasker. To make good this response, he would have to defend some account of (at least roughly) how many people’s moral motivation needs to be undermined in order for God’s plan for creation to be compromised, and to what degree. Then he would have to show that the reasonable belief that God prevents all gratuitous evil really would lead this many people’s moral motivation to be undermined to this degree. Both steps seem rather 37  See Keller (2007, 13), Rhoda (2010, 291), and Himma (2011, 132). There are other ways to argue that moral motivation would not be compromised. For example, Jada Twedt Strabbing has suggested to me that one might argue that theists could anchor their moral motivation in their love for God and their desire to love and desire what God loves and desires. Space does not permit an examination of this suggestion. 38  Note that Hasker sets aside as irrelevant those individuals who “through lack of intelligence or opportunity, or simply because of sloth” fail to draw the relevant morality-undermining inferences (1992, 44, n. 28).

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

46  Klaas J. Kraay daunting, to say the least. The former requires a more detailed account and defense of God’s plan for creation than Hasker has given to date, and of course, such accounts are notoriously controversial. As for the latter step, this seems like a matter for experimental philosophy, and I suspect that it will be very difficult to formulate and execute an experiment to properly test Hasker’s hypothesis. Moreover, even if it were to turn out that enough people would find their relevant motivations sufficiently undermined after reflecting on (1), this would surely raise some new and vexing questions. If God cares about our moral motivations so much, why did he make them (or allow them to become) so easy to undermine? Why didn’t he simply strengthen our moral motivation, rather than permitting gratuitous evil?39

6. Conclusion I have argued that Hasker’s argument is not threatened by the first three objections in the previous section. The fourth objection, on the other hand, can be developed in ways that place considerable pressure on Hasker to give a rather more robust defense of (b) and (c) than he has offered to date. Perhaps this can be done. If not, however, there is one further strategy open to Hasker. In most of his writing on this topic, Hasker has generally concentrated on the psychological question of whether various motivations would be undermined by our knowledge (or, I have added, reasonable belief) that God prevents all gratuitous evil. In one place, however, he briefly claims that if (1) is true, morality itself would be undermined (1992, 29–30). Hasker’s idea is that on either consequentialist or deontological moral systems, principles prohibiting the infliction of harm would be undermined by (1), rendering these systems philosophically inexplicable or incoherent.40 Since this kind of undermining is philosophical, rather than psychological, Hasker could develop this version of his argument and thereby entirely sidestep the worries raised in the previous section about his claims concerning the undermining of various motivations. This is not the place for a complete evaluation of how such an argument could go, so I will simply report my 39  Dustin Crummett suggested this to me. 40 Rowe (1991, 82–3) briefly engages this argument. He expresses some sympathy for Hasker’s claims concerning consequentialist moral systems, but objects to Hasker’s claim that deontological moral systems would be philosophically compromised by (1). Note also that Hasker seems to retreat a bit from this idea in his 2008, p. 192. n. 23.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Theism, Pro-Theism, Hasker, and Gratuitous Evil  47 sense that it has considerable promise.41 But since Hasker has said so little about it to date, and since it is bound to be highly controversial, rather more needs to be said about how, exactly, the truth of (1) would undermine these moral systems. My overall conclusion, then, is somewhat inconclusive. It seems to me that Hasker has developed a very important challenge to (1), but that more defense of his premises (b) and (c) is needed. Perhaps Hasker can provide this. If not, as I have just sketched, there is another—and more promising— way for him to challenge (1). The success of either argument would be an important result for the debate about the axiological consequences of God’s existence, insofar as it would undermine one line of apparent support for pro-theism. The success of either argument would also be an important result for the debate about whether God exists, insofar as it would seriously compromise arguments from evil that invoke (1).

Appendix: The Argument From Excessive Gratuitous Evil Hasker claims that if his criticism of (1) is plausible, “the evidential problem of evil” can be rejected (2004b, 91; 2008, 197). But this is too bold. For one thing, as noted, the only people who can justifiably resist (1) are those who reasonably believe that God’s goals for creation are (or would be) as Hasker describes. Moreover, and more importantly, (1)–(3) is certainly not the only “evidential” argument from evil to atheism. Someone who is persuaded by Hasker to reject (1), but who nevertheless wishes to argue from evil to atheism, might well be tempted by the following argument: (4)  If God exists, it is false that gratuitous evil occurs far in excess of what must be permitted by God in order to achieve his goals. (5)  Gratuitous evil occurs far in excess of what must be permitted by God in order to achieve his goals. Therefore, (6)  God does not exist. Rowe, as it happens, briefly suggests this move (1991, 88, n. 20), as have several other philosophers, although none develop it in detail.42 41 Crummett  (2017) defends Hasker’s claim that the truth of certain patient-centered restrictions on the permission of evil would undermine consequentialist and deontological moral systems. I think he is right about this, but my focus here is on the more general, nonpatient centered requirement expressed by (1). 42  See Keller (1989, 163–6; 2007, 14–16); Chrzan (1994, 135); O’Connor (1995, 391; 1998, 69–70); and Reichenbach (2010, 214).

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

48  Klaas J. Kraay What should we make of this argument? Premise (4) is modest—significantly more modest than premise (1) in the original argument. Moreover, it is prima facie plausible, given God’s attributes. It seems that God would be irrational or morally blameworthy (or both) if he were to permit far more gratuitous evil than is needed to achieve his goals. At the very least, it’s difficult to see how such a being could be essentially unsurpassable in rationality and goodness. Furthermore, (4) is evidently invulnerable to Hasker’s original objection. Perhaps there are ways to attack it, but I suspect that those who wish to resist this argument will concentrate their fire on premise (5) instead. Hasker ventures a few remarks suggesting that he would resist (5). Hasker appears committed to the view that one can justifiably assert (5) only if one is confident both that (i) there is a particular amount of gratuitous evil that God needs to permit in order to achieve his goals, and that (ii) we can identify this amount fairly accurately (1992, 33–6, 43; 1995, 424–5). But he deems it very implausible that “ . . . there is some particular amount of the kind of evil in question, such that if God permits that amount of such evil to exist morality is maintained, but if he permits any less then morality is undermined” (1992, 33).43 Moreover, Hasker thinks that even if there were such a particular amount, it is very doubtful that we would be able to identify it accurately. So Hasker seems to think that (5) cannot reasonably be asserted. I disagree about what is required to reasonably assert (5). One of Hasker’s early interlocutors offered the following analogy: One can reasonably judge that certain amounts of mashed potatoes are far too much to serve at a dinner for four, without being able to judge what is precisely the right amount.44 This seems exactly right, and I would add that we could reasonably believe that some amounts of mashed potatoes are too much even if there is no such thing as the exact right amount.45 In short, this analogy shows that one can reasonably believe that some amount is ‘too much,’ even while failing to know exactly how much is ‘just right,’ and even if there is no amount that is ‘just right.’ An analogy in the moral domain may help to drive the point home. Suppose, as seems plausible, that parents can sometimes be morally justified in permitting their children to experience some preventable pain, on the grounds that they will benefit more than they will be harmed by the pain. (Consider, for example, teaching your

43  Here Hasker refers only to GME, but his point could also be expressed, mutatis mutandis, with respect to GNE. Keller (1989, 163–4; 2007, 13) also expresses doubts about (i). 44 Hasker cites John Glenn as offering this example in a 1990 conference commentary. Hasker rejects the analogy as inapt, by claiming that the amount of gratuitous evil God needs to achieve his goals is “fairly sharply” and “fairly clearly” defined, in contrast to the amount of potatoes needed to serve four (Hasker 1992, 43, n. 27). 45  Chrzan (1994, 135) offers two further analogies, both of which, as Hasker (1995, 425) rightly points out, are inapt, since they turn on (i) there being an exact right amount and (ii) this amount being easy to discover.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Theism, Pro-Theism, Hasker, and Gratuitous Evil  49 daughter to ride a bicycle: The expected benefit of mastering this important skill is generally thought to outweigh the harm incurred by the entirely predictable, entirely preventable skinned knees.) It also seems very plausible to hold, however, that parents would not be morally justified in permitting far more preventable pain than is necessary to attain the relevant benefit. And surely this is so even if parents cannot know how much pain is ‘just right’ in this context, and even if there is no such amount. So, whether or not Hasker is right to deem (i) and (ii) implausible, if these analogies are apt (as I think they are), they show that one can assert (5) without being committed to these claims. Accordingly, Hasker cannot resist (5) in this way. The most obvious—and, I think, most plausible—way to resist (5) is to invoke some form of skeptical theism. In the contemporary literature on the problem of evil, it is a matter of considerable controversy whether the inscrutability of any instance of evil provides good reason to think that the evil in question is gratuitous. As noted earlier, skeptical theists typically reject (2) in the original argument from evil by alleging that the inscrutability of some evil fails to constitute adequate evidence for its gratuitousness. But (5) is rather more ambitious than (2): It alleges not just that (probably) some particular instance of evil is gratuitous, but that (probably) enormously many instances of evil are gratuitous. So, friends of skeptical theism will think themselves in a stronger position to resist (5). But here’s the rub: Hasker cannot himself employ this strategy, since he is a fierce critic of skeptical theism. Hasker thinks it highly plausible that a great deal of gratuitous evil in fact occurs, contra what the skeptical theist asserts (2004a, 53–4; 2010b, 18–19). Hasker thinks that skeptical theism ineluctably leads to an untenable skepticism about induction (2010b, 19–21), and to a repellent moral skepticism (2004a, 51–2; 2010b, 21–7), and that it courts incoherence (2010b, 27–9). Now is not the time to rehearse and evaluate Hasker’s arguments for these conclusions; the mere fact that he draws them suffices to show that he, at least, cannot rely on skeptical theism to resist (5).46, 47

46  Here is a more modest way of putting the point: if Hasker wishes to resist (5) by appeal to skeptical considerations, he will need to take care to not say anything inconsistent with his own criticisms of skeptical theism. 47  Remote ancestors of this paper were presented at the University of Glasgow’s Philosophy of Religion Seminar (May 24, 2012), the Midwest Division Meeting of the Society of Christian Philosophers (March 27, 2014), Ryerson University’s 7th Annual Philosophy Symposium (May 14, 2014), and at the Canadian Society of Christian Philosophers Annual Conference (May 26, 2014). A more recent ancestor was presented at the Rutgers Workshop on Value Theory and Philosophy of Religion (August 12, 2015). I am grateful to Dean Zimmerman and Philip Swenson for inviting me to the Rutgers workshop and to the audiences at all these events for their feedback. Special thanks to Kirk Lougheed, Chris Dragos, and Dustin Crummett for very helpful discussions of these issues, and to an anonymous OSPR referee for helpful comments. Finally, I am grateful to the John Templeton Foundation for supporting my research in 2011–12, and again in 2013–15, and for sponsoring the aforementioned Rutgers Workshop.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

50  Klaas J. Kraay

References Alston, W. (1991). “The Inductive Argument from Evil and the Human Cognitive Condition.” Philosophical Perspectives 5: 29–67. Azadegan, E. (forthcoming). “Antitheism and Gratuitous Evil.” The Heythrop Journal. Chisholm, R. (1968). “The Defeat of Good and Evil.” Proceedings of the American Philosophical Association 42: 21–38. Chrzan, K. (1994). “Necessary Gratuitous Evil: An Oxymoron Revisited.” Faith and Philosophy 11: 134–7. Crummett, D. (2017). “Sufferer-Centred Requirements on Theodicy and AllThings-Considered Harms.” Oxford Studies in Philosophy of Religion 8: 71–95. Davis, R. and Franks, W. P. (2015). “Counterpossibles and the ‘Terrible’ Divine Command Deity.” Religious Studies 51: 1–19. Davis, S.  T. (2014). “On Preferring that God Not Exist (or that God Exist): A Dialogue.” Faith and Philosophy 31: 143–59. Dougherty, T. (2011). “Recent Work on the Problem of Evil.” Analysis 71: 560–73. Dougherty, T. (2014). “Skeptical Theism.” Stanford Encyclopedia of Philosophy (Winter 2016 Edition), ed. Edward  N.  Zalta. https://plato.stanford.edu/ entries/skeptical-theism/. Dumsday, Travis (2016). “Anti-Theism and the Problem of Divine Hiddenness.” Sophia 55: 179–95. Gelinas, L. (2009). “The Problem of Natural Evil II: Hybrid Replies.” Philosophy Compass 4: 560–74. Gellman, J. (2010). “On God, Suffering, and Theodical Individualism.” European Journal for Philosophy of Religion 2: 187–91. Hasker, W. (1989). God, Time, and Knowledge. Ithaca, NY: Cornell University Press. Hasker, W. (1992). “The Necessity of Gratuitous Evil.” Faith and Philosophy 9: 23–44. Hasker, W. (1995). “Chrzan on Necessary Gratuitous Evil.” Faith and Philosophy 12: 423–5. Hasker, W. (1997). “O’Connor on Gratuitous Natural Evil.” Faith and Philosophy 14: 388–94. Hasker, W. (2004a). “The Sceptical Solution to the Problem of Evil.” In Providence, Evil, and the Openness of God, 43–57. New York: Routledge.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Theism, Pro-Theism, Hasker, and Gratuitous Evil  51 Hasker, W. (2004b). “Can God Permit ‘Just Enough’ Evil?” In Providence, Evil, and The Openness of God. New York: Routledge, 81–94. Hasker, W. (2008). The Triumph of God over Evil: Theodicy for a World of Suffering. Downer’s Grove, IL: InterVarsity Press. Hasker, W. (2010a). “Defining ‘Gratuitous Evil’: A Response to Rhoda.” Religious Studies 46: 303–9. Hasker, W. (2010b). “All Too Skeptical Theism.” International Journal for Philosophy of Religion 68: 15–29. Himma, K. (2011). “Review of Hasker’s The Triumph of God over Evil.” Religious Studies 47: 129–40. Howard-Snyder, D. and Howard-Snyder, F. (1999). “Is Theism Compatible with Gratuitous Evil?” American Philosophical Quarterly 36: 115–30. Jordan, J. (2003). “Evil and van Inwagen.” Faith and Philosophy 20: 236–9. Jordan, J. (2004). “Divine Love and Human Suffering.” International Journal for Philosophy of Religion 56: 169–78. Kahane, G. (2011). “Should We Want God to Exist?” Philosophy and Phenomenological Research 82: 674–96. Keller, J. (1989). “The Problem of Evil and the Attributes of God.” Faith and Philosophy 26: 155–71. Keller, J. (2007). Problems of Evil and the Power of God. Aldershot: Ashgate, Chapter 2. Kraay, K. (2013). “Method and Madness in Contemporary Analytic Philosophy of Religion.” Toronto Journal of Theology 29: 245–64. Kraay, K. (Ed.) (2018). Does God Matter? Essays on the Axiological Consequences of Theism. New York: Routledge. Kraay, K. and Dragos, C. (2013). “On Preferring God’s Non-Existence.” Canadian Journal of Philosophy 43: 153–78. Linford, D. and Megill, J. (forthcoming). “Cognitive Bias, the Axiological Question, and the Epistemic Probability of Theistic Belief.” In Ontology of Theistic Beliefs: Meta-Ontological Perspectives. Ed. Miroslaw Szatkowski. Berlin: de Gruyter. Lougheed, K. (2017). “Anti-Theism and the Objective Meaningful Life Argument.” Dialogue 56: 337–55. Lougheed, K. (forthcoming a). “The Axiological Solution to Divine Hiddenness.” Ratio. Lougheed, K. Ed. (forthcoming b). Five Views on the Axiology of Theism: What Difference Does God Make? (Bloomsbury).

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

52  Klaas J. Kraay Luck, M. and Ellerby, N. (2012). “Should We Want God Not to Exist?” Philo 15: 193–9. McBrayer, J. (2010). “Skeptical Theism.” Philosophy Compass 5: 611–23. McLean, G. R. (2015). “Antipathy to God.” Sophia 54: 13–24. Maitzen, S. (2009). “Ordinary Morality Implies Atheism.” European Journal for Philosophy of Religion 2: 107–26. Maitzen, S. (2010). “On Gellman’s Attempted Rescue.” European Journal for Philosophy of Religion 2: 193–8. Mawson, T. (2011). “Theodical Individualism.” European Journal for Philosophy of Religion 3: 139–59. Mawson, T. (2012). “On Determining How Important it is Whether or Not there is a God.” European Journal for Philosophy of Religion 4: 95–105. Moser, P. (2013). “On the Axiology of Theism: Reply to Klaas J. Kraay.” Toronto Journal of Theology 29: 271–6. Mugg, J. (2016). “The Quietist Challenge to the Axiology of God: A Cognitive Approach to Counterpossibles.” Faith and Philosophy 33: 441–60. O’Connor, D. (1995). “Hasker on Gratuitous Natural Evil.” Faith and Philosophy 12: 380–92. O’Connor, D. (1998). God and Inscrutable Evil: In Defense of Theism and Atheism. Lanham, MD: Rowman and Littlefield, Chapter 3. Penner, M.  A. (2015). “Personal Anti-Theism and the Meaningful Life Argument.” Faith and Philosophy 32: 325–37. Penner, M.  A. (forthcoming). “On the Objective Meaningful Life Argument: A Reply to Lougheed.” Dialogue. Penner, M. A. and Arbour, B. (2018). “Arguments from Evil and Evidence for Pro-Theism.” In Does God Matter? Essays on the Axiological Consequences of Theism. Ed. K. Kraay, 192–202. New York: Routledge. Penner, M. A. and Lougheed, K. (2015). “Pro-Theism and the Added Value of Morally Good Agents.” Philosophia Christi 17: 53–69. Peterson, M. (1982). Evil and the Christian God. Grand Rapids, MI: Baker Book House. Plantinga, A. (1974). The Nature of Necessity. Oxford: Clarendon Press. Reichenbach, B. (2010). “Review of Hasker’s The Triumph of God Over Evil and Keller’s Problems of Evil and the Power of God.” Faith and Philosophy 27: 212–18. Rhoda, A. (2010). “Gratuitous Evil and Divine Providence.” Religious Studies 46: 281–302. Rowe, W. (1979). “The Problem of Evil and Some Varieties of Atheism.” American Philosophical Quarterly 16: 335–41.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Theism, Pro-Theism, Hasker, and Gratuitous Evil  53 Rowe, W. (1988). “Response to Hasker’s ‘The Necessity of Gratuitous Evil.’” Unpublished manuscript, Purdue University, cited extensively in Hasker (1992). Rowe, W. (1991). “Ruminations about Evil.” Philosophical Perspectives 5: 69–88. Rowe, W. (1996). “William Alston on the Problem of Evil.” In The Rationality of Belief and the Plurality of Faiths, ed. T.  Senor, 71–93. Ithaca, NY: Cornell University Press. van Inwagen, P. (1988). “The Place of Chance in a World Sustained by God.” In Divine and Human Action: Essays in the Metaphysics of Theism, ed. T. V. Morris, 211–35. Ithaca, NY: Cornell University Press. [Reprinted in van Inwagen, P.  (1995). God, Knowledge, and Mystery: Essays in Philosophical Theology, 42–65. Ithaca, NY: Cornell University Press.] van Inwagen, P. (1991). “The Problem of Air, the Problem of Evil, the Problem of Silence.” Philosophical Perspectives 5: 135–65. [Reprinted in van Inwagen, P.  (1995). God, Knowledge, and Mystery: Essays in Philosophical Theology, 66–96. Ithaca, NY: Cornell University Press; and in D. Howard-Snyder (Ed.) (1996). The Evidential Argument from Evil, 151–74. Indianapolis: Indiana University Press.] van Inwagen, P. (2001). “The Argument from Particular Horrendous Evils.” American Catholic Philosophical Quarterly 74: 65–80. van Inwagen, P. (2006). The Problem of Evil. Oxford: Oxford University Press. Wykstra, S. (1984). “The Humean Obstacle to Evidential Arguments from Suffering: On Avoiding the Evils of ‘Appearance.’ ” International Journal for Philosophy of Religion 16: 73–93.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

4

God and Gratuitous Evil A Response to Klaas Kraay William Hasker

1. Preliminaries My sincere thanks to Klaas Kraay for his unusually thoughtful and perceptive comments on my necessity-of-gratuitous-evil defense against the evidential problem of evil (the NGE defense).1 He has been extremely thorough in his research, and his interpretations of my views are generally quite ­accurate. I hope to be able to respond in a way that will move the discussion of our topic forward. First, however, there are a few preliminaries that need attending to. I shall say only a little about the axiological debate between “pro-theism and anti-theism” which Kraay uses to frame his discussion. He claims that all of the cells in his table represent coherent possibilities, but I don’t think this is correct.2 In particular, I believe the only coherent possibility for a theist is pro-theism. The theist, in considering the matter, will of course be wondering how it would be if, contrary to what she believes to be the case, God did not exist. It will immediately occur to her that the existence of God is itself an incomparably great good, one which very likely outweighs by a large margin any goods or evils that might obtain in the created universe, and this threatens to reduce the entire discussion to triviality. But suppose she agrees to set this aside and consider just the goodness or badness of the rest of reality, not counting the existence or non-existence of God. It will then occur to her that if there were no God, there wouldn’t be anything else either. So the only way either anti-theism or indifferentism could be the correct answer for her, is if the net value of the created universe is either negative or zero. That, however, is impossible: a good God would not create

1  See this volume, Chapter 3.

2  See Chapter 3, p. 32 for the table.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

God and Gratuitous Evil: A Response to Klaas Kraay  55 a universe with zero or negative value. So the only coherent response for her is to check the “pro-theism” box. For agnostics and atheists, on the other hand, all of the cells may in principle be available. (Though I find it hard to see how indifferentism could be the correct response for anyone.) For them, the answer will depend on three factors: (1) How they assess the way the universe is, or would be, on the assumption that there is no God. (2) How they assess the way the universe would be, or perhaps is, on the assumption that there is a God. (3) What scale of values is used to assess these different possibilities. My suspicion is that there are too many independent variables here for the discussion to arrive at any clear result. Something needs to be said about the dialectical context of our discussion. First, there is the question as to what can be expected from a philosophical argument. Like other philosophers, I am attracted to what may be termed “universally compelling” arguments. That is, arguments which are convincing to anyone who is aware of the generally accessible facts assumed by the argument, is capable of following the argument’s reasoning, and is not so blinded by prejudice that she is unable to accept the argument’s conclusion. I am attracted to such arguments, but, like many other philosophers, I have come to see that arguments of this strength are rarely if ever available. That is to say, they are rarely available for interesting positive conclusions; some refutations may meet the standard. I must then agree with George Mavrodes that many philosophical arguments must be “personrelative”:3 they will be convincing to some persons, and some groups of persons, depending on which premises, and which modes of reasoning, they are antecedently disposed to accept. Sometimes, to be sure, differences of these sorts are themselves subject to reasonable argument, but experience shows that consensus often fails to result. With regard to the problem of evil in particular, it is important that this problem is usually presented in an attempt to show that theism, or some particular variety of theism, is either incoherent, or improbable given the evidence, or in some other way rationally deficient. It follows from that that, in order for the argument to succeed, it must appeal to premises which the adherents of the variety of theism in question either accept, or reasonably ought to accept. And in defending the coherence and rational acceptability of their position, theists are entitled to appeal to any propositions which are integral to, or strongly consonant with, the theistic view in question. 3  See George Mavrodes, Belief in God: A Study in the Epistemology of Religion (New York: Random House, 1970).

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

56  William Hasker The fact that atheists or agnostics may be disinclined to accept certain ­propositions to which the theist appeals is certainly of interest, but it has no obvious bearing on the coherence or reasonableness of the theistic views at issue. (Propositions of this sort may of course be vulnerable to attack from further arguments.) It should also be noted that even “good” arguments—arguments that proceed by acceptable modes of inference from premises one accepts—vary considerably in their force. Some may be completely convincing, but others provide only some lesser degree of support. This may be because the premises are held with less than full conviction, or because the premises confer a degree of reasonableness or plausibility on the conclusion, but fall short of establishing it with certainty. Other things being equal, the more convincing an argument is the better, but we may often have to settle for something less than the ideal. Finally, a word about the argument under discussion. I have decided to term this the “necessity-of-gratuitous-evil” defense, or more briefly, the “NGE defense” against the argument from gratuitous evil.4 In calling it a defense, I mean to say that it seeks to block a certain type of argument for the non-existence of God (roughly, a Rowe-type argument from gratuitous evil), but does this without specifying God’s reason(s) for allowing the evils in question. A theodicy, in contrast, does offer at least a tentative proposal concerning God’s reasons for allowing those evils. Serious distortions can result if a proposed defense is treated as if it were a theodicy: a prime example here is found in the many responses to Plantinga’s free will defense that treat it as a theodicy, in spite of his protestations that it is nothing of the sort.

2.  The NGE Defense In Sections 2 through 5.3 of his paper, Kraay sets out my defense, considers certain objections against it, and provides answers for those objections. As I have noted, his readings of my views are in general quite accurate. In view of this, I will not restate my argument on these points, but will simply endorse (and applaud) Kraay’s formulations of my position. There is one 4  See William Hasker, The Triumph of God Over Evil (Downers Grove, IL: InterVarsity Press, 2008), p. 198.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

God and Gratuitous Evil: A Response to Klaas Kraay  57 small addendum I would make. In Section 4 he considers a situation which I maintain to be possible (though not actual) in which there is no gratuitous moral evil, not because God prevents this, but simply because humans voluntarily refrain from the sorts of actions that would create such evil. He then asks, if God’s preventing gratuitous evil would result in the undermining of our moral motivation, wouldn’t that be the case here also? His reply is that, in the case described, it would be very difficult (in effect, impossible) for human beings to know that there is no gratuitous moral evil. That is correct so far as it goes, but there is a further point to be made. In the case as described there is no gratuitous moral evil, but this is so only because of the voluntary actions of human beings. So when someone is confronted with a moral choice, the only guarantee of the continued absence of gratuitous moral evil is precisely the action to be taken by that person. In this context, there is no reason for moral motivation to be undermined, no reason for the person to feel that “it won’t make any real difference” whether one acts in a morally responsible way or not.

3.  The Problem of Moral Motivation Finally in Section 5.4 Kraay comes to a point at which he thinks my argument needs further support. His questions here concern my claim that if we knew (or reasonably believed) that God prevents all gratuitous evil, our moral motivations would be undermined. (The distinction between moral and natural evil, while important in itself, does not play a role in my argument at this point, so I will ignore it for the time being.) At this point I think it will be helpful for me to quote some of the things I have said in developing this part of my argument; after this we can attend to Kraay’s objections. In setting out the argument, I stated a principle of divine moral intention, as follows: (DMI)  It is an extremely important part of God’s intention for human persons that they should place a high priority on fulfilling moral obligations, and should assume major responsibility for the welfare of their fellow human beings.5

5  “Can God Permit ‘Just Enough’ Evil?” in Providence, Evil, and the Openness of God (London: Routledge, 2004), p. 82.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

58  William Hasker I also stated a principle of harm and the significance of morality: (HSM)  An important part of what leads human beings to attribute great significance to morality is the perception that pointless harm and suffering very often result from morally objectionable behavior.6 I further asserted an offsetting good principle that will hold if God prevents all gratuitous evil: (OG)  Any harm resulting from a morally wrong action will be offset by a “greater good” that God could not have obtained without permitting the evil in question.7 Given these principles, the result follows in straightforward fashion. If an agent knows (or reasonably believes) that God prevents all gratuitous evil, she will infer that any harm resulting from an action of hers will be offset by a greater good. It then follows, by (HSM), that an important part of her motivation to attribute great significance to her acting morally will be undermined. (At least, this will be true in general; there may be exceptional cases where it does not hold.) And this is just what is needed for my argument. At this point Kraay presses his objection: we could “hold that God could—and would—simply block us from making whatever inferences are subsequently needed to undermine the motivations mentioned” in (HSM). This is certainly an interesting suggestion, but it is hardly unproblematic. There is something troubling in the idea that God would, as a matter of policy, systematically inhibit the proper functioning of our cognitive faculties in this way. But the objection does suggest a way in which God could prevent all gratuitous evil without the negative consequences for moral motivation. Stated more formally, the principle of divine inhibition of inference goes like this: (DII)  If God exists, and if the prevention of gratuitous evil threatens to conflict with moral motivation in the way urged by the NGE defense, God would prevent such conflict by preventing people from inferring from his prevention of gratuitous evil that any actions they perform would do no harm overall.

6  “Can God Permit ‘Just Enough’ Evil?” 7  “Can God Permit ‘Just Enough’ Evil?” p. 83.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

God and Gratuitous Evil: A Response to Klaas Kraay  59 Kraay acknowledges that, as I have argued, “lamentably, some theists have indeed allowed their belief that God prevents all gratuitous evil to undermine their moral motivation, and . . .this has led to very deleterious results.” Does this not, then, constitute a counterexample which demonstrates that (DII) does not hold? Kraay replies, “Unfortunately, no. The existence of such people counts against the idea at issue (namely, that God would block us from forming the motivation-undermining inferences) only on the assumption that theism is true—and this assumption is, of course, illicit in the context of a debate about an argument for atheism.” (He credits Philip Swenson and Dean Zimmerman for helping him to see this point.) So far, this is surely correct. But this last reply brings out another rather severe objection to the objection—that is, a reason why a theist who has her wits about her will refuse to accept the principle, (DII), on which the objection is based. Namely this: the principle in question lends itself to a quick and easy disproof of God’s existence! The argument goes like this:   (I)  If God exists, and if the prevention of gratuitous evil threatens to conflict with moral motivation in the way urged by the NGE defense, God would prevent such conflict by preventing people from inferring from his prevention of gratuitous evil that any actions they perform would do no harm overall (DII).   (II)  The prevention of gratuitous evil threatens to conflict with moral motivation. (III)  There are many people who have not been prevented from making the inference in question. (Examples here include me and some other philosophers,8 but more i­mportantly, the persons mentioned (and I believe there are many of them) whose moral motivation has in fact been adversely affected by the belief that in God’s providence “all is for the best.”) Therefore, (IV)  There is no God. This argument, I maintain, constitutes a powerful reason for theists to reject (DII), a principle which has at best only a certain mild plausibility going for it. 8  An interesting example here is Stephen Maitzen. In “Ordinary Morality Implies Atheism” (European Journal for Philosophy of Religion 12 (2009), pp. 17–26), Maitzen takes a principle similar to (1) which he claims is an entailment of theism, points out the conflict between this principle (assuming God exists) and ordinary morality, and draws the conclusion stated in his title.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

60  William Hasker Kraay, however, introduces yet another reason to doubt that our knowing that God prevents all gratuitous evil would undermine moral motivation: Several authors have held that the motivation to follow the dictates of a deontological moral system would not be compromised by the knowledge or reasonable belief that God prevents all gratuitous evil. This challenge can be made vivid by appeal to actual theists, many of whom hold both that God prevents all gratuitous evil and that they are indeed subject to a deontological moral system.

There are indeed such persons. Many of them, I suspect, simply have not thought through the implications of holding that God prevents all gratuitous evil, as urged by the NGE defense. Others are aware of the line of argument in question but have devised spurious responses to it.9 The members of both these groups are in a certain way irrational in their approach to these issues; I think it is unseemly to suppose that the fulfillment of God’s purposes, as specified by (DMI), depends on this sort of irrationality. Finally, there may indeed be some individuals who are fully cognizant of the logic of the ­situation but who still maintain undiminished concern for fulfilling what they perceive as their moral obligations. I suspect, however, that there are relatively few such individuals. I say this, in the light of the strong resistance I have encountered, on the part of otherwise highly rational philosophers, to acknowledging the implications of the NGE defense. Many people seem to feel a strong cognitive dissonance when they consider seriously that acts of mercy they feel they ought to perform will not make the world any better than it would be without those actions. None of this, however, serves to overcome the thrust of the line of reasoning I have proposed. I have no need to maintain that every single individual would be affected in the way postulated by that defense; it is enough that a great many individuals would be so affected, unless this is prevented by the sorts of irrationality we have discussed. And this, I believe, is evident. Finally, let me point out that none of the non-teleological alternatives proposed does well in dealing with the most difficult cases, those involving persons remote in time and space who are not known to us personally. Examples would include future generations, as well as groups of people in

9  See Eleonore Stump, Wandering in Darkness: Narrative and the Problem of Suffering (Oxford: Oxford University Press, 2010), pp. 412–13; for my reply, see “Light in the Darkness? Reflections on Eleonore Stump’s Theodicy,” Faith and Philosophy 28:4 (October 2011), pp. 449–50.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

God and Gratuitous Evil: A Response to Klaas Kraay  61 distant locations who are threatened by natural disasters or disease epidemics. What sense can be made of deontological principles requiring us to help such people, when our doing so will result in equal or greater suffering for some other group of people? Or why suppose that God would especially love these people and desire that they be helped, when so doing would involve equal or greater suffering for some other group of people?10 It seems unlikely that good answers will be forthcoming.11 At this point I would like to quote some things I said earlier concerning the force of my argument: It is freely acknowledged that these arguments do not constitute conclusive disproof of NGE, the principle that God prevents all gratuitous evil,12 even for theists. It is conceivable that, in spite of all that I have said, the incoherence that has been pointed out is insignificant in practice and poses no real difficulty. It is also conceivable that the tension between NGE and DMI is just a fact of life which must be accepted, with DMI having to yield (since NGE is a necessary truth) when the two principles come into conflict. Perhaps, then, those who urge our responsibility in cases involving disease epidemics, famines, and natural disasters in remote parts of the world, and those who insist that we have a responsibility to take thought for the needs of future generations, are just misguided. I must confess, however, that neither of these alternatives strikes me as at all plausible. On the contrary, I believe the conflict is real and important. And it seems evident to me that DMI, the principle of divine moral intention, is far more deeply entrenched in the theistic worldview than is NGE, so that the tension between them is an indication that NGE should be abandoned.13

10  Additional discussion of why this result would follow will be found in the following section. 11  Kraay suggests that the emphasis on moral motivation could be supplemented or replaced by an argument to the effect that “if (1) is true, morality itself would be undermined.” I do indeed suggest an argument to this effect in “The Necessity of Gratuitous Evil” (Faith and Philosophy 9 (1992), pp. 29–30). However, I have come to believe that it will be very difficult for this sort of argument to succeed. (1) does imply that moral requirements cannot be grounded in judgments about the total amount of good or evil in the universe, or the total consequences of particular actions. But other sorts of morality are possible that do not depend on such judgments– for example, a purely deontological morality based on divine commands. So this sort of argument is not pursued in “Can God Permit ‘Just Enough’ Evil?” 12 In this context, “NGE” abbreviates “no gratuitous evil,” whereas in “NGE defense” it abbreviates “necessity of gratuitous evil.” I apologize for the awkwardness. 13  “Can God Permit ‘Just Enough’ Evil?” p. 88f.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

62  William Hasker

4.  The Problem of Excessive Gratuitous Evil In the Appendix to his article, Kraay proposes yet another Rowe-type ­argument, one he thinks is potentially forceful even if the NGE defense as presented thus far is successful. The argument goes like this: (4)  If God exists, it is false that gratuitous evil occurs far in excess of what must be permitted by God in order to achieve his goals. (5)  Gratuitous evil occurs far in excess of what must be permitted by God in order to achieve his goals. Therefore, (6)  God does not exist. This may well be the most formidable challenge to the NGE defense; at any rate, it is a challenge that I have not thus far explicitly addressed. Kraay points out that several philosophers have suggested this as a problem for my argument. However, the issue also arises, perhaps even more naturally, in the context of Peter van Inwagen’s response to Rowe-type arguments. Van Inwagen argues against (1) on grounds of vagueness. In very many cases, he urges, there is no determinate answer to the question, how much evil must be permitted in order to obtain a certain good. An example: A thousand children are suffering from a potentially fatal disease. Medication is available that may lead to a cure; unfortunately, there is not enough for all of the children. Dividing the medicine equally between all insures that all of them will die, so some of the children must be chosen by lottery to receive the medication. But whatever number are given the medicine, it is plausible that by very slightly diminishing the dose for each, one more child could have been given the chance at recovery. So, how many children should be given the medicine?14 It seems that there is no determinate answer to this question; analogously, there is no determinate answer to the question as to how much evil must be permitted by God in order to achieve his goals. But while (1) may thereby be discredited, this leaves open the possibility of an argument, not just from gratuitous evil, but from excessive gratuitous evil.15

14  Peter van Inwagen, The Problem of Evil (Oxford: Oxford University Press, 2006), p. 109. 15  See John Martin Fischer and Neal Tognazzini, “Exploring Evil and Philosophical Failure: A Critical Notice of Peter van Inwagen’s The Problem of Evil,” Faith and Philosophy 24 (2007), pp. 458–74.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

God and Gratuitous Evil: A Response to Klaas Kraay  63 Kraay cites remarks of mine to the effect that “one can justifiably assert (5) only if one is confident both that (i) there is a particular amount of gratuitous evil that God needs to permit in order to achieve his goals, and that (ii) we can identify this amount fairly accurately.” Kraay is right that I said this, and he is also right that I was mistaken to say it. The NGE defense itself does not depend on this assumption, and I have abandoned it in more recent statements of the defense, so this objection to (5) is not open to me.16 How then might I respond to (5)? Kraay believes the most promising response is that of skeptical theism, but he also notes my own intransigent opposition to skeptical theism. This intransigence, however, does not mean that I am unable to resist (5). Consider the following propositions: (5a)  There exist an extremely large number of evils, each of which is such  that God could prevent it without being prevented thereby from achieving his goals. (5b)  There exist an extremely large number of evils, such that God could prevent most or all of them without being prevented thereby from achieving his goals. I do endorse (5a), but it is obvious as a matter of logic that (5a) does not entail (5b), which is what is needed for the argument from excessive gratuitous evil to go through. The difference is illustrated nicely by an example quoted by Kraay, of a large audience which is enthusiastically applauding a musical performance. If any one member of the audience is prevented from applauding, there will still be enthusiastic applause for the performance. But if even half of the audience is kept from applauding, the overall effect will be quite different. (A critic reporting on the occasion might say the audience had a “mixed response” to the music.) This of course is a contrived example; is there any reason to suppose that a parallel situation might exist in real life, so that (5a) might be true but not (5b)? There are at least two such reasons. One of these I will term (forsaking elegance of expression), the “it-has-to-get-really-bad” problem. Anyone who follows current affairs is aware of examples of the following kind: A process 16  But why did I take this line in the first place? I believe my thinking was roughly as follows: According to (1), it is logically necessary that God prevent all gratuitous evil. Logic, however, is an exact science, so if there is a logical requirement of this sort the requirement must be precise rather than vague. This reasoning, however, ignores the freedom of God. God is logically constrained by his own nature to do only what is good and wise, but this does not entail that in any given situation there is exactly one good and wise act that God must perform.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

64  William Hasker is underway that, if allowed to continue, will inevitably produce extremely bad results, causing harm to a great many people. The process could be halted or reversed, but only at considerable cost, and only by taking actions that are perceived as adverse to the interests of an influential group of ­people. What typically happens is a lengthy delay in addressing the problem, even though this delay results in far greater costs, and a worse final result, than if the problem had been addressed sooner. “It has to get really bad,” so that the harmful effects of the process are obvious to everyone, before the needed steps can be taken. The prime example at present is the long-delayed, and still inadequate, response to the threat of global climate change, but additional examples can easily be found. If we assume, plausibly enough, that God’s desire would be that we should address the problem and prevent the catastrophic outcome that is threatening, then this purpose of God’s would actually be frustrated by his preventing the bad consequences of the process as they unfold. Another possible objection to (5) is found in the “too-much-intervention” objection. For an example of this, take cancer. It is a fact of biology that sometimes cells in our bodies run amok, multiplying without regard to their normal functions and causing havoc in various parts of the body. Many cases of cancer undoubtedly fit the definition of gratuitous evil; the pain, suffering, and premature death caused by them far outweigh any possible good results of the process. Most individual cases, then, could presumably be cured by God without endangering any of God’s overall goals for the world and for human life. But to cure all of them, or even a large proportion, would require a very large number of separate divine interventions—probably several millions each year in a country the size of the United States, and many times more than that worldwide. Now it is supposed (plausibly, I think) in some accounts of divine providence that one of God’s purposes is that the natural world should for the most part function naturally, with the various entities it contains operating according to their own inherent natures and capacities—to be sure, conserved in being and energized by the sustaining power of God. Miracles, however, should be relatively rare, as indeed they are portrayed as being in the Bible. So the many millions per year of special divine interventions that would be required to cure the many cases of cancer would represent “too much intervention” to be consistent with the stated purpose. Similar situations may obtain for many of the other evils that oppress us. In the light of these problems, what should we make of (5)? Is it true or false? My answer is that I don’t know. It may be that the “it-has-to-getreally-bad” problem and the “too-much-intervention” problem, together

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

God and Gratuitous Evil: A Response to Klaas Kraay  65 with others we haven’t explored, mean that (5) is false—that there are not in fact a very large number of instances of evil, all of which could be prevented by God without preventing him from obtaining his objectives. Or, it could be that in spite of those problems and others, there is all the same a large excess of gratuitous evils, as claimed by (5). I simply don’t know which of these is the case, and I think it will be hard to make out a compelling argument one way or the other. These reflections on (5) may well leave us dissatisfied. It may be difficult to establish the truth of (5) convincingly, and to that extent the argument from excessive gratuitous evil has been neutralized. On the other hand, (5) certainly has not been disproved, and to that extent the argument remains a threat. This may lead us, then, to give further consideration to (4), the analog to (1) in Rowe’s original argument. Kraay observes, correctly, that (4) is considerably more modest than (1). According to (1), any gratuitous evil at all is sufficient to disprove God’s existence, whereas (4) has this result only if there is a large excess of such evil. Nevertheless, I will argue that (4) succumbs in the end to the same objection that was made to (1). Let me explain. The ultimate objection to (1) is that (1) results in a “moral trade-off,” in which any action taken to reduce or eliminate one occurrence of evil will in effect be neutralized by an increase in such evil (or a decrease of good) somewhere else in the system. But the very same problem afflicts (4); it’s just that the problem doesn’t kick in quite as soon. When we ask about the implications of God’s implementing (4), we have to assume that God is actively working to keep the level of gratuitous evil within the “acceptable zone” where it has not become excessive. (Thus, we assume that it is not the case that humans, by their own willing cooperation with God’s good purposes, keep the level of evil below what would trigger a concern about (4). That just is not the sort of world we live in.) Now, what would happen when a person is considering some action that might prevent suffering that would otherwise occur? In most cases, (4) has little bearing on the situation. Whether this action is carried out or not, the amount of gratuitous evil presumably will remain within the rather broad “acceptable” range, and no specific divine response will be called for. (This contrasts with the situation with (1), where any such action will immediately trigger a divine response, in order to keep good and evil appropriately in balance.) In such situations, there is no reason why the agent’s moral motivation should be compromised. However, the situation changes when we consider actions that will have a very large effect. Consider, for example, the 2014 epidemic of the Ebola virus in West Africa. This was by any account a major health disaster, resulting in over 11,000 deaths and a great deal of other damage to the lives and

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

66  William Hasker welfare of the people in the affected areas. The epidemic was met with a far-from-perfect but nevertheless vigorous response from many different governments and non-governmental organizations, and this response kept the levels of death and destruction far below what they might otherwise have been. Judging from what has happened in other historical epidemics, the number of fatalities had the epidemic gone unchecked might well have been in the hundreds of thousands or even millions. Now, consider this situation in the light of the assumption that God governs the world in accordance with (4). If the epidemic had been allowed to proceed unchecked, this would have led to an enormous amount of gratuitous evil, without doubt enough to unbalance the situation as called for by (4). This imbalance, however, must be compensated for, in order that the strictures imposed by (4) not be violated. So if no effective measures had been taken to halt the epidemic, there would of necessity have been, somewhere else in the system, either a large decrease in some kind of evil, or an enormous bonus of goodness, enough to rectify the imbalance caused by the terrible suffering from the unchecked epidemic. And by the same token, those who led the effort to halt the epidemic were insuring that there would be either a great deal of additional gratuitous suffering elsewhere in the world, or that there would not be some commensurately great good that would otherwise have occurred. (It’s not that their actions would do harm; rather, the point is that something else just as good would have happened if they had not intervened to suppress the epidemic.) Needless to say, no thoughts along this line were in the minds of those who rallied, sometimes at great personal cost, to halt the epidemic—nor should they have been. Such, however, are the implications of (4), when applied to situations such as this. Similar reasoning, of course, will also apply to the adoption of public policies designed to alleviate suffering and to improve the general welfare. Ironically, the more effective such policies might be in preventing harm to persons and in enhancing the ­general good, the more is lost because of them elsewhere in the world-system. I submit that these consequences give us good reason to reject (4) just as we have seen reason to reject (1). It occurs to me, however, that some readers might find the reasoning depicted in the preceding paragraph cold and repellent. They may find it morally objectionable, even in a hypothetical discussion, to talk about terrible harms to vast numbers of human beings as if the people involved were counters in some sort of game. If so, I fully share in this repugnance. The need to reason in this way, however, is not something created by me; rather, it is imposed by the framework of assumptions with which we are operating,

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

God and Gratuitous Evil: A Response to Klaas Kraay  67 assumptions crystallized in such principles as (1) and (4). If the way these principles require us to reason is found to be repugnant, the remedy is to reject the principles. And that, I say, is just what we ought to do.

5. Conclusion The conclusion reached thus far, that we should reject the claim that God necessarily prevents all gratuitous evil, will no doubt meet with some resistance. Rejecting it is not, however, an especially radical move; this principle has been rejected by such distinguished philosophers as John Hick, Austin Farrer, Michael Peterson, Marilyn Adams, Jeffrey Jordan, and James Sterba, among others. Or, to restate this rejection in slightly different terms, we can assert that God is not a utility-maximizer. It is not clear why this should occasion surprise. We are familiar with objections to utilitarianism as an account of human obligation, so why should it seem obvious that this is the best way to describe God’s perfect goodness? To be sure, this negative conclusion is not all that needs to be said. A defense, however successful, is not a theodicy, nor is it an account of divine providence. So we need a better account of the goodness of God, and a better account of divine providence, that can take the place of those that have been rejected. Those topics, however, must wait for another occasion.17

17  Some things that may be helpful along these lines are offered in my Providence, Evil, and the Openness of God, and also in The Triumph of God Over Evil.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

5

Moral Indulgences When Offsetting is Wrong Rebecca Chan and Dustin Crummett

0. Introduction Suppose that it would be wrong for Scott to eat meat, absent a successful defeater. (Roughly, a successful defeater is a factor that nullifies or overrides the considerations that normally make the act wrong.1) Scott realizes this. But Scott also realizes that there are many very effective, underfunded, animal welfare charities. For instance, Animal Charity Evaluators (2017) estimates that one of their top-rated charities, the Humane League, may do the equivalent of saving as many as 180,000 animals from being factory farmed for every thousand dollars it receives through advocacy work. Scott realizes that even if this estimate is off by a few orders of magnitude,2 he could easily do more good by making fairly small donations to such charities than he does harm by eating meat. (This is especially true if he limits himself to eating beef: since cows are very large and produce a lot of meat, eating beef contributes to far fewer animal deaths than, say, eating chicken or shrimp.) And he prefers making such donations to giving up meat altogether. So he does that.3 1  This is McMahan’s (2001, 236) gloss on moral defeaters. See also footnote 7. 2  A critical discussion of Animal Charity Evaluators’ methodology can be found at https:// medium.com/@harrisonnathan/the-problems-with-animal-charity-evaluators-in-briefcd56b8cb5908. 3  Scott is named after Scott Alexander, author of the blog Slate Star Codex, who (inspired by another blogger, Katja Grace) once advocated a course of action similar to the one Scott takes. His post is what inspired us to consider the permissibility of the course of action described (2015a). After writing this paper, we learned that, in a separate post, Alexander similarly puzzled over the moral distinction between pollution and murder offsetting, which we discuss in the next paragraph (2015b). There, Alexander considers that universalizability might be an objection to offsetting meat-eating, but does not defend our claim that it might be a solution to explaining the moral difference between these different types of offsetting. He also addresses the question of the morality of offsetting in Alexander (2017).

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Moral Indulgences: When Offsetting is Wrong  69 Presumably, there is a point at which such donations aren’t otherwise obligatory. (If Scott was obligated to donate part of his money anyway, suppose he already donated that part, and we’re talking about additional donations; we assume that Scott isn’t obligated to give all of his disposable income to charity.) Of course, Scott could donate the money without eating meat, and that would be better than what he does. But Scott could also permissibly refrain from either donating or eating meat, and that would be worse. Since Scott does better than he was obligated to do, he figures that he does nothing wrong—he figures, that is, that his donation defeats what would otherwise be the wrongness of his eating meat. But does it? Notice that while there may be uncertainty with respect to the permissibility of Scott’s actions, there is little uncertainty with respect to some other, structurally similar actions. Consider Billy Kincaid, who Sherlock Holmes discusses in episode two of season three of the BBC’s Sherlock. When Watson asks Holmes to be his best man, Holmes misunderstands and thinks Watson is asking who the best man he knows is. Holmes responds: Billy Kincaid, the Camden garrotter. Best man I ever knew. Vast contributions to charity, never disclosed. Personally managed to save three hospitals from closure and ran the best and safest children’s homes in North England. Yes, now and then there would be some garrottings, but stacking up the lives saved against the garrottings, on the balance . . .

For obvious reasons, Sherlock’s response is played for laughs. Suppose that Billy performed his good works with the intention of offsetting his garrottings. Billy’s reasoning might then be pretty similar to Scott’s, but Billy’s actions seem clearly impermissible. One shouldn’t murder another person, even if one then saves two lives by donating a few thousand dollars to an anti-malaria charity. Even acts of heroism don’t make up for a “garrotting now and then.” While Billy Kincaid acts impermissibly, there are structurally similar actions that appear permissible. Consider Pauline, who as part of an otherwise modest lifestyle, likes to take a lot of nice vacations to foreign countries. This travel generates a lot of pollution. In fact, it generates far more than her fair share—an amount of pollution that it’s wrong to generate, absent a defeater. However, Pauline is conscientious about her pollution, and donates money to pollution fighting efforts with the intention of offsetting the pollution she creates. Her donation is otherwise supererogatory, has a positive effect that outweighs the badness of her pollution, and accomplishes a good outcome that could have also been accomplished without her

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

70  Rebecca Chan and Dustin Crummett traveling. Pauline’s actions have the same structure as Billy’s, but unlike Billy’s, her actions seem permissible. Indeed, purchasing pollution offsets is a common practice amongst factories, and we generally regard this practice as a morally good thing. We might have reservations about the particular details of some pollution offset system—about, say, whether the offset mechanisms involved are really effective, whether it distributes the burdens of pollution fairly, etc.—but few people have a problem with such a system in principle.4 Finally, consider God. Suppose that, for any world God could make, there’s another, arbitrarily better world which God could make. (Perhaps God could always make a world much better by adding many additional good things.) And suppose that there’s at least one world—call it W1— which God can permissibly make (cf. Kierland and Swenson 2013). Suppose the world God does make is like W1, except for two features. First, it contains some gratuitous evil, i.e., evil that exists either for no particular reason or for some morally insufficient reason (maybe God allows some terrible atrocity, knowing it will produce some good literature). Second, God adds a bunch of additional good things in order to ensure that the world God does make is much better than W1. Does God do anything wrong? This chapter is about the above actions and others that are structurally similar to them. Specifically, it explores the permissibility of courses of action like these, ultimately with the aim of shedding light on the wellknown question of what God should do if worlds just get better and better. First, it introduces the concept of moral indulgence to characterize sets of actions that bear the structure of those above (Section 1). Roughly speaking, an agent is morally indulgent when they do something which, absent a defeater, is wrong, and, in order to offset this, do something which is supererogatory and which is more good than the bad action was bad.5 Second, this chapter proposes an explanation of when and why being morally indulgent is permissible (Section 2). As our above cases suggest, some morally indulgent sets of actions are permissible while others are impermissible. We seek to find an explanation for why this is. Our proposed explanation appeals to universalizability of the sort found in certain forms of Kantianism, contractualism, and rule consequentialism. Finally, we explore 4  There are exceptions who take a significantly more negative view than we do (see, e.g., Sandel 2012), but they’re in the minority. 5  A similar concept is independently developed in a currently unpublished paper, “Moral Offsetting,” by Tyler John and Amanda Askell.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Moral Indulgences: When Offsetting is Wrong  71 the implications of our account for the question of what God should do, and why God should do it, if there are no unsurpassable possible worlds (Section 3). This discussion of divine morally indulgent actions contains an interesting upshot: while the Kantian, contractualist, and rule consequentialist principles underlying universalizability converge in non-divine cases of moral indulgence, they come apart in the divine case (and perhaps other structurally similar cases).6 Thus, our discussion of divine indulgence, in addition to its interest for philosophers of religion, may provide a reason to prefer some of these theories over others.

1.  Moral Indulgence Scott, Billy, and Pauline are morally indulgent. An agent is morally indulgent if: (a)  an act A is wrong, absent a defeater for its wrongness; (b)  there is some other act B, such that: (i) B is supererogatory (unless it is obligatory only because one performs A), (ii)  the goodness of B-ing outweighs the badness of A-ing, and (iii)  the goodness of B-ing can be realized independently of A-ing; (c)  a person performs B with the intention of offsetting the badness of A. For instance, eating meat is, we supposed, wrong, absent a defeater; donating to animal charities is supererogatory, outweighs the badness of eating meat, and can be done independently of eating meat; and Scott donates in order to offset the badness of his eating meat. So another way to put the

6  We can imagine extremely unrealistic cases where a non-divine agent is placed in a situation similar to the one in which we imagine God being. For instance, an extremely powerful genie might be able to choose between options which are relevantly similar to creating W1, W2, or W3, or might give a human being the opportunity to do so. But there are at least two reasons for focusing specifically on the divine case. The first is that the question of what worlds God ought to create if there is no best world is of independent interest to philosophers of religion, in light of the fact that God may actually be in this situation. (Similarly, theological fatalism could be reframed as oracle fatalism, but theological fatalism is still of special interest since it’s the subject of a live debate in philosophy of religion.) The second is that many people think that in moral theorizing, we ought to be more confident about, and give more evidential weight to, judgments about realistic cases as opposed to extremely unrealistic ones. Assuming that there’s at least some real possibility that God may exist and be in the situation we describe, this may be a reason for ethicists qua ethicists to pay this case special attention, even ­independently of its interest for philosophers of religion.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

72  Rebecca Chan and Dustin Crummett question we’re interested in is whether, when, and why actions which meet condition (b) can defeat what would otherwise be the wrongness of actions which meet condition (a) when performed with an intention that satisfies (c). A few features of this definition deserve comment. First, saying that “A is wrong, absent a defeater” means what it sounds like. As mentioned, defeaters are factors that nullify or override the considerations that normally make the action wrong; without them, the normally wrong act remains wrong.7 Regarding (b)(i), the clause “unless it is obligatory only because one performs A” is needed because we might think there are cases where, given that we perform A, we are obligated to perform B in order to defeat the wrongness of A. The point is just that, if we hadn’t performed A or B, we wouldn’t have done anything wrong. Regarding (b)(ii), the “goodness of B-ing” and the “badness of A-ing” are intentionally left broad, and can mean whatever plausible thing we want them to mean. The most obvious way to interpret the terms is as referring to the value of the states of affairs which result from the actions, but if we wanted, we could instead interpret them in terms of, say, the virtuousness or viciousness of the actions. Additionally, we talk about how B-ing outweighs A-ing because we have in mind cases of harm outweighing rather than harm blocking.8 Finally, note that (b)(iii) is essential to differentiating moral indulgences from other cases which have received more attention from ethicists. In general, there’s nothing odd about the idea that performing a supererogatory action might defeat what would otherwise be the wrongness of a different action: for

7  Nullifying defeaters are undercutting: what would normally count as a reason for action does not count as a reason at all, given the presence of a nullifying defeater. For instance, that an action would promote pleasure is generally a reason to perform that action. But if an agent will derive pleasure from torturing someone, perhaps promoting this sadistic pleasure does not even count as a reason in favor of carrying out the torture—the reason is nullified. Overriding defeaters are rebutting: reasons for action still carry normative weight, but are outweighed by other reasons. For example, it is normally wrong to cause a child to experience pain by poking them with a needle, but if one is doing so to provide the child with a vaccination that will prevent even more severe suffering, causing the pain is permissible because it is overridden by the associated good. Here, there is still a factor that counts against vaccinating the child, but other, countervailing considerations are more important. This distinction will not be too important in what follows: we are concerned only with whether the wrongness is defeated (so that performing an action meeting condition (a) is permissible), not with the nature of this defeat. (It does seem clear that Pauline has a reason, even given her making the donation, not to do the travel—namely, that she will cause a lot of pollution. It may be less clear whether she has deontic moral reason—the sort of reason that creates moral obligations—in light of having made the donation. For an explanation of the difference between deontic moral reasons and other reasons, see Wallace 2013.) 8  In other words, if B-ing defeats the wrongness of A-ing by preventing A-ing from having any bad effect at all, we don’t count that as a case of moral indulgence.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Moral Indulgences: When Offsetting is Wrong  73 instance, suppose I fail to make my promised lunch date with you because I’m busy supererogatorily saving someone from a burning building. But here, breaking the promise is necessary for saving someone, so that this is just an ordinary case of moral aims conflicting. However, in moral indulgence cases, there is no conflict; the agent could realize both moral aims. They just don’t. It’s this feature that makes them deserving of special attention. Notice that the above conditions for moral indulgence were given as sufficient, rather than necessary, conditions. That’s because we want to allow an agent to be morally indulgent by performing only one act which possesses, in the right way, both right- and wrong-making features.9 Specifically, an agent can also be morally indulgent if:   (a*)  in virtue of possessing property p, an act A is wrong, absent a defeater for its wrongness; (b*) A also possesses a property p* such that: (i*)  performing an act with p* is supererogatory (unless it is obligatory only because one performs an act with property p), (ii*)  p*’s goodness outweighs p’s badness, and (iii*)  the goodness of performing an act with property p* can be realized without performing an act with property p;   (c*)  The agent performs A with the intention of having their performance of an act with p* offset their performance of an act with p. Most ordinary cases of moral indulgence involve performing two separate actions. However, the single act version will be important in Section 3.2. Perhaps God’s actualizing a world consists in a single act which fulfills condition (a*) (in virtue of actualizing gratuitous evil) as well as condition (b*) (in virtue of containing lots of additional goodness). If so, we want to treat God’s act as structurally analogous to what Scott does, even though it involves only one act, rather than two. There is one more thing to note before moving on. Scott harms animals by eating meat, and then helps animals by donating money. It seems natural to say that there’s some sense in which Scott’s bad action and good action affect the same cause in a way that they wouldn’t if, say, he ate meat and later donated to an anti-malaria charity. Call this “cause-matching.” 9  Also, we want to remain neutral with respect to how to individuate actions. So having one- and two-action conditions for being morally indulgent allows us to characterize action(s) as morally indulgent without having to dive into the metaphysics of act individuation.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

74  Rebecca Chan and Dustin Crummett Cause-matching often seems natural and appropriate, but we face at least two questions about its moral significance. First, why match causes, instead of just donating to whatever cause will produce the most good? Second, is there some metaphysically privileged way of individuating causes? (If we think of the problem to which Scott contributes as the deaths of the particular animals he eats, then his good act doesn’t help the same cause. If we think of it as the fact that bad things happen, then the anti-malaria donation would cause-match, too.) If there is no metaphysically privileged way of individuating causes, then, since whether someone cause-matches is relative to which of infinitely many arbitrary ways of carving up causes we consider, any moral significance which attaches to cause-matching may also seem unacceptably arbitrary or relativistic. We don’t treat cause-matching as essential to moral indulgence, or to its permissibility. However, we do share the sense that it sometimes seems natural and appropriate. Fortunately, our account will be able to explain both why this is and how, for relevant purposes, to individuate causes. We’ll return to this point in Section 2.4.

2.  The Moral Status of Moral Indulgence 2.1  Two desiderata and what can’t meet them As we noted in the Introduction, it seems clear that morally indulgent agents sometimes act wrongly, and sometimes don’t: Billy’s murders are wrong, while Pauline’s polluting isn’t. Meanwhile, the moral status of Scott’s actions at least doesn’t seem very clear to us; we expect that individual intuitions about it will differ, and that they will usually be much less strong than in those concerning the cases of Billy and Pauline. In providing our own account of the moral status of moral indulgence, we’ll attempt to meet at least two desiderata. First, the explanation should do justice to our intuitions about the paradigmatic cases. It should yield the result that Billy acts wrongly while Pauline acts permissibly. It also should yield the result that the moral status of Scott’s action is less clear than Billy or Pauline’s. Second, the explanation should not be ad hoc. Rather, whatever it appeals to as constituting the morally important differences between the cases should be things which we have independent motivation for thinking are morally important. We don’t claim that the explanation we provide is the only account which meets these two conditions. Probably, many different accounts can do this,

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Moral Indulgences: When Offsetting is Wrong  75 so that selecting between them will ultimately need to be done on other grounds (for instance, various reasons for accepting one moral theory or another which have nothing to do with moral indulgence). But notice that meeting these conditions isn’t trivial. Some moral theories, at least in their most common forms, won’t be able to do so. Famously, maximizing act consequentialism is unable to accommodate supererogation. This will make moral indulgence impossible, since no actions will be able to meet condition (b). Of course, this also seems intuitively wrong: for instance, of note here is that it implies that Pauline acts wrongly. One way for an act consequentialist to accommodate supererogation is to adopt a satisficing view, on which I’m morally required only to produce enough (not necessarily the most possible) good. Actions which produce more than enough good, when it was possible for me to instead produce just enough good, would then be supererogatory. However, since normal forms of satisficing act consequentialism only care about staying above the line, they’ll allow too many cases of moral indulgence: Billy’s actions (and others relevantly like them)10 are permissible, just like Pauline’s are. (In fact, that satisficing consequentialism might license gratuitous murder or assault, provided one stays above the satisficing line, is a classic objection against it; see Kagan 1989.) But in other cases, we will be unable to accommodate the intuitive judgments about these cases for reasons which have nothing essentially to do with supererogation. Consider a view which holds that, when an action has both good and bad effects, it is a necessary condition for its being 10  We can imagine two types of satisficing view. One says that the sum total of all our actions must fall above a line. Another says that every one of our individual actions must fall above a line. Philip Swenson pointed out to us that the latter view would not allow actions like Billy’s, since murdering would fall below the line. However, this view is still too permissive in morally equivalent single-action indulgence cases. If Billy is permitted to donate only a certain amount, $X, could make a donation of $X+Y which would save several more lives than a donation of $X, but instead makes a donation which saves these additional lives but also somehow kills his rival, this is clearly impermissible, but will still fall above the satisficing line (which was met by a donation of just $X). Further, this view will probably be too strict in other cases. Assuming that, when I can avoid it, it forbids my performing actions which are not merely suboptimal but are actually fairly bad, it will rule out Paula’s going on her trip, which is fairly bad taken on its own, despite her offsetting donation. (If it doesn’t rule this out, then it would allow me to do nothing but needlessly somewhat bad actions throughout my entire life, which doesn’t seem correct.) Finally, it’s a little hard to see why an act consequentialist would be drawn to the individual-action form of satisficing: Why care about the effects of individual actions, if the total effects of two series of actions are the same? So we think the individualaction form of satisficing act consequentialism is less inherently plausible while also facing counterexamples in moral indulgence cases similar to those faced by other forms of act consequentialism.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

76  Rebecca Chan and Dustin Crummett permissible that it meet the criteria laid out by the doctrine of double effect. Generally, these criteria are taken to be something like the following from McConnell (2003, 880): 1. The act itself must be morally good or at least indifferent. 2. The agent may not positively will the bad effect but may permit it. If the agent could attain the good effect without the bad effect, they should do so. The bad effect is sometimes said to be indirectly voluntary. 3. The good effect must flow from the action at least as immediately (in the order of causality, though not necessarily in the order of time) as the bad effect. In other words the good effect must be produced directly by the action, not by the bad effect. Otherwise the agent would be using a bad means to a good end, which is never allowed. 4. The good effect must be sufficiently desirable to compensate for the allowing of the bad effect. Of course, there are some slight variations to the doctrine of double effect, but they won’t make a difference here. Suppose that the good (enjoyment, neat Facebook pictures, etc.) achieved by Pauline’s travel is not itself sufficient to justify causing the pollution (which is clearly possible, provided travel causes enough pollution). Her action then fails condition 4, and so is wrong regardless of her later donation.11 Since this is a counterintuitive result, we seem to have a counterexample to the principle of double effect which has no clear counterpart outside of moral indulgence cases. Perhaps 11  We could imagine a more complicated case in which Pauline ensures that the action of traveling and the action of making the donation are the same act. In that case, the act would meet condition 4, since, we supposed, the donation was much more desirable than the pollution was undesirable. However, for the principle of double effect to be very useful, it must be interpreted as requiring that the agent, if possible, achieve the good without the evil, even if that means sacrificing other goods which are not proportional to the evil. (For instance, suppose I can bomb the dictator now, ending the war but killing hundreds of civilians who have gathered to hear his speech, or I can the bomb dictator in an hour, killing only him but causing me to miss the season premiere of The Good Place. I must wait an hour, achieving the proportional good—ending the war—even though this means missing out on a less important good— seeing the season premiere.) McConnell builds this into condition 2 (“If he could attain the good effect [i.e., the one which, as per condition four, is proportional to the evil] without the bad effect he should do so”), while Walzer (1977, 151–6) builds it into condition 3. Either way, Pauline’s action will fail this condition, since she could achieve the good—donating—without the pollution, even though it would cost her the travel which, we supposed, was not proportional to the pollution. We could say the same if Pauline claims that, in the ordinary version of the case, her donation is actually a good effect of her travel, since she wouldn’t make it without traveling. (Of course, that claim might also be problematic for other reasons.)

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Moral Indulgences: When Offsetting is Wrong  77 there are other moral views which similarly cannot meet the desiderata. But rather than dwelling on those, we’ll turn to our own explanation.

2.2  Universalizability and the second desideratum We propose that whether a case of moral indulgence is permissible depends, roughly speaking, on whether it would be acceptable if people generally believed and were motivated to comply with principles that allowed them to be morally indulgent in the relevant way.12 This is an appeal to a certain understanding of universalizability. This account fulfills our second desideratum of not being ad hoc, since a number of leading moral theories accept universalizability, i.e., that an act is permissible if and only if it would be acceptable for people to generally believe and be motivated to comply with principles that allowed them to perform the act in those circumstances. We’ll focus on three moral theories that accept some form of universalizability here: rule consequentialism, contractualism, and a certain form of Kantianism. Roughly speaking, rule consequentialists will understand acceptability in terms of producing good consequences, and so think an act is permissible if and only if in accordance with principles which are such that they would have the best (or at least sufficiently good) consequences if accepted by everyone (e.g., Hooker 2000). Contractualists will understand acceptability in terms of being justifiable to all, where that is understood in a certain way, so that an act is permissible if and only if in accordance with principles for the regulation of behavior which everyone in some kind of fair decision situation could not reasonably reject (e.g., Rawls 1971; Scanlon 1998). And Kantians of a certain stripe will understand acceptability of terms of what the agent, or all agents, could rationally will, and so will think an act is permissible if and only if it is in accord with principles which the agent, or all agents, could rationally will to be universally accepted.

12  Notice that this is somewhat different from a view which focused on which principles would be acceptable if people in general successfully complied with the relevant principles. Focusing on acceptance means that we must take into account the possibility that people might believe and be motivated to comply with principles without actually complying with them— due to ignorance, self-deception, weakness of will, or whatever—and so entails that acceptable principles will be determined in a way that minimizes the possibility of failures like this. We think, and it has been argued elsewhere, that it is better for people who accept moral theories of this general sort to go with our route, rather than the perfect compliance route (cf. Hooker 2000, chs. 3.2 and 4.2).

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

78  Rebecca Chan and Dustin Crummett We should say a bit about the kind of Kantianism we’ll focus on, because it differs importantly from orthodox Kantianism. For Kant himself, there are two ways a maxim might be such that one could not will that it be a universal law. First, it might be such that it “cannot even be thought without contradiction as a universal law of nature” (Kant 1997, 33). So, famously, I  supposedly cannot conceive without contradiction of a situation where a maxim allowing false promising is universalized, since “in accordance with such a law there would properly be no promises at all, since it would be futile to avow my will with regard to my future actions to others who would not believe this avowal or, if they rashly did so, would pay me back in like coin,” so that “my maxim, as soon as it were made a universal law, would have to destroy itself ” (Kant 1997, 15). Of course, there is a question here about what, exactly, not being able to conceive of such a situation without contradiction comes down to (cf. Korsgaard 1985). Second, it might be that, though there is no “inner contradiction” in the maxim’s being universalized, I nonetheless cannot rationally will that it be universalized because this would conflict with something else which I necessarily will, so that such a will would necessarily “contradict itself ” (Kant 1997, 33). So, for instance, when someone considers universalizing a maxim allowing them to spend their life in leisure without developing their talents, they can conceive this situation without contradiction (“a nature could indeed always subsist with such a universal law”), but nonetheless “cannot possibly will that this become a universal law,” since a rational being “necessarily wills that all the capacities in him be developed” (Kant 1997, 33). Of course, there is a question here about what, if anything, agents necessarily will, and why. Derek Parfit (2011) suggests some important modifications to Kantianism. He suggests (chs. 1–2, 12–14) that we should simply start with the claim that we have reasons to want certain things (that we avoid pain, for instance), and understand what we can rationally will in terms of what’s in accordance with those reasons. This allows us to avoid claims about what we can or can’t conceive without contradiction, or what we do or don’t necessarily will. And he suggests that we should consider what everyone could rationally will to be a universal law, not just what the agent could will, in order to avoid worries about whether it could be rational for an agent to will a clearly immoral maxim which would harm others but benefit themselves (2011, 334–8). Accordingly, he suggests that Kantians should adopt the principle that an act is wrong “if and only if, or just when, such acts are disallowed by some principle that is . . . one of the only principles whose being universal laws everyone

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Moral Indulgences: When Offsetting is Wrong  79 could rationally will” (2011, 412–13), where what everyone can rationally will is understood in the way just described. We think Parfit’s Kantianism is more plausible than Kant’s Kantianism, so, from now on, when we discuss Kantianism, we’ll really have in mind Parfit’s modification of it. Of course, if it turns out that Kant’s Kantianism can also yield the right results about the paradigm cases, this only helps us, since it only strengthens our claim that the feature we’re appealing to is found in many different moral theories. With this kind of Kantianism in mind, we can now turn to the way in which the three theories we’re focusing upon converge—or, as the case may sometimes be, diverge. There is some dispute over whether the acceptability conditions posited by each of these theories wind up yielding co-extensional results. For instance, are the principles that could be justified to everyone those that would have the best consequences? Or are the principles that have the best consequences those that every agent rationally could will? Parfit (2011) argues at great length that the best forms of rule consequentialism, contractualism, and Kantianism will turn out to be co-extensive with one another with respect to their judgments about the permissibility of actions. Unsurprisingly, others disagree.13 We don’t accept the claim that the three theories converge; in fact, we’ll argue in Section 3.3 that this claim is false. However, we do think that the best forms of these theories exhibit a great deal of overlap. Fortuitously for us—or so we argue—they overlap in a way that yields agreement when it comes to the permissibility of moral indulgence in the paradigm cases which we wish to explain. (Thus, we won’t need to worry too much about what the right way to flesh out what “acceptability” is, even in Section 2.3.) There are two reasons one might doubt this, which have to do with the fact that our examples involve, respectively, pollution and non-human animals. We’ll address those in turn. Some people think that nature has some sort of intrinsic value which is harmed by pollution. If this is right, rule consequentialism will recognize a prima facie duty not to pollute which is grounded in this value. It’s not obvious that the same can be said for contractualism and Kantianism. For instance, Scanlon (1998, 218–23) argues that contractualists should not recognize “impersonal” values, which are not directly tied to the interests of any person, as directly giving rise to moral obligations. The thought is that in the contracting situation, no one will have standing to reject principles 13  Parfit’s argument comprises Volume I of On What Matters; for critical responses, see the essays by Susan Wolf, Allen Wood, Barbara Herman, and T. M. Scanlon at the beginning of Volume II.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

80  Rebecca Chan and Dustin Crummett on the grounds that they permit acting against these impersonal values, since no one directly has a stake in their being respected. Whatever intrinsic value nature has is supposed to be an example of such an impersonal value. If this is right (we make no claim about whether it is), then contractualists will not recognize the prima facie duty which rule consequentialists recognize here. However, both contractualists and Kantians will still recognize a duty against pollution which is grounded, not in the intrinsic value of nature, but in the interests of others who rely on nature for survival, enjoyment, etc., and who will be harmed by disease (e.g., in the case of toxic pollutants) or natural disaster (e.g., in the case of CO2, which contributes to global warming) if pollution is excessive. Accordingly, rule consequentialists, contractualists, and Kantians can all agree that some amount of polluting will be wrong absent a defeater, even if they disagree somewhat about what that amount is or how strong the relevant obligation is. Accordingly, they will agree that it’s possible to set up a case like Pauline’s, where someone pollutes so much that they act in a way which is wrong, absent a defeater. And the explanation we give in Section  2.3 for why acceptable principles allow Pauline’s indulgence won’t be affected by the differences between the theories. The issue with animals is a little more complicated. Provided that the flourishing and suffering of animals has value and disvalue, consequentialists think we have obligations to them. But contractualists and Kantians don’t always agree: for instance, Scanlon (1998, 177–87) is “inclined” (184) to think that animals should be excluded from the contracting situation, and Kant (2001, 240) thinks we have no obligations to them since they lack a rational nature. This would be an important difference, when it comes to Scott’s case. Such people might agree that we have reasons, derived from our obligations to humans, not to do certain things to animals: maybe we shouldn’t eat them for environmental reasons, or shouldn’t torture them so as not to cultivate cruel dispositions. But they might also think that Scott’s case isn’t morally much different than Pauline’s case, whereas we thought it was much less obvious that what Scott does is permissible. For our part, we think it’s obvious that we have obligations to animals. Fortunately, there are various ways contractualists and Kantians might explain this. For instance, contractualists might give animals trustees in the contracting situation to whom we must justify the principles we select (cf. Rowlands 1997; Scanlon 1998, 183–5). Kantians might say that non-human animals themselves possess the kind of agency necessary to be worthy of treatment as ends in

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Moral Indulgences: When Offsetting is Wrong  81 themselves, and so to have moral status within the Kantian framework (Korsgaard  2004). Since we think it’s obvious that we have obligations to animals, we think the best forms of contractualism and Kantianism will be compatible with this fact. (Anyone who agrees with our judgment about Scott’s case will probably agree.) We’ll therefore assume that we’re discussing animal-friendly versions of contractualism and Kantianism in what follows.

2.3  Application of universalizability and the first desideratum Our explanation involving universalizability, which appeals to a component of many popular moral theories, is not ad hoc, and so meets our second desideratum. It also meets the first. Consider Billy, our indulgent, philanthropic murderer. It’s hard to imagine that a society could last very long if people thought it was acceptable to murder others, provided that they donated a few thousand dollars to the Against Malaria Foundation. The results would be catastrophic, probably leading to the collapse of society altogether. People would live in fear, distrust, and resentment of each other. They might feel a need to preemptively strike against their enemies, and already competitive arenas like the academic job market would become fraught with danger.14 Notably, this line of reasoning is similar to a classic rule consequentialist explanation of why one shouldn’t butcher a patient in order to save five others with the patient’s organs (cf. Parfit 2011, 363–4). If doctors generally did this, society would be filled with fear, distrust, and resentment, and people would fail to seek needed medical treatment. The resulting situation would be unacceptable by any plausible standard. For instance, it would have bad consequences, and some people would be  able  to reasonably reject it and unable to rationally will it. Thus, if

14 One might wonder why “‘don’t freak out and dissolve society when people morally indulge in murder” couldn’t be added to the rule so that offsetting murder with saving many would be universalizable. It’s important to remember that in evaluating rules, one of the things we should consider is the effort needed to get people to accept and be motivated to comply with these rules (e.g., Hooker 2000, ch. 3). If a rule makes acceptance and motivation to comply much harder or impossible, that’s a reason to think it isn’t universalizable. Given contingent facts about human psychology, it’s hard to see people accepting and being motivated to comply with such a rule. See also footnote 9.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

82  Rebecca Chan and Dustin Crummett universalizability is necessary for permissibility, Billy’s donations don’t defeat the wrongness of his killing. We get the intuitively correct result. Meanwhile, consider Pauline. A certain amount of pollution inevitably will be created by any modern society, and there are great benefits to producing some. The question for acceptable, universalizable principles is not whether to pollute, but how to regulate the pollution that does occur. It’s important that we not produce so much that the marginal costs exceed the marginal benefits, and, if someone produces more than their fair share, this will be wrong, absent a defeater. (Principles which allowed people to just produce as much pollution as they wanted, without needing a defeater for doing so, would result in too much pollution being created, with the attendant catastrophes.) However, people do sometimes have reasons for wanting to produce more than their fair share—economic reasons, say, or, in Pauline’s case, personal enjoyment. If someone is willing to use resources which they otherwise could permissibly have kept to do things which more than offset their additional pollution, it seems acceptable to allow this to serve as a defeater for the wrongness of their polluting: they get what they want, and the pollution situation is better than if they’d done nothing. Further, unlike with murder, it is generally not possible to target particular individuals by polluting. (Polluting may impose a very small risk—say, of getting a disease—on a very large number of people, but that would be an extremely ineffective method of targeting one’s rivals.) Therefore, the same reasons for fear, preemptive strikes, and avoiding the academic job market are not applicable to pollution offsets. So, under the right circumstances, offsetting will promote value, and no one has reason to object. As mentioned, we may have issues with particular pollution offset systems (e.g., they may be ineffective or unfairly disadvantage poor people), but having a fair and effective system in place to manage pollution seems, in principle, universalizable. Again, we get the intuitively correct result. Finally, Scott’s situation is more complicated. On the one hand, unlike Billy’s case, allowing people to behave as Scott behaves wouldn’t spread distrust and resentment, provoke preemptive strikes, or, generally speaking, do anything else that might lead to the collapse of civilization. Further, it would produce some good, insofar as it might promote donations and other good acts which, in individual cases, more than offset the harm of eating meat. On the other hand, we might worry, for instance, that it would be impossible to cultivate and sustain the attitudes needed to eliminate the widespread abuses of animals which exist in our society as long as people feel free to eat them, even if they feel obligated to make up for doing so. And, of course, an

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Moral Indulgences: When Offsetting is Wrong  83 outright prohibition on eating meat would have the b ­ enefit of reducing the killing of animals for meat. So the question winds up being complicated. We won’t try to settle the question here, though we hope we’ve provided a framework in which one might do so, and think it’s worth doing, given the question’s practical relevance. What matters for our purposes now is that, again, our explanation gives the result we were looking for—Scott’s case is less clear than either Billy’s (which is rendered impermissible) or Pauline’s (which was permissible).

2.4 Cause-matching As seen above, universalizability yields the correct result in all three cases and meets the two desiderata. Our account also provides a plausible account of the significance of what we earlier called “cause-matching,” or directing your good action towards helping the same issue that your bad action exacerbated. Recall that, on the one hand, cause-matching often seems appropriate, while, on the other, it seems puzzling for two reasons. First, it isn’t immediately clear why one would cause-match instead of just doing whatever was best. Second, attempts at cause-matching might seem arbitrary, absent some metaphysically privileged way of individuating causes (and it might seem implausible that there is such a way). Consider the first puzzling feature. There are various reasons why acceptable principles might recommend or sometimes even require cause-matching, and these help explain why it sometimes seems especially appropriate. For instance, cause-matching might help ameliorate coordination problems. There are reasons to do things that cause pollution, but it’s important that we not cause too much, and whether we cause too much depends, in part, on individual actions performed by lots of different people. One reasonable way to help make sure that we don’t cause too much, and to help make sure that people causing excess pollution aren’t doing it for totally trivial reasons, is to make individuals responsible for offsetting their pollution if they cause more than their fair share. (Of course, in practice, this will also need to be supplemented with other measures.) Or consider the importance of indicating concern for various issues (cf. Crummett forthcoming). If we learn that Scott eats meat, we might reasonably suppose that he either rejects strong claims about the value of animal lives, or accepts but doesn’t act in accordance with them. Since people take their cues about what’s acceptable partly from other people, it will be very hard to cultivate and sustain good

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

84  Rebecca Chan and Dustin Crummett attitudes towards animals if many people send such signals.15 Perhaps, though, one way for Scott and others to offset this impression is to ensure that they do other actions such that they benefit animals on the whole. Acceptable principles might therefore recommend signaling in this way, if they allow what Scott does at all. With all that being said, while causematching is often appropriate, we don’t claim that it’s essential to the permissibility of moral indulgence. Perhaps there are cases where considerations like these don’t apply, or where they’re outweighed by the potential for doing much more good by not cause-matching. And, again, this seems like an intuitively reasonable position. Now consider the second puzzling feature: individuation. Given our understanding of why cause-matching is appropriate, there doesn’t need to be any metaphysically privileged way of individuating causes. Instead, whether we should classify actions as affecting “the same cause” for purposes of cause-matching will depend on contingent considerations about the results that principles mandating that I cause-match in a certain way will have in practice. The coordination problem in which Pauline is involved has to do with limiting the amount of pollution. If there is no deep metaphysical reason to classify “pollution” as a single cause without classifying “pollution and criminal justice reform” as one, it’s nonetheless the case that Pauline’s donating to criminal justice reform won’t help us with the relevant coordination problem. Or in Scott’s case, that he eats meat makes it natural to suppose that he must not care too much about farmed animals; it’s much harder to draw inferences from this to his views about, say, nuclear disarmament. So cancelling the inference by signaling concern for animals is what’s morally salient. This view provides a way of individuating causes based on their morally salient features, regardless of whether individuating them based on morally salient features is more metaphysically joint-carving than alternatives. As we said earlier, we don’t take cause-matching to be an essential feature of morally indulgent actions. However, it does often seem appropriate, and our view explains why this is. Furthermore, our view offers natural ways of responding to the two puzzles associated with cause-matching. In light of this, we do take the ability of our view to account for cause-matching as a major advantage of our view. 15  Would the appeal to universalizability render this consideration irrelevant, since, in e­ valuating principles, we assume that people generally accept and attempt to follow them? No, see footnotes 12 and 14.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Moral Indulgences: When Offsetting is Wrong  85

3.  Divine Indulgence 3.1  Background assumptions In Section 2, we argued that universalizability gives plausible answers about which cases of moral indulgence are permissible and which are not. Furthermore, this kind of universalizability is an independently credible candidate for being morally important, since it is found in certain forms of Kantianism, contractualism, and rule consequentialism. In this section, we turn to the case of divine indulgence. Interestingly, while Kantianism, contractualism, and rule consequentialism converge in their verdicts on the morally indulgent actions we’ve discussed thus far, they diverge when it comes to the divine case (and any structurally similar imaginary cases). Thus, the discussion of whether it’s permissible for God to perform morally indulgent actions might, in addition to being interesting on its own, provide a reason to prefer some of those theories over others. Before getting underway, there are two background assumptions about God’s obligations which we would like to make explicit. First, we assume that God has moral obligations. Second, we assume that the same fundamental principles that govern God’s obligations also govern our obligations (so that the account we developed to handle human moral indulgence can also be applied to God). These assumptions are not wholly uncontroversial (e.g., Adams  2017; Murphy  2017). But they comprise the dominant view, and, anyway, we don’t have space to defend them here. We’ll also make two more assumptions about the nature of modal space and its impact on God’s obligations. First, we assume that for each world that God could make, there is an arbitrarily better world that God could make instead. Here are two ways that might be true. First, perhaps God could improve on any world by adding an additional happy person, while holding everything else fixed. (This requires that two things be true: First, the modal claim that, for any world, there’s another world which is just like it except that it has one more happy person, and, second, the axiological claim that this additional person makes the world better.) God could then make any world arbitrarily better by adding an arbitrarily large number of additional happy people. Second, perhaps there is no limit on how well-off a person can be, and God could improve on any world by making every person in that world slightly better off—for instance, maybe there is no metaphysical limit on how much pleasure you can experience, or to how strong a desire can be when it’s satisfied or to how many desires you can have which are satisfied, or maybe an objective list theory of well-being is true, and there’s no limit on

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

86  Rebecca Chan and Dustin Crummett how many objective goods you can be positively related to. (This requires that two things be true: First, the modal claim that, for any world, there’s another world which is just like it except that everyone is better off, and, second, the axiological claim that making everyone better off makes the world better.) God could then make any world arbitrarily better through an arbitrarily large number of improvements to individual well-being. Again, these claims are not uncontroversial (e.g., Kraay 2010; Climenhaga 2017), but they are widely accepted and seem fairly plausible to us. We also assume that even if the prior claim is true, there is at least one world which God can permissibly make. (Given that at least this world exists, the theist better hope that it was created permissibly!) Some philosophers (e.g., Rowe 2004, ch. 6) have argued that if there are no unsurpassable worlds, God cannot exist, since a perfect being could not make a world which was less than the best. Brian Kierland and Philip Swenson (2013) argue that this claim is false, since it contradicts a true version of the oughtimplies-can principle. There must be at least one world which God can actualize, consistent with God’s perfection. Assuming that God’s perfection entails God’s not violating any moral obligations, it follows that there’s at least one world God can permissibly make. We think Kierland and Swenson are correct, and so accept this implication.

3.2  Divine indulgence With all this in the background, we now can construct two cases of divine indulgence. Here’s the first case. Imagine three possible worlds. The first, W1, is a totally paradisiacal world. W1 is good enough that God can permissibly make it, despite the fact that God could have made a better world by adding more happy people. The second, W2, is like W1, with two exceptions. First, it contains Bob, an innocent person who suffers horribly and pointlessly throughout his life before being annihilated. Second, the world also contains 500 trillion additional happy people. There are so many additional eternally happy people that, overall, W2 is better than W1.16 Finally, the 16  One might claim that W2 can’t really be better than W1: Bob’s terrible suffering lexically outweighs any possible set of additional good things which God might add. You might think this is so because, in actualizing W2, God would wrong Bob, and divine wrongdoing would be so horrible as to lexically outweigh any set of created goods. If so, set that aside. Claims about the supposed disvalue of the wrong involved in making W2 can’t be involved in the answer to the question we’re concerned with—whether and why it’s wrong to make W2—on pain of circularity. Maybe you instead think W2 can’t be better than W1, not because of any wrongdoing,

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Moral Indulgences: When Offsetting is Wrong  87 third world, W3, is like W2, except that Bob also gets an eternal, happy life. W3 is better than W2. Given that W1 is a world that God permissibly can make, W3, which is a Pareto improvement upon W1, should also be a world that God permissibly can make. Notice that creating W2 seems to fit the single action model of moral indulgence. Doing so has a property—namely, giving Bob a pointlessly horrible existence—that makes performing it wrong, absent a defeater. Furthermore, since Bob’s horrible existence is totally pointless, nothing serves as a defeater, unless creating more goodness than was otherwise necessary does. Creating W2 also has another property—namely, causing the 500 trillion happy people to exist—which is such that (i) it’s supererogatory, (ii) it outweighs the badness of giving Bob a horrible existence, and (iii) in light of the possibility of W3, the relevant good could have been achieved without the relevant bad, or anything similar to it. Finally, creating W2 might also meet the intention criterion; assuming God intends to act in accordance with God’s obligations, God’s creating more goodness than was otherwise necessary will presumably be partly intended to offset God’s creating gratuitous evil. If this is right, then the creation of W2 fits the structure of morally indulgent actions. So, could God permissibly create W2, condemning Bob to a life of nothing but pointless, unimaginable torment? If so, this would seem to undermine any version of the problem of evil which appeals to the claim that God would act wrongly in allowing evil.17 After all, for any gratuitous evil we observe, it might, for all we know, be offset by something wonderful somewhere else. On the other hand, the permissibility of W2 might have unnerving implications—what if God treats us like Bob? For our part, we think the creation of W2 seems clearly impermissible: If God wants the but because Bob’s suffering just directly outweighs any possible additional goods. We think this claim is implausible. After all, there are people on earth who live lives full of horrible suffering. If philosophical naturalism is true, their earthly lives represent their entire existences. But very few naturalists think this alone makes the existence of life on earth regrettable (even if they think it is, in fact, regrettable due to the proportion of people who lead such lives). And taking this view would have implausible implications for certain ethical issues. For instance, it would imply that people who don’t believe in an afterlife should think procreation generally makes the world vastly worse, since, unless your line quickly dies out, it’s virtually certain that at least one of your descendants will have a life of misery. But thinking this on this ground alone would be very odd. 17  Other versions might, for instance, appeal to the claim that God loves us, and therefore wouldn’t allow certain evils, rather than the claim that God would act wrongly in allowing those evils. For a criticism of versions which appeal to something other than God’s moral obligations, see Murphy (2017, ch. 2).

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

88  Rebecca Chan and Dustin Crummett 500 trillion additional happy people, God should just make W3 and give Bob a nice life, too. (Note, too, that if God were to turn the creative decision over to us, it would seem clearly impermissible to create W2 rather than W3. Combined with our earlier assumption that God’s obligations are similar at the fundamental level to our own, this suggests that it would also be impermissible for God to create W2 rather than W3, unless there is some morally important disanology between ourselves and God which is relevant to this case.) If this judgment was in tension with our best theory of the permissibility of moral indulgence, then perhaps we would need to rethink it. On the other hand, if a theory can account for this judgment, that might be a reason to accept the theory. We’ll return to this point shortly. Here’s a second moral indulgence case. The case of moral indulgence involving W1–W3 employs the first procedure for improving worlds by adding additional happy people. A morally indulgent case can also be constructed out of the second procedure involving increasing the happiness of existing people. (This second example is important primarily because a certain form of rule consequentialism will have something different to say about this case than about the prior case.) Again consider three worlds: W4, W5, and W6. W4 is like W1. It’s a totally paradisiacal world that God can permissibly make, even though God could have made everyone much better off. W5 is like W4, except for two features. First, it contains Jack, an innocent person who suffers horribly and pointlessly throughout his life before being annihilated. Second, it contains large improvements in everyone else’s (already great) lives—improvements that are sufficient to outweigh the disvalue of Jack’s suffering. Finally, W6 is like W5, except that Jack has a life as eternally wonderful as everyone else’s. Like W4, W6 is a world that God can permissibly make. Creating W5 intuitively appears impermissible for essentially the same reasons that creating W2 did. After all, it contains Jack, an individual who has a pointlessly horrible life. Given W6, it’s clear that Jack’s suffering is not necessary for achieving some greater good. Nevertheless, W5 is significantly better than W4, which is permissible to make. In light of this, creating W5 looks morally indulgent, again for basically the same reasons that creating W2 does.

3.3  Universalizability and divergence on the divine So what does our account imply about the permissibility of these cases of moral indulgence? Unlike in the cases of Scott, Billy, and Pauline, we think

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Moral Indulgences: When Offsetting is Wrong  89 there will be a deep, essential difference depending on which particular moral theory’s account of universalizability we employ. Start by considering rule consequentialism. To know what rule consequentialism says about the cases in question, we must distinguish between particular forms of it more precisely than we have so far. Specifically, we must make two distinctions. The first is between maximizing and satisficing rule consequentialism. Maximizing rule consequentialism tells us to follow the rules that would have the best consequences if people generally believed and were motivated to comply with them. Satisficing rule consequentialism tells us we need only follow rules which would have sufficiently good consequences if people generally believed and were motivated to comply with them. Next, distinguish between person-affecting and total rule consequentialism (cf., e.g., Parfit 1984, ch. 18; Singer  1993, 103–4). Person-affecting rule consequentialism tells us that what matters is the sum total of consequences for people who actually exist. It denies that there is any obligation to create additional happy people. Total rule consequentialism claims that what matters is just the overall value of the world. It recognizes an obligation to create additional happy people, if one can do so without causing offsetting harm.18 These are cross-cutting distinctions, so that one can be a maximizing person-affecting, maximizing total, satisficing person-affecting, or satisficing total rule consequentialist. The maximizing total view implies that, if there are no unsurpassable worlds, God cannot permissibly create any world. For any world God creates, a principle which forbade creating that world and instead required creating a better world would produce a better result. This entails that W1 and W4 are impossible (since they were defined as being permissible), and that the other worlds mentioned are also impossible (since they were defined relative to W1 or W4). Since we thought that, given ought-impliescan, there should always be at least one world which God can permissibly ­actualize regardless of what modal space looks like, this seems to us like a bad result. Meanwhile, if there is a limit to how well-off an individual can be, then the maximizing person-affecting view implies that W1 and W3 are possible. That God could create additional happy people doesn’t count against their 18  Singer (1993, 103–4) defines the total view as implying that “that it is good to increase the amount of pleasure in the world by increasing the number of pleasant lives, and bad to reduce the amount of pleasure in the world by reducing the number of pleasant lives,” and the personaffecting (or, as he calls, the “prior existence”) view as denying “that there is value in increasing pleasure by creating additional beings.” Our usage tracks his, except that we put things in terms of what’s obligatory rather than what’s valuable.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

90  Rebecca Chan and Dustin Crummett permissibility, since there’s no obligation to create happy people. This view also implies that W2 is impermissible, since it fails to maximize the wellbeing of the people who actually exist. By making W3, God could increase Bob’s well-being without hurting anyone else. We agree with all these judgments. However, if there is not a limit to how well-off an individual can be, then this view implies that God cannot permissibly create any world. For any world God creates, a principle which forbade creating that world and instead required creating an otherwise similar world where the inhabitants were better off would produce a better result for the people who actually exist. This implies that W4–W6 are impossible. Since we thought that there should always be at least one world which God can permissibly actualize regardless of what modal space looks like, this again seems to us like a bad result. Meanwhile, any form of satisficing view probably implies that all of W1–W6 are permissible. Consider the satisficing total view. If all we need is for our principles to result in a world which is good enough, they can allow God to create worlds like W1, W3, W4, and W6. However, since W2 and W5 are better than W1 and W4, respectively, presumably they will also allow the creation of those worlds. After all, these worlds are better than W1 and W4, respectively, which were good enough, and what mattered was that things were good enough. Of course, because, on the rule consequentialist picture, we need to consider the consequences of people generally believing that a certain set of principles is the correct one, we will also need to factor in the consequences of people believing that God is permitted to allow horrible, pointless evils to befall them. This might cause some anxiety, and so on. However, God can always make up for this by adding more good to the world. Principles which tell God to make a world which is good enough and which forbid God from allowing horrible, pointless suffering will not produce better consequences than principles which tell God to make a world which is good enough, factoring in any pointless evils God allows and any bad consequences which result from people knowing God can allow pointless evils. Both will satisfice. Since we thought W2 and W5 were intuitively impermissible, we think this is a bad result. Similar things can be said about the satisficing person-affecting view. Since this view doesn’t recognize an obligation to produce additional happy people, this possibility is no threat to the permissibility of W1 or W3. Since it only requires that God produce enough good for actually existing people, there will presumably be permissible worlds like W4 and W6. However,

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Moral Indulgences: When Offsetting is Wrong  91 since it only requires that the sum total of good for existing people be good enough, it will allow the misfortune of Bob and Jack to be offset by additional goods to others, and so will allow W2 and W5. Satisficing rule consequentialism here seems to license too much, just as satisficing act consequentialism licensed too much when it allowed Billy’s moral indulgence. We saw that every form of rule consequentialism gave what we regarded as implausible judgments of one form or another. If this was a general feature of attempts to explain the permissibility of moral indulgence by appealing to universalizability, it might threaten our account. If this was a feature of every plausible attempt to explain the permissibility of moral indulgence, it might cause us to rethink our initial judgments. However, both contractualism and Kantianism seem to yield what we regarded as the intuitively correct judgments about the two cases of divine moral indulgence we presented. First, consider contractualism. Plausibly, contractualism renders both W2 and W5 impermissible, since, plausibly, Bob and Jack could reasonably reject principles allowing them to be given pointlessly horrible lives. If their suffering was necessary for the additional goods found in their worlds, perhaps one could argue that reasonableness required them to take one for the team. However, given that isn’t so, intuitively, it seems perfectly reasonable for them to insist that, if God wants the additional goods involved, God should make a world like W3 or W6 instead. What about W1 and W3? Generally, contractualists have thought that we have an obligation to act in ways which are justifiable to all and only actual people, without having any obligation to bring merely possible people into existence (cf. Scanlon 1998, 177–87). If this is the right route, then contractualism would be more like person-affecting than total consequentialism, and the possibility of additional happy people would be no threat to the permissibility of these worlds. What about W4 and W6? One might suppose that, if there is no limit to individual well-being, individuals could reasonably reject whatever God did, since they always could have been made happier. But the contracting procedure is supposed to model a decision-making process among reasonable agents who are attempting to reach agreement on principles which are acceptable to all parties. In such a situation, agents who will object to any principle put before them don’t seem to be acting in good faith. It therefore seems plausible to think that some lives will be good enough for God to give us, even though it may be difficult to say exactly

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

92  Rebecca Chan and Dustin Crummett where the line is. (And like we said, it seems intuitive to think the line will rule out a life consisting of nothing but horrible torment.) Contractualism therefore gives what we thought were the correct answers about the permissibility of W1–W6. Now consider Kantianism (of the Parfit-flavored sort we have in mind). Could Bob and Jack rationally accept their treatment? Answering “no” makes a slightly stronger claim than saying that they could rationally reject it (it could be that they could rationally accept or reject it). However, given the availability of W3 and W6, it still seems fairly clear to us that they shouldn’t accept the treatment they get in W2 and W5. As in contractualism, what matters is plausibly what actual people can rationally accept, so there is no threat to the permissibility of W1 from the possibility of additional happy people. Can the people in W4 accept their position, despite the fact that they could have been happier? Consider the famous case of EverBetter Wine, which gets better and better over time without limit (Pollock 1983). Fortunately, I am immortal, so I can wait as long as I like to drink the wine. Unfortunately, whenever I drink the wine, I would have been better off had I waited. It seems to us that there will be a point at which I’m rationally permitted to drink the wine—presumably, after it’s gotten really, really, really good—even though it’s hard to say exactly where that point is, and even though I forsake arbitrarily great gains by drinking it at that point rather than later. If I’m rationally able to give myself a benefit which is good enough in the wine case, it seems plausible that I’ll also be rationally able to accept God’s giving me a life which is good enough. Therefore, W4 and W6 will be permissible, which, again, seems to us like the right result. To sum up: Our initial judgments were that God should be permitted to actualize some world or other, even if there are no unsurpassable worlds, but that God shouldn’t actualize worlds where someone has a pointlessly horrible existence. The forms of rule consequentialism we considered each violated one of these claims. However, Kantianism and contractualism both implied them. That we can justify these judgments on a theoretical level may be a reason to hold on to them. That Kantianism and contractualism can do so while rule consequentialism cannot may be a reason to favor these theories over rule consequentialism. And that our universalizabilityfocused account of the moral status of moral indulgence can, when interpreted in the Kantian or contractualist way, give us plausible judgments in these cases may be a reason to accept it.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Moral Indulgences: When Offsetting is Wrong  93

4.  Closing Thoughts We close by noting an important question for future research. A feature of Bob’s and Jack’s cases was that they not only suffered horribly, but had horrible existences: their suffering was uncompensated and their lives were bad for them on the whole. However, not every case of moral indulgence needs to be like this. Suppose there is no limit to how well-off a person can be, so that, for any life they actually have, they could have had instead an arbitrarily better life. And suppose that, in such a situation, there is nonetheless at least one life God is permitted to give us. We can now construct a case of moral indulgence where we start with L1, a fantastically good life which God is permitted to give us, then consider L2, a life which is like L1 except that it contains (i) some horrible pointless suffering and (ii) offsetting goods so that we’re better off than in L1, and finally consider L3, which is like L2 except without the pointless suffering. Can God give us L2? If so, this would be hugely significant. Just like the claim that W2 and W5 are permissible, it would seem to undermine any version of the problem of evil which appeals to the claim that God would act wrongly in allowing evil. After all, for any gratuitous evil we observe, it might, for all we know, be offset for that person by some heavenly bonus. And the potential implications may not be as disturbing as those which follow from saying that W2 and W5 are permissible. We might suffer pointless evils, but we might at least be assured of not pointlessly being denied an existence which is extremely good on balance. On the Kantian and contractualist views, whether God is permitted to give us L2 will depend on whether we could, respectively, rationally accept or not reasonably reject God’s giving us L2 when we were fine with God’s giving us L1. Whether this is so will depend on whether we have reason to care about harms apart from their impact on our overall well-being (since, with L2, we are harmed but better off overall). One might suppose that we do. Return to the EverBetter Wine. Suppose someone else needs to decide when to open the wine and give it to me. One might think that, intuitively, it would be rational for me to accept their opening the wine now, and rational to accept their opening it a month from today, but not rational to accept their opening it a month from today while slamming my hand in a door, even if I’m better off in the final scenario than in the first. On the other hand, it might also seem somewhat odd in itself to think that I should care about harms apart from their impact on my overall well-being, and perhaps there is a way to argue that our intuitions about door slamming and similar

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

94  Rebecca Chan and Dustin Crummett things aren’t appropriately responsive to the bizarre features of the case. We don’t think we have a decisive way to settle this issue here. However, given its potentially great significance, it deserves further attention.19

Bibliography Adams, Marilyn McCord (2017), “A Modest Proposal? Caveat Emptor! Moral Theory and Problems of Evil,” in Ethics and the Problem of Evil, ed. James Sterba, Indiana University Press, pp. 9–26. Alexander, Scott (2015a), “Vegetarianism for Meat-eaters”, Slate Star Codex. http://slatestarcodex.com/2015/09/23/vegetarianism-for-meat-eaters/. (accessed August 19, 2018). Alexander, Scott (2015b), “Ethics Offsets,” Slate Star Codex. http://slatestarcodex.com/2015/01/04/ethics-offsets/ (accessed August 19, 2018). Alexander, Scott (2017), “Contra Askell on Moral Offsetting,” Slate Star Codex. http://slatestarcodex.com/2017/08/28/contra-askell-on-moral-offsets/ (accessed August 19, 2018). Animal Charity Evaluators (2017), “The Humane League.” https:// animalcharityevaluators.org/charity-review/the-humane-league/#rf1-12708 (accessed March 15, 2018). Climenhaga, Nevin (2017), “Infinite Value and the Best of all Possible Worlds,” Philosophy and Phenomenological Research, 97(2), pp. 367–92. Crummett, Dustin (forthcoming), “Expression and Indication in Ethics and Political Philosophy,” Res Publica. Herman, Barbara (2011), “A Mismatch of Methods,” in On What Matters, Vol. 2, ed. Samuel Scheffler, Oxford University Press, pp. 83–115. Hooker, Brad (2000), Ideal Code, Real World, Oxford University Press. Kagan, Shelly (1989), The Limits of Morality, Oxford University Press. Kant, Immanuel (1997), Groundwork of the Metaphysics of Morals, trans. and ed. Mary Gregor, Cambridge University Press. Kant, Immanuel (2001), Lectures on Ethics, trans. Peter Heath, ed. Peter Heath and J. B. Schneedwind, Cambridge University Press. Kierland, Brian and Philip Swenson (2013), “Ability-Based Objections to No-Best-Worlds Arguments,” Philosophical Studies, 164: 669–83. 19  We are thankful to the participants at the 2017 Theistic Ethics Workshop, Philip Swenson, and an anonymous referee for helping us improve and sharpen the ideas in this paper. We also owe special thanks to Father Raphael Mary Salzillo for preventing embarrassingly false things from being said about the Catholic practice that inspired the title of this paper.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Moral Indulgences: When Offsetting is Wrong  95 Korsgaard, Christine (1985), “Kant’s Formula of Universal Law,” Pacific Philosophical Quarterly 66(1–2): 24–47. Korsgaard, Christine (2004), Fellow Creatures: Kantian Ethics and Our Duties to Animals, Tanner Lectures on Human Values. Kraay, Klaas (2010), “Theism, Possible Worlds, and the Multiverse,” Philosophical Studies 147(3): 355–68. McConnell, F.  J. (2003), “Principle of Double Effect,” in New Catholic Encyclopedia, Vol. 4, The Catholic University of America Press. McMahan, Jeff (2001), The Ethics of Killing, Oxford University Press. Murphy, Mark (2017), God’s Own Ethics: Norms of Divine Agency and the Argument from Evil, Oxford University Press. Parfit, Derek (1984), Reasons and Persons, Oxford University Press. Parfit, Derek (2011), On What Matters, Vol. 1, Oxford University Press. Pollock, John (1983), “How do you Maximize Expectation Value?” Noûs 17(3): 409–21. Rawls, John (1971), A Theory of Justice, Harvard University Press. Rosen, Gideon, (2010), “Might Kantian Contractualism be the Supreme Principle of Morality?”, in Essays on Derek Parfit’s On What Matters, ed. Jussi Suikkanen and John Cottingham, Wiley-Blackwell, pp. 97–115. Rowe, William (2004), Can God be Free?, Clarendon Press. Rowlands, Mark (1997), “Contractarianism and Animal Rights”, Journal of Applied Philosophy, 14: 235–47. Sandel, Michael (2012), What Money Can’t Buy: The Moral Limits of Markets, Farrar, Straus, and Giroux. Scanlon, T. M. (1998), What We Owe to Each Other, Harvard University Press. Scanlon, T. M. (2011), “How I am not a Kantian,” in On What Matters, Vol. 2, ed. Samuel Scheffler, Oxford University Press, pp. 116–40. Singer, Peter (1993), Practical Ethics, Cambridge University Press. Wallace, R. Jay (2013), “The Deontic Structure of Morality,” in Thinking About Reasons: Themes from the Philosophy of Jonathan Dancy, ed. David Bakhurst, Brad Hooker, and Margaret Olivia Little, Oxford University Press, pp. 137–67. Walzer, Michael (1977), Just and Unjust Wars, Basic Books. Wolf, Susan (2011), “Hiking the Range,” in On What Matters, Vol. 2, ed. Samuel Scheffler, Oxford University Press, pp. 33–57. Wood, Allen (2011), “Humanity as an End in Itself,” in On What Matters, Vol. 2, ed. Samuel Scheffler, Oxford University Press, pp. 156–68.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

6

Time of Trial Frances Howard-Snyder

Some philosophers have argued that there are tragic choices, situations in which every choice available to an agent is wrong. For example, Agamemnon has to either abandon his army or kill his daughter. In the movie Sophie’s Choice, Sophie is forced to choose which of her two children will be killed by the Nazis. If an agent acts wrongly, he or she is somehow tainted by the act and fails to be morally perfect. So, for such agents, moral taint is inescapable. Recently, some other philosophers have argued that value and modality might be such that even God, an omnipotent being, might be forced into a tragic choice, and hence fail to be morally perfect. In my story, “Time of Trial,” I imagine a young man, Seth, who, through no fault of his own, is forced into an intolerably difficult moral situation. While some tragic situations force an agent to choose between two equally awful choices (for example, Sophie’s choice), some others force an agent to  choose between two options that are both intolerably awful, but one of which is clearly better than the other. For example, the option of one of Sophie’s children dying is clearly better than both dying. Seth is faced with the second sort of choice—between allowing a child to be killed, on the one hand, and burning a cat alive himself, on the other. The death of a child is clearly worse than the death of a cat. Seth regards the burning and killing of the cat as a totally unacceptable thing to do. His tormentors clearly view this action as an action that will get his hands dirty, i.e. corrupt him. Nevertheless, he does choose to do it, but only after offering himself as a sacrifice instead. What I want to show is that his moral integrity is unblemished by his action, because of his justified, well-supported belief that the only alternative to his action is the death of a young child. Although he feels deeply ashamed, he is not morally vicious for having done this. Obviously, the reader may disagree.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Time of Trial  97

Time of Trial Lily dangled a piece of straw in front of her cat’s nose and laughed when Coal batted it. She barely noticed the smell of old sheep—wet wool and pee—but shivered at the wind that knifed through gaps in the planks of the barn’s walls. A low metallic groan outside: a bus, the high school bus. Lily heard loud male voices and then the barn door banging open. “Sh,” she whispered and pulled the cat into a tight embrace. If they kept quiet, she could watch without being noticed. The boys wouldn’t think to look up towards the loft. Lily inspected her brother’s new companion with interest. A boy with a real haircut and ironed pants, a boy who must have a mother. Not the sort Mike typically spent time with. “We have to make the bridge out of these bits of balsa wood, and nothing else?” Mike asked. “Nothing except glue,” the other boy said. “And how much does it have to hold again?” “Twenty-five pounds.” “Geez, that’s almost as much as my little sister. You sure of that, Seth?” Seth laughed. “Mr. Grisham said it can be done. Lots of kids do it. Hey I found some YouTube videos that show how. Here . . .” He pulled a phone out of the back pocket of his clean khaki pants. “Damn! I’m out of cellular data. You get WiFi here?” “No!” Mike sounded a little annoyed, like he didn’t want the rich boy to think they were poor. “I’ll have to use your landline later,” Seth pushed the phone back into his pocket. “I have to call my mom for a ride. Anyway . . .” He started describing the videos he’d seen. “They said the key was tresses over arches.” From the way Mike moved around, rubbing the toe of his boot in the straw Lily could tell he was irritated. But Seth had no idea. He probably didn’t have brothers. She wondered whether the teacher had assigned them to work together or whether Mike had picked Seth. The two had nothing in common. Mike had pimples and greasy hair and iron muscles—while Seth had smooth skin, shiny hair, and a little puppy fat. They started working. The details bored Lily so she stretched out and tried to spread Coal’s furry body across her chest for warmth. She could go inside but she didn’t want the boys to know she’d been spying on them.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

98  Frances Howard-Snyder To pass the time, she told herself the story about running away with Coal over the mountains chased by the bad guys, to freedom in a new land. In the story Coal kept wondering off chasing insects and Lily had to carry her most of the way. She was heavy but Lily had promised never to leave her. She was all the girl had. Lily noticed a trail of white smoke across the grey-white sky through the hole in the roof. Her older brothers were burning trash, banging about and hollering. Seth sniffed the air and made a disgusted face—probably at the smell of burning rubber. Mike shoved the bridge with his finger. “When do we get to test this thing out?” “We have to wait till the glue dries,” Seth said, with a laugh like this should have been obvious. “I have an idea. Let’s go to my place. We can play Call of Duty and get a snack. My mom always bakes cookies on Thursdays. Today she’s making ginger snaps and oatmeal raisin.” Lily wanted to ask if she could come too, but she knew the boys wouldn’t want a nine-year-old girl tagging along. Mike shook his head. “I don’t like cookies.” The way he said it made Lily’s muscles get tight. “You’re kidding,” Seth said. “And I like mothers even less.” Careful, she wanted to warn Seth. Mike’s weird about mothers—especially since our mother left with the vacuum salesman. Got sucked up in a romance, the boys joked. But they didn’t like other people talking about mothers. “Oh, come on Mike. Some mothers are all right. My mom for instance—” Mike turned fast and knocked the bridge off the bench. There was no sound of breaking. Maybe it was light enough or maybe the straw cushioned it. “Hey,” the boy protested. “That was a dick move.” He bent down but Mike had jammed his boot into the little bridge. This time she heard the wood splintering. Seth stood up and punched Mike lightly on the arm. It couldn’t have hurt more than a cat paw swatting your finger. But Mike reacted like he’d been attacked, snarling and swinging his fist into Seth’s gut. Seth grunted. Coal struggled but Lily held her tighter. Seth bent over holding his belly. When he looked up he blinked like he was trying not to cry. He backed towards the door and fumbled with the handle behind him. Lily held her breath. Was Mike going to let him go? Mike followed him outside. A deeper voice asked, “What’s going on, Mikey?”

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Time of Trial  99 “This pussy is planning on running home and telling his mama that I whipped him—even though he was the one who started it.” “And you’re going to let him do that?” Then all three brothers and Seth were in the barn. Judd and Garth brought the smell of smoke with them. The soot on their faces along with their beards and big shoulders, made them look like bears. Seth kept moving backwards until he was in the middle of the room. His face was very pale, his eyes moving wildly. The brothers moved around him like wolves. Lily gripped Coal tighter and kissed her ears to keep her from being scared. “Brothers need to stick together,” Judd said. Then he punched Seth. They took turns. Lily heard the thuds and the groans and the crying. After a bit Seth was on the dirty floor and they were kicking him—in the gut at first and then when he rolled up to protect his nuts, in the back. One time when they took a break, Seth sat up and tried arguing with them. “My parents can pay you.” “You’ll get in big trouble if you don’t stop.” And then he was just pleading. “Please, please, stop. You’re hurting me.” Lily looked up at the broken roof. The sky had changed color, slowly like a teabag in milky water. She needed to pee but she didn’t want to let them know she was there. When they got into the hitting game they sometimes didn’t know how to stop. “Mike! Judd! Garth! Lily!” an adult voice called from outside. “It’s 5:30. Time for chores and dinner.” Their Uncle Ted lived with them and was in charge, while their father was away visiting his sick mother. The brothers stopped what they were doing and looked at each other. They couldn’t just ignore their uncle. But they couldn’t leave Seth either. Let him go, Lily prayed silently. Please let him go. He would drag himself home and his mom would wash his face and cover it with kisses like on TV and let him sleep in clean sheets and maybe sit by his bed and watch over him. Then she’d call the police and make trouble for Mike and Judd and Garth. Lily would be okay with that. They deserved trouble. But they would not be okay with it. “He’s hurt pretty bad,” Garth said. “No shit, Garth,” Judd said. They were always telling Garth he was stupid. “He’ll run to the pigs if we let him go.” Well, no shit, Judd, Lily wanted to say, you shoulda thought of that before you broke his teeth and his nose. But she kept her mouth shut. “I won’t,” Seth said. His words were all blurry. “I swear I’ll keep quiet. If anyone asks I’ll tell them I was in a fight. And I started it.”

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

100  Frances Howard-Snyder Mike laughed. “You probably figured out that Garth here isn’t the sharpest knife in the drawer, but you know me. I’m in physics, just like you. I’m no dummy.” Seth lifted his head and pulled back his shoulders. “I understand why you hit me. And I forgive you.” Mike cocked his head and moved around Seth slowly, sort of inspecting him, and then he started laughing, a mean coyote laugh like you hear sometimes when the sheep are out for the night. Just then the door opened. In the light from outside Lily could see Uncle Ted’s shape. “Hey, you little fuckers. I don’t call twice. Get your skinny asses . . .” He paused. “What’s this?” He flicked the switch and the single bulb lit up. Seth looked worse even than you would have guessed. Uncle Ted came into the barn, chest first. Judd stepped forward like this wasn’t his problem, like he and Ted were two adults dealing with a couple of naughty kids. “Mike’s really fucked up this time.” Uncle Ted looked back and forth between Mike and Seth. He and Mike were about the same size, shorter than the other men in the family, thin, but with muscles like chains. Uncle Ted had scars where Mike still had pimples. He stared at Mike and moved his fingers, curling and uncurling like they were itching for a fight. “He said this place stinks,” Mike said. “He made fun of me for not having a mom.” Ted moved in front of Seth. “Is that true?” Seth swallowed and shook his head. “No, sir.” “He thinks he’s better than us,” Judd said. “Yeah, he thinks he’s a saint or something,” Mike said, speaking slowly like the words were very strange. “He said he would forgive us.” Uncle Ted scratched his head. “Maybe we can cure him of that and persuade him not to run to the police at the same time . . .Kill two birds with one stone.” All four boys waited to hear his idea. He took his time telling it. “I bet that shirt was clean when he came in here,” he said softly. “It’s easy to get dirty,” Uncle Ted said. “Isn’t it, Mikey?” Mike shrugged. “You just gotta get his soul as dirty as his shirt and then he won’t be telling no tales and he won’t be thinking himself better than you. You gotta get him to do something really bad and then he’ll have to shut his piehole about the rest of it.”

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Time of Trial  101 Mike laughed. Then the boys were all talking at once about bad things they could make Seth do—like kill someone or rape an old lady. Lily shivered and squeezed Coal so tight she bit her and took off, scrambling down the ladder. She was headed for the door but Uncle Ted scooped her up faster than an eagle and held her dangling by the skin of her neck. “Lovely as those ideas are, boys. I have no interest in going back to jail. Killing and raping are felonies.” “But it wouldn’t be on us. It would all be on him,” Mike said, jerking his head in Seth’s direction. “Sure, but the county prosecutor has a way of putting it on guys like me. No, let’s not do any killing or raping—least not of human beings. But a cat . . .” Here he held Coal up to the light. She wriggled, but he had a tight grip on her. “No one cares about a dumb animal.” Mike shrugged like he was disappointed. “Get that burn bin from outside,” Uncle Ted said to Garth, who set off without question. When the metal container was standing in the middle of the barn, Uncle Ted tossed Coal into it. She meowed and scratched the sides in a desperate effort to get out. They’re going to burn my cat! Lily realized. “No!” she screamed and flung herself down the ladder and towards the bin. If she could knock it over maybe Coal could escape between their legs. Even if Lily got a whipping, it would be worth it. But Mike tripped her up. Her knee hit the bare earth hard. “Stay!” Uncle Ted said without looking at her. Then he went into the far corner of the barn and came back with a rusty can. He swung it back and forth to make a sloshing sound. He held out his hand to Seth and pulled him to his feet. “Here,” Ted pushed the can towards him. “Pour it on the cat and then set it on fire.” “No,” Seth said. Uncle Ted laughed. You didn’t say ‘no’ to him. He hit Seth high in the chest and they heard a crack. Seth staggered back and started crying. Uncle Ted offered him the can again. They all waited. Then Seth forced himself to stand up straight. “No,” he said louder this time—like some sort of TV hero, like he was going to take whatever pain Ted gave him and just refuse to do the terrible thing Ted wanted him to do. Mike hit him and then kicked him when he fell. “Stop,” Ted told him. “You’ll just kill him and then where’ll we be?”

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

102  Frances Howard-Snyder Seth struggled to his feet. If Lily had been a little older she would have been in love with him. He wasn’t going to hurt her cat no matter what they did to him. There was no way they could make him. “I have a better idea,” Uncle Ted said. He stepped over, grabbed Lily’s arm, lifted her and then shoved her against the wooden pole that held up the roof. “Get some rope Garth. There’s got to be some around here somewhere.” Garth shuffled his feet. “What you planning, Uncle Ted? I thought you said no people get killed.” “Shut up and do what you’re told. I’m pretty sure no people are getting killed tonight.” Garth found two ropes and he and Mike used them to tie the girl to the pole—at the ankles and chest. Ted held out the can of gas to Seth again. “Cat or kid. Your choice.” Seth met his eye. “Your own niece. You’d burn your own niece?” Garth wailed, “No Uncle Ted, we can’t hurt her. She’s just a little baby.” Judd said, “Shut up. You idjit. This saint won’t let a little kid burn. She’ll be fine.” “But what if he decides he’s too good to do it? What if he leaves it all on us?” Uncle Ted came in close to Seth, shoulder to shoulder, breathing in his face. “What’s it to be, huh? Cat or kid? Keep your hands clean or get them dirty?” Seth took a step back, looking at Lily and then at the cat. Lily was breathing slowly, panting almost, with her mouth open like an old dog on a hot day. They were talking about killing her. Seth was breathing in the same way, slow and heavy like he couldn’t believe the horror of this. Lily struggled against the ropes. But the ropes just hurt her shoulders. Seth shook his head. “What?” Ted asked. “‘Both’ you say. I’m taking that head shake for a ‘both’…unless you set me straight.” Uncle Ted unscrewed the lid from the can and held it over Lily’s head and then tipped it slightly. She could smell the gas. He was going to pour gas on her and set it on fire. She’d burned her fingers on the stove before and knew how much burns hurt. These burns would be all over her body. “No!” She started to cry and thrash about, pleading with him. Then she felt the pee come out of her, hot on her leg and down to her shoe. Garth giggled a little. But this was nothing to worry about by comparison with what they were planning. “Don’t hurt her!” Seth yelled. Ted held the can towards him. “Your choice.”

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Time of Trial  103 Seth shook his head again. Ted brought the can up again over Lily’s head. This time he splashed a couple of drops on her hair and fiddled in his shirt pocket and brought out a little Bic lighter. Seth swallowed. “You won’t do it. I know you won’t. You won’t set a kid on fire.” It was like a game of chicken. No one wanted to be the one who burned something alive. But people got killed playing chicken. Lily and her brothers had a cousin that died like that. Mike laughed, a deep, not funny laugh. He held up his arm and pulled down his sleeve, revealing a long bubble-gum smear of a scar. “You think he won’t do it, do you? You don’t know him like I do.” Lily had never known how that scar happened. Mike always got angry when she asked but she figured that it must have been Uncle Ted’s doing. She’d seen enough of Uncle Ted’s rages to know that he could throw hot oil on a kid’s arm or push a kid’s head under water for a minute or hang a kid out of the window of a moving car, just for the thrill of it. She’d never seen him kill anyone though. Maybe he would just burn her a bit and then stop. She started crying again. “You said you didn’t want to go back to prison,” Seth said. “Oh yeah, I said that. And I never change my mind, do I boys?” He looked around at Judd and Mike and laughed hard. Seth put his hand out and took the gas and the lighter. He put his other hand over his mouth like he was going to throw up. You could see that he was not a boy who would torture a cat and not a boy who would let a girl be murdered. Now he was forced to be one or other of those boys and he couldn’t bear it. “This is crazy,” he said to Ted. “I have money—nearly $300 saved up in my bank account. I could run over and draw it out for you. That would do you more good than a dead cat.” Uncle Ted said, “No. This is my final offer. Take it or leave it.” Seth closed his eyes, and when he opened them, said, “If you like pain.” He was breathing hard. “You could hurt me instead.” Uncle Ted laughed. “I could hurt you as well, you mean. That’s true and maybe I will. But right now, what I want is to see you burn that kitty.” Seth bent his head and closed his eyes. “What? You praying now? Praying that God will come down and strike us dead?” Garth moved uneasily. “Maybe we should leave this be, Uncle Ted.” “Shut your mouth, you big pussy,” Mike muttered.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

104  Frances Howard-Snyder Uncle Ted looked at them both and chuckled. “God ain’t coming, Garthy. Don’t worry. I thought he might come before—I even called on him once or twice myself, but He never came, not to help or to hinder.” Seth didn’t move. Would he just be unable to be the person who burned a cat? Would he leave it all on Ted’s shoulders? Would he let Lily be burned alive? “Please Seth,” she screamed. Seth looked at her. Their eyes met. She hated the thought of her cat burning but she had to let him know that it was better than herself burning. Seth walked closer to the metal container, slow as a zombie. Maybe something would happen at the last moment, the way it did in movies. Uncle Ted made a threatening move and Seth lifted the can and started pouring. Coal thrashed about and banged the sides of her prison. Mike put his boot against the barrel to steady it. Then Seth lit the lighter. He was about the throw it in when Ted stopped him. “Here. Put a little trail of gas down the inside of the bin, and light that. Otherwise the lighter might go out before it hits the gas.” Seth obeyed. In less than a second there was a whoosh of flame and a horrible screeching. Seth covered his ears and closed his eyes. Lily couldn’t cover her ears but she set up a screaming of her own so she wouldn’t hear Coal. Then the screeching stopped. Ted looked in the burn bin and laughed. “All right,” he said to Seth. “You can go now.” “What if he tells?” Mike asked. “What’s to tell? Two kids had a fight. One of them burned a cat. Waste of police time to investigate that.” “You’ll let her go?” Seth asked, pointing at me. “My little darling niece? Of course. I’d never let any harm come to her.” He untied the ropes and patted Lily’s head. His hands felt like snakes and scorpions but she didn’t flinch. She imagined Seth behind her limping away in the other direction. She loved him and she hated him but she knew better than to pay him any attention in front of the others. Inside the house she said she just wanted milk and crackers and to go to bed early. Her brothers watched her in a way that made her think they felt bad about scaring her and killing her cat. They wouldn’t bother her for a while. She climbed the stairs in a daze, wishing she could sleep forever and get the sound and smell of that killing out of her head. She hated Uncle Ted and Mike for making Seth do it and she hated Seth.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Time of Trial  105 But that wasn’t fair. Did she wish Seth had let her be burned instead? Uncle Ted made out afterwards that he was just kidding. So if Seth had refused, nobody would have died. But Uncle Ted had seemed so sure he would do it. How could you tell? How could Seth have told? Seth tried to get Uncle Ted to hurt him instead. That was really brave when you think about it. What would Lily have done in his place? She remembered what she had screamed at him. She’d told him to do it. She couldn’t hate him. She suddenly thought of him walking through the fields in the dark, with all those broken bones, terrified and hating himself. Maybe he’d just fall down and not be able to get up. Maybe he would be so ashamed and sad that he’d just lie there and die of cold. She lay back on her bed and held her pillow against her chest and petted its back and told herself a story about climbing out of her window with a blanket and walking down the dirt road through their property. Seth would have a big head start but she would be moving faster. She would find him and tell him she forgave him and that he was still a good person. But that didn’t feel right. It wasn’t enough to lay back and think about being brave, not this time. That was what a coward would do. She needed to get out and find him. She opened her sash window quiet as she could and listened. She heard Mike and Garth arguing downstairs. Their sounds didn’t change. They hadn’t heard her. Then she pushed out the torn screen and carefully placed it on her floor. She stuck her head out, listening for night sounds. She could hear crickets and frogs but no coyotes. The moon was almost full—good for seeing your way, not so good if you didn’t want to be seen. Her mind was a muddle. Uncle Ted will smack me around if he knows I snuck out. But Seth needs me. Maybe his mom will give me cookies if I help him. Maybe I’ll never find him. Maybe I’ll get eaten by some critter in the night. All these fears and hopes banged around in her head like balls in a pinball machine. Then she heard the wail and saw the gleam of red and blue in the night. Seth’s mother must have called the police; the police must have called the school and found out who Seth had been assigned to work with. They were here now to save Seth and to make trouble for Lily’s family. Lily knew where she needed to be. She climbed out the window and down the tree and started running on bare feet across the cold ground towards the wails and the lights.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

7

Why Did The One Not Remain Within Itself? Mark Johnston

Nothing is added to God’s goodness from [creatures], just as a point adds nothing to a line. Cardinal du Four, Quaestiones De Rerum Principio, question 4, article i, section 3.1 [It] is necessary to undertake a serious examination of the latter question concerning the truth of the transfinitum, for were I correct in asserting its truth . . . then there would be (without doubt) a certain danger of religious error for those of the opposite opinion, since: error circa creaturas redundat in falsam de Deo scientiam (Summa Contra Gentiles; II, 3). Georg Cantor, letter to Fr. Ignatius Jeiler2 The transfinitum, with all its richness of shapes and forms, points with necessity toward an Absolute, toward the true infinite (das wahrhaft Unendliche). This Absolute allows for no addition or subtraction to its quality and must therefore be seen as an absolute maximum. It transcends to a certain degree the human ability to comprehend and stands in particular beyond mathematical determination. Georg Cantor, Gesammelte Abhandlungen mathematischen und philosophischen Inhalts.3

1 The quotation is taken from William  E.  Mann’s translation to be found in Being and Goodness, Scott MacDonald (ed.) (Ithaca, NY: Cornell University Press, 1991), pp. 276–7. 2 G. Cantor, Nachlass VI, p. 169. 3  E. Zermelo (ed.) (Berlin: Springer, 1932), p. 405.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  107

I. Introduction Concerning the aboriginal explanatory ground of all things, Plotinus articulated the deepest enigma: “Why did the One not remain within itself?” (Enneads V, 1 [10] 6, 6). It is a question one would consider asking only if one thought that reason had a chance of reaching down into the ground of things, and supposed that the ground of things might be sufficiently unified to be an intelligible object of inquiry. Even granting such a vigorously confident rationalism, there is some obvious appeal in a broadly Spinozistic deflation of Plotinus’s question to this effect: “No, no; it did remain within itself!” For the Spinozist, Plotinus’s question has a false presupposition. Everything other than the One is an attribute or mode of the One, directly or indirectly following from its essence. What we see around us is then just an infinitesimal sliver of the infinite variegation of the Unique Substance, which Spinoza titles “Deus sive Natura.” Since that Unique Substance could not but exist, and the sort of following from in question is not contingent, necessitarianism is entailed: Everything exists necessarily, and must be as it is, and nothing else could have been. In keeping with Books 4 and 5 of the Ethics, Spinozists have tried to build necessitarianism into a form of consolation, perhaps even salvation; but for most of us, necessitarianism remains a stultifying conceit. Some who find the necessitarian answer stultifying may still feel the intellectual appeal of an ultimate ontological explanation in terms of some sort of fundamental ground. What we wanted is an answer to this question: “How did the One give rise to the contingent world we see around us, a world whose contingencies we can add to, as a result of our own acts of free-will?” Here, the traditional theistic answer seems to have been made by a bespoke tailor. The One, whom we know of as God, gave rise to the contingent world around us by an aboriginal act of libertarian free choice, which we know of as creation. The bespoke tailoring gives us something that is at the same time blatantly anthropomorphic and intellectually satisfying: The fundamental contingency of reality arises from the same kind of source we seem to find in ourselves, namely genuinely free choice. What should be made of that? Our finding it intellectually satisfying or “intuitive” is, of course, predicted by the fairly uncontroversial core of the psychology of religion, with its empirically justified emphasis on our natural disposition to

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

108  Mark Johnston favor animistic explanations. Yet, once we set the genetic fallacy aside, and accordingly distinguish causes and reasons, the question remains: Does the idea, in some form or other, have any chance of being true? And what would follow if it were true? A theist need not deny the charge of anthropomorphism, so long as it is distinguished from the charge of anthropocentrism, with its Feuerbachian overtones of the projection of human features onto a supposed divinity. One long-established strand in theism is the “Imago Dei” doctrine, namely that we, or at least our rational wills, were created in the image and likeness of God, where this includes sharing in his capacity for libertarian free choice. So the anthropomorphic character of the theistic model is not a bug, but a feature. However, even if we grant the prima facie coherence of the theistic form of ultimate explanation, a version of Plotinus’s question immediately arises: Why did God not leave well enough alone? Things were already perfect, were they not? Why mess with that? What could possibly have led God to create? What objective (or normative) reasons were the considerations which “inclined without necessitating” God’s rational will in favor of creating? In the absence of some kind of answer to this question, the theology of traditional theism presents us with a mere placeholder for a certain kind of ultimate explanation, an explanation which it does not itself deliver. How then should we fill in the placeholder? What was God’s reason for creating?

II.  A Framework For Our Inquiry To get anywhere with the question of why God created, we will have to make some initial assumptions about the nature of God, and simply explore the question of his4 reason for not remaining within himself relative to those assumptions. The assumptions that follow are close to central, and enduring, though sometimes controverted, elements in one traditional theistic conception of  God. These assumptions are pressed into service here because their 4  I understand something of the worry about the pronoun, but the same worry applies to “she” and to “he or she”. It is a limitation of the language that we have no standardized nongendered third person singular personal pronoun. Still, as a second best, anyone who wants to substitute “Ἁγία Σοφία” for “God” throughout the text and follow up the Greek feminine with “she” certainly has my blessing. In fact, God probably provides the best case for “zee.”

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  109 implications are well understood, thanks to a long history of sophisticated thought and commentary. If true, the assumptions express de re necessary truths concerning God and creation. In the background is the basic picture: While God exists necessarily, creation is a contingent operation. The contingency of creation is not mere per se contingency; it is straightforward modal contingency. It is not just that creation’s existence is not guaranteed by the nature of creation considered in itself. There might have been no creation at all. I do not say that what follows is the core or essence of Judeo-Christianity nurtured as it was by Neoplatonism, let alone the core or essence of theism. That kind of claim is extremely problematic, perhaps even insultingly preemptory. Still, we have to work with assumptions, and what follows are simply the assumptions your author has some grip on thanks to his exposure to a Neoplatonically tinged Thomism. I present them, not as the best scholarly understanding of Thomas, but as characterizing the best “God of the Philosophers” that I happen to know. (I am happy to be shown a better one.) Even relative to these assumptions, finding an answer to our question will prove extremely difficult. Without them, or some other set of equally constraining and historically well-understood assumptions, our question is best passed over in silence. My initial thesis is that these assumptions are individually and collectively consistent, and so should be assigned a non-zero epistemic probability. Given that, their collective probability might be raised by the following line of argument. First, the law-like frame of the material universe, insofar as it is revealed by cosmology and quantum field theory, appears contingent. (If instead, one regards this frame as essential to a material universe, then the existence of matter, rather than some other universe-composing stuff, appears contingent.) Second, the demand for an ultimate ontological explanation of contingency—an explanation of contingency that does not represent it as a disguised necessity—is legitimate. Accordingly, the explanatory character of the assumptions, the fact that they account for contingency when developed along the lines set out below, raises their probability, at least in the absence of an alternative. Appended to that familiar line of reasoning is a new idea, developed near the end of the paper. Given that a non-zero probability attaches to these hypotheses, we are practically required to hope that they are true. For otherwise, so the argument will go, it is unclear how we could then actually occupy a significant realm of practical reason, rather than being just desiredriven things who rationalize our impulses.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

110  Mark Johnston Here, then, are the assumptions: 1. God is Subsisting Existence, i.e., the Form or Pre-eminent Exemplar of Existence, with respect to which all other existents are derivative participants. This account of what God is has consequences for what is properly predicated of him: It lies in his nature or essence to exist.5 So the question of the ground of his existence, the question of why it is that a thing with his nature or essence exists, does not arise. (Which is not to say that we have, in the fashion of the ontological argument, an a priori basis for asserting the existence of God. Instead, the situation is this: If God, so conceived, exists then he is an autonomous existent, i.e., fact of his existence does not require a ground.) 2. God is the Good,6 the Form or Pre-eminent Exemplar of Goodness, with respect to which everything else that is good is a derivative participant. This account of what God is has consequences concerning what is properly predicated of him and of ensembles, i.e., wholes or mere pluralities, involving him.7 He has by his essence every positive value or perfection it is possible for him to have simply (i.e., not in virtue of some relation to other things) and he has these to a degree that is unsurpassable. Moreover, his degree of goodness taken as a whole is unsurpassable, and not just unsurpassable by him; no other thing or plurality of things could have a degree of goodness that surpassed it. (In what follows, what mostly will matter is the predication of unsurpassable infinite goodness of God, and not the Neoplatonic identity statement expressed at the head of this paragraph.) 3. God is Power Itself, the Form or Pre-eminent Exemplar of Power; thus he is able to do anything ontologically possible, i.e., anything consistent with the consequences of his essence and the essences of other things. (Of course, the contemporary philosophical conception 5  Part of the thought is that he can’t simply satisfy the predicate “exists” because of something not entailed by his nature. 6  Part of this Neoplatonic thought is that he can’t just be predicatively good, as if the standards of goodness and badness are not grounded in him but stand over and above him and just give him the highest possible mark. For then as a rational will his choices would be prescribed by what is good or bad; he would be under an axiological constraint that does not derive from his nature. 7  Forms are particulars not properties, and the transition from the identification of God with the unsurpassable infinite good to the predication of the corresponding property of him is not trivially guaranteed. But there are various intermediate premises embedded in the tradition that might underwrite it. For example: To be good to some degree is to participate to that degree in goodness itself, where identity is the highest degree of participation.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  111 of what is ontologically possible may be transformed by insight into God’s reason for creating. For example, arbitrary segmentations of possible creations need not be possible creations. A creation that includes rocks is an ontological possibility, since it is actual; but not a creation consisting of just one of those rocks.) 4. God is Knowing Itself, the Form or Pre-eminent Exemplar of Knowing; so that, for example, his modal knowledge includes ­knowledge of all realizable possibilities, and he foreknows, or better “overknows”—knows from the perspective of eternity—all contingent outcomes, even those for which his will is not the ground or source. 5. God is Rational Willing Itself, the Form or Pre-eminent Exemplar of Rational Willing, and so his rational will is perfectly responsive to reason and value. 6. Since God is not fundamentally in any limiting medium such as time, his “acts” of rational willing are not successive, but are aspects of a  single comprehensive act of will, whose content covers all eventualities. 7. Creation was ex nihilo; it was not some operation on preexisting materials, whose natures placed an antecedent limitation on God’s will. (So, in particular, it did not consist in “making concrete” already existing necessary beings, as on the view that is now known as “necessitism.”8) 8. God’s creating was contingent; there could have been no creation at all.9 Whether they are true or false, those eight assumptions are consistent (or so I would, and partly will, argue). Further, if such a God existed then

8  For a recent defense of necessitism, see Timothy Williamson, Modal Logic as Metaphysics (Oxford: Oxford University Press, 2013). A necessitist need not be a necessitarian, for she may hold that the necessary beings that figure in the fixed domain of modal logic have some contingent properties. However, if the trope theory correctly describes the ground for the truth of predications then there is an argument from the thesis that every particular is necessary to necessitarianism. Of course, if God exists then the whole structure of modality is different. If it is not true that something satisfying assumptions 1 through 8 could not exist, how can modal logic be the guide to modal metaphysics? 9  This is the conviction Aquinas aims to defend in Book 1 of the Summa Contra Gentiles, Chapters 75–86. Still, there is an undeniable tendency, shown in those very chapters, toward implying that creating something or other is necessary for God. As I try to show, in order to explicate Aquinas’s best insights, certain thoughts from Cantor are helpful, and lead us away from the idea that some creation or other is necessary.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

112  Mark Johnston he would be an ideal candidate to be an autonomous ground of everything there is, autonomous in the sense of not needing further explanation or grounding.10,11 Furthermore, if we can consistently tease out from these assumptions an account of just how, in creating, God had a libertarian free  choice, a choice not even coerced by the force of the reason to create, then we would have a model of fundamental ontological explanation in which such an explanation does not imply even the conditional necessity of what it explains. If the idea of God having a completely adequate but non-coercive reason to create represents a genuine conceptual possibility, then sense can be made of the idea that there are sufficiently explanatory reasons for other things being as they are, while these other things (or all of the propositions to the effect that they exist) are not necessitated by those reasons and are not necessarily as they are. This opens up a space in which there is an attractive version of the Principle of Sufficient Reason—one that does not require the necessity of God’s creating, and so leaves creation contingent, although adequately explained. Admittedly, the attraction will be felt only by those who hope that reason gets a foothold at the base of things, a limited audience to be sure. The version of the Principle of Sufficient Reason I have in mind is this: The basic idea  All non-autonomous facts are explicable, in the sense that there must be completely adequate reasons for them.

10  I take this notion of autonomy from Shamik Dasgupta, specifically from his “Metaphysical Rationalism,” Noûs 48 (2014), pp. 379–418. The idea is that some facts, such as the fact that two is even, and some entities, namely those that exist necessarily, do not require grounding or ontological explanation. In that paper, Dasgupta argues that the ultimate explanation of all other facts in terms of autonomous facts entails necessitarianism. The present paper gives a model which shows that there is no such entailment. Does this show that a ground need not necessitate what it grounds? The old talk of ground arose from the PSR, but the new talk of ground is tied to “making it the case” where, if this is not mere causal making, it is necessitating what is made the case. The language does not matter. A theorist should feel free to hold to the modern usage. But she should then admit that there are metaphysical explanations, in particular successful applications of the PSR, which are not grounding explanations. 11  In the very recent discussion of ground, a clear distinction is made between the efficient cause of a particular and that particular’s ground. The existence of {Socrates} is grounded in, but not efficiently caused by, the existence of Socrates. It may seem that in speaking of God’s free choice to create as a contingent ground of finite (and transfinite) existence this distinction is being muddied. Not so. Recall the famous Thomistic taglines: “After creation there are more beings, not more Being” and “God is the to be of things.” The idea is that as a result of contingent creation, existence takes on finite forms. The relation between finite particulars and the creative God is thus not just causal; they participate in his being. Nonetheless they remain finite and contingent, for their finite forms are contingently realized.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  113 The crucial gloss  However, these reasons can take different forms. Sometimes such reasons are found in a necessitating ontological ex­plan­ ation, which would show how derivative things must be as they are, given more fundamental things; but alternatively there can be completely adequate but non-coercive practical reasons for choice and intentional action, reasons which provide a non-necessitating explanation of the choice and the corresponding action via those antecedent reasons.12 As some of my readers will know, as yet I cannot see my way clear to being a traditional theist, with all its usual entailments.13 Why then go to all this trouble? First, to make the purely ontological point that fundamental ontological explanation need not be necessitating; fully reasonable rationalists need not be necessitarians! In today’s jargon, there are ontological ex­plan­ ations that do not necessitate. Second, it may be helpful to expose a blind spot in contemporary theistic philosophy of religion, precisely in respect of appreciating the consequences of the very possibility of an infinitely good God; more exactly, a blind spot concerning God’s available reasons for creating, and its consequences for theodicy, and for what I call the “melioristic” way of life. Third, it will come as a surprise and caution to some that there is a corresponding blind spot in a certain kind of atheism that merely treats it as an obvious contingent fact that the traditional theistic God does not exist, a blind spot which leaves the very idea of a non-religious ethics open to being blindsided by infinitarian considerations. So I hope to show that just what is most deeply at issue between traditional theism and atheism, in particular the form of atheism that embraces reductive naturalism, has not yet come wholly into view. And finally, although I do not have the space to develop this here, important parts of the theistic picture outlined 12  As it happens, this goes against the grain of some very interesting recent revivals within ontology of the Principle of Sufficient Reason, where necessitation is taken as a consequence of ontological explanation. Along with Dasgupta, “Metaphysical Rationalism,” see Martin Lin “Rationalism and Necessitarianism,” Noûs 46 (2012), pp. 418–48, and (forthcoming) “The Principle of Sufficient Reason in Spinoza,” in The Oxford Handbook of Spinoza (Oxford: Oxford University Press). Then there is Michael Della Rocca’s body of work including “Spinoza’s Substance Monism,” in O.  Koistinen and J.  Biro (eds), Spinoza: Ontological Themes (Oxford: Oxford University Press, 2003), and his “PSR,” Philosophers’ Imprint, No. 7 (2010), pp. 1–18, and “A Rationalist Manifesto: Spinoza and the Principle of Sufficient Reason,” Philosophical Topics 31 (2004), pp. 75–93. Alexander Pruss, on the other hand, seems to be holding out for something like the principle as formulated above in his The Principle of Sufficient Reason: A Reassessment (Cambridge: Cambridge University Press, 2006). 13 See Mark Johnston, Saving God: Religion after Idolatry (Princeton, NJ: Princeton University Press, 2011), a self-conscious attempt, not to do philosophy, but rather to give expression to a particular non-traditional religious sensibility.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

114  Mark Johnston below can ultimately be understood as removable scaffolding, which when removed leaves what I take to be the heart of the matter.14

III.  Why Did God Create? The bare voluntarist answer to the effect that God just willed creation without having any reason to create ignores the fact that God’s willing, as opposed to his merely effecting or emanating, is the operation of a completely rational will; a will that is always consonant with the intellect’s appreciation of a completely adequate reason that favors the choice in question. So what then was God’s reason for creating? In asking this question, we are precisely not looking for a rationally coercive reason for creating. For a central aim of the present enterprise is to explain how God, in creating, genuinely had the option to create or not to create. On the other hand, we are not looking merely for prima facie or pro tanto considerations that might favor creation but would not in themselves provide a completely adequate reason to create.15 What we are after is illumination as to the kind(s) of reason(s) which could make up a completely adequate reason to create, something that could make rational sense of creation, even while allowing for the rational option of refraining from creation. 14  One key to unlocking unnecessary scaffolding begins with explaining, and perhaps partly culling, the striking Neoplatonic identities which assumptions 1 through 5 together entail. How can the pre-eminent exemplar of existence also be that of goodness, of power, of knowing, and of rational willing? The other key is the theologically plausible point, taken as uncontroversial by both Maimonides and Aquinas, that when we say that God exists and is good and is powerful and has an intellect and a will we are inevitably talking in a fashion that cries out for further explication. The untenable theological position is that the ground of the truth of these predications is the same as the ground of the truth that human beings exist, and are good, and powerful, and have an intellect and a will. But as to just how these keys might unlock and help discharge much of the ontological scaffolding surrounding traditional theism, that is a story for another occasion. A hint: What can be explained by appeal to a will that is wholly responsive to goodness might also be explained by a developed axiarchy. 15  If one has a prima facie reason to do something then unless the prima facie reason is able to be undermined, i.e., shown not to be a reason at all, that reason is then a pro tanto reason, i.e., one that contributes to a complete case for doing the thing in question. Pro tanto reasons differ from mere prima facie reasons in that they retain their force, even if outweighed, whereas mere prima facie reasons may be undermined and so may make no contribution to a complete objective case for performing the act. A complete case for performing an act may be compatible with having rational options, i.e., there may be other acts in the choice situation such that the agent has a complete case for performing them as well. So complete reasons may not thereby determine the rational will; they need not be “coercive” reasons.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  115 To summarize: If God had a reason to create then it would be a completely adequate reason, a reason that would justify his creating. This reason could justify his creating, even if it was not a coercive reason, i.e., a reason that required him to create. Creation would thus be explained by a free choice of God’s, but it would not thereby be necessitated by God’s nature. How exactly are we to think of a non-coercive but completely adequate reason? The standard exposition begins with some sort of Buridan’s ass case. Say you need a fountain pen and there you are at the pen shop equidistant from two Jacques de Molay mediums, both priced the same and indistinguishable one from the other. You take one of the Jacques de Molay mediums to the counter and pay for it. There was no coercive reason to buy the pen you bought, rather than the other; but there is a completely a­ dequate reason for buying that pen: You needed an affordable fountain pen and this—the one you bought—is an affordable fountain pen. You also had a completely adequate reason for buying the other pen instead; namely, you needed an affordable pen and that—the other one—is an affordable fountain pen. You have adequate reasons for choosing either pen, and you are rationally free to go either way. The case serves to highlight a general distinction between completely adequate reasons and coercive reasons, but this is a distinction with application beyond the restricted class of Buridan’s ass cases, where Buridan’s ass cases are those where the reasons balance because they are tokens of exactly the same type of reason. In the pen example you had a justifying reason for choosing the pen you did, which is shown by the fact that if the other pen had not been in the shop you would have had a coercive reason to choose the pen you did. The justifying reason, namely that it was an affordable and appealing pen, was a completely adequate reason for choosing the pen, given your background needs. That remains true, even though it does not provide a contrastive rational explanation of your choosing the pen you did, rather than the other one. So let us say that completely adequate reasons for an act are justifying reasons that are not outweighed by the reasons for any alternative to the act; whereas coercive reasons for an act are justifying reasons that are weightier than the reasons for any alternative act.16 The interesting issue concerns the 16  Though coercive reasons are, by this definition, a species of completely adequate reasons, when I write of completely adequate reasons without qualification, I shall have in mind reasons that are not coercive.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

116  Mark Johnston variety of ways in which justifying reasons can fail to be outweighed. In particular, are any of these cases in which the adequate reasons provide a contrastive rational explanation for the choice they support? Three other sorts of cases immediately come to mind. There are cases where, within the limits of the agent’s capacity for discrimination, her conflicting desires seem to her equally strong. Insofar as episodic desires provide, or better, disclose, reasons, then cases in which the desires in play conflict but are, within the limits of the agent’s discrimination, “equally strong” may call for a choice, a choice which is not rationally coerced either way but which is nonetheless fully rationalized or made fully intelligible by the desire that favored the choice. So when I find the pâté and the oysters equally appealing, and opt for the oysters, the appealing features of the oysters disclosed by my desire for them provide a completely a­ dequate reason for my choice.17 That the oysters have those desirable features is a completely adequate reason for choosing them; it is justifying, which is shown by the fact that it would have been coercive if neither the pâté nor anything equally desirable had turned up on the menu. And furthermore, so we are supposing, it is not outweighed by any reason for the alternative choice of the pâté. We need not think of this as a case in which the oysters and the pâté promise exactly the same sort of gustatory pleasure; in both cases the pleasure could be “intentional” so that the pleasure taken in eating the oysters and the pleasure taken in eating the pâté need not share a pure hedonic common factor. (Try as I might, I can’t find the hedonic common factor.) If that is true then there seems to be room for a contrastive rational explanation. I went for the oysters rather than the pâté because I anticipated that specific intentional pleasure. Another type of completely adequate but non-coercive reason arises from the fact that sometimes, when a desire provides or discloses a reason, the reason is wholly conditional on the persistence of the desire in question. As a result, the agent is often faced with a choice either to let her desire wane or to act on it. For example, suppose that I am suddenly inhabited by a desire to play three-cushion billiards, a now rare but beautiful game, played on a large, heated table with no pockets. The nearest full-size table in regular use is in New Brunswick, and it takes some trouble to get there. I have a 17  For a development of this account of the conditions of intelligibility of desires as reasons for choice and action see my “The Authority of Affect,” Philosophy and Phenomenological Research 63 (2001), pp. 181–214.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  117 choice. I could use my knowledge of how my desires come and go, and distract myself by working in the garden, in order to allow the urge to play three-cushion billiards to wane. Or I could get in the car, and take the trouble to get to New Brunswick and play. Wanting to avoid the stop-and-start traffic on US Route 27 gives me reason on the side of letting my desire wane, while the desired delights of the anticipated game weigh in on the side of going with my desire. I have a choice, and it may be that neither side presents me with a coercive reason. Still, if I choose to get in the car and head to the game then my reason for that choice—that the game offers the desired delights—fully rationalizes or justifies the choice. The reason is completely adequate, though not coercive, as is shown by two facts taken together: (i) it is justifying, for it would have been coercive if some crucial element in the opposing reason had not held up, e.g., if the trip to New Brunswick was not at all irritating, and (ii) it is not outweighed by the reasons for the alternative choice. Moreover, if I choose to go and play, it would be right to say that I chose to do that rather than to go to the distracting garden because of my desire to play threecushion billiards. There is still another sort of case, more philosophically interesting, in which one might find oneself with completely adequate but not coercive reasons. Some choice situations may present conflicts of incommensurable values, or more generally incomparable values—to drop the implication that it is commensuration or quantitative comparison of the relevant values that is really at issue. These would be values which are of such different kinds that it seems artificial to regard them either as equivalent in weight or rank, or as involving one value that is weightier than, or to be ranked more highly than, the other. E. M. Forster once announced: “If I had to choose between betraying my country and betraying my friend, I hope I should have the guts to betray my country.” But we can imagine degrees of friendship, degrees of national legitimacy and degrees of betrayal all varied on both sides, so that there is an imaginable case where even Forster might have found himself with a genuine ethical dilemma. It may be that such a dilemma was faced by Wilfred Blunt, the brother of the now disgraced Anthony Blunt, the so-called “fourth man,” who had once spied for the Soviets, and who posed some ongoing risk to British national security. It might have been the case that the fact that Wilfred was Anthony’s brother was a completely adequate reason for Wilfred’s not

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

118  Mark Johnston turning him in, while it was also the case that the fact that his brother had betrayed, and might again betray, his country was a completely adequate reason for turning him in. This choice that Wilfred, so we imagine, may have faced would have been agonizing, and rationally so, partly because it was un-coerced by reason. Even having chosen not to betray his brother, there might have remained the worrying thought that he could rationally have gone the other way, and not joined one’s brother in betraying one’s country. This phenomenology of agonized choosing, and perhaps of residual guilt associated with either way of going, is very different from a Buridan’s ass choice, where there is no rational sense of loss with respect to the shunned option, precisely because the shunned option simply offered an equal helping of the very same sort of value as the chosen value. The ass, or the pen-buyer, just “plumps for” the bale of hay or the pen he moves toward; whereas the agent facing a choice in the face of incomparable values can cite a distinctive non-coercive reason for the choice he made. There is no contrastive rational explanation of why the pen-buyer bought the one pen rather than its duplicate. But Wilfred, as we imagine him, can offer a contrastive rational explanation for not turning his bother in, namely “He is my brother.” Creation was obviously not a Buridan’s ass choice, for the value of remaining and the value of creating do not exhibit equal helpings of the very same sort of value. So we shall then have to investigate whether God’s un-coerced rational freedom to create must arise from an “offsetting” reason, i.e., either from the counterbalancing weight of some reason not to create, or from the incomparable values presented by the option of remaining within himself as opposed to the option of creating. I do not deny that it could have arisen from either balancing or incompar­ able values, perhaps in the following way. Suppose, as is argued below, that God’s reason to create was to manifest his own nature, i.e., his goodness, power, intellect, and uncoerced free will. In acting, he always affirms the Good, and his own Good, for he is the Good and there is, consequently, no distinction between his Good and the Good. But any finite manifestation of free rational willing, however perfect, will face a choice of fundamental orientation—a choice God cannot face—namely whether to subordinate its own good to the Good, or to subordinate the Good to its own good. The latter is the choice definitive of a morally evil will. Suppose then that any adequate manifestation of God’s nature involves the creation of free rational wills, and that God can foreknow that in any such manifestation, some of these creatures will make the choice definitive of an evil will. Then God has a reason against creating, namely that in any adequate manifestation of his

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  119 nature moral evil will arise.18 In more familiar terms, the foreknowable Fall might have provided God with a reason against creating, a reason whose weight might have offset the weight of God’s reason to create. In a different vein, Alexander Pruss imagines God as akin to a creative artist who realizes that any manifestation of himself in his chosen medium will be imperfect. Pruss supposes that this anticipatable imperfection itself may provide a non-coercive reason not to create in that medium. Could the appreciation of that sort of imperfection, in God’s case the infinite distance between his own goodness and any finite (or transfinite) manifestation of it, have provided God with an offsetting, i.e., either counterbalancing or incomparable, reason not to create?19 Perhaps it could have, perhaps not. But here we are merely speculating about the offsetting character of reasons as they appear “in the mind of God.” It would be better if we could avoid such speculations in favor of an account of how God could have a non-coercive but adequate reason to create that need not be offset by some reason not to create. For although certain (mostly negative) facts about the mind of God can be deduced from our assumptions, there is no deducing the offsetting potential of various reasons as they present to God. Likewise, to say that the reason to avoid moral evil and the reason to manifest one’s nature are for God incomparable reasons is mere further speculation, not itself constrained by what the Neoplatonic Thomistic tradition already specifies as God’s nature. The serious intellectual demand is to find the minimal speculative basis consistent with our finding ourselves in a workable place within the space of practical reason. (As the personite considerations suggest, Naturalism does not provide that basis.) Instead of speculating about offsetting reasons in either form, I aim to defend this view: The reasons on both sides, the reasons for creating and the reasons for remaining, emerge from the value of God’s affirming his own infinite goodness in this or that manner, and as a result do not differ in weight. The extra reason to create does not produce a more weighty rational case for creating rather than remaining. There is in the infinite case an unprecedented fifth way in which there can be completely adequate but non-coercive reasons. God’s infinite goodness implies that his possessing an extra reason to create, whatever that exactly turns out to be, does not entail his having more of a reason, in the sense of a weightier reason, to create rather than remain within himself. This can be so even if God’s extra reason to create is not offset by a reason to remain within himself! 18  See my “How did Evil Come into the World?” ms. of a talk delivered at Oxford, 2018. 19  Pruss, “The Cosmos as a Work of Art,” ms.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

120  Mark Johnston But what exactly is this extra reason? That will occupy most of our attention. But first, we will have to defend the abstract possibility of God’s having a completely adequate but non-coercive reason, in the face of an aporia or apparent paradox. Then we shall attend to the question of what specific kind of completely adequate but non-coercive reason God in creating could have had.

IV.  Could God Have Had Any Completely Adequate But Non-Coercive Reason To Create? In the next sections I will be writing as if value and practical reason “take in each other’s washing.” Given the values, the weights of various practical reasons can be understood as measuring the comparative adequacy of various forms of rational responsiveness to the values. Perhaps it is also true that something is a value, i.e., is to be valued in some way, if and only if the weight of practical reason comes down in favor of valuing it in that way. In setting out the aporia and the response, we need not assume any priority on the part of reason or value. Here is the aporia or apparent paradox attending the idea that God could have any kind of completely adequate but non-coercive reason to create, as opposed to refraining from creating and simply remaining within himself. If that reason for creating, whatever it might be, is non-coercive then it cannot outweigh the reason God had to remain within himself. For if it did, then given our definition of a coercive reason, God would have a coercive reason to create, and therefore would create by ontological necessity.20 So if, as tradition has it, God was free to remain within himself, there can be no difference in weight between the separate reasons God has for these two acts: (1) Creating (2)  Refraining from creating (i.e., “remaining within himself ”). 20  This last step has been denied; notably by the seventeenth-century Spanish Jesuit, Diego Ruiz de Montoya and by his student Diego Granado. For an eye-opening discussion of these important precursors to Leibniz and his version of the Principle of Sufficient Reason, see Michael J. Murray “Pre-Leibnizian Moral Necessity,” The Leibniz Review 14 (2004), pp. 1–28. As far as I can see, neither Montoya nor Granado have a response to the argument that an essentially perfectly rationally being must go with a coercive reason by ontological necessity. Notice that it is not responsive to this point to maintain that God has a certain kind of freedom qua merely omnipotent being, and that this freedom includes the freedom to ignore coercive reasons. Qua merely rational being, I am not essentially embodied, but so what?

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  121 Now the aporia emerges. In the absence of an offsetting reason against creating, a reason that is either incomparable with the reason to create, or whose weight counterbalances the weight of the reason to create, the rational case for God to create must have been coercive. So, given that God is the supreme rational will, creation must have been necessary. Yet finite reality appears contingent. Therefore conceiving of a contingent creation as the operation of a pre-eminently rational will seems paradoxical, once thought through. It then would follow that there is nothing in the vicinity answering to the idea of an attractive alternative model of fundamental ontological explanation—one where the ultimate “ground” is not necessitating—being illustrated by God’s relation to his choice to create. Much of the ontological interest of our investigation would then have evaporated.

V.  The Solution to the Aporia of Contingent Creation As against this, the theistic tradition embodied in our initial assumptions already contains the resources to address the aporia. It can explain just how God can have a completely adequate but non-coercive reason to create. It can do that without resorting to speculations about either the incomparability or the counterbalancing weight of God’s reasons not to create. This brings with it a second advantage, aside from that of avoiding unconstrained speculation. The reasons against creating gestured at above do not clearly fit the bill. On the face of it, and in the absence of an articulated theodicy, the foreknowledge that evil wills shall arise can look like a coercive reason against creating, whereas the mere comparative imperfection of any creation seems like a less-than-adequate reason against creating, one that would be outweighed by any adequate reason to create. The first traditional resource concerns the non-discursive character of God’s rational will (assumption 6), and the second concerns the relation between his will and his own unsurpassably infinite goodness (assumptions 2 and 5). Since God is not fundamentally in time, his will is not discursive; he does not will this thing for that reason, and later will that thing for this reason, and so on and so forth. The more correct model has him willing all that he wills “at once,” i.e., in one, single, comprehensive act, with its own comprehensive, though partly conditional, content (assumption 6). This, of course, is best understood against the background of God’s comprehensive modal

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

122  Mark Johnston knowledge; his comprehensive knowledge from an eternal point of view of all that could possibly happen. His single act of will covers, though it does not generate, all contingencies, including any further contingencies introduced by his creatures’ acts of will. To offer a debased image, God already has the perfect chess book, anticipating all the moves, to whatever play, of unreliable supporters and less than ultimately effective enemies. His single act of will is thus fully determinate with respect to anything that he alone brings about directly, and only consequentially determinate in respect of any contingency creatures might bring about as a result of the operation of their free wills. So, for example, it includes the salvific work of the Incarnation consequential upon our freely chosen fall from grace. Indeed, his will covers all the myriad ways that branch out from the exercise of human (and angelic) freedom.21 Secondly, God necessarily wills or affirms his own existence cognized as infinite goodness. In our terms, he has a coercive reason to affirm it, since it cannot be equaled by, let alone bettered by, any other good. Indeed, if one were to assign a weight to that reason, it would have to be an unsurpassably infinite weight, if there were such a thing (assumption 2). Taking these two thoughts together, it is then natural to think of God’s willing (affirming) whatever else he wills (affirms) besides his own goodness in adverbial terms. Everything else God wills or could will is just a determinate manner of his willing (affirming) his own goodness, a manner which cannot be merely parsed out as a further willed conjunct there alongside his willing (affirming) his own goodness. So God’s options are not acts (1) and (2) above, not simply creating or refraining from creating, but rather: (3)  Affirming his own goodness in the manner that has him remain within himself 22 (4)  Affirming his own goodness in the manner that involves expressing it in creation. Think of it this way: God’s willing his own goodness is not a conjunctive common factor in acts (3) or (4) any more than being primary-colored is a

21  On the complexities introduced by the so-called “angelic wills,” see my “How did Evil Come into the World?” ms. 22  And perhaps giving expression to it within the Trinitarian structure of his inner life.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  123 conjunctive common factor in being red and in being blue.23 God’s willing his own goodness is a determinable with (3) and (4) as its more determinate versions. (Of course, (3) and (4) are not themselves completely determinate; the full determinacy is given by the content of God’s single comprehensive act of will.) Now there seems to be room to reinstate the idea of God’s freedom from coercive reasons to create. Although God has a coercive reason to will or affirm his own goodness—to perform some determinate act of that type— he does not have a coercive reason to will his own goodness in the manner that has him remain within himself. Nor does he have a coercive reason to will his own goodness in the manner that involves expressing it in creation. Crucially, it is not as if God first affirms his own goodness and then looks around to see if reason favors creating, or favors remaining. If that were the  case, then, it would follow that unless his reason to create somehow ­balances, or is incomparable with, his reason not to create, he then has a coercive reason to create. In the absence of such a reason not to create, the aporia returns. Instead of any such conjunctive conception, we need the adverbial conception, suggested by assumption 6, if we are to make any progress. However, there also has to be something quite distinctive about just how God’s manner of affirming his goodness by creating incorporates the weight of his reason to create. It must, as it were, swallow up the weight of the reason without thereby obliterating the fact that there is a reason to create. The adverbial conception of creation opens up that possibility but does not itself explain it. We require some further account of how his extra, i.e., not offset, reason to create, whatever it is, does not give him more of a reason to create than to remain. Having said that, the adverbial conception is required, and it imposes a constraint on our search for God’s actual reason for creating. That reason must be such that God’s acting on it would be a manner or way of God’s affirming his own goodness. The import of this constraint will become clear when we come to the specific proposal as to God’s reason to create.

23  Recall the distinction between cases where we can invoke some non-trivial and positive differentia in order to vindicate the conjunctive structure, as in “man is the species that is both rational and animal,” and cases of determinates under a determinable as in “ being red is a way of being colored” where there is no non-trivial positive differentia that we can conjoin with being colored to specify being red.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

124  Mark Johnston

VI.  The Need for a Further Traditional Resource It may still seem that we can generate a coercive reason for God’s willing his own goodness in the manner that expresses his goodness in creation. (Again, we are setting aside any speculative resort either to incomparability or to the idea that God could have had some counterbalancing reason to remain within himself, a reason with some less-than-coercive weight, that balances the weight of the reason favoring creating.) Let the combined weight of the reasons favoring the act described by a, be given by some measure WRF (a). The situation may then appear to be contradictory. If God has genuine rational freedom either to create or to remain then it is neither the case that WRF (God’s willing his own goodness in the manner of expressing his goodness in creation) > WRF (God’s willing his own goodness in the manner of remaining within himself), nor vice versa otherwise God’s choice would be coerced, one way or the other. Let WRF (God’s willing his own goodness in the manner of remaining within himself) = w0; let the unoffset and adequate reason for creating have some weight w1. In the absence of an offsetting disvalue associated with creating, all that would be valuable in the act of God’s affirming his own goodness by remaining is also found in the act of God’s affirming his own goodness by creating. Since the weight of reasons, particularly God’s reasons, perfectly track value then WRF (God’s willing his own goodness in the manner of expressing his goodness in creation) is at least w0. But now, if God’s unoffset and adequate reason for creating has some weight w1 then WRF (God’s willing his own goodness in the manner of expressing his goodness in creation) = w0 + w1. Since w0 and w1 are both positive measures, then w0 + w1 > w0, contradicting our assumption of God’s freedom to remain. That is, in the absence of an offsetting reason not to create, then whatever rational grounds there were for God to remain within himself, the weight or force of those grounds must be surpassed by the weight or force of God’s reason to create. It follows that if God has a completely adequate reason to create (and so a reason

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  125 with some positive weight) then that reason is coercive. Since God is the supremely rational will, and so is impeccable in his response to reasons, he must create. As it happens, the doctrine that God must create is explicitly designated as a heresy—a dangerously erroneous deviation from the teaching of the Church on faith and morals—by the First Vatican Council (1870). Against the advocates of the doctrine, the Council pronounces an excommunication lata sententia, meaning that the terrible sentence was thereby being carried out, in every case to which it applies, without any further need of ecclesial action: If anyone . . . holds that God did not create by his will, free from all necessity, but as necessarily as he necessarily loves himself . . . let him be cut off from the community of the faithful and the saving grace of the sacraments.24

How can the dreadful anathema then be avoided? How could God, as the pre-eminently rational being, have a completely adequate reason for creating which was not thereby also a coercive reason, a reason which thereby necessitated his creating? As we shall see, the answer lies in God’s unsurpassably infinite goodness. It is there that we find the ground of his libertarian freedom to create or to remain. This ground suffices for that freedom, so we need no recourse to further speculations either about incomparability or balancing.

VII.  Need the Weights of Various Reasons Add for the Problem to Arise? We often have the experience of sequentially appreciating the normative reasons in favor of a course of action, so that our reasons to perform the act strengthen, then become completely adequate and finally appear (correctly) to be coercive. For example, faced with two initially qualitatively indistinguishable plates of pasta puttanesca, and already feeling pretty full, I may have a weak prudential reason to choose arbitrarily the plate on the right. As my hunger increases, I may come to have a completely adequate reason to choose 24  Session 3 of Vatican 1 “Dogmatic Constitution on the Catholic Faith: 1. On God the Creator of All Things” Canon 5 (online text, Eulogos, 2007).

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

126  Mark Johnston arbitrarily the plate on the right. But then I find out that only the plate on the right has capers in the mix. Given my appreciation of capers and what they do for a pasta puttanesca, I thereby come to have a coercive reason (and hence a stronger or “weightier” reason) to choose the plate on the right, over the plate on the left. This last transition in thought is an appreciation of an extra normative reason of the prudential sort, a reason that somehow increases or enhances the weight of the case for choosing the plate on the right. However, there is no reason to suppose that, where, The weight of the reason deriving from the fact that the plate on the right is pasta puttanesca = (some less than coercive weight) wi The weight of the reason deriving from the fact that the plate on the right includes capers = (some less than coercive weight) wk it follows that, in the context at hand The weight of the reason deriving from the fact that the plate on the right is pasta puttanesca and includes capers = a weight measured by wi + wk as if there were context independent “rational grams,” i.e., units of weight by which we could measure and commensurate the weight of all reasons. That goes far beyond the datum; namely that given my appreciation of the fact that the plate on the right has capers in it, I thereby come to have a coercive reason (and hence a stronger or “weightier” reason) to choose the plate on the right, over the plate on the left. For one thing, the idea of additivity holding everywhere is utterly implausible. Even for an extravagant lover of capers, adding capers to a banana split does not improve the banana split to the same degree as it does a puttanesca. So, at the very least, one of the relevant modes of enhancement of a rational case for choosing some option is not to be modeled by the addition of the weights associated with various reasons. Accordingly, in our discussion of the weight of God’s reasons, the simple idea of additivity should be replaced by the idea of some more general enhancement function which measures how one reason taken together with another in a certain context enhances the case provided by just the first reason alone. This is just what we should expect from ordinary thought and talk. When we talk of a newly discovered reason adding to the case in favor of a course of action, we often mean no more than that it enhances the case. We rarely, if ever, mean that the weight of the new reason sums with the weight of the old reasons.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  127 To develop this idea, let the variable “+*” range over enhancement functions, functions which capture the various ways in which the weights of various reasons could interact or combine to produce a combined rational case with a weight that enhances their individual degrees of value or weights. Let “x >c y” mean that the degree or weight associated with x by a contextually relevant enhancement function +*c is greater than or is at least ordinally ranked above the degree or weight associated with y by +*c. And let “x =c y” mean that the degree or weight associated with x by a contextually relevant enhancement function +*c is equal to or is at least ranked at the same ordinal rank as the degree or weight associated with y by +*c. In some simple contexts +*c may take on the value +. But an enhancement function could be multiplicative, or even more mathematically complicated. It could also be a function that measures the weight of ­reasons arising from the enhanced value of an organic whole whose individual values separately provide reasons to act, as with the caper-bearing pasta puttanesca. More realistically in the case of the pasta—where the datum is just that given my appreciation of the fact that the plate on the right has capers in it, I thereby come to have a coercive reason (and hence a stronger or weightier reason) to choose the plate on the right, over the plate on the left—the enhancement function could be such as to give merely ordinal comparisons among the weights of various reasons and combinations of reasons. Accordingly, the overall constraints on being an enhancement function in a given context where reasons are somehow combining, i.e., the overall constraints on being an admissible value of +*, should be minimal, and arguably just these: (i) +* is commutative, i.e., in any context c (vi +*c vj) =c (vj +*c vi) and (ii) in any context c, where vi and vj are positive, (vi +*c vj) >c vi. The second condition requires that there be some enhancement, and the first condition states that the order in which the two reasons are considered should not affect how one reason enhances the other. In the context of the pasta plate on the right, the reasonable claim is just that there is some enhancement function +*p which is such that (w i + *p w k ) > p w i where, wi = the weight of the reason deriving from the fact that the plate on the right is pasta puttanesca and wk = the weight of the reason deriving from the fact that the plate on the right includes capers. It seems hard to deny that there is some such enhancement function.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

128  Mark Johnston Likewise, in the case of God’s creating, all that is required for God’s having a coercive reason to create, a reason the ignoring of which would be a defect in the operation of a rational will, is that there be some contextually relevant enhancement function +*g, such that

(w 0 + *g w 1 ) >g w 0

where w0 is a measure or ranking of the weight of the reason in favor of God’s God’s remaining within himself, and w1 is a measure or ranking of the weight of the reason in favor of God’s creating. Is there such a contextually relevant enhancement function? That is, does God’s extra, i.e., un-offset, reason to create make for a weightier rational case in favor of creation? If so, God creates necessarily, and not from libertarian freedom. Only if God’s extra, i.e., un-offset, reason to create does not make for a weightier rational case in favor of creation, is there conceptual space for a non-necessitarian model of ultimate explanation and for the applicability of our version of the Principle of Sufficient Reason. Clearly, the move from a simple summative conception of how the weights of reasons combine (one that implies that the weights can be unitized in a way that allows for addition) to a more inclusive conception of enhancement does not itself make it easier to see just how God could have had a non-coercive reason to create. If anything, it makes it harder. For as long as there is some enhancement function that serves to model some kind of enhancing influence of the reason to create, the aporia remains.

VIII.  The Way Forward The way forward is to explain how God’s extra reason to create, a reason that is not offset by some reason not to create, does not give him a more weighty reason to create. That is, to explain how it can be true that WRF (God’s affirming his own goodness in remaining within himself) = WRF (God’s affirming his own goodness in the manner of expressing his goodness in creation) = w0 even though w1, the weight of God’s reason to create rather than remain is some positive weight. So that either w0 + w1 = w0, or more realistically, (w0 +*g w1) =g w0

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  129 or, at least, neither “w0 + w1 > w0”, nor “(w0 +*g w1) >g w0” is true, because neither claim expresses a well-formed proposition. How can we find a path to this result? Cantorianism to the rescue: Since God’s goodness is absolutely infinite and since as the pre-eminent rational will he is perfectly responsive to value in his own analog of reasoning, it follows that the weight of the reason for God’s willing or affirming his own goodness, in whatever determinate manner, is itself absolutely unsurpassable, and so has neither some cardinal measure that “w0” could pick out, nor some ordinal rank that “w0” could pick out. So “w0” does not denote. Hence there is no meaningful question of the form: Is w0 + w1 (or w0 +*g w1) greater than, less than or equal to w0? The weight of the reason God has to will or affirm his own goodness is thus, as Cantor put it, “beyond mathematical determination,” for any cardinal measure or ordinal determination of rank admits of a greater measure or rank. Though there is an extra reason for creating, this does not give God a weightier rational case for creating, i.e., a weightier rational case for affirming his own goodness in the manner that involves expressing it in creation, rather than for affirming his own goodness in the manner that has him remain within himself. It is in virtue of God’s unsurpassable goodness that God’s extra reason to create does not give him a more weighty case in favor of creation. Still, that extra reason for creating may be invoked in a contrastive rational ex­plan­ ation of his creating rather than remaining. God created rather than remained because of this reason, and soon we shall see just what this reason has to be.

IX.  The Transfinite Alternative Alternatively, though this is in the end only an apparent possibility to be discharged, if God’s goodness admits of some transfinite cardinal measure and w0 is just that, then since God is perfectly responsive to value, w0 will also be the measure of the weight of the reason for God’s willing his own goodness. So where w1 is the weight of the reason in favor of creating, then there is, indeed, a proposition expressed by

w 0 + w1 > w 0 .



OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

130  Mark Johnston However, that proposition is false, since the value of creation cannot have a greater cardinal measure or ordinal rank than the rank or measure of God’s goodness, and God’s reason is perfectly responsive to value, so that the measure w0 is the highest measure for any available reason. So the intended transfinite addition will be, in the familiar Cantorian way, “uneventful,” and will simply deliver back w0 as its result. Thanks to the relative “uneventfulness” of cardinal arithmetic (explained below), and since, inevitably, it is not the case that w1 > w0 (thanks to the unsurpassibility of God’s goodness) it follows that w 0 + w1 = w 0 . So the weight of the extra reason for creating cannot give God a weightier reason to create rather than to remain. Again, there being an extra reason to create rather than remain, even one that is not offset by a reason not to  create, does not leave God with a weightier rational case in favor of creation. What then of non-additive combinations of the weights w0 and w1? To the extent that those combinations are mathematically underspecified, and could include exponentiation and the like, we cannot appeal to the uneventfulness of transfinite arithmetic to settle questions of relative weight. Given that, couldn’t there be some enhancement function +*g such that

( w 0 + *g w1 ) > g w 0 ?

No, there could not be! To see why, let’s make explicit something that we already have been relying upon. The reader will have noticed that throughout the present discussion we have been deliberately conflating—we have been moving back and forward between—the value of God’s willing or affirming his own goodness and the weight of his reason to do this, and the value of creation and the weight of God’s reason to create. The conflation is harmless because we have been thinking of value as the correlate of practical reason and we have been conceiving of God as by his nature perfectly responsive to value. So suppose that (i) the value of God’s goodness admits of some trans­ finite cardinal measure, which corresponds to the weight of the reason God has to affirm his own goodness, and w0 is just that transfinite cardinal measure, and (ii) w1 is the measure of the weight of the reason God has to create, which corresponds to a measure of the value of him creating and (iii) w1

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  131 measures some lower, or certainly not higher, transfinite or finite value than does w0, as it must be since God’s goodness is unsurpassable. Then, with the relevant comparison “>g” understood as holding among cardinal measures of value, there is no enhancement function +*g such that

( w 0 + *g w1 ) > g w 0 .



( w 0 + *g w1 ) > g w 0 .

For given that w0 is the measure of God’s goodness, nothing else, not even including the ensemble of God and creation, could have a higher transfinite measure of value than w0. That is because nothing can exceed God in value, even if he only exhibits some transfinite cardinal degree of value. Thus the minimal second condition on an enhancement function, i.e., that the second reason/value enhances the rational case provided by the first, is not satisfiable when the first reason/value has unsurpassably infinite weight. Therefore there is no such enhancement function as +*g. Likewise, if w0 corresponds to the ordinal rank of the value of God’s goodness then nothing can be ordinally ranked higher than w0. So, with the relevant comparison “>g” understood in merely ordinal terms, there is no enhancement function +*g such that

Again the minimal second condition on an enhancement function, i.e., that the second reason/value enhances the rational case provided by the first, is not satisfiable when the first reason has unsurpassably infinite weight. Formalism aside, the point is that the case for creating over remaining could not be enhanced by God’s extra, or un-offset, reason for creating, whatever that reason might be, and whatever form of enhancement is in play. To summarize: Either way, either in virtue of God being absolutely good in Cantor’s sense, or in virtue of God exhibiting some transfinite cardinal degree of goodness (or ordinal level of goodness) that could not possibly be surpassed by the goodness of anything else, even including an ensemble of God and any other thing, God’s reason to create, whatever that reason might be, was non-coercive. That result is of some interest, since it can be seen to follow not from any speculation about the possible offsetting reason not to create, but merely from the unsurpassibility of God’s goodness, whichever form that takes. And arguably, God’s goodness has to be unsurpassable if he is in the relevant

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

132  Mark Johnston sense the Highest One; that is, the only One, the absolute worship of whom would be without any tincture of idolatry.25 Before we examine the specific kind of non-coercive reason for creating that is available to God, let us turn to the relevant details of Cantorianism, and just how they undermine the claims that (w0 + w1) > w0 and (w0 +*g w1) >g w0.

X.  Cantor on the Transfinite and the Absolute One of Cantor’s great achievements was to resolve the longstanding antinomies of infinite size, which had plagued the philosophy of mathematics from Aristotle to Bolzano. Cantor thereby helped remove the impression of an a priori basis for the all-too-influential Aristotelian animus toward the actual or completed infinite. Secondly, as part of the philosophical background for his mathematical work Cantor makes a crucial distinction between the “transfinitum,” the totality of ever increasing infinite cardinal measures of how many or how much, and the Absolute “which allows for no addition or subtraction to its quality and must therefore be seen as an absolute maximum” that “stands in particular beyond [any] mathematical determination” which would assign it a measure of weight or how much along any dimension of quantitative comparison. Cantor shows that the ideas of increase and diminishment do apply to transfinite cardinal measures, though in a restrictive way, a way that makes transfinite arithmetic mostly uneventful. This is in explicit contrast to the Absolutely Infinite, with respect to which nothing is to be made of the ideas of increase or diminishment. Cantor’s unending hierarchy of cardinal numbers, understood as potential measures of how many or how much, begins with all finite “cardinal” numbers, i.e., all finite whole numbers 0, 1, 2 . . . understood, not as markers of positions in a rank order, but as measures of how many or how much. It continues with the transfinitum, the unending sequence of all transfinite numbers; which can be represented as measures of the size of sets, and their power sets, and the power sets of their power sets, and so on. The power set of a set is the set of all proper and improper subsets of the set, including the empty set {}. So the power set of the set {x,y} is the set

25  For more on idolatry in this extended sense, see Johnston, Saving God: Religion after Idolatry.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  133 {{}, {x}, {y}, {x,y}}. And the power set of {x, y, z}, is {{}, {x}, {y}, {z}, {x, y}, {x, z}, {y, z}, {x, y, z}}. In general if a set has n members then the power set has 2n members. So if the size or cardinality of a set is measured by n, the cardinality of its power set is measured by 2n. Starting with the natural numbers and considering the sequence of power sets we get:26 ‫א‬0 = the cardinality or measure of the size of the set of natural numbers ‫א‬1 = the cardinality of the power set of ‫א‬0. ‫א‬2 = the cardinality of the power set of ‫א‬1 ‫א‬3 = the cardinality of the power set of ‫א‬2 … … … And so on without end, where in general ‫א‬j = 2‫א‬j − 1 One of Cantor’s seminal achievements was to provide a consistent criterion for measuring the relative size or “power” of infinite sets. Two sets have the same power or cardinality if and only if they can be put into one-to-one correspondence. This leads to the “uneventfulness” of transfinite arithmetic (i.e., of addition, subtraction, and multiplication, but not, of course, exponentiation), which is manifested by the following equations that fall out of Cantor’s system. Finite Uneventfulness ‫א‬j + f = ‫א‬j, where f is any finite number, however enormous. ‫א‬j − f = ‫א‬j, where f is any finite number, however enormous. ‫א‬j × f = ‫א‬j, where f is any finite number, however enormous. Infinite Uneventfulness ‫א‬j + ‫א‬i = ‫א‬j where j > i or j = i ‫א‬j × ‫א‬i = ‫א‬j where j > i This arithmetic uneventfulness holds throughout the transfinite hierarchy of cardinal measures of how many or how much. As if all this were not enough, Cantor presses on in the realm of the infinite. He tells us that if there is an Absolute, an unsurpassably infinitely good being, then its goodness is not capable of being measured by any cardinal measure of how many or how much. 26  Assuming the Generalized Continuum Hypothesis, i.e. for any infinite cardinal x, there is no cardinal y, such that y lies between x and the cardinality of the power set of x, for ease of exposition. Otherwise, we could proceed just in terms of the Beths, putting “‫ ”ְב‬for “‫ ”א‬throughout.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

134  Mark Johnston Indeed, we can see just why the Absolute’s goodness cannot be measured by any cardinal number. For if it were measurable, the measure would have to correspond to the greatest cardinal number, and there is none. For suppose there was a greatest cardinal, then consider one of the sets whose cardinality it measures. Take the power set of that set. The resultant set would have a larger cardinality than the supposed infinite cardinal. So the supposed greatest cardinal would not be the greatest cardinal. So there is no greatest cardinal.27 Hence the very idea of increase or diminishment of Absolutely Infinite Goodness simply does not get a foothold, anywhere, ever, under any con­ ceiv­able circumstances. That is the real Cantorian basis for the claim that it is not true that w0 + w1 > w0, and more generally that it is not true that (w0 +*g w1) >g w0. There is no adequate infinite cardinal measure or ordinal rank that could be associated with “w0”, the purported measure of the weight of the reason to affirm Absolutely Infinite Goodness. Absolutely Infinite Goodness is outside of the sphere of the quantifiable, of what can be measured by any quantity or relative rank. The same point applies if we consider non-Cantorian measures of infinite quantities. Suppose we follow John Conway28 and Philip Ehrlich,29 and recognize that the system of surreal numbers is the most inclusive system of measures, and note that in the system of surreals infinite arithmetic is not invariably uneventful, so that there are more fine-grained transfinite distinctions than appear in Cantor’s system of measures. Even so, there is no greatest surreal number, by which we could measure God’s goodness. God’s goodness, to paraphrase Cantor, is “an absolute maximum” that is “beyond mathematical determination” and this remains so even if we move outside Cantor’s own system of transfinite measures to the surreals. God’s goodness is beyond even surreal determination. (The same point applies to hyperreal measures.) Since God is perfectly responsive to goodness, the magnitude of his reason to affirm his own goodness is likewise “an absolute maximum,” “beyond mathematical determination” as Cantor puts it. 27  Nor, we might add, can the comparative value of the Absolute relative to any other values be ranked ordinally higher than any such value. For just as there is no largest cardinal, there is no highest ordinal. That is, the sequence of ordinals 1, 2 . . . ω, ω + 1, ω + 2, . . . , ω·2, ω·2 + 1, . . . , ω2, . . . , ω3, . . . , ωω, . . . , ωω to the power ω, . . . has no end. 28 J. H. Conway, On Numbers and Games (New York: CRC Press, 2001). 29  Philip Ehrlich, “The Absolute Arithmetic Continuum and the Unification of All Numbers Great and Small,” The Bulletin of Symbolic Logic 18(1) (2012), pp. 1–45. See also Philip Ehrlich, “The Absolute Arithmetic Continuum,” Synthese (forthcoming).

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  135

XI.  The Aporia Resolved Either by way of his Absolute Goodness or by way of his instantiation of an unsurpassable transfinite measure of goodness, the goodness of God puts him in a position to have a completely adequate but non-coercive reason to create. Though the reasoning that leads to the paradox of contingent creation also fails if God is merely transfinitely good, it must be said that from the theistic point of view there is something defective in thinking of God as exhibiting merely some transfinite measure of goodness. For that implies that God is not absolutely good, is not the Good, and that a better is indeed conceivable. So the real theistic solution to the paradox of contingent creation lies in the observation that there is an incoherence in the central thought driving the aporia or apparent paradox. If God has some reason to create that is not “offset,” i.e. either counterbalanced by, or incomparable with, his reason not to create, then his reason to create will have some weight w1, which when combined with the weight w0, provides a weightier reason, w0 +*g w1, to create, rather than to remain within himself. The resultant reason with weight w0 +*g w1 is thus a coercive reason to create, one that a supremely rational being must follow. So creation is necessary. Since God is Absolutely Infinitely Good, there is no measure that could be associated with the designators “w0” and “w0 +*g w1”. Nor is there any sense in which his un-offset reason to create makes for a stronger rational case to affirm his goodness in the manner of creating. In short, the observation that when it comes to God’s reasons to create, extra does not imply more, arises not from the familiar absorption properties of Cantorian transfinite measures, but instead from the Cantorian conception of being “an absolute maximum.” As such, the observation remains in place even if we adopt non-Cantorian transfinite measures, such as the hyperreals or the surreals. Even if we were to allow a transfinite measure of God’s goodness, there still would be no sense to be made of God’s extra reason to create, whatever it is, providing him with more of a reason to create than to remain within himself. For God’s reason to affirm his own goodness in either determinable manner is unsurpassable in weight. His own goodness is unsurpassable in value; and, as the pre-eminent rational will, his reasons and their weight perfectly match the values in play. It follows that there is no assignment

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

136  Mark Johnston function that coherently counts w0 +*g w1 as having a higher measure or rank than w0. The resultant situation in respect of contrastive explanation is interesting. Given that God has an extra and adequate reason to create rather than remain, there is a contrastive explanation of his creating rather than remaining within himself, namely one which cites the reason in question, the reason whose content we will open up in the discussion below. There is an interesting asymmetry here. Suppose that God had not created, would there then have been a contrastive explanation of his remaining within himself rather than creating? If God has no offsetting reason to not create then even though he has a completely adequate reason not to create, and is free not to create, there would be no contrastive explanation, no citing of a reason as to why God remained within himself rather than creating. There would simply be a reason to remain, namely that in doing so he was affirming his own goodness, the reason for which was not outweighed by any other reason.

XII.  The Upshot for Theology If this diagnosis of the aporia is right then something new enters into the familiar framework of natural theology. It is God’s unsurpassably infinite goodness that explains his complete freedom to create, a freedom that includes freedom from coercive reasons in the matter of creation. Likewise, it is God’s unsurpassably infinite goodness that allows him to contingently ground created things by way of a free choice based on completely adequate but not coercive reasons. Moreover, it is thanks to God’s unsurpassably infinite goodness that there is an applicable and attractive version of the Principle of Sufficient Reason, namely: The basic idea  All non-autonomous facts are explicable, in the sense that there must be completely adequate reasons for them. The crucial gloss  However, these reasons can take different forms. Sometimes such reasons are found in a necessitating ontological explan­ a­tion, which would show how derivative things must be as they are, given more fundamental things; but alternatively there can be completely adequate but non-coercive practical reasons for choice and intentional action, reasons which provide a non-necessitating explanation of the choice and corresponding action via those antecedent reasons.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  137 Of course, all of this could be taken merely conditionally, as just a consistent model of fundamental ontological explanation in which that notion is freed from the idea of necessitation. That ontological result comes by way of thinking of the onto theology as a model, and only that. A certain sort of theist may go further. The stated version of the Principle of Sufficient Reason is magnetic in the following sense: Any rationalist metaphysician worth her salt would be happy to see that it could be satisfied by some plausible total conception of reality. For suppose we turn away from the principle, truncating our questioning by resting with some unexplained contingency, such as the Big Bang, or an endless contingent past. Is this not a second best option, intellectually felt as such? It is also very plausible to claim that we, our chosen acts, along with the events in which we swim, indeed the law-like frame of material reality and even the existence of matter itself, are all contingent. That claim and the magnetic principle are rendered understandably consistent if there is a creator or fundamental ground who has total freedom to create or remain, because he is unsurpassably infinitely good. Our theist is here offering us an argument to an unsurpassably infinitely good God as an explanation of how two commitments of the intellect can be responsibly held together. In the absence of a better alternative ex­plan­ ation, it stands as a prima facie argument for the existence of such a God. Of course, we must look at the counterarguments, such as the several arguments known under the heading of “the” argument from evil. But once we take up the Cantorian point of view on das wahrhaft Unendliche, those arguments look quite different. For one thing, in creating, God could have no rational interest in questions of better or worse worlds. For reality was already absolutely unimprovable and undiminishable in value! God cannot be impeached for not creating the best world; he can only be impeached if what purports to be creation is not an adequate expression of his Absolutely Infinite Goodness. The terms of impeachment are radically transformed, so that atheology, as yet, may not have begun to formulate the relevant argument. Outside of theology and atheology, another thing which will look quite different is the landscape of practical reason, in particular the range of reasons we have to act or refrain from acting. In acting, we can have no practically rational interest in questions of better or worse outcomes! And, as we will see below, this surprising result remains in place so long as the existence of unsurpassable goodness has a non-zero epistemic probability.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

138  Mark Johnston

XIII.  An Inventory of Kinds of Normative Reasons Let’s now turn to the question of just why God did not remain within ­himself. Given that God could have had some kind of non-coercive but completely adequate reason to create; that is, to affirm his own unsurpassable goodness in the manner of creating, what kind of reason was it? Our investigation so far, in particular the adverbial consideration, has placed a significant constraint on the character of that reason: It has to be a reason that represents creation, insofar as it is something favored by that very reason, as a manner of God’s affirming his own goodness. This constraint will figure significantly in what follows. But the first thing to see is that many of the sorts of reasons which move finite rational wills could not be among God’s reasons to create. There is an obvious sense in which God’s mind must remain opaque to us, but it is also true that his distinctive nature rules many things out as potential reasons or motives. Accordingly, we can have significant negative knowledge of the mind of God. Could we say that God just wanted to create, and leave it at that? Desires as psychological conditions may be motives, but this still leaves the question of just what kinds of reasons for action they disclose to the subject. In what respect do they present an act as worth performing? “In respect of satisfying the desire in question, so long as the desire persists” is an answer, but it does not in fact give a reason to act on a desire rather than to act to remove it. (Recall Buddhism, which centrally includes a systematic therapy aimed at eliminating desire.) In any case, God is not a proper subject of desire in the psychological sense. Desires arise, but no Godly attitude arises, i.e., appears in time. Talk of God’s desires is best understood as metaphorical talk about the content of his rational will. And of course, given that God is omnipotent, there is no ontological possibility of there being a time at which his will is not fulfilled. “He desired to create” is at best then a way of indicating that he chose to create, not a citation of a reason for his creating. While the psychological conditions that motivate us to act are many and various, and are the proper object of cognitive science, the objective normative reasons for acting, the considerations which justify or make intelligible performing an act rather than refraining from it, or performing some other positive alternative act, may, as a first pass, be sorted into seven kinds or types of normatively significant propositions.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  139 1. That the act would improve reality, understood as a maximal totality of states of affairs bearing value impersonally considered, overall; that is, taking into account all dimensions of value—including hedonic value, aesthetic value, the value of any natures that exist or come to exist, and the value of desire or interest satisfaction—which states of affairs considered in themselves, and without regard to just whom they involve, can have. Alternatively, that the act prevents the diminishment of such value in reality overall.30 Call the reasons stated by propositions of that type “melioristic reasons,” for their force as reasons turns on the rational appeal of making reality as a whole better, or at least preventing reality as a whole from getting worse.31 2. That the act would makes things better for someone or for a community of people to whom the agent has a personal attachment or bears special responsibility (as when one is the assigned caretaker in charge of increasing welfare within the given community), or the act prevents something which would make things go worse for him or them. Propositions of that type state reasons deriving from personal attachments and special relationships.32

30  Notice that this idea of making reality better overall is more general than the idea of increasing the additive sum of value. We could reject simple additive aggregation as the method of calibrating improvement and yet still keep to the idea of improving reality overall. We could for example recognize the formation of an organic unity, a unity whose overall value is more than a sum of the value of its parts, as a kind of improvement. Or we could recognize only ordinal comparisons of being better than, so that the best is not determined by its measure being the greatest sum but by its premier ranking over all other goods. Likewise, the theological conclusion that the impersonal value of reality overall is absolutely unimprovable and undiminishable is a stronger conclusion than the idea that nothing could increase the value of reality overall by mathematically summing to the measure of that value and nothing could decrease that value by mathematically subtracting from the measure of that value. That is to say that the problem to come for the idea that God created for a melioristic reason is not to be met by rejecting the simple additive model of aggregation. 31  Under conditions of uncertainty such reasons present to us as intimations of different degrees of expected utility; but this is a complication that does not arise for God. 32  To highlight the distinction between impersonally better and better for imagine a case where there is a threat of Sally being instantaneously replaced by an exact duplicate. You prevent that from happening. The outcome you secure is better for Sally but it is not impersonally better; there would have been the same number of persons exhibiting the same value-making features before and after the replacement. If Sally is your friend or a family member or . . . (specify the personal attachment) then you obviously have a pro tanto reason to secure the outcome. Making that point is not yet to endorse a strong welfarist perspective and recognize standing reasons for all to aim at what is good for people in general.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

140  Mark Johnston 3. That the act would advance the perfecting of one’s own nature, or the cultivation of one’s talents. Propositions of that type state perfectionistic reasons. 4. That the act would advance one’s own anticipatable future interests. Under this heading we find prudential reasons, which do not derive from reasons of self-love or self-attachment, since they apply even under conditions of self-loathing and self-detachment. 5. That the act complies with an authoritative command, as in the case of legal and administrative reasons, or with the implicit commands provided by ethical exemplars, where the command is not constituted by a speech act but a course of exemplary action.33 6. That the act is morally required; that there is a moral reason to do it. 7. That the act is the expression of a virtue relevant in the situation of acting; for example it is, in the situation, the generous or the kind or the courageous or the just thing to do. Propositions of that type state reasons deriving from virtue.34 Obviously, this inventory of the kinds or types of normative reasons is not anything like the last word about their nature. Philosophical theorists might either try to deny that there are normative reasons of one or another of the seven kinds, or try to reduce one kind of normative reason to another. So both the Divine Exemplar Theory and the Divine Command Theory of moral obligation treat moral reasons as reasons of the fifth kind, i.e., reasons of compliance with an authoritative command, namely God’s command, where the command is either explicit or arising from his exemplary nature.35 33 We often note that the exemplary performance of an individual player “inspired the whole team.” But in the few cases in which I have been on the receiving end of this the mode of inspiration the notable thing was the obliging or commanding character of the performance, namely its carrying the imperative: Act like this! 34  It is under this heading of reasons of virtue that I think it best to accommodate the reasons favoring prioritarian and other global distributional preferences. The crucial point is that unlike purely melioristic reasons whose force arises from doing the impersonal best one can in a straightforward impartial way, the reasons favoring prioritarian and other biased global distributional preferences, such as lexamin and maximin arise from another source. After all, such schemes do not support maximizing overall impersonal value, nor some non-biasing function of it, such as average value. Overall impersonal value along with avoidance of bias should not be given the sole foundational role in determining what is best. Rather these theories represent what happens when distributional questions are influenced by considerations of justice, compassion, or humanitarian virtue. 35  For a Divine Exemplar variant on the Divine Command Theory, see Linda Zagzebski “Morality and Religion,” in William J. Wainwright, ed., Oxford Handbook of the Philosophy of Religion (Oxford: Oxford University Press, 2005), following up on her insightful Divine

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  141 Kantians might be tempted to reduce reasons of the seventh kind to reasons of the sixth kind. And a thoroughgoing Impersonal Consequentialism would try to account for all reasons in terms of reasons of the first kind. It is best to prescind from these ambitious philosophical proposals, and simply take the seven kinds of normative reasons at face value. For what follows is an argument by elimination, which shows that none of these kinds of reasons could be God’s reason for creating. Any reductionist account of one kind of reason in terms of another, and any outright denial of the existence of one or another kind of reason, would just make the argument that follows easier, in ways the opponents of the reduction or elimination in question then would find tendentious. In this classificatory scheme, the normative reasons for an act are taken to be true propositions that favor the act in question. Given some true proposition proposed as a normative reason for a given act, our taxonomy of reasons invites us to specify just what kind of reason it expresses. For the same proposition considered in different contexts may express different kinds of normative reasons. Indeed, the same proposition may express different kinds of normative reasons for different agents, depending on their relations to the persons, groups, or things figuring in the proposition. For example, the proposition This act will benefit Martha may be, for someone who barely knows Martha, a weak, prima facie meliorisitic reason for performing the act in question, in the sense that on the face of it, the truth of the proposition appears to go some way, but not very far, in establishing that the act will make reality better overall. For Martha’s husband, however, the same proposition may be a strong pro tanto reason of personal attachment to perform the act. Even if other reasons outweigh this reason, it remains a genuine reason of personal attachment, and a strong reason of that kind. It is Martha’s husband’s relation to Martha that

Motivation Theory (Cambridge: Cambridge University Press, 2004). Zagzebski styles her view “Divine Motivation Theory” because she emphasizes God’s motives as the exemplary aspects of God. Though she introduces her own theory as a successor to Divine Command Theory, I believe any exemplarist theory of the genuine ground of reasons should treat the exemplar’s behavior as implicitly commanding—as if it were saying “Act like this”—where the exemplar’s behavior is also the source of the authority of the implicit command. This, at any rate, is how the one who takes the exemplar as an exemplar sees it. But nothing that follows depends on this way of assimilating the two views. See footnote 47 on whether Zagzebski’s view helps with identifying the reason God created.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

142  Mark Johnston provides the basis for the proposition’s being a strong pro tanto reason of personal attachment for Martha’s husband to perform the act.

XIV.  Immediate Elimination of Four Kinds of Normative Reasons God, as traditionally conceived, is the pre-eminent rational will. When he acts, i.e., when his rational will operates, he acts not from subjective confusion but from objective normative reasons, reasons that are not just prima facie or pro tanto reasons, but are as completely adequate as reasons can be, given the presence of rational alternatives. When we ask “Why did God create?” we are asking: What was his completely adequate reason behind creating? And we can immediately see that reasons of the third, fourth, fifth and sixth kind can be set aside. On traditional theistic views, God in creating was not perfecting his nature. Nor was he advancing his own self-interest; i.e., fulfilling some need of his which otherwise would have been unmet. Nor was he under an authoritative command (explicit or implicit) to create, or if he was it can only have come from himself, which simply pushes the question back to the reason for that command. Nor was he morally obliged to create. To whom? By whom? Kantians might reject those two questions, by urging that the source of moral obligation is not in directed duties toward others but in the generality requirements of rational willing; so that moral requirements are the upshot of clearheadedly giving the law to oneself as a free being. But on such a Kantian view, in the case of a pre-eminently rational will such as God, moral reasons would coercively support whatever they support.36 So if God had a moral reason to create, creation would not be contingent. But we are here trying to make real sense of the traditional notion that creation is contingent. So why did God create? The answer, it may seem, has to cite a reason of the first, second, or seventh kind. 36  One exception to the principle that moral reasons are coercive may be thought to arise in “Buridan’s ass” moral cases, as when two strangers are in trouble, and I can only save one, with same odds of success on either side. Here, one might think that I have a moral reason to save the one, and a moral reason to save the other, and neither is coercive. My alternative diagnosis is that here I have a coercive moral reason to save someone. In any case, this kind of “exception,” if it is an exception, is irrelevant in the case of God and creation, which obviously is not a Buridan’s ass case.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  143

XV.  Did God Create out of Reasons of Personal Attachment? As already noted, reasons of personal attachment, though they take the form of propositions concerning how the act in question would benefit some given person, require a basis in virtue of which they count as reasons of personal attachment. The basis involves some significant personal relationship between the agent and the beneficiary. The basis must be in place in order for there to be a reason of this kind in play. In that sense, the basis must be in place prior to the act in question being a reasonable act to perform in the light of the reason of personal attachment. The same applies to reasons of special responsibility; the incurring of the special responsibility must be in place prior to the reasons arising from what is good for some group of people. Let’s concentrate on reasons of personal attachment; the considerations evinced will apply mutatis mutandis to reasons of special responsibility. The basis requirement is then this: The basis B for some proposition P being a reason of personal attachment for an agent to perform some act will involve some personal relationship between the agent and the potential beneficiary or beneficiaries described in P. The basis B must be in place prior to the act in question for P to be a reason of personal attachment for that act. Typically, but not always, the required priority is realized by temporal priority; i.e., the basis B is in place before the act emerges as an option. If the act has to be already performed for the basis to be in place then the relevant proposition is not a reason of personal attachment. However, the requirement of priority, the requirement that is usually realized by temporal priority, is inherently ontological. The act cannot be ontologically prior to the basis. That is, the particular basis B for some proposition’s being a reason of personal attachment for an agent to perform some act cannot be ontologically dependent on the agent’s performance of the act.37 Let me illustrate this rather abstract requirement by way of examples. Suppose an excellent judge of human nature tells me that if I befriend Matthew, whom I presently barely know, this will benefit him. And suppose that the judge is right; Matthew and I meet and as a result we become fast 37  For a useful discussion of the notion of ontological priority, see Kit Fine, “Ontological Dependence,” Proceedings of the Aristotelian Society 95 (1995), pp. 269–90.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

144  Mark Johnston friends, to our mutual benefit. Suppose further that the judge is so widely known to be an excellent judge in such matters that I could come to know, simply on the basis of the excellent judge’s testimony, that if I befriend Matthew that will benefit him. (All the friendships he has set up have been rewarding for both parties.) Even supposing all that, the proposition Befriending Matthew will benefit him is not a reason of personal attachment for me to befriend Matthew. This is because it could only be a reason of personal attachment if I have some significant personal relation to Matthew; but my significant personal relation to Matthew—my friendship with him—is originally constituted by my act of befriending him, and so is ontologically dependent on that act. So my resultant friendship with Matthew cannot be the basis for the proposition Befriending Matthew will benefit him being a reason of personal attachment for my attempting to perform the act of befriending Matthew. At most, the proposition is some weak, prima facie, melioristic reason to befriend Matthew. Suppose now that I can save one of Martha and Matthew from harming themselves by sympathetically listening to her or his woes, and for some reason I can’t save both. Suppose moreover that I know that in the very act of sympathetically listening to his woes, I will befriend Matthew; and I know that sympathetically listening to Martha will have no such effect. Do I then have a reason of personal attachment to sympathetically listen to Matthew’s woes but not Martha’s? No, I have at most an extra melioristic reason to sympathetically listen to Matthew’s woes but not Martha’s. For, in this case too, my coming to have a relationship of friendship to Matthew ontologically depends on my befriending him. Turn now to God and Martha. God’s loving Martha ontologically depends on Martha’s existing, and Martha’s existing ontologically depends on God’s creating Martha. So by transitivity, God’s loving Martha ontologically depends on God’s creating Martha. It follows that God’s loving Martha, or indeed his having any personal attachment to Martha, cannot be the basis for the proposition Creating Martha would benefit her being a reason of personal attachment for God to create Martha. The upshot is that God cannot have reasons of personal attachment to create particular persons, and this is so even if he has complete foreknowledge

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  145 from eternity of just whom he creates and of his subsequent personal attachments to them. For those very personal attachments, which God foreknows (or better “over-knows”) from eternity, are ontologically dependent on God’s creating the persons in question. To summarize: Given that the particular basis B for some proposition’s being a reason of personal attachment for an agent to perform some act A cannot be ontologically dependent on the performance of A, God could not have created out of reasons of personal attachment. Mutatis mutandis for reasons of special responsibility. Could it be, nonetheless, that God somehow creates out of reasons of love or reasons of generosity?

XVI.  Creating from Reasons of Love At this point, we might do well to consider a standard evangelical account of the reasons for creation, one explicitly provided, for example, at Christianitytoday.com. In accord with traditional theism, the author is adamant that creation did not satisfy some need of God, be it for company, or to have admirers to praise him, or to occupy himself with something other than himself—as it might be a universe-wide diorama or field of play; not even, I might add, a play properly called Hiding Atman, where creation is the Matrix, but all the way down, including the appearance of separate individual personalities. (That noble game, described in versions of Advaita Vedanta, incoherently requires that God be susceptible to boredom, and to the relief of boredom.) Before giving his answer as to why God created, our author writes, It wasn’t because he needed us: “The God who made the world and everything in it . . . is not served by human hands, as if he needed anything” (Acts 17:24–25). And he didn’t make us because he was lonely. Long before we were here, God already had “company” with his Son and the Holy Spirit, referred to in Genesis 1:26, “Let us make man in our own image.” And he didn’t make us because he needed his ego fed. It’s not like God made us to satisfy some craving to be worshiped. God is totally secure in who he is—without us.38

38  Dawson McAllister, “Why Did God Create Us?” Christianity Today. https://www. christianitytoday.com/iyf/advice/faithdoubt/why-did-god-create-man.html.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

146  Mark Johnston Then follows an account of why God created, a not atypical evangelical answer: [However] despite not needing us, God chose to create us anyway, out of his great love: “I have loved you with an everlasting love” (Jeremiah 31:2). Yes, God loved us before he even created us. It’s impossible to get our heads around that idea, but it’s true; that’s what “everlasting” love means.39

It would be uncharitable to interpret this evangelical proposal as supposing that God is (atemporally or non-successively) in a practical condition analo­gous to reasoning of the form I love Martha, therefore, Let Martha be created. That reasoning is incoherent precisely because the putative reason is ontologically dependent on the occurrence of the act for which it is supposed to be a reason. However, there is available a different, and somewhat metaphorical, model of just how God’s love might be relevant in creation. Suppose that you are having some architects design your house and they take you through two very different designs in quite incredible detail. You might then say, gesturing at the second design, “I love this house,” namely the house that would come into being if the second design were implemented. Of course, it is very likely that there is not some single particular house that would come into being on any way of implementing the architect’s second design. No matter how much detail the architects provide, it is unlikely to amount to a fully individuating specification of a definite particular house, i.e., a specification such that if it were followed in any possible situation, it would yield the very same particular house. Now God is traditionally conceived as having a full repertoire of just such individuating specifications or ideas (a consequence of assumption 4). The traditional thought is that God has an individuating specification or idea, which would be an idea of Martha, if he actually creates Martha. And so on for everything else which he could create. (Alvin Plantinga develops the outer regions of this idea in the context of his theory of individual essences.40 39  McAllister, “Why Did God Create Us?” 40  See his The Nature of Necessity (Oxford: Clarendon Press, 1974) and “Existentialism,” Philosophical Studies 44 (1983), pp. 1–20.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  147 As far as modality goes, Plantinga’s picture will serve well enough for the purposes of this paper.) Compare this thought: Perhaps it is possible that there is one and only one person who would come into existence if a specific sperm fertilized a specific egg in an utterly determinate particular way. But suppose this never happens. We need not say that the individuating specification describes a possible, non-existent person, i.e., the one who satisfies the specification. For that would be to commit ourselves to the extraneous thesis of Alexius Meinong; the utterly surprising and controversial thesis that something can have a property or stand in relations without existing. Instead, we need only say that the individuating specification would have specified a particular person if the specification had been realized, i.e., if the sperm had fertilized the egg in just that way. Now just as with the architect’s second design, it is prima facie con­ceiv­ able that God might “love,” i.e., affirm the goodness of, the very individual specification that would be an individual specification of Martha, if it were realized. Likewise for all that he creates.41,42 In any case, when it comes to talk of God’s creating from reasons of love, we should distinguish reasons of personal attachment, which cannot be in play, from reasons that might arise from God’s affirmation of the goodness of realizing particular individuating specifications. It is quite easy to hover indecisively over these two quite different sorts of considerations, and then end up talking as if we might have had a personal relationship with God “before” we were created, a thought whose narcissism is surpassed only by its incoherence. 41 Perhaps this—the realization of those individuating specifications that would be the specification of us if he creates us—is all that is required to make sense of God “calling us” into existence, to use Paul’s phraseology at Romans 4:17. There Paul adopts what is, in effect, a Meinongian idiom—“καλοῦντος τὰ μὴ ὄντα ὡς ὄντα”—suggesting that the Creator calls on non-existent things to come into being. But it is only an idiom. The same could be said of Jeremiah, 1:5 “Before I formed you in the womb, I knew you.” That can be read to suggest the existence of the person prior to embodiment. Or it can be read as a dischargeable Meinongian idiom, dischargeable in terms of a Plantinga-style theory of would-be individual essences. Meinongianism is not a Christian, or more generally theistic, doctrine, nor is it a consequence of theism’s adequate articulation. It is a highly controversial thesis about the nature of predication, one whose coherence is doubtful. 42  A central caution here is that there is no valid route to Meinongianism from the truth of remarks like “Some Greeks loved Athena.” “Loved” and its cognates are intensional transitive verbs. Despite the fact that “loved” is a relational predicate, no worldly relation to anything is implied by such truths. That there was something that the Greeks, therefore, loved may seem to provide a road to Meinongianism. But that road is closed by the proper linguistic semantical understanding of this sort of use of “something.” (Here I have been helped by discussions with Friederike Moltmann.)

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

148  Mark Johnston Now we can give a sense to the idea that God realized certain individual specifications because he “loved” them. It means: because he affirmed the goodness of realizing those individual specifications. “Loved” is here either a metaphor or a loose polysemy, as is shown by the strangeness of literally saying that he failed to love the individual specifications he did not create. Even so, what were God’s reasons for affirmation of the goodness of realizing the individuating specifications of just those beings he created? If God has reasons to realize or instantiate what would otherwise have the status of mere would-be individual specifications then, assuming that our inventory is exhaustive, the reasons to realize or instantiate them would be either of the first or of the seventh kind. Either they would be melioristic reasons, having to do with improving reality as a whole, or they would be reasons arising from virtue, where the relevant virtue, as we are sometimes told, is God’s graciousness or generosity.43

XVII.  Did Generosity Provide God with Reason to Create? Could the completely adequate reason for creating be a reason arising from the virtue of generosity? That an act would be a generous thing to do is a reason to so act, though ordinarily and at least for us, it is not in general a coercive or even a completely adequate reason, i.e., a reason that all by itself, even if other reasons 43 Thus John Finnis in his essay “Philosophy and God’s Nature: Second Thoughts,” in Religion and Public Reasons: Collected Essays, Volume 5 (Oxford: Oxford University Press, 2011), quoting from his own Aquinas: Moral, Political, and Legal Theory (Oxford: Oxford University Press, 1998), writes, since God, being pure actuality, can have no lack or need of any kind whatsoever, all the good in the created universe—creatures and everything from which any creature ever benefits—must be given out of God’s sheer generosity (p. 198). And John M. Cooper, in Panentheism: The Other God of the Philosophers (Grand Rapids, MI: Baker Academic, 2006), in describing what he takes to be the advantage of classical theism over panentheism (God is in all and all is in God) along the dimension of capturing God’s sovereignty observes: God is the Lord not only because he rules the world but also because he decides whether and which world exists. He knows all possible creatures in all possible worlds from all eternity because he has complete knowledge of his power to create and sustain them. In this ideal sense, our world and all possible worlds are ‘in God’ from all eternity. But they are not part of him, and his nature does not entail that any be created. An important implication of God’s ontological independence is that his act of creating is truly agapic— entirely loving and gracious in giving creatures existence.  (p. 325) Notice the lapse into a Meinongian idiom here, which can easily be avoided, without affecting Cooper’s point, by reverting to talk of all possible creaturely individual specifications.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  149 were to the contrary, would justify the performance of the act. So  the proposition that Befriending Matthew would be a generous thing to do might be a pro tanto reason to befriend Matthew. Likewise, there might be pro tanto reasons of generosity to befriend Martha, Elizabeth, and so on and so forth. But there is not thereby a completely adequate reason to do each of these things. While exhibiting some generosity is an overall requirement on being a good or virtuous person, the requirement of generosity is not a requirement governing every possible case in which one could be generous. Nor is it rational to be generous in every case in which one could be generous. This appears to be because of the sacrifices involved in unrelenting generosity, and the way those sacrifices can cost one the chance to live a life of one’s own. What determines how one should manifest one’s generosity? Here it is natural to cite exigencies such as the resources one has at hand, the need of the potential beneficiary, and the degree to which one’s aid will benefit him or her. Procreating can be a generous, indeed an extraordinarily generous, thing to do. But as with befriending people, because of the demands involved, no one is required to perform as many of these generous acts as he or she can. Generosity does not require procreating as much as one can; nor does it require befriending as many people as we can. That is due to the personal sacrifices involved. Because of the human cost of such sacrifices, generosity is an overall requirement on a life, and not a requirement in every case where it is possible to be generous. But creating involves no such sacrifices for God as conceived by traditional theism, for nothing can make God worse off. He is, you will recall, unsurpassable and undiminishably infinite goodness. Although for us, generosity is a generic or “imperfect” duty, for God it would be a universal or “perfect” duty.44 So, on this line of thinking, a perfectly generous God would be required to create everyone whose life would be worth living. But if it is from such reasons of generosity that God creates, then God is required to create in general, and indeed is required to create each particular person who exists. Since God exists necessarily, and necessarily does whatever is ethically 44  An imperfect duty, such as the duty of beneficence, is a duty with respect to which the agent has discretion as to (i) how much to do in furtherance of the duty and (ii) which of the opportunities to discharge his duties he acts upon. God has no duties, at least with respect to creating, but it remains to be seen whether in God’s case the virtue of beneficence or generosity has this discretionary aspect.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

150  Mark Johnston required of him (if anything is) then on this line of thinking, you and I are necessary existents.45 But we are not. If generosity is for God an analog of a perfect duty then God could not have helped but (i) create, and (ii) create us. And this idea is at odds with the traditional theistic idea of contingency with which we began. So God, even if he is extraordinarily generous, did not create from reasons of generosity. There is a more systematic consideration to the same effect. Generosity towards creatures, like the love of creatures, is a virtue only exercisable “after” creation. There must be prior subjects to whom one could be generous for the reasons deriving from the virtue of generosity to be in play. It is mere metaphor to say that God could be generous to individual specifications. And it is a kind of proleptic fallacy to suppose that the persons who came into being as a result of God’s realizing individual specifications, came into being as a result of God’s prior responsiveness to reasons deriving from the virtue of generosity. What then was God’s reason for creating? The only reasons that remain from our inventory are melioristic reasons, at least if we have exhausted the scope of reasons of virtue.46 45  And we would be necessary beings not just in the “proxy” sense allowed by necessitists such as Bernard Linsky, Edward Zalta, and Timothy Williamson. In Karen Bennett’s deft way of putting it, we would necessarily be in the “showcase” of the actual world and not merely necessarily among the “stock” shared by all worlds. The latter is all Linsky, Zalta, and Williamson affirm. 46 What of Linda Zagzebski’s illuminating emphasis on divine virtue within her divine motivation theory? Does that provide an independent basis for God’s reason in creating being a completely adequate reason deriving from virtue? It looks like it does not, precisely because her theory is an exemplarist theory. In a nice summary of her impressive work on these matters in “Morality and Religion,” Zagzebski writes: The overall structure of the theory is exemplarist. Moral properties are defined via reference to an exemplar of goodness. God is the ultimate exemplar, but there are many finitely good human exemplars. In this respect, the theory is similar to that of Aristotle, (Nicomachean Ethics bk. 2, ch. 6, 1107a), who defines virtue as what would be determined by the person with phronesis (practical wisdom), and morally virtuous acts as acts that the phronimos would do in the circumstances in question.  (p. 357) And later adds: Here is a brief outline of divine motivation (DM) theory. The paradigmatically good person is God. Value in all forms derives from God, in particular, from God’s motives. God’s motives are perfectly good, and human motives are good insofar as they are like the divine motives as those motives would be expressed in finite and embodied beings. Motive-dispositions are constituents of virtues. A virtue is an enduring trait consisting of a good motive-disposition and reliable success in bringing about the aim, if any, of the good motive. God’s virtues are paradigmatically good personal traits. Human virtues are those traits that imitate God’s virtues as they would be expressed by human beings in human circumstances. The goodness of a state of affairs is derivative from the goodness of

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  151

XVIII.  There Could Not Be Melioristic Reasons to Create As our original inventory had it, a melioristic reason for an act is either a true proposition to the effect that the act would improve reality overall or a true proposition to the effect that the act would prevent the diminishment of value in reality overall, taking into account all dimensions of impersonal value, including hedonic value, aesthetic value, the value of any natures that exist or come to exist, and the value of desire or interest satisfaction. (Recall that melioristic value is impersonal in respect of making nothing of just who is involved.) There is a highly intuitive mistake about the extent of reality that grounds melioristic considerations, namely that it is always inevitably the future, or more generally how things go causally downstream from one’s acts, which alone determines or grounds such considerations. The concentration just on the future is valid when the past is, evaluatively speaking, finite. And this has almost always been assumed to be the case. When the past exhibits only finite value, then a future improvement will make things better overall. Relative improvement will guarantee improvement period. The heuristic of looking just to consequences to see how the melioristic reasons stand is thus a perfectly good one. Yet if there is already an infinite totality of good, as with a backwards infinite cycle of eternal return, then the heuristic will fail us. Adding good to reality is then, precisely, otiose. As Vital Cardinal Du Four presciently observed 700 years ago, that is just like adding a point to a line. And preventing the diminishment of good is like preventing a point from being taken from a line.47 Neither the removal nor the addition makes any difference to the length of the line. the divine motive. Outcomes get their moral value by their relation to good and bad motivations. For example, a state of affairs is a merciful one or a compassionate one or a just one because the divine motives that are constituents of mercy, compassion, and justice, respectively, aim at bringing them about. Acts get their moral value from the acts that would, would not, or might be done by God in the relevant circumstances.  (pp. 358–9) Suppose that God freely creates from some completely adequate divine motivation, then the true account of the rational adequacy of that motivation cannot simply appeal to it being God’s exemplary motive in creating. God is not, in any interesting sense, an exemplar for himself. 47  Like, but not exactly like, supposing that the set of points in a line has the cardinality of the reals. The antecedent infinite good could not have that cardinal measure, or indeed any such cardinal measure, since all cardinal measures are surpassable. Even so, the same “uneventfulness” of infinitary arithmetic applies; unsurpassable goodness is just that—unsurpassable and so unimprovable.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

152  Mark Johnston That is to say that not even a pro tanto melioristic reason for an act is to be found when the combined value of pre-existing reality and of the act and of its consequences is just the same as the value of pre-existing reality period. In such a case an act that brings about more good in the sense of bringing about extra goods does not bring about more good in the sense of making reality better overall. The infinite case thus forces us to distinguish these two ideas of “more good.” And then we see that melioristic reasons, reasons to the effect that this or that act makes the world better, need not get a foothold in every case in which this or that act promises to bring about extra goods. In the infinite case the mere bringing about of extra goods (whose measure is inevitably of less than or equivalent cardinality to those already in existence) will be otiose. It will not make reality better overall. The upshot is that if God is unsurpassably infinitely good, and undiminishably so, then he can have no melioristic reason to create, in particular he can have no meliorisitic reason to realize this or that individual specification. That could never make things impersonally better. That is just the upshot of considering the value of reality as a whole. For whether creation takes place or not, reality as a whole includes unsurpassable and undiminishable goodness. Again, the very idea of increase or diminishment of unsurpassable goodness simply does not get a foothold, anywhere, ever, under any conceivable circumstances. We are simply not in the realm of increase or diminishment of impersonal value. Recall our initial characterization of a melioristic reason. A melioristic reason in favor of an act is (i) a true proposition to the effect that the act would improve reality, understood as a totality of states of affairs, overall, or (ii) a true proposition to the effect that the act prevents the diminishment of value in reality overall. So, if God exists then since he is Absolutely Infinite Goodness or, at least, since his goodness is unsurpassable by any whole or plurality, then any such proposition is false. It follows that there are no pro tanto or contributory reasons deriving from considerations of increased overall value (or from considerations of avoiding any decrease of overall value), since their inherent manner of contributing as reasons is by way of their promising to increase overall impersonal value (or avoiding any decrease of overall impersonal value). Hence not only are there no completely adequate reasons deriving from considerations of impersonal value. Reasons of impersonal value make no contribution to any completely adequate or coercive set of reasons. They are always mere prima facie reasons, which a knowledgeable agent will always set aside.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  153 Someone could say: “I just care about making the future better.” That is a perfectly good piece of autobiography. But that pattern of concern is not thereby expressive of a melioristic reason to make the future better. The form of a melioristic reason has to do with making reality better overall. One might as well say “I just care about making things better in New Zealand.” For a New Zealander, or for someone who has deep attachments to that country, that remark might be intelligible as some sort of response to reason, but the reason in question would precisely not be a melioristic reason. Going beyond the merely autobiographical, someone might say: “Decision theory, when reasonably generalized away from the agent’s own proprietary preferences, shows that it is rational to focus on maximizing expected future utility.” But what is the ground of the prima facie rationality of focusing on maximizing expected future utility? It is inevitably something like the following. a. It is obvious that if you can make reality better overall you should, i.e., that is what reason requires, at least in the absence of moral considerations to the contrary. b. In acting we usually can only affect the future; that is our only way available to us of making things better overall. c. But we cannot know the future, in particular we cannot know the impact of our acts on long-term future utility, we can at most have reasonable expectations about this. d. So in being guided by a. our only method of implementing this guide is to aim to maximize expected future utility. The thing to see is that if God exists then a. is useless advice. For then we can’t make reality better overall. In creating, God could not have missed that point.

XIX. Ameliorism So if God exists it is not just that Impersonal Consequentialism has no capacity to distinguish the rightness or wrongness of acts. There are no pro tanto or contributing melioristic reasons to act. We can never rationally be guided by melioristic considerations in and of themselves! “Ameliorism” is the obvious neologism.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

154  Mark Johnston The meliorisitic life is then not a rational life. Yet, even we non-­ consequentialists seem embedded in just such a form of life. Even if the idea of making things better overall is not the only ethical idea at our disposal, it remains a central one. But if God exists it is not a workable idea at all! Of course, this intermediate conclusion is not itself the conclusion that there is no reason to care about either the quality of one’s acts or the consequences of one’s acts; for there may remain non-melioristic reasons to do just that. We should clearly distinguish two different uses of “better overall” as it applies to some act and its outcomes, a narrow use relied on in the discussion that follows, where being better overall implies an improvement from the melioristic point of view, and a more general use where it only implies that the weight of reasons, considered quite generally, favors the act and/or its outcomes.

XX.  “To Show Forth His Goodness” Our attempt to find God’s reason for creating has so far failed. God has no melioristic reasons to create. Furthermore, God has no reasons of personal attachment to create, and no completely adequate reasons deriving from the virtues. Nor in creating was he perfecting his nature. Nor was he advancing his own self-interest; i.e., fulfilling some need of his which otherwise would have been unmet. Nor was he under an authoritative command to create. Nor was he morally obliged to create. We have exhausted our initial inventory of reasons. Where does that leave us? The same anathema quoted in Section 6 from the documents of the First Vatican Council continues as follows: If anyone . . . denies that the world was created for the glory of God, let him be cut off from the community of the faithful and the saving grace of the sacraments.48

And the Baltimore Catechism presents the following dialog: Q. “Why did God make us?” A.  “God made us to show forth His goodness and to share with us His everlasting happiness in heaven.”49 48  Session 3 of Vatican 1 “Dogmatic Constitution on the Catholic Faith: 1. On God the Creator of All Things” Canon 5 (online text, Eulogos, 2007). 49  Baltimore Catechism No. 2 at Project Gutenberg, Question 3.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  155 These remarks, offered for the guidance of the faithful, have been interpreted as though they were citing either reasons deriving from an as-it-were proleptic personal attachment to us (the proleptic fallacy), or perfectionistic reasons involving God’s self-development, or melioristic reasons, or reasons of generosity, or some amalgam of these. But none of these kinds of reasons could be among God’s reasons to create. So what are the faithful being encouraged to believe? The doctrine actually seems clear enough, whatever one goes on to make of its truth or falsity. God created for this reason: to show forth his goodness, to manifest his glory, inter alia to us, in this life and in the heavenly life.50,51 But can we make any sense of this as an eighth kind of potentially ­adequate reason to create? Is sheer manifestation a reason in and of itself?

XXI. Manifestationism Suppose you are a painter abandoned on a desert island, certain you have no means of escape. Still you might paint, simply for the reason that the activity manifests or gives expression to your sensibility and skill. In each case, a reason for choosing to paint the particular thing you paint in the way that you paint it may be that the painting in question manifests your sensibility and skill. You may have other reasons as well, for example that continuing with painting gives your life a shape, a reason of prudential self-interest, 50  In his commentary on the sentences of Peter Lombard, Bonaventure remarks that God created all things “not to increase his glory, but to show it forth and [important conjunction] to communicate it.” See Commentaries on the Sentences of Peter Lombard: Works of St. Bonaventure: Translation, Introduction and Notes by Robert  J.  Karris (Franciscan Institute, 2012), II Sent. I, 2, 2, 1. 51  Compare also the emphasis on creation’s expressing and representing the divine perfection in this passage from John Finnis (2012), in “Philosophy and God’s Nature: Second Thoughts,” quoting from his own Aquinas: Moral, Political, and Legal Theory (Oxford: Oxford University Press, 1998), pp. 311–12: On the one hand, then, God cannot have chosen to create the universe to meet any need; lacking nothing, reality as a whole could not be improved by God’s choice to create. On the other hand, creating, directing, and sustaining the universe need not, and could not, have been pointless, lacking in intelligibility; like every other aspect, so to speak of the single divine act of being, the idea and act of bringing into being this universe cannot be deficient in intelligibility. [So] the point, the common good, of the universe must be the expressing, representing, and communicating (somehow sharing) of the divine perfection of actuality by bringing into being a universe of creatures which are each like . . . God in having actuality, perfections, intelligibility, and so forth. . . . By [their] flourishing, creatures and systems of creatures become more similar to . . . and more apt to represent . . . God, each in its own way; and all together they can express the inexhaustible divine perfection by their plurality and diversity . . . . (p. 196)

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

156  Mark Johnston and that it is pleasurable, a melioristic reason. The pure reason to manifest one’s nature is neither of these, nor is it essentially communicative. (Think of Caruso singing in the shower.) You might have lost all hope of your paintings ever being discovered. That might undermine your motivation to paint; but still it seems that there is some sort of objective reason to manifest your skill and sensibility by painting. Indeed, I believe that reason is still in place, even if manifesting your skill and sensibility by painting is not that good for you. (Imagine the pigments available on the island are toxic, and your lung collapses as a result of your painting. Still, the reason to manifest your skill and sensibility remains.) Could such a reason to manifest, a reason to the effect that this painting done in this way manifests one’s sensibility and skill, be, in and of itself, a completely adequate reason to go on painting? In part that depends upon the significance of what is manifested and in part on the adequacy of the manifestation. If you have little skill and less sensibility, or if the pigments available on the island are extremely limiting, then it is hard to see much point in going on painting, unless that point comes from other reasons, such as giving some structure to your life. In God’s case there is no question of the significance of what is to be manifested and no question of the adequacy of the manifestation being limited by lack of sensibility, skill, or resources. So, upon reflection, it does seem as if God might have an eighth kind of completely adequate reason to create, namely a pure reason of manifestation. If that is right, creation is not there to improve things in any impersonal way; there is no sort of melioristic consideration that would recommend it. Nor is creation an expression of his love for us, except in that metaphorical sense which invites the proleptic fallacy. Instead, God’s reason to create is to manifest his glory, to show forth his infinite goodness in a finite, created medium.52 Returning to the constraint, implied by our earlier discussion on something counting as a reason for God to create, recall our adverbial conception of creation as a manner of God’s affirming his own goodness. The truth of that conception turned out to be a necessary condition on the possibility of God having a completely adequate but non-coercive reason to create, even in the absence of an offsetting reason not to create. For on the alternative, 52  Is it a moral fault if he does manifest his goodness by creating a world that has an impersonally better variant? I cannot see that at all.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  157 conjunctive, conception God affirms his own goodness and then, as it were, considers the further act of creating or not creating, so that in the absence of an offsetting reason against creating he has a coercive reason to create (given there is some reason to create). The adverbial conception means that any proposed reason for God’s creating must be able to be understood as a reason that represents creation as a manner of God’s affirming his own goodness. The manifestationist view represents creation as God’s chosen expression of his goodness. God’s choosing to create to manifest his goodness does seem to be a determinate manner in which God can affirm his own goodness. Just as the painter who never expects to be rescued may reasonably affirm his own skill and sensibility by manifesting it in his painting, God may reasonably affirm his own goodness by manifesting it in creation. Perhaps here is the point to enter a caution about the implications of Manifestationism.53 It is an account of God’s reason for creating. In no way is it at odds with the idea that God is loving or generous or just toward his creatures. The point is only that those attitudes cannot be the grounding reasons for his creating, since they themselves are partly grounded in the existence of his creatures. Manifestationism does imply that we are among the means by which God gives expression to his goodness. But, crucially, that does not imply that as creatures we are mere means in any ethically loaded sense. We are not God’s mere playthings or display pieces. Part of the expression of God’s goodness is to create free beings “in his own image and likeness,” beings worthy of respect, and so beings which cannot be used by God in the way that some parents use their children, say, as trophies or display pieces. Indeed, once God’s creatures appear, they are objects of God’s overwhelming love and generosity. God can be loving and generous, without it being the case that his reason for creating was that it was the loving or the generous thing to do.

XXII.  How Mistaken We Have Been About What Matters If anything close to the foregoing is right then we have radically overestimated the importance of melioristic reasons (by thinking there were any) 53  What follows is partly in response to Meghan Sullivan’s helpful commentary at the NYU History of God Conference.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

158  Mark Johnston and massively underestimated the importance of reasons of manifestation (in part by conflating them with other kinds of reasons). Accordingly, our expectations as to how creation should turn out are likely to be way off, so that many of our objections to how it has turned out will simply be misplaced. Clearly this will be encouraging news, if it is news, to skeptical theists. In particular, any melioristic conception of God aiming to improve things by looking to bring about the best of all possible creations must be set aside. Instead, the relevant condition is that there be no defect in God’s act of self-expression that is his creating, so that creation is an adequate expression of God’s nature. Still, if all we have is that God created to manifest his goodness we are still in the dark as to his plot and his style of self-manifestation and even as to his intended genre of self-manifestation—tragedy or comedy, thriller or mystery, farce or horror tale—evidently it is none of these, but something else which we cannot, on our own, quite discern.54 Another part of what is called the hiddenness of God may consist in our inability to translate his expressive language into ours; and perhaps the very attempt at translating or decoding is in the end mistaken, as Nathaniel Hawthorne suggests in a striking passage describing his first sightings of Perugia and its environs: The lower part of the road was on the edge of the hill, with a narrow valley on our left; and as the sun had now broken out, its verdure and fertility, its foliage and cultivation, shone forth in miraculous beauty, as green as England, as bright as only Italy. Perugia appeared above us, crowning a mighty hill, the most picturesque of cities; and the higher we ascended, the more the view opened before us, as we looked back on the course that we had traversed, and saw the wide valley, sweeping down and spreading out, bounded afar by mountains, and sleeping in sun and shadow. No language nor any art of the pencil can give an idea of the scene. When God expressed himself in the landscape to mankind, He did not intend that it should be translated into any tongue save his own immediate one.55

54  If we were not on our own in the relevant sense, i.e., if there were an authoritative revelation, then a developed theology based upon it might illuminate the plot, genre, and style of God’s self-manifestation. Here one thinks, for example, of the theology of Hans Urs von Balthasar, in particular his The Glory of the Lord: A Theological Aesthetics (Edinburgh: Ignatius Press, 1982–91). 55 As quoted in Brian Harding’s introduction to The Scarlet Letter (Oxford: Oxford University Press, 1990), p. 6.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  159 Having said all that, it must be admitted that the face of the material universe does not have the features one would naïvely expect it to have, if it were the manifestation of God’s goodness. On its face, and for the overwhelming great part, the material universe is, at best, an immensely boring— and, at worst, a malignant—progression. (Cosmology and evolutionary theory have made this evident in ways that were unimagined by previous generations.) Moreover, the free rational wills we know best have evolved with a tendency toward moral evil, i.e., the subordination of the Good to their own good. That is just to say that manifestationism is not itself a theodicy. Indeed it is not itself even an etiology, either of natural malignancy or of moral evil. However, I believe that it provides a framework for the most promising ­etiology and consequent theodicy, which must of course be filled in by relying on authoritative revelation, if any there be. For it forces on us the question as to what made possible the freely chosen self-seeking orientations of the wills of the first fruits of God’s self-expression, those manifestations of his own free rational and powerful nature which were as rationally coherent and powerful as they could have been, compatible with their creaturehood.56

XXIII.  How Corrosive is Ameliorism when it Comes to Other Sorts of Reasons? If God exists and is unsurpassably infinitely good, then there are no melioristic reasons for God to create, or indeed to create in this fashion or that. But the deeper reality behind this is that if God exists and is unsurpassably infinitely good, then there are no melioristic reasons period; i.e., no melioristic reasons for any act, let alone creation. Infinite goodness in the beginning implies ameliorism ever afterwards. There are no pro tanto or contributing reasons to the effect that this or that act would, or would have a chance of, improving reality overall. Any kind of consequentialism which focuses on improving reality overall is just a nonstarter if God exists and is at least infinitely good to some cardinal degree greater than anything we can bring about or make probable. Just like the rest of us, but in a more focused and formal way, consequentialists and decision theorists who concentrate on impersonal value and the expected

56  See my “How did Evil Come into the World?” ms.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

160  Mark Johnston impersonal value of outcomes, have been habitually making mistakes by assuming that things are impersonally improvable. Reasons of manifestation aside, one has to wonder whether and to what extent other kinds of reasons remain in force given ameliorism, i.e., if there are no pro tanto melioristic reasons. Even if one rejects any consequentialist reduction of the other kinds of reasons to melioristic reasons, one may legitimately worry whether several of the other kinds of reasons come to lose some, if not all, of their importance. A reason is important if and to the degree that it is worthy of our attention and consideration in relevant contexts of action. If in acting we can never improve reality along the dimension of impersonal value, impersonal in the sense that the value in question does not turn on just who is benefitted, then the importance of other reasons may be lessened, and to that extent they may lose their apparent weight or force, and so their capacity to advance the rational case in favor or against actions. For example, if the flourishing of human beings adds nothing to the goodness of reality, how important are the virtues or dispositions of character whose whole point is to promote human flourishing? Is the virtuous life then any more than an idle kind of psychic athleticism or skill-development, where, as with pushpin, the exercise of the skill never makes anything better? The worry about importance may even apply to a straightforwardly deontological conception of moral reasons as categorical demands. Yes, they remain categorical oughts, and the consequentialist reduction of moral reasons is a non-starter, but do they have the importance we supposed if their observance never makes reality better as a whole? (Kant, who understood the nature of the categorical demands of morality as well as anyone, was deeply worried by this kind of thought, where the form of “betterness” that concerned him was justified happiness.) Moreover, even if one adopts the position of someone who has an antecedent reason of special responsibility to care about what is good for every person in some group, the reasons arising from the perspective of particular persons may seem less important. If doing what is good for Sally and for Frank and for . . . does not make things better overall, then it may seem like a somewhat indulgent fixation, unless there is something else, some other sort of backing reason, that favors it. That last observation won’t be convincing to philosophers who insist that facts concerning what is good for this one and that one are not grounded in what is good period. On such a conception, the reasons that

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  161 arise from what is good for a person remain in force even when reality as a whole is unimprovable, i.e., even when there are no normative reasons to the effect that benefitting some person makes things better overall. That is a tenable view, but only so long as we do not think of the relation good for as merely locational, i.e., as just a way of locating in the life of a person a good independently grounded as such. A consideration that might push one to the locational conception is the double-counting worry. There do appear to be facts about the location of impersonal goods in this or that life; if there are also independent facts about what is good for a person then we seem to have an odd sort of double counting. In any case, the theism I have been discussing fits most naturally with the locational view. God is unsurpassable infinite goodness; finite goods are manifestations of this unsurpassable infinite goodness, and something is good for a person because and insofar as it (i) orients him or her toward infinite goodness or (ii) introduces some one or other of these manifest finite goods into his or her life. In creating, God chooses among individuating specifications, in part on the basis of just how the goods that are specified in them would contribute to the overall manifestation of his goodness. On such a locational conception, and given ameliorism, it is hard to understand how there can be normative reasons deriving simply from facts about what is good for particular people. So in the theistic context ameliorism can be quite corrosive when it comes to the importance of non-melioristic reasons. And again, infinite goodness in the beginning implies ameliorism ever after. The naïve atheist might say: “Well, so much the worse for the idea that infinite goodness exists!” But a more sophisticated atheist will see that if the idea of infinite goodness in the beginning has so much as a positive nonzero probability, however small, then given Cantorianism it will follow that every act we could possibly perform makes the same contribution to the overall expected value of reality; that is, no contribution whatsoever.57 Call this consequence “E(xpected) V(alue)-ameliorism.” EV-ameliorism is quite 57  Suppose God has a non-zero probability of existing. Where N is huge, let it be 1/N. In a seminal paper, Alan Hayek notes that if God is so much as epistemically possible, i.e., if it is rational to assign his existence a non-zero probability, then within the Cantorian framework of measures every act has zero impact on overall expected value. For given Cantorian transfinite multiplication, overall expected value will be always be infinite: 1/N x Aleph-Whatever = Aleph-Whatever. Which we certainly can’t improve on! Decision theory is then useless as an account of what we should do. Unless, Hayek suggests, we move to non-Cantorian measures of utility such as the hyperreals or the surreals. See Hayek, “Waging War on Pascal’s Wager,” Philosophical Review 113 (2003), pp. 27–56.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

162  Mark Johnston corrosive even for those who believe that there are reasons that do not consist in or supervene on considerations to the effect that performing a given act as opposed to some alternative act positively affects overall expected value. Setting aside reasons of manifestation, the remaining non-melioristic reasons may now seem much less important. From the clearheaded atheist’s point of view, the importance of the practical reasons available to us may depend on the incoherence of the very idea of existing infinite goodness. If that is right, when Thomas Nagel wrote I don’t want there to be a God; I don’t want the universe to be like that58 he did not go far enough. He should have said I don’t want God to be epistemically possible; I don’t want the space of epistemic possibilities to be like that. Is that a rational desire or hope? More generally, would it be rational to treat both corrosive ideas—the idea of absolutely infinite value and the idea of unsurpassable transfinite value—as incoherent and so assign them zero probability? The problem with taking the idea of greater than finite value to be incoherent is that the idea has other, non-theological, sources; for example, in certain infinitary cosmological models and in the version of “the problem of personites” that arises if time is continuous.59 If it is rational to assign zero probability to the idea that greater than finite goodness exists then, bizarrely, we should thereby be able to rationally alter our credences concerning the extent of the universe and the nature of time. Strictly speaking, Hayek’s anxiety and his proposed solution are either mathematically misplaced or a sign of “transfinite idolatry.” God is not really in the picture here. Nothing with a merely transfinite degree of value, measurable by an Aleph, could be God, for its goodness would then be less than some measure of goodness. There is no greatest transfinite cardinal. Yet God’s goodness is unsurpassable. It has no measure in any reasonable number system, including the hyperreals or the surreals. It is, as Cantor put it, “an absolute maximum.” So contrary to Hayek’s suggestion there is no way of calculating God’s impact on expected utility. Some probability times what exactly? Still, this observation, though appropriate, does not get to the heart of the matter. I feel that those of us who are embedded in the melioristic form of life are not off the hook. My thought is that even though the mathematical argument that worried Hayek can be set aside, things can’t be better as far as the usefulness of decision theory understood as a normative guide goes, if there is already an absolute infinity of value rather than a merely transfinite value. With that thought goes another observation: The resort to non-Cantorian measures such as the hyperreals or the surreals will not save us from the threat posed by God to decision theory understood as a discriminating normative guide to which acts should be preferred. 58  The Last Word (Oxford: Oxford University Press, 1997), p. 130. 59  See my “The Personite Problem: Should Practical Reason be Tabled?” Noûs 50 (2016), pp. 617–44.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Why Did The One Not Remain Within Itself?  163

XXIV.  One Extant Solution to the Corrosive Effect of Ameliorism Whatever one’s assessment of the ultimate plausibility of the JudeoChristianity of the synoptic gospels might be, it is clear enough that it provides the basis for a distinctive reply to the kind of worry ameliorism raises concerning the importance of the various kinds of practical reason. God’s commands60 are genuinely reason-giving, even if all other sources of reasons have dried up.61 Moreover, some of these authoritative commands give importance and hence independent support to other kinds of reasons, such as reasons of love or personal attachment.62 So it is with the two “new” commandments: Love the Lord your God with all your heart and with all your soul and with all your mind. This is the first and greatest commandment. And the second is like unto that: Love your neighbor as yourself. (Matthew 22:37–8)

The second commandment is the mind-boggling and terrifying command to extend relations of agapic love to whoever turns up;63 as such, it underwrites what looks like a total commitment to universal welfare. The first command is inter alia a command to make the glorification of God the centerpiece of one’s life. Those who understand these commands should, I think, admit some sympathy with the young disciple “who trusted in riches” and who walked away, upon being invited to give up all he had (Mark 10:17–31). From the ordinary melioristic point of view, it looks as if things will very likely go terribly wrong if one does not walk away. That is only to say that those of us who understand the commands, and yet still trust in riches in the sense of the possibility of making things better, have not as yet found a space from which we can see that “the form of this world” namely the extraordinary prima facie appeal of the melioristic way of life, “has already passed away” (1 Corinthians 7:31). 60  Including the implicit command that is the exemplary character of the life of the “Son of Man.” 61  Here I assume that something like Robert Adams’s well-known proposal to deal with the Euthyphro dilemma is the correct theistic response to that problem. 62  Or do they instead provide surrogates for these, i.e., reasons which cover the same range of cases? 63  As the parable of the Good Samaritan, offered in response to the question “Who is my neighbor?” suggests.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

164  Mark Johnston The argument has been that it is the sheer consistency of the non-idolatrous idea of the infinitely good God that undermines the melioristic point of view. God’s existence is not required for that, but only for our finding a place in the space of reasons, after the undermining has been fully grasped. And here is a final surprising fact, which may serve to widen the “eye of the needle” for some of the clearheaded “rich”: If there is some epistemic possibility that time is continuous then theism’s major contemporary opponent, reductive naturalism, itself combines with a couple of nearirresistible premises to imply EV-ameliorism.64 Yes, the consistency of the idea of the infinitely good God implies EV-ameliorism and so undermines the melioristic perspective from which first-century Judeo-Christianity is unworldly madness. But reductive naturalism also has the same antimelioristic consequence! If that is right, our reflections not only provide a rationalist with a prima facie argument for the existence of God. They also suggest a kind of “reverse-Pascal” argument. If EV-ameliorism does corrode the importance of other sources of reason then rational hope, in the sense of the hope that we actually occupy a significant realm of practical reason, and are not just desire-driven beings who rationalize our impulses, may require the hope that the idea of an infinitely good and commanding God is not only consistent, but corresponds to reality.65

64  For the argument, see my “The Personite Problem,” and “Personites, Maximality and Ontological Trash,” in Philosophical Perspectives 30 (2016), pp. 198–228. 65  Versions of this paper were given at the NYU History of God Conference and the Rutgers Conference on God and Value, both held in November 2015, and as part of my Brackenridge Lectures at UTSA in October 2017. Thanks to Bob Adams, Michael Almeida, Ralf Bader, Paul Boghossian, Andrew Chignell, David Chalmers, Shamik Dasgupta, Cian Dorr, Kenny Easwaran, Philip Ehrlich, Johann Frick, Sherif Girgis, John Haldane, Paul Horwich, Klaas Kraay, Adam Lerner, Paul Lodge, Timothy O’Connor, Meghan Page, Mary Perret, Tim Pawl, Alexander Pruss, Gideon Rosen, Meghan Sullivan, Nick Stang, Nat Tabris, Larry Temkin, Peter Vallentyne, and Dean Zimmerman for their comments on the present thoughts, as well as to the John Templeton Foundation for sponsoring the aforementioned Rutgers Conference.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

8

Organic Unities, Summation, and the Problem of Evil Noah Lemos

1.  The Principle of Organic Unities and the Thesis of Universality There are different ways to approach the problem of evil. We might think, for example, that an omnipotent, omniscient, and morally perfect God must create the best possible world, or at least a world such that none is better than it is. We might wonder whether such a world could contain anything that was evil or intrinsically bad. Alternatively, we might think that God need not create such optimal worlds. Still, we might wonder whether any world created by an omnipotent, omniscient, and morally perfect God could contain evil. Is the existence of evil compatible with the world’s being created by such a being? No doubt there are other ways to formulate the problem. Many philosophers have thought the principle of organic unities relevant to the problem of evil. The principle of organic unities was, of course, brought to the fore in moral philosophy in G. E. Moore’s Principia Ethica. There Moore states two important theses concerning intrinsic value. The first is the principle of organic unities. Moore writes, The value of a whole must not be assumed to be the same as the sum of the values of its parts.  (1903, p. 28)

Moore denies that the values of some wholes are “mere sums” of the values of their parts. In keeping with this view, Moore held that what is i­ ntrinsically bad need not act as a “minus” or “negative” that lowers or detracts from the value of any whole to which it belongs. Indeed, a bad part can actually enhance the value of some whole to which it belongs. Similarly, something neutral might either enhance or lower the value of a whole to which it belongs and something that is good might lower the value of a whole of which it is part.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

166  Noah Lemos In addition to the principle of organic unities, Moore accepts the thesis of universality. He writes, The part of a valuable whole retains exactly the same value when it is, as when it is not, a part of that whole. If it had value under other circumstances, its value is not any greater, when it is part of a far more valuable whole; and if it had no value by itself, it has none still, however great be that of the whole of which it now forms a part.  (1903, p. 30)

According to the thesis of universality, the intrinsic value of a thing does not change from whole to whole or context to context. If an intrinsically bad part enhances the value of that whole, it does not cease to be intrinsically bad. It does not, as it were, enhance the value of the whole by ceasing to be intrinsically bad and becoming intrinsically good. When we consider aesthetic value something akin to the principle of organic unities seems true. Consider, for example, a rest in a piece of music. Considered in itself it has no beauty. It is simply the absence of notes. But a rest can greatly enhance the aesthetic value of a musical piece. Or consider some magnificent painting of the Crucifixion. The face of the crucified man might be an ugly face, bruised, bloody, and distorted in agony. But now imagine that we remove that part of the painting that contains the ugly face and replace it with the smiling, beautiful face of actress Scarlett Johansson. It does not seem that we have improved the aesthetic value of the painting. On the contrary, we have diminished its value. The presence of the more beautiful part made the whole less beautiful. If the principle of organic unities is true, then we should be at least suspicious of the view that the intrinsic value of a possible world is a sum of the value of its parts or even of its “basic” intrinsic value states.1 But, more importantly, the principle of organic unities leaves open the possibility that the presence of some evils or some intrinsically bad states of affairs do not detract from the world’s value and that some even enhance it. If the value of the world is not a mere sum of the value of the parts, then perhaps, for all we know, the evils in the world do not make the world worse and some might even make it better than it would be without them. Such a view has been attractive to many philosophers from A. C. Ewing and John Wisdom to more recently Robert Audi and Michael Almeida.2

1  I consider some of the difficulties with such a view in Lemos (1993, 1994, and 2010). 2  See, e.g., Wisdom (1935); Ewing (1947, 1953); Audi (2011); Almeida (2012).

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Organic Unities, Summation, and the Problem of Evil  167

2.  Chisholm on Defeat and the Problem of Evil Few philosophers, however, have done more to advance our understanding of the relevance of the principle of organic unities to the problem of evil than Roderick Chisholm. In his 1968 Presidential Address, “The Defeat of Good and Evil,” and later in his Brentano and Intrinsic Value, Chisholm develops an account of organic unities and discusses its implications for the problem of evil. The view he defends is squarely in the Moorean tradition insofar as he accepts both the principle of organic unities and the thesis of universality. I think it would be useful to consider his views in some detail. Both the principle of organic unities and the thesis of universality refer to parts. What is the relevant notion of a “part”? Chisholm takes states of affairs, or, more precisely, states of affairs that obtain to be the bearers of intrinsic value. So, the relevant notion of a part is one pertaining to states of affairs. Chisholm proposes: D1:  p is a part of q =Df. q is necessarily such that (i) if it obtains, then p obtains, and (ii) whoever entertains q entertains p.  (1986, p. 73) To entertain a proposition is to hold it before one’s mind. Entertaining a proposition is different from believing it. There are a great many propositions that I now believe that I am not entertaining. A few minutes ago, I believed, for example, that some elephants are gray, but I was not entertaining that proposition. Also, I can entertain a proposition without entertaining everything it implies. So, for example, I can entertain the proposition John is tall without entertaining the propositions, something is tall or something is something. Given D1, we may say that something is red and round has as a part, something is red. Similarly, I know that there are dogs has as a part, there are dogs. On the other hand, John believes that there are green monkeys does not have as a part, there are green monkeys, since the former is not such that if it obtains, the latter obtains. According to D1, every state of affairs is a part of itself. We need, however, the concept of a “proper part.” Let us say: D1’:  p is a proper part of q =Df. p is a part of q and q is not a part of p. Henceforth, when I refer to parts I shall be referring to proper parts. Chisholm illustrates and supports the principle of organic unities by appealing to various sorts of examples. One type of example concerns “pleasure in the bad.” Consider the following state of affairs: (1)  S is pleased that John is suffering.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

168  Noah Lemos Chisholm suggests that (1) is intrinsically bad. Still, he notes that (1) has a part that is intrinsically good, namely, (2)  S is pleased. Chisholm claims that (1) is an organic unity because it is a bad whole with a good part and no bad parts. Its intrinsic value is not equal to the sum of the intrinsic value of its parts. “But,” one might object, “isn’t John is suffering a bad part of (1)?” No, since (1) does not imply that John is suffering. S might simply believe ­mistakenly that John is suffering. One might also object, “I am not sure that (1) is intrinsically bad. Does S believe that it is bad that John is suffering? If he does believe that it is bad that John is suffering and he does take pleasure in it, then that state of affairs seems bad.” I am inclined to think that (1) is intrinsically bad whether or not S believes that John’s suffering is bad. Such pleasure is inappropriate or unfitting to its object whether or not S believes that it is. Still, suppose we hold that (1) is neutral. The state of affairs, S believes that John’s suffering is bad also seems intrinsically neutral. If we hold that the wider state of affairs, S is pleased that John is suffering and S believes that John’s suffering is bad, is itself intrinsically bad, then we still seem to have an organic unity since the wider whole is intrinsically bad, but it only has neutral parts. Still, I am inclined to think (1) is intrinsically bad and is an example of an organic unity. Another type of organic unity concerns “displeasure in the bad.” Consider (3)  S is displeased that John is suffering. Chisholm suggests that (3) is intrinsically good. Still, (3) has as a part, (4)  S is displeased, that is intrinsically bad. If this is right, then (3) is an organic unity insofar as it is an intrinsically good whole with a bad part and no good parts. The intrinsic value of (3) is not equal to a sum of the value of its parts. Finally, consider the following example of “merited suffering”: (5)  S is suffering and S deserves to suffer. One might hold that (5) is intrinsically good or, at least, not intrinsically bad. (5) may also be thought to be an example of an organic unity since it has a part that is intrinsically bad, namely S is suffering, and it has no ­intrinsically good parts.3 3  It might be that that certain states of affairs involving free action are also examples of organic unities. So, for example, it might be the case that (p) S performs x freely is intrinsically

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Organic Unities, Summation, and the Problem of Evil  169 In his discussion of organic unities, Chisholm introduces the concept of “defeat.” He suggests that the intrinsic value of a state of affairs can, in some cases, be defeated by a wider whole of which it is a part. So, for example, he suggests that the goodness of (2) is defeated by the wider whole, (1), of which it is a part. But what does it mean to say that the goodness of a state of affairs is defeated? Chisholm defines the general concept of defeat this way: D2:  Some of the goodness of G is defeated by W = Df. (i) G is a good part of W and better than W; and (ii) if W has a bad part that is worse than W, then that bad part is a part of G. (1986, p. 82) We may say, following Chisholm, that the goodness of (2) is defeated by (1). The first part of D2 is satisfied insofar as (2) is a good part of (1) and better than (1). The second part of D2 is satisfied trivially since (1) has no bad parts. Again, to say that the goodness of (2) is defeated by (1) is not to imply that (2) ceases to be intrinsically good or has a different intrinsic value when (1) obtains. Indeed, as D2 implies, the goodness of (2) is not defeated by (1) unless (2) is a good part of (1). The badness of a state of affairs may also be defeated. Chisholm says: D3:  Some of the badness of B is defeated by W = Df. (i) B is a bad part of W and worse than W; and (ii) if W has a good part that is better than W, then that good part is a part of B.  (1986, p. 83) Chisholm would tell us that the badness of (4) is defeated by (3). The first part of D3 is satisfied insofar as (4) is a bad part of (3) and worse than (3) and the second part of D3 is satisfied trivially insofar as (3) has no good parts. Chisholm would also tell us that the badness of S is suffering is defeated by (5) for similar reasons. Chisholm says that a state of affairs is defeasibly good if there is some wider state of affairs which would defeat its goodness. A state of affairs is indefeasibly good if there is no wider state that would defeat its goodness. A state of affairs is defeasibly bad if there is some wider state of affairs which would defeat its badness. A state of affairs is indefeasibly bad if there is no wider state of affairs that would defeat its badness. indifferent. It might also be the case that (q) x is generous is an intrinsically good state of affairs. Suppose, however, that the whole that consists of p and q is intrinsically better than q. If this is so, then the intrinsic value of p and q would not be a mere sum of its parts. We might contrast the previous case with the following. Consider (r) x is stingy. We might suppose that the whole which consists of p and r is worse than r. If this is so, then p and r would also seem to be an organic unity. Thus, one might think that freedom enhances the value of virtuous actions and aggravates the badness of vicious ones.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

170  Noah Lemos As we have seen, Chisholm suggests that the intrinsic goodness of (2)  S is pleased is defeated by (1)  S is pleased that John is suffering. Of course, other philosophers might take a different view. A Benthamic hedonist, for example, might hold that the goodness of (2) is indefeasibly good. Different axiologies will pick out different states of affairs as ­defeasibly or indefeasibly good or bad. Chisholm holds that propositions about whether a state of affairs is intrinsically good or bad are necessary and that the kind of knowledge we have of them is a priori. He also holds that propositions about whether a state of affairs is defeasible are also necessary and that our knowledge that they are defeasible is also a priori. However, propositions that the value of a state of affairs is actually defeated, as distinguished from being defeasible, are a posteriori and contingent. As Chisholm says “to know that the value of a state of affairs is actually defeated is to know, with respect to some state of affairs that would defeat its value, that that state in fact obtains” (1986, p. 95). Chisholm adds, “There is no absurdity in supposing that a rational being might know a priori, with respect to some state of affairs p, that p is good or that p is bad, and yet not know whether the value of p is defeasible” (1986, p. 95). This might happen if, for example, every state of affairs that might defeat the value of p is such that he or she had never even conceived of it. “Suppose, for example, that the goodness of pleasure would be defeated by the pleasure’s being undeserved but not defeated in any other way. And suppose there were a person who knew what pleasure is but who didn’t know what it is to deserve anything or what it is not to deserve anything. That person might know that pleasure is intrinsically good and believe ­mistakenly that it is also indefeasibly good” (1986, p. 96). I fear that all I have been able to do here is sketch in a small way some of what Chisholm says about the theory of value, organic unities, and the concept of defeat.4 His work on value theory is, I think, some of the most important in the twentieth century and rewards careful study. In any case, how do the concepts of organic unity and defeat bear on the problem of evil? Chisholm writes, “We encounter the problem of evil when 4  Chisholm distinguishes other kinds of organic unities than those involving the defeat of good and evil. In Brentano and Intrinsic Value, he distinguishes ten types of organic unities, including those which involve “enhancement” and “aggravation.”

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Organic Unities, Summation, and the Problem of Evil  171 we try to answer the simple question: ‘Is it possible for a totality that is ­better than any other possible totality to include anything that is intrinsically bad?’ ” (1986, p. 98). Chisholm suggests that a world that is at least as good as any other could contain states of affairs that are intrinsically bad if the badness of each of these states were defeated. Chisholm asks, “What if the evils of the world were defeated by a certain state of affairs q such that q is good and such that any possible state of affairs entailing q is better than any possible state of affairs not entailing q?” (1968, p. 37). Any world that contained q would be better than any world that did not. q would thus be a part of any optimal world. Moreover, if q defeats the evils of the world, then q has various evils as parts. It would follow, therefore, that any optimal world contains evil. Is there such a state of affairs that defeats all the evil of this world and, if so, what could it be? Chisholm writes, What possible state of affairs could thus serve to defeat the evil that is in the world? The wise theodicist, I should think, would say that he doesn’t know. Is it at least logically possible with respect to the evil that does exist that it is defeated? The most the theodicist has a right to say, I believe, is that it is epistemically possible. It may be, for all we know, that the evil of the world is defeated by some state of affairs that is absolutely good. And, it may also be, for all we know, that the goodness of the world is defeated by some state of affairs that is absolutely evil.  (1968, pp. 37–8)

Chisholm does not claim that that there is some state of affairs that defeats all the evil of this world. He does not even claim that it is logically possible that there is such a state of affairs. The most the theodicist has a right to claim, he suggests, is that, for all we know, there could be. Whether Chisholm is right, that this is the most the wise theodicist has a right to say, I do not know. Some philosophers, however, will think that the wise theodicist is not entitled to say even this much. This is because they reject the principle of organic unities. They think it is either incoherent or false. Since the concepts of an organic unity and defeat figure in so many discussions of the problem of evil, I wish to consider some prominent objections to them in the following sections.

3.  Is the View That There Are Organic Unities Incoherent? The Moorean position holds a part that is intrinsically bad or neutral can enhance or increase the intrinsic value of a whole to which it belongs. This

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

172  Noah Lemos is illustrated in (3) and (5) above. In this respect, the principle of organic unities holds that a thing can contribute to the whole a value which it does not have. Jonathan Dancy holds that this view is not only false, it is incoherent. According to Dancy, the Moorean view “severs the connection between ­values and reasons.” Dancy writes, The general idea is that where there is value there are reasons of certain sorts—reasons to protect, promote, cherish, respect, tend, approve, defend, and so on. So to say that a part with no value can contribute value that it has not got commits one to saying, it seems, that though there is no reason to preserve the part as a part, there is a reason to protect the whole, and that reason derives from the presence of the part. Now this does sound incoherent. Surely we have a reason to protect the part here, if it is contributing value. So its presence is of value, it would seem, on breaching the link between values and reasons.  (2003, pp. 630–1)

Why does Dancy think the Moorean view is incoherent? This argument is not as clear as one might hope. Still, I take Dancy to hold that the Moorean is committed to the following possibility: (6)  p has no value and p contributes value that it has not got. I assume that Dancy takes (6) to imply both,   (7)  There is no reason to preserve p, and   (8)  There is a reason to preserve p. I take Dancy to hold that (7) follows from the first part of (6), and (8) follows from the second part of (6). Dancy suggests it is incoherent to hold all three claims. I think Dancy is right. (6) does lead us to an incoherent position. If p has no value at all, no intrinsic value, instrumental value, no contributory value, if p has absolutely no value whatsoever, then there is no reason to preserve it. The problem with (6) is that it asserts that p has no value and it contributes to value. In other words, it asserts that p has no value and that p has contributory value. That is what makes it incoherent. I see no reason, however, to attribute (6) to Moore or to believe that that the Moorean position is committed to it. Instead of (6), the Moorean accepts the following possibility: (6’)  p has no intrinsic value and p contributes value that it has not got.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Organic Unities, Summation, and the Problem of Evil  173 But note that the first part of (6’) does not imply (7). It is possible that p has no intrinsic value and there is a reason to preserve p. Perhaps p has instrumental value, in which case there would be a reason to preserve it. Perhaps p has no intrinsic value, but it contributes intrinsic value to some whole to which it belongs. In that case, too, there would be a reason to preserve p. (6’) does not lead us into the incoherent position of (6). Of course, the claim that a thing has no positive intrinsic value does imply that there is no reason of a certain sort. I would say that if p has no positive intrinsic value, then the contemplation of just p does not require one to favor it for its own sake or as such. But, again, it is perfectly consistent to say that there is no reason to favor p for its own sake and there is a reason to preserve p (because, say, p has instrumental or contributory value). In any event, I do not think Dancy’s argument shows that the Moorean view severs the connection between reasons and values or that the Moorean position is incoherent. Even if Dancy has not shown the Moorean view to be incoherent, the view that something can contribute a value that it does not have might still seem false or puzzling. One might wonder, “How can something that is intrinsically neutral or intrinsically bad, enhance the intrinsic value of a whole to which it belongs? How can something contribute to the intrinsic goodness of a whole if it has none itself?” If we allow ourselves a metaphor, we might think in terms of one person giving money to another. If I have no money, then I cannot contribute any to you. In the same way, one might think, that if a thing lacks a certain value, it cannot contribute to some other thing’s having that value. Let us call this view “the principle of poverty.” If the principle of poverty is true, then something that lacks intrinsic goodness cannot contribute to the intrinsic goodness of some other thing, such as a whole to which it belongs. I think the principle of poverty is false. Consider the evaluative property of beauty. It has already been noted that that things which are not themselves beautiful can contribute to the beauty of wholes of which they are a part. Musical rests or pauses which are not themselves beautiful can contribute to the beauty of musical works, and images that are not themselves beautiful, such as suffering faces, can contribute to the beauty of paintings. It is false that something which itself lacks beauty cannot contribute to the beauty of some other thing. The principle of poverty does not hold in the case of beauty. Consider also epistemic justification. To say that a belief is epistemically justified is to make a positive evaluation of it. Epistemic justification is a kind

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

174  Noah Lemos of epistemically evaluative property. I assume that a belief can be justified in virtue of, say, the fact that one has a certain sensation or the fact that one’s belief is the product of a reliable process. Such facts can contribute to or confer justification on beliefs, but they themselves do not have justification. The fact that one has a certain sensation does not itself have epistemic justification, though it may confer justification on some belief. Similarly, the fact that a belief is the product of a reliable cognitive process may confer epistemic justification on that belief, but the fact that the belief is the product of a reliable cognitive process does not itself have epistemic justification. If this is so, then some facts that do not themselves have the evaluative property of epistemic justification can confer that property on other things, such as beliefs. The principle of poverty does not hold, then, for the value of epistemic justification. Finally, I assume that the property of being morally right is an evaluative property of actions. I also assume that some actions can be right in virtue of things which are not themselves actions, that do not themselves have the property of being right. Some actions can be right in virtue of their consequences and the consequences of an action need not include other right actions. The consequences of actions can sometimes contribute to or confer rightness on actions, even if they themselves do not have it. So, again, the principle of poverty does not hold for the evaluative property of moral rightness. I think that these examples show that the principle of poverty is false. They show that it is false that if a thing that lacks a certain value, then it cannot contribute to some other thing’s having that value. To the extent that one’s doubts about organic unities rest upon the principle of poverty, they rest upon something false.

4.  Does the Moorean View Lead to Evaluative Schizophrenia? Some critics charge that this Moorean view leads to a sort of “evaluative schizophrenia.” Consider again both: (1)  S is pleased that John is suffering, and (2)  S is pleased. As we have seen, Chisholm says that (1) is intrinsically bad and (2) is ­intrinsically good. (1) is a bad whole with a good part, (2), and no bad parts.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Organic Unities, Summation, and the Problem of Evil  175 Jonas Olson holds that the Moorean must say about a wicked pleasure such as [1] that while it is bad as whole it is good qua containing pleasure, since John’s being pleased has the same (positive) final value in [1] as in [2]. But this is problematic . . . .Worse still, on the familiar assumption that there is an intimate tie between attitudinal responses and values, it seems that we, according to the intrinsicalist [Moorean] picture, ought to take up a positive attitude (of rejoice, say) towards the whole, [1], qua its being an instance of pleasure, and a negative attitude (of condemnation, say) towards the very same whole qua its being an instance of wicked pleasure.  (2004, p. 42)

According to Olson, the Moorean is committed to holding that (1) is ­intrinsically bad and it is good qua containing pleasure. He also says that the Moorean is committed to holding that, given the assumption that there is an intimate tie between attitudinal responses and values, one ought to have both a positive attitude toward (1) and a negative attitude toward (1). Such a position “evokes a kind of evaluative schizophrenia in the evaluator” (2004, p. 42). I do not think the Moorean is committed to holding that (1) is both intrinsically good and intrinsically bad. He is certainly not committed to holding that (1) is intrinsically good because it contains pleasure or has (2) as a part. What some Mooreans are committed to is that (1) is intrinsically bad and that it has a part, (2), that is intrinsically good. But the fact that (1) has a part that is intrinsically good does not imply that it itself is i­ ntrinsically good. Since the Moorean denies that (1) is intrinsically good he is not committed to the claim that we ought to take up a positive attitude of rejoicing toward the whole. He need not hold, therefore, that one ought to both rejoice in (1) and condemn it. He is not, therefore, committed to evaluative schizophrenia. Still, Olson’s objection does raise a question for the Moorean position. Suppose we hold that (2) is a good part of (1) and (1) is intrinsically bad. What should be our attitude toward the whole and the part? Should we say, for example, that we ought to disfavor or condemn (1), but favor or rejoice in (2)? Or should we disfavor both (1) and (2)? In considering the question, we should recall a distinction that Chisholm draws between an antecedent emotion and a consequent emotion (1986, pp. 20–2). An antecedent emotion is one directed at a state of affairs when that state of affairs is considered in isolation. A consequent emotion is one directed at that state of affairs when it is considered in association with other states of affairs. So, for example, God antecedently disfavors that some men be damned, but when He contemplates some wider state of affairs that

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

176  Noah Lemos includes their sin and lack of repentance, He favors it. In such a case, He may be said to favor their damnation “implicitly,” though not for its own sake. I suggest we may say something similar about (1) and (2). We may say that considering just (2), S is pleased, in isolation requires that one have a positive attitude toward it. The appropriate antecedent attitude is one of favor. If, however, we are considering the wider state of affairs (1), S is pleased that John is suffering, the contemplation of this wider state of affairs requires that we disfavor (2). In such a case we would say that the appropriate consequent attitude toward (2) is one of disfavor. If this is right, then we may say that the appropriate attitude we should take toward (2) depends upon whether we are considering it in isolation or whether we are considering it as part of some wider whole. I suggest that when we consider (2) as part of (1) we should disfavor both the whole and the part. “But if contemplation of the wider state of affairs requires us to have a negative attitude toward (2), then doesn’t that mean that (2) is, in that wider context, intrinsically bad? And would this not violate the Moorean thesis of universality?” No, for we may say that the intrinsic value of (2) depends upon what would be an appropriate attitude toward (2) considered in isolation, as such, apart from any wider wholes of which it is a part. Even if the contemplation of a wider whole requires that we disfavor (2) “implicitly,” it is still the case that the contemplation of just (2) as such, in isolation, requires that we favor it. Consequently, the intrinsic value of (2) does not change. Consider the state of affairs, my child is in pain. The contemplation of just that state of affairs in isolation requires that one disfavor it. The appropriate antecedent attitude is one of disfavor. But now suppose that your child has been in a terrible accident and has suffered a spinal injury. The doctor tells you that if, when he awakes, he feels pain in his legs that is a good sign and it means he will probably be able to walk again. Now imagine that your child awakes and feels pain in his legs. You are overjoyed that he feels pain. Such a response seems, under the circumstances, entirely appropriate. In such a case you are considering a wider state of affairs, such as, my child is in pain and his being in pain is a sign that he will walk again. We may say that consideration of that wider state of affairs requires that one favor my child is in pain implicitly. In this case, the appropriate consequent attitude is one of favor. Still, from the fact that consideration of the wider state of affairs requires that one favor it, it does not follow that one desires for its own sake, my child is in pain, or that it is appropriate to desire it for its own sake or that it is intrinsically good.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Organic Unities, Summation, and the Problem of Evil  177 We should not simply assume, therefore, that if something is intrinsically good, then we ought to favor it and if something is intrinsically bad, then we ought to disfavor it. In some cases, it is appropriate to favor what is intrinsically bad and disfavor what is intrinsically good. To put the point more precisely, in some cases the appropriate consequent attitude to what is intrinsically good may be disfavor and the appropriate consequent attitude toward what is intrinsically bad may be favor. This can be the case, even if the appropriate antecedent attitudes are the opposite.

5.  The Objection from Evaluative Inadequacy Michael J. Zimmerman offers an important objection to the principle of organic unities. Zimmerman’s objection makes use of the concept of “evaluative inadequacy.” To say that a state of affairs is evaluatively inadequate is to say that it has no intrinsic value. It is neither intrinsically good, nor bad, nor indifferent. If a state of affairs is evaluatively inadequate, it cannot be ranked in comparison with other states of affairs. It is not intrinsically better or worse than any other state of affairs and it does not have the same intrinsic value as any other state of affairs, since it has none. To illustrate the concept of evaluative inadequacy, consider again the state of affairs: (2)  S is pleased. (2) does not tell us how much S is pleased. It does not tell us the duration of S’s pleasure. It does not tell us what S is pleased about. Zimmerman says that (2) does not contain enough information to warrant any particular attitude. He claims that the contemplation of just (2) does not require favor, disfavor, or indifference. Instead, Zimmerman says that the contemplation of an evaluatively inadequate state of affairs requires that we withhold any attitude (2001, p. 145). The appropriate antecedent attitude toward an evaluatively inadequate state is one of withholding. In order to support his view that (2) is not intrinsically good, Zimmerman asks us to consider the broader state of affairs, (1)  S is pleased that John is suffering. If we imagine that (1) obtains, he asks, “what reason is there to think that anything good has happened?” (2001, p. 145). Again, he asks,

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

178  Noah Lemos What reason could there be to declare [(2)] intrinsically good, what reason could there be to say that the contemplation of it as such requires that one favor it, given that it can happen that such a state occur in virtue of [(1)], contemplation of which (we are assuming) requires that one d ­ isfavor it? Surely what the contemplation of [(2)] requires is not favor, disfavor, or indifference. What contemplation of such a state requires is that one withhold any such attitude. The state is evaluatively inadequate.  (2001, p. 145)

Again, if a state of affairs is evaluatively inadequate we should withhold favor, disfavor, and indifference. We should also withhold preference. We should not prefer it to any other state of affairs and we should not prefer any other state of affairs to it. How does the concept of evaluative inadequacy bear on our examples of organic unities? Consider again: (1)  S is pleased that John is suffering. Chisholm says that (1) is an organic unity because it is an intrinsically bad whole that has an intrinsically good part, (2) and no intrinsically bad parts. But Zimmerman would say that (2) is not intrinsically good. It is evaluatively inadequate. So, he would say, we do not have an example of a bad whole with a good part and no bad parts. Something similar could be said about, (3)  S is displeased that John is suffering. Chisholm assumes that (3) is an organic unity since it is a good whole with no good parts and a bad part, namely (4)  S is displeased, and no good parts. Zimmerman would say that (4) is not intrinsically bad. He would say that it is evaluatively inadequate. Therefore, contra Chisholm, (3) is not a good whole with a bad part and no good parts. It is not, therefore, an organic unity. Similar considerations would apply to (5)  S is suffering and S deserves to suffer. (5) might appear to be an organic unity insofar as it is not intrinsically bad even though it has an intrinsically bad part and no good parts. But again, Zimmerman would urge that the state of affairs, S is suffering, is not ­intrinsically bad. It is evaluatively inadequate.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Organic Unities, Summation, and the Problem of Evil  179 I think Zimmerman’s objections to these examples are mistaken. There are four reasons why I think so. First, Zimmerman claims that (2) is evaluatively inadequate. If (2) is evaluatively inadequate, then it is not i­ ntrinsically better than any other state of affairs. But consider the following state of affairs: (6)  Ten billion innocent people are in extreme agony for years. I think that (2) is intrinsically preferable to (6). It seems to me that the contemplation of just (2) and (6) requires that one prefer (2) as such to (6). If this is right, and (2) is intrinsically better than (6), then (2) is not evaluatively inadequate. To put the point in a slightly different way, if (2) is evaluatively inadequate, then the contemplation of just (2) and (6) requires that one withhold preferring (2) as such to (6). But it seems false that this is so. Therefore, it is false that (2) is evaluatively inadequate. Similar considerations apply to (4). Zimmerman says that (4) is evaluatively inadequate. But consider: (7)  There being ten billion people enjoying innocent and deserved pleasures. I am inclined to think that (7) is intrinsically better than (4). The contemplation of just (7) and (4) requires that one prefer (7) as such to (4). Again, if (7) is intrinsically better than (4), then (4) is not evaluatively inadequate. Second, suppose that (1)  S is pleased that John is suffering obtains. Now suppose two men, A and B, each know that (2)  S is pleased. But suppose neither has any idea what S is pleased about. They have no beliefs about what S is pleased about. They simply believe that S is pleased. Imagine, though, that A takes intrinsic pleasure that S is pleased. A enjoys as such the fact that S is pleased. He is glad that S is pleased. B, on the other hand, takes intrinsic displeasure that S is pleased. It makes him miserable that S is enjoying something. He hates in itself the fact that S is pleased. On Zimmerman’s view, A’s and B’s attitudes toward (2) are inappropriate. Each of them should be withholding favor and disfavor toward (2). Moreover, since he holds that (2) is evaluatively inadequate, having no intrinsic value whatever, it seems that their attitudes are equally inappropriate.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

180  Noah Lemos But such a view does not seem right. It seems to me that A’s attitude toward (2) is appropriate and B’s is not. Given that neither has any idea what S is pleased about, A’s attitude seems generous and fitting and B’s seems mean spirited and petty. If this is right, then it is false that (2) is evaluatively inadequate. Third, Zimmerman holds that states of affairs such as (2) are evaluatively inadequate because they do not contain enough information to warrant any particular attitude. (2), for example, does not contain any information about the intensity, duration, or object of the pleasure. Since it does not contain enough information, the appropriate stance is to withhold any attitude toward it. Since it is not an appropriate object of intrinsic favor or disfavor or preference, it has no intrinsic value. One difficulty with this view is that almost all the things toward which people ordinarily take various intrinsic attitudes would seem to be evaluatively inadequate according to Zimmerman’s account. For example, I find that I am pleased that my son is happy, I am displeased that a co-worker is sad, I prefer for its own sake that a stranger be happy rather than sad. When I am pleased that my son is happy, the object of my pleasure is that he is happy. It is not that he is happy to degree ten for six hours about the outcome of a basketball game. I typically do not know the intensity, duration, and object of his happiness. When I am displeased that a co-worker is sad, the object of my displeasure is that she is sad. It is not that she is sad to degree nine for thirteen hours that her dog is ill. I hardly ever know the intensity, duration, and object of her sadness. Most of the objects of my ordinary attitudes of pleasure and displeasure, favor and disfavor, and preference would seem to lack the amount of information that would make them, according to Zimmerman, evaluatively adequate. Most of the objects of my ordinary attitudes would thus seem to be evaluatively inadequate by his account and, according to his account, the appropriate stance I should be taking to the objects of my ­ordinary attitudes is that of withholding any attitude. I think something is wrong with such a view. I think we can distinguish between a state of affairs being evaluatively inadequate and its being evaluatively indeterminate. If a state of affairs is evaluatively indeterminate, it lacks a determinate degree of value, but it can still be intrinsically good or bad and it can be intrinsically better or worse than some other state of affairs. If a state of affairs is evaluatively indeterminate, the contemplation of it can still require that we favor or disfavor it as such or that we prefer it as such to another state of affairs. So, for example, the state of affairs, my son

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Organic Unities, Summation, and the Problem of Evil  181 being happy, may be such that the contemplation of just that state of affairs requires that I favor it. It might be that the state of affairs does not contain enough information to have a determinate value, but it can still be an appropriate attitude of favor. Similarly, we may say that (2) is evaluatively indeterminate. It does not have a determinate degree of intrinsic value, but it is still intrinsically good. Moreover, it is intrinsically better than (6) insofar as the contemplation of just (2) and (6) requires that one prefer (2) as such to (6). Finally, Zimmerman denies that (1) and (3) are organic unities and that they are genuine cases in which the goodness or badness of a state of affairs is defeated. He denies this because he holds that that (2) and (4) are evaluatively inadequate. Still, he seems to grant, for example, that (1) is ­intrinsically bad. But now consider (1). Could it not be the case, for all we know, that there is some state of affairs, p, such that some of the badness of (1) is defeated by p? Couldn’t (1) be defeasibly bad? Perhaps there is, as some theists believe, some state of affairs that defeats all the evils in the world. Perhaps there are simply one or more states of affairs that would defeat some of the badness of (1) even if they do not defeat all the evils of the world. What might they be? Perhaps the badness of (1) would be defeated by each of the following: p’ S is pleased that John is suffering and John believes that S is pleased about his suffering and John is not angry. p’’ S is pleased that John is suffering and S is punished for being pleased that John is suffering. p’ ’’ S is pleased that John is suffering and God is entitled to punish S for being pleased that John is suffering and God does not punish S. Perhaps there are other states of affairs which would defeat the badness of (1). If this is right, then it seems to me that we have further examples of organic unities. But, what is more, since Zimmerman seems to agree that (1) is intrinsically bad, he cannot appeal to the evaluative inadequacy of (1) to deny that these are examples of organic unities. Still further, if (1) has an intrinsic value, but one that might be defeated by a wider state of affairs, then why should we not think that same of (2) and (4)? Why treat them differently? There is, I think, no good reason to do so. In conclusion, I believe that the principle of organic unities can be supported by plausible examples and that none of the arguments or ­considerations above give us good reason to doubt it. Perhaps Chisholm is

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

182  Noah Lemos right in suggesting that the best the wise theodicist can say is that it is ­epistemically possible that all the evils of the world are defeated. Perhaps it is not. I, for one, do not know.5

Bibliography Almeida, Michael J. 2012. Freedom, God, and Worlds. Oxford: Oxford University Press. Audi, Robert. 2011. Rationality and Religious Commitment. Oxford: Oxford University Press. Brentano, Franz. 1969. The Origin of Our Knowledge of Right and Wrong. Translated by Roderick Chisholm and Elizabeth Schneewind. London: Routledge & Kegan Paul. Chisholm, Roderick. 1968. “The Defeat of Good and Evil.” Proceedings and Addresses of the American Philosophical Association 42: 21–38. Chisholm, Roderick. 1981. “Defining Intrinsic Value.” Analysis 41: 99–100. Chisholm, Roderick. 1986. Brentano and Intrinsic Value. Cambridge: Cambridge University Press. Dancy, Jonathan. 2003. “Are There Organic Unities?” Ethics 113: 629–50. Ewing, A. C. 1947. The Definition of Good. New York: Macmillan and Co. Ewing, A. C. 1953. Ethics. New York: Free Press. Lemos, Noah. 1993. “Higher Goods and the Myth of Tithonus.” The Journal of Philosophy 90: 482–96. Lemos, Noah. 1994. Intrinsic Value: Concept and Warrant. New York: Cambridge University Press. Lemos, Noah. 2010. “Summation, Variety, and Indeterminate Value.” Ethical Theory and Moral Practice 13: 33–44. Moore, G. E. 1903. Principia Ethica. Cambridge: Cambridge University Press. Moore, G. E. 1912. Ethics. Oxford: Oxford University Press. Olson, Jonas (2004). “Intrinsicalism and Conditionalism about Final Value.” Ethical Theory and Moral Practice 7: 31–52. Wisdom, John (1935). “God and Evil.” Mind 44: 1–20. Zimmerman, Michael  J. 2001. The Nature of Intrinsic Value. Lanham, MD: Rowman and Littlefield. 5  An earlier version of this paper was presented at the August 2015 Rutgers Workshop on God and Value. I would like to thank the John Templeton Foundation for sponsoring that workshop.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

9

Organic Unities and the Problem of Evil A Reply to Lemos Michael J. Zimmerman

I  G. E. Moore’s best-known rendition of the principle of organic unities goes as follows: The value of a whole must not be assumed to be the same as the sum of the values of its parts.  (Moore 1903: 28)

Moore believes this principle to be of great importance to ethics (so great that he presents it entirely in italics), and many philosophers appear to agree with him about this, including, of course, both Roderick Chisholm and Noah Lemos. I remain unconvinced. Before the truth or falsity of the principle of organic unities can be determined, its meaning must first be made clear. Here there are several difficulties, on which I will touch only briefly. First, the sort of value with which Moore is concerned is what he calls intrinsic value. He uses this term primarily because he believes that whatever has this sort of value has it in virtue of its intrinsic properties alone. Lemos agrees with Moore about this, but others have disagreed (e.g., Korsgaard  1983; Kagan  1998; Rabinowicz and Rønnow-Rasmussen  1999), claiming that the sort of value that is at issue—the value that something has “for its own sake”—can sometimes supervene at least partly on its bearers’ non-intrinsic properties. I will go along with Lemos and Moore (and Chisholm) on this, in part because Lemos’s endorsement of the principle of organic unities presupposes both this thesis of intrinsicality and the closely related thesis of universality, according to which a whole’s intrinsic value is invariant under all circumstances. My aim is to cast doubt on the principle of organic unities even when these theses are granted.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

184  Michael J. Zimmerman Second, the sort of value with which Moore is concerned is an impersonal kind of value. It has to do with how things are “in the world,” so to speak; it is not the kind of value, for example, with which personal welfare has to do. As a possible example of this distinction, consider Schadenfreude, a phenomenon on which I will be commenting below. Let us suppose, in keeping with Moore and Lemos (and Chisholm), that such pleasure is intrinsically bad. What I take this to mean is, roughly, the following: the world is, insofar forth, a worse place for the occurrence in it of Schadenfreude. (Here, the phrase “insofar forth” is crucial and, as we will see, tendentious.) So saying is perfectly consistent with also saying that a person is, insofar forth, better off for the occurrence of Schadenfreude in his life—that such pleasure enhances his welfare. Third, just what sort of thing can be the bearer of intrinsic value is hotly disputed. (This issue is of course closely related to the question whether intrinsic value supervenes only on its bearers’ intrinsic properties.) Here Lemos parts company with Moore. At times, Moore talks of individual objects as having intrinsic value; at others, he talks of types of individual objects as having intrinsic value; at still others, he talks of the existence of individual objects as having intrinsic value; and, at yet others, he talks of states of individual objects as having intrinsic value. Lemos, following Chisholm, is much more circumspect, talking only of states of affairs as having intrinsic value. (More particularly still, Lemos talks only of states of affairs that obtain as having intrinsic value. He says that this is Chisholm’s view, too. I’m not sure that this is so—Chisholm seems to me to vacillate on the question—but I won’t pursue the point here.) Although I am inclined to a somewhat different account of what sort of thing can bear intrinsic value, for present purposes I will once again go along with Lemos on this matter. Fourth, there is the question what the part–whole relation comes to, when the bearers of intrinsic value are understood to be states of affairs. Here Lemos adopts Chisholm’s account, according to which a state of affairs p is a part of a state of affairs q if and only if q entails p (where entailment is understood to involve strict implication coupled with an entertainment condition). Again, I will follow suit (for the time being; but see Section IV below). As Lemos notes, it is clear that Moore intends the principle of organic unities to be concerned with a whole’s proper parts in particular. On the present approach, we may understand p to be a proper part of q just in case q entails p but p does not entail q. Fifth, there is the question whether and how it is meaningful to talk of summing the intrinsic values of different parts of a whole. If this is not

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Organic Unities and the Problem of Evil  185 meaningful, then of course the principle of organic unities is true, but only trivially so. But if it is meaningful, then there is reason to think that that fact alone suffices to show the principle false. Erik Carlson has recently argued that a statement to the effect that the value of a certain thing is, or is not, equal (or even just proportional) to the sum of the values of certain other things is meaningful only if the values in question can be measured on a ratio scale, but also that, if they can be so measured, then it follows that the value of a whole is identical to the sum of the values of its parts (Carlson  2015: 288–9; see Lemos  1994: 61 ff. regarding further problems with summation). Rather than get entangled in this question, I propose that we relax our conception of what constitutes an organic unity. As Lemos emphasizes, the crucial question is whether the intrinsic value of a state of affairs can be defeated by another state of affairs. He claims that such defeat can and does occur, whereas I doubt that that is so. (This is not to say that whatever is an organic unity must involve such defeat. Chisholm maintains otherwise (1986: 85 ff.). But it is to defeat in particular that theodicists who invoke organic unities must appeal.)

II  One objection to the view that there are organic unities is that posed by Jonathan Dancy (Dancy  2003) and discussed by Lemos. I should say straightaway that I believe Lemos’s response to Dancy succeeds. Here I will briefly reconstruct portions of this response in order to explain where it is that my bafflement at the notion of an organic unity has its source. Consider the following proposition regarding two states of affairs, p and q: (1)  p has no intrinsic value and q has intrinsic value and p contributes its intrinsic value to q. This proposition is clearly incoherent. No state of affairs can contribute value that it does not have. Now consider: (2)  p has no intrinsic value and q has intrinsic value and p contributes to the intrinsic value of q. This proposition is not incoherent. As Lemos notes, it seems clear that p can contribute to the intrinsic value of q even if p has no intrinsic value itself.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

186  Michael J. Zimmerman Having said that, I should add that it seems that “contribute” must mean something different in (1) from what it means in (2). If p contributes its intrinsic value to q, then q will derive at least some of its intrinsic value from p. If p contributes to the intrinsic value of q even though p has no intrinsic value of its own, then q won’t derive any of its intrinsic value from p; rather, q’s intrinsic value will supervene on p (or something to do with p). However exactly the relations of derivation and supervenience are to be understood, it’s clear that they are distinct relations. After all, even something that has value non-derivatively must be such that its value supervenes on something. Now consider: (3)  p has intrinsic value and q has intrinsic value and p is a part of q and p does not contribute its intrinsic value to q. This proposition gives expression to the idea that p’s intrinsic value is (totally) defeated by q. It implies that p’s value is ineffectual—that its value is completely disconnected from the value of q, despite the fact that it is a part of q. I’m not sure whether this proposition is incoherent, but it is here that my bafflement arises. I don’t understand how value can be ineffectual in this way. (I said above that the claim that Schadenfreude is intrinsically bad can be understood, roughly, as the claim that the world is, insofar forth, a worse place for the occurrence in it of such pleasure. I also said that the phrase “insofar forth” is both crucial and tendentious. That it is crucial is easily seen. Suppose that, on some occasion, an episode of Schadenfreude happens to have such good consequences that the world is better than it would have been, had this episode not occurred. Then this episode will turn out to have been instrumentally good; that of course does not alter the fact that it was intrinsically bad. That the phrase is tendentious can be seen from the fact that it presupposes that the intrinsic value of Schadenfreude contributes to the value of the world and is thus not defeated. Someone who accepts the possibility of organic unities will therefore reject this characterization of the sort of impersonal value at issue.)

III  The second objection to the view that there are organic unities discussed by Lemos concerns their involving a kind of “evaluative schizophrenia.” The objection presupposes (as does Lemos, and as will I) a so-called “fitting attitudes” view of intrinsic value according to which the contemplation of an intrinsically good state of affairs requires that one display a kind of favor

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Organic Unities and the Problem of Evil  187 toward that state of affairs, while the contemplation of an intrinsically bad state of affairs requires that one display a kind of disfavor toward it. The complaint is that, in those cases in which, say, the intrinsic badness of some state of affairs is supposedly defeated by some wider state of affairs, whoever contemplates the states of affairs in question would be required to display a form of misplaced schizophrenia (Olson 2004: 42). Lemos responds by saying that, although in such cases ambivalence would indeed be required, it would not be misplaced. He gives the example of his son’s being in pain being a sign that he (his son) will walk again, and he notes that it is surely appropriate for him to welcome his son’s pain for the sake of that intrinsically good state of affairs of which it is a sign, even though it is also appropriate for him to disfavor his son’s being in pain for its own sake. This response has much merit. Ambivalence of the sort mentioned by Lemos is indeed often appropriate. Many situations call for it. Another common kind of case is one in which the obtaining of some intrinsically bad state of affairs has intrinsically good consequences. Consider the case, just mentioned, of Schadenfreude that has good consequences. Or consider the expression, “Once bitten, twice shy.” A painful lesson might well be a worthwhile one. Nonetheless, I don’t think that Lemos’s response wholly disposes of the worry at issue. In cases of the sort just mentioned, in which disfavor of some state of affairs, p, for its own sake is appropriately coupled with favor of that same state of affairs for the sake of some other state of affairs, q, to which it is in some way related, it seems that we can correctly say, “If only q had obtained without p! Then things would have been better still.” But it seems that we cannot correctly say this in a case in which the intrinsic badness of p is supposedly defeated by q. In particular, it would be very odd for a theodicist to declare, “If only this world had obtained without the evil in it!” Not only is this exclamation of dubious coherence, it fails to acknowledge what strikes me as a troublesome incongruity—a kind of schizophrenia—that is involved in lamenting any part of a world that is good enough for God to have created it.

IV  The third objection that Lemos discusses concerns what, in Zimmerman 2001, I called evaluative inadequacy. Consider a particular episode of Schadenfreude, such as John’s being pleased at Mary’s pain. Call this state of

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

188  Michael J. Zimmerman affairs q, and let us agree that it is intrinsically bad. Then, I suggested, the state of affairs, p, that is John’s being pleased and is a proper part of q should not be declared intrinsically good (or intrinsically bad, or intrinsically neutral) but rather declared evaluatively inadequate; that is, it should be said to have no intrinsic value at all. My reason for saying this was that contemplation of p would appear to require not directing favor (or disfavor, or indifference) toward it for its own sake but, rather, the withholding of any such attitude. I went on to suggest that many purported instances of organic unities could be treated in like fashion (Zimmerman 2001: 142 ff.). Although he doesn’t raise the issue in his contribution to the present volume, in an earlier paper Lemos charged me with making a fundamental mistake here (Lemos 2006: 79). The mistake concerns the proper application of the isolation method for determining the intrinsic value of something. Moore is well known for saying the following: In order to arrive at a correct decision on the [question “What things have intrinsic value?”], it is necessary to consider what things are such that, if they existed by themselves, in absolute isolation, we should yet judge their existence to be good.  (Moore 1903: 187)

The key idea is that, to assess something’s intrinsic value, we must focus on the value that it has for its own sake, rather than for the sake of some other thing to which it may be related, and to do that we must isolate it from any such other thing. Lemos endorses this idea, as do I. In Lemos 1994: 10–11, he draws a crucial distinction between two versions of the isolation method. He calls Moore’s version “ontological isolationism” and declares it “not very helpful.” He is surely right about that. Moore’s version is in fact incoherent, since nothing can exist by itself, in absolute isolation. Following Chisholm, Lemos maintains that the correct version to use is that of “intentional isolationism.” This is the form of isolationism to which he alludes in the present paper. It involves ensuring that, when attempting to assess the intrinsic value of some state of affairs, we contemplate or consider just that state of affairs as such. For example, when attempting to assess the intrinsic value of John’s being pleased, we must ensure that it is that state of affairs alone that we are considering, and not the fact that it is embedded in the wider state of affairs that consists in his being pleased at Mary’s pain. I agree entirely, but putting this principle into practice can be a tricky business. When arguing, on the basis of the assumption that q (John’s being pleased at Mary’s pain) is intrinsically bad, that p (John’s being pleased) should be

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Organic Unities and the Problem of Evil  189 declared evaluatively inadequate, I noted that p could be part of some other state of affairs, r (John’s being pleased at Peter’s good fortune), which, for the sake of argument, we should take to be intrinsically good. That is, I  noted that p could be contained both in wholes, the contemplation of which as such requires that we disfavor them for their own sakes, and in wholes, the contemplation of which as such requires that we favor them for their own sakes. Given this “flexibility” in p, and given the fact that, on any occasion on which p obtains, it will and must do so as part of some wider state of affairs, I suggested that the contemplation of p as such requires that we withhold any attitude toward it for its own sake. Lemos complained that this argument violates intentional isolationism because it bases the ­evaluation of p on a consideration of q and r. I understand the charge but deny it. My purpose in mentioning q and r was not to say that the ­evaluation of p should take these other states of affairs into consideration, but rather to draw a contrast between p and another state of affairs, p*, which consists in John’s being (as I put it) just pleased. By “being just pleased” I meant something along the lines of being pleased in a way that is neither appropriate nor inappropriate, neither deserved nor undeserved, etc. My point was simply this. It may be that the contemplation of p* as such requires that we favor it for its own sake, but this of course does not imply that the contemplation of p as such requires that we favor it for its own sake. Further, while p is of course a part of q (and of r), p* is not. My suggestion was that, if it seems tempting to say that, although q is intrinsically bad, p is nonetheless intrinsically good, that is because p is being confused with p*. I expect that Lemos would in turn deny the charge that he is confusing p with p*, and maybe he isn’t. But if he isn’t, then I don’t know why he wants to declare p intrinsically good. Moreover, I continue to think that he might be confusing the two. Again, implementing intentional isolationism can be really tricky. Consider, by way of analogy, another case that Lemos mentions in which something like an organic unity might be thought to exist: the striking case of Scarlett Johansson’s beautiful face diminishing the value of a painting of the Crucifixion. Should we declare her face beautiful as such, out of context? That is not at all clear to me. Johansson also has beautiful shoulders, but imagine her nose being replaced by one of them. Her face would no longer be so beautiful. Would we, should we, still declare the shoulder beautiful as such? (As what? As a shoulder? As a nose? As itself?) I find it hard to know what to say. I have raised these questions about what successfully putting intentional isolationism into practice involves because it seems to me that some of

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

190  Michael J. Zimmerman Lemos’s objections to my view might be rebutted by charging him, rather than me, with the misapplication of this method. Consider, first, the comparison of p, John’s being pleased, with s, ten billion innocent people being in extreme agony for years. Lemos says that the contemplation of just p and s (as such) requires that we prefer p (for its own sake) to s (for its own sake), and thus that it cannot be that p is evaluatively inadequate. But I wonder. Let us agree that s is extremely bad, intrinsically, and thus that contemplation of just s as such requires that we disfavor it intensely for its own sake. What of p? Here I think the matter is not so clear. When addressing the matter of evaluative schizophrenia, Lemos invokes Chisholm’s distinction between an antecedent and a consequent attitude. He would say that the contemplation of p as such requires that we favor p for its own sake (an antecedent attitude), whereas the contemplation of p as being a part of q requires that we disfavor p for the sake of q (a consequent attitude). I think we should invoke this same distinction in the present context. Suppose that we agree that there is no intrinsically bad state of affairs of which p can be a part whose intrinsic badness comes anywhere close to the intrinsic badness of s.1 Then the degree of any consequent disfavor of p that could possibly be required when contemplating p would not come close to matching the degree of antecedent disfavor of s that is required when contemplating s. It is in this way, I think, that we could correctly say that p is indeed preferable to s, even if p is evaluatively inadequate. Something similar may be said in response to Lemos’s second and third objections. If I see that John is pleased but have little or no idea what it is that he is pleased about, then a certain generosity of spirit might prompt me to assume that his pleasure is not inappropriate or undeserved. In such a case, it seems perfectly appropriate for me in turn to be pleased at his being pleased, not, strictly, for its own sake, but for the sake of whatever wider state of affairs it happens to be a part of. Such generosity of spirit may on occasion be misplaced, however. If John has a long history of Schadenfreude and I know this, that should surely cause me to hesitate to take pleasure in his being pleased. (I agree that there is a distinction between evaluative inadequacy and evaluative indeterminacy and that the latter does not imply 1  To be acceptable, this claim requires qualification. There are, I think, three ways in which one state can be a part of another. The first involves the latter having a constituent property that entails the constituent property of the former. The second involves the latter being a conjunctive state of which the former is one of the conjuncts. The third involves a combination of these two relations. It is only the first of these that is in play in my account of evaluative inadequacy. See Zimmerman 2001: 58 ff. and 142 ff.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Organic Unities and the Problem of Evil  191 the former. It may indeed be that the intrinsic value of some evaluatively adequate states is indeterminate, a point that I concede at Zimmerman 2001: 179–80.) In recent correspondence Lemos has indicated that this response doesn’t go to the heart of his concern. Suppose that two men, A and B, give no thought whatsoever to what it is that John is pleased about or to whether his pleasure is appropriate or deserved. They simply contemplate p (John’s being pleased) as such, and A favors p for its own sake whereas B disfavors it for its own sake. Isn’t it clear that, although B’s attitude is inappropriate, A’s attitude is not, and doesn’t this show that p is not evaluatively inadequate? I am inclined to agree that A’s attitude is not inappropriate, whereas B’s is, but I don’t think this implies that p has intrinsic value. My official position is that, on the assumption that Schadenfreude is not intrinsically good, contemplation of p as such requires both withholding favor of it for its own sake and withholding disfavor of it for its own sake. This implies that withholding these attitudes is appropriate, but it doesn’t imply that the failure to withhold them is inappropriate. On my account (Zimmerman 2001: 92 ff.), (in)appropriateness comes in degrees, whereas requirement does not. More precisely, what is required is that which is most appropriate, and so it remains possible that an attitude that one is not required to hold—indeed, an attitude that one is required not to hold—is nonetheless not ­inappropriate; in fact, it could even be appropriate to some less-than-maximal degree. In the present case, I am inclined to think that favoring p for its own sake is neither appropriate nor inappropriate, whereas disfavoring it for its own sake is positively inappropriate. This verdict matches one that I reached when discussing a similar question, namely, the question of what attitude is required when one contemplates an intrinsically neutral state of affairs. In that case I suggested that, although indifference is required, favor would not be inappropriate, whereas disfavor would be (Zimmerman 2001: 109 ff.). In even more recent correspondence Lemos has indicated that this response still does not go to the heart of his concern. Consider yet another comparison, this time between p, John’s being pleased, and t, Mary’s being pained, and imagine that the mean-spirited B contemplates each of p and t as such and prefers t for its own sake to p for its own sake. If not only p but also t is evaluatively inadequate, then, given what I have just said, I seem committed to saying that, although B is required neither to favor nor to disfavor either p or t, it is not inappropriate for him to favor either one. But surely, even if it is not inappropriate for him to favor p for its own sake, it is

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

192  Michael J. Zimmerman indeed inappropriate for him to favor t for its own sake, and hence ­inappropriate for him to prefer t to p. I am inclined to agree with Lemos here. Such agreement requires me either to say that, although it is not inappropriate to favor some evaluatively inadequate states of affairs, it is inappropriate to favor others, or to say that t is not evaluatively inadequate but intrinsically bad. (It does not require me to retract the claim that p is evaluatively inadequate.) Since I don’t know of a principled way of defending the first option, I am inclined to embrace the second. This might seem to be a fatal concession, though. Doesn’t it require that I admit defeat (as it were)? Consider u, Mary’s being pained at Sam’s pain. Isn’t u, of which t is a part, intrinsically better than t, and doesn’t that mean that u is an organic unity, and thus that at least one such unity exists? And if one, why not many? Let us agree (as, for present purposes, I am prepared to do) that t is a part of u and that u is intrinsically better than t. That requires me to declare u an organic unity only if I cannot identify another part of u whose value can be balanced against the disvalue of t in such a way as to account for the value of u without recourse to declaring t’s value to have been defeated by u. I suspect that, on the Chisholm–Lemos account of the part–whole relation, it is indeed the case that no such part can be found. But that may simply provide reason to adopt a different account of this relation (as indeed I do at Zimmerman 2001: 58 ff.). For notice that there is obviously a feature that u possesses but t lacks, and that is that Mary’s pain in u is appropriate. Consider the state of affairs v, Mary’s experiencing an appropriate emotion. It is arguable that v is intrinsically good. Indeed, I suspect that Lemos would agree that v is intrinsically good. If, contrary to the Chisholm–Lemos account,2 v should be declared a part of u, then the case for declaring u an organic unity is weakened considerably. Finally, let me note that it is only for the sake of argument that I have granted that an episode of Schadenfreude such as John’s being pleased at Mary’s pain is evaluatively adequate. I readily acknowledge that this might in fact not be the case, and that only wider states of affairs are evaluatively adequate (Zimmerman  2001: 146). My claim is simply that evaluative ­adequacy emerges only at a certain level of complexity in the constituent properties of states. Axiologies will often differ on what that level of

2  Why contrary? Because it seems clear that u is not necessarily such that whoever entertains it entertains v (even if u is necessarily such that, if it obtains, then v obtains).

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Organic Unities and the Problem of Evil  193 complexity is, and I leave entirely open the question which axiology we should accept (Zimmerman 2001: 151).

V  I used to think it mattered a great deal whether the principle of organic unities was true, but I have now come to regard the debate between its ­proponents and its opponents as something of a tempest in a teapot. Both proponents and opponents typically accept the same evaluations of the wholes in question, and that is where the practical implications lie. (Should we teach our children that it is inappropriate to take pleasure in other people’s misfortune, for example? I believe that we should, and I expect that Lemos would agree.) Moreover, in the absence of a cogent argument against the very possibility of organic unities (and I certainly don’t claim to have provided one), it is unclear just what theoretical significance is to be attributed to the debate over whether the principle of organic unities is true. I  once thought that finding a principled way to reject the principle (by invoking the concept of evaluative inadequacy, for example) would facilitate the computation of intrinsic value, but that now seems to me a pipedream. For unless and until a method for precisely measuring the intrinsic values of parts of states is found, evaluating the wholes that comprise them will continue to be a matter of intuition rather than computation. Having said that, I should now append a qualification. There is one area of inquiry in which, it seems to me, it matters considerably whether the principle of organic unities is true: the problem of evil. Theists and atheists do not typically accept the same evaluation of that particular whole that is the actual world. Although, contrary to Chisholm (1986: 100 ff.), I don’t think that a theodicist must invoke the concept of defeat in order to have any hope of succeeding at his task, it is surely the case that that task would be rendered easier if it were reasonable for him to invoke this concept, and it would indeed be reasonable for him to do so if it were reasonable to claim, as Lemos does, that organic unities are ubiquitous. But if, as I think, it is not so reasonable to make this claim, then the theodicist’s task is not so easy after all.3 3  I am very grateful to Dean Zimmerman and Noah Lemos for their comments on a previous draft and, more generally, to Noah for his persistent and penetrating opposition over the years. An earlier version of this paper was presented at the August 2015 Rutgers Workshop on God and Value, sponsored by the John Templeton Foundation. I would like to thank the Foundation for its support and the other conference participants for their feedback.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

194  Michael J. Zimmerman

References Carlson, Erik (2015). “Organic Unities.” In The Oxford Handbook of Value Theory, ed. I. Hirose and J. Olson (Oxford: Oxford University Press): 285–99. Chisholm, Roderick  M. (1986). Brentano and Intrinsic Value (Cambridge: Cambridge University Press). Dancy, Jonathan (2003). “Are There Organic Unities?” Ethics 113: 629–50. Kagan, Shelly (1998). “Rethinking Intrinsic Value.” Journal of Ethics 2: 277–97. Korsgaard, Christine (1983). “Two Distinctions in Goodness.” Philosophical Review 92: 169–95. Lemos, Noah (2006). “Indeterminate Value, Basic Value, and Summation.” In The Good, the Right, Life and Death, ed. K.  McDaniel et al. (Aldershot: Ashgate): 71–81. Lemos, Noah (1994). Intrinsic Value: Concept and Warrant (New York: Cambridge University Press). Lemos, Noah (2019). “Organic Unities, Summation, and the Problem of Evil.” In Oxford Studies in Philosophy of Religion, Volume 9, ed. L.  Buchak, D.  W.  Zimmerman, and P.  Swenson (Oxford: Oxford University Press): 165–182. Moore, G. E. (1903). Principia Ethica (Cambridge: Cambridge University Press). Olson, Jonas (2004). “Intrinsicalism and Conditionalism about Final Value.” Ethical Theory and Moral Practice 7: 31–52. Rabinowicz, Wlodek and Rønnow-Rasmussen, Toni (1999). “A Distinction in Value: Intrinsic and For Its Own Sake.” Proceedings of the Aristotelian Society 100: 33–52. Zimmerman, Michael  J. (2001). The Nature of Intrinsic Value (Lanham, MD: Rowman and Littlefield).

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

10

Unrestricted Actualization and Divine Providence Bruce Langtry

Michael Almeida has recently made interesting and highly ingenious contributions to philosophical discussion of divine providence. Some of them have involved his making use of his novel claim that necessarily, God can unrestrictedly actualize various states of affairs, actions, events, and possible worlds.1 (From now on, except in quotations, I will abbreviate “unrestrictedly actualize” by “U-actualize,” and “possible world” by “world.”) This chapter introduces the core idea, draws attention to its bearing on some current issues in the theory of goodness, and discusses whether Almeida provides good reasons for believing that (C)  Necessarily, God has available an infallible method by which God can bring about whatever undetermined events God chooses, except those which it is metaphysically necessary or accidentally necessary that God does not bring about. Almeida works with a libertarian view of free choice and action, but rejects Molinism, a variety of libertarian theism which implies that some worlds containing free human actions and other undetermined events are feasible for God, in the sense that contingently, God has available an infallible way of bringing it about that any one of them is actual; other worlds, however, are not feasible for God, even though it is metaphysically possible that God has an infallible way of bringing it about that one of these is actual. Which worlds are 1  I take a world to be a possible state of affairs S such that for every other state of affairs S*, either S* or not-S* is contained in (i.e., is a part of) S. The possibility involved will always be metaphysical possibility, i.e., what some people call broadly logical possibility. The actual world is the world that contains all and only the actual states of affairs. The proposition that p is true in a world S if and only if S contains the state of affairs Its being true that p. Necessarily, if God brings it about that p then the actual world contains the state of affairs God’s bringing it about that p. So necessarily, God can bring it about that a world S is actual only if S contains God’s bringing it about that S is actual. Whether God can do anything that it is metaphysically possible that God does is a controversial matter, to which Almeida’s ideas are directly relevant.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

196  Bruce Langtry feasible for God is said to depend on which counterfactuals of creaturely freedom (CCFs) are contingently true, a matter which is not up to God.2 Thus Molinism implies that (C) is false. Open Theism, another variety of libertarian theism, implies that no worlds containing undetermined events are feasible for God in the foregoing sense; if God aims to bring it about that one of these worlds is actual, whether he succeeds is not completely within his control. Thus Open Theists agree that God cannot do everything such that it is metaphysically possible that God does it. Almeida describes himself as a Moderate Anselmian. That is, he believes that God is a necessary being and is necessarily omnipotent, all-knowing, and perfectly good, yet (unlike Traditional Anselmians) does not believe that these truths about God are knowable a priori.3 Almeida is also a libertarian. For brevity’s sake, in this chapter I’ll speak as if both Anselmian theism and libertarianism are true.

1.  Interpreting Almeida’s Core Ideas about U-actualization Almeida (2012, chapter 4) declares that as well as strong actualization and weak actualization, there are two other approaches to world-actualization open to God: restricted actualization and U-actualization.4 He explains U-actualization this way: It is also true that, necessarily, God can strongly actualize the state of affairs T that includes the state of affairs of God’s having predicted or prophesied that En will perform A.5 But if, necessarily, God can predict 2  This view is expounded and defended by Flint 2006, amongst others. Almeida (2012, pp. 113–14) explains “counterfactual of creaturely freedom”—and says that Molinists regard CCFs as extending to non-free, chancy events as well as free, undetermined actions. Perhaps there is some philosopher who holds that necessarily, for every possible free creaturely action (not already ruled out by prior divine activity) God has available a method of infallibly bringing it about, on the grounds that in every world there are true CCFs suitable for his doing so (though these differ from world to world). Since any good argument for this view would need premises which are all necessarily true, the prospects of finding one seem poor. 3  I will assume that Almeida holds that God is in time, and that his being all-knowing involves his having foreknowledge of undetermined events; everything that he says is compatible with this assumption, and much that he says strongly suggests it. 4  The expressions “strongly actualize” and “weakly actualize” seem to have been introduced by Alvin Plantinga, who uses them, for example, in Plantinga 1974. Almeida’s explanations of their meaning are, respectively, at 2012, pp. 56, 108, and at pp. 57, 109. 5  Almeida describes En as an “instantiated essence”; the expression is explained in 2012, pp. 59–61, which in turn reflects Plantinga 1974, ch. 5. Since instantiated essence En performs actions, we can safely say that instantiated essences are persons. Although “En” is intended to

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Unrestricted Actualization and Divine Providence  197 that En performs A then it is true in every world that God can bring it about that En performs A without causing En to perform A.  Call that unrestricted actualization. Unrestricted actualization ensures that God can strongly actualize a state of affairs T such that necessarily, T only if God actualizes a morally perfect world.6 And God can actualize T in every possible world unrestrictedly, simply by making a suitable prediction.7

Almeida normally uses “T,” in relation to any given world W, to denote the largest state of affairs which God strongly actualizes in W, i.e., the state of affairs which includes all and only the other states of affairs which God strongly actualizes in W. Although in typical cases a counterfactual conditional whose antecedent is T and whose consequent is En will freely perform A is a CCF, if T includes God’s having predicted or prophesied that En will freely perform A, then, given Almeida’s view that God is necessarily allknowing, the counterfactual conditional is a necessary truth, and is thereby disqualified from counting as a CCF. If T includes God’s having predicted that En will freely perform A, then Almeida’s sentence “God can actualize T in every world unrestrictedly” must be understood as expressing the proposition Necessarily, God can actualize T unrestrictedly—not For every world W, God can bring it about unrestrictedly that W contains T. For consider some world W1 in which En does not exist, or exists but does not freely perform A. Given that God is necessarily all-knowing and necessarily avoids telling lies, it is impossible that God brings it about that W1 contains T, and therefore impossible that T is true-in-W1. Similarly the proposition Necessarily, God can U-actualize En’s freely performing A does not entail For every world W, God can bring it about that W contains En’s freely performing A. If Almeida’s account of U-actualization is correct, then for every world V which does not contain En’s freely performing A, it is true in V both that God can U-actualize En’s

denote a particular individual, distinct from (say) En+1, it is plain Almeida does not envisage a situation in which two people are both free with respect to reciting a poem and God can U-actualize the first one’s freely reciting a poem but cannot U-actualize the second one’s doing so. Accordingly Almeida is to be understood as making claims about what is necessarily true concerning all free created agents and all actions such that it is neither necessary nor (at the relevant times) accidentally necessary that the agent does not freely perform A. 6  “Morally perfect world” is defined at Almeida 2012, p. 58. Almeida 2016 discusses issues related to worlds that are both morally and naturally perfect. 7  Almeida 2012, p. 9; similar passages appear on pp. 102, 110, 116. Plainly, Almeida is tacitly employing a convention such that “T” serves both as a placeholder for a singular term and as a placeholder for a corresponding sentence.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

198  Bruce Langtry freely performing A, and that if God were to do so then the actual world would be some world other than V.8 Almeida uses both “strongly actualizes” and “causes” as equivalents of “is causally sufficient for” or “causally determines.”9 In this narrow sense of “cause,” the striking of a match never causes it to light. Presumably Almeida would say that in typical cases in which someone lights a match, the match’s being struck, the match’s dry condition, and the presence of oxygen each made a causal contribution to the lighting, without causing the lighting. Although the second sentence of the quotation relies on a distinction between causing and bringing about, in such a way that one might think that “bring about” is intended to denote a weaker relation than “cause,” Almeida has in mind divine infallible bringing about—i.e., necessarily, God’s attempt to bring something about is successful. In that case, Almeida must hold that necessarily, if God predicts an event or state of affairs without qualifications such as “almost certainly,” or “unless I revise my plans,” then the predicted event or state of affairs occurs. The expressions “bring about” and “predict,” in the quoted passage and elsewhere, should be interpreted accordingly. “Strongly actualize God’s prediction that p” means “strongly actualize God’s predicting that p.” Here I have recast part of the quotation as an argument: (1)  Necessarily, God can predict that En will perform A.10 [premise] (2)  If, necessarily, God can predict that En performs A then, necessarily, God can bring it about that En performs A without causing En to perform A. [premise] (3)  Necessarily, God can bring it about that En performs A without causing En to perform A. [from (1), (2)] 8  Similarly, Almeida’s premise (1) (see below) does not entail the false proposition In every world W, God can predict that in W En will freely perform A. After all, W might be one of the worlds (W1, say) in which En does not exist. If in W1 God can predict that En will perform A, then in W1 God does not predict that En will perform A, but there are other worlds, suitably related to W1, in which God does. 9  This clearly emerges, for example, from statements in Almeida  2012, pp. 56, 108, 114. Almeida 2016, footnote 9, uses the expression “probabilistic causation”; but if “cause” is understood in the way it must be intended when in the opening paragraph of the paper Almeida says “Not even God can cause moral agents to act freely,” probabilistic causation does not count as a kind of causation. 10  Since “T” does not appear in the antecedent of the conditional embedded in (2), I have omitted talk of “T” from the formulation of (1). I have omitted talk of prophesying because Almeida gives it little or no attention and it plays no significant role in his argument.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Unrestricted Actualization and Divine Providence  199 Almeida evidently regards the foregoing argument as very strong. For in his next chapter (2012, p. 140), he says flatly, “We have shown in Chapter 4 that God can unrestrictedly actualize a morally perfect world. It follows that, necessarily, God can actualize a morally perfect world.” Obviously, there is a sense of “bring it about” in which it often happens that one human brings it about that another other human freely does suchand-such without strongly actualizing the latter’s doing so; here the method the first human uses can fail, but in this instance does not fail, to bring about its intended outcome. Open Theists hold that in the foregoing sense God often brings it about that a human freely performs some specific action. The quoted passage, however, shows that Almeida is not an Open Theist. Nor is he a Molinist.11 If (3) is true, then there are worlds in which God brings it about that En freely performs A without causing it. In those worlds, how does God bring about En’s action? Almeida does not explicitly tell us. But there is little doubt that he holds that God brings about En’s action by predicting it. Here are three reasons why: • Almeida offers no suggestion as to what other steps God might take in order to U-actualize a created being’s action, and without some suggestion his ideas concerning U-actualization, while incompatible with Theological Determinism, Molinism and Open Theism, would not sit alongside them as a candidate theory of divine providence. • He distinguishes four “senses” in which God can bring it about that En performs A: strongly actualizing, weakly actualizing, restrictedly actualizing and U-actualizing En’s performing A. The first two, at least, are proposed ways in which God might bring about E n’s action, rather than senses of the sentence “God brings it about that E n freely  performs A”.12 So it is easy to see restricted actualization and U-actualization as proposed ways. The supposed difference between U-actualization and restricted actualization of a specific action is not merely that the former occurs when there is a divine prediction and 11  If this truth was not already obvious from the quotation, it has emerged by the end of Almeida 2012, section 4.5. Nevertheless, Almeida seems to believe that it is possible that there are true CCFs. 12  Similarly, handing the waiter two $10 bills, using one’s credit card, writing a check, and so on, are different ways of paying $20 to a restaurant rather than different senses of “paying $20 to a restaurant.”

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

200  Bruce Langtry the latter when there is a divine announcement: it is that if God U-actualizes a free action he does so by predicting it (and in every world he can U-actualize the action), while if God restrictedly actualizes a free action he does so by announcing it (and only in some worlds can he restrictedly actualize the action). • The quotation implies that there is at least one course of action C, namely, making a suitable prediction, such that in every world it is true that if God were to take C then God’s target undetermined state of affairs St would be actual.13 Almeida does not say: There is at least one state Sk that God can be in, namely, having infallible knowledge that matches St, such that in every world it is true that if God were to be in Sk then St would be actual. If Almeida holds that if God infallibly actualizes St then he does so by predicting it then there is an obvious explanation of why Almeida does not make the alternative statement: neither he nor anyone else thinks that if God were to infallibly actualize St then he would do so merely by infallibly knowing that St is actual. Thus Almeida’s proposition (3) can be strengthened to read: (3a)  It is true in every world that God can, by predicting that En will freely perform A, bring it about that En freely performs A without God’s causing En to do so. A corresponding adjustment to Almeida’s premise (2) can easily be made. It is likely that Almeida would accept two further strengthenings of his (3).14 The first is inclusion of the phrase “at all candidate times” in (3a), immediately before the first occurrence of “God”; I stipulate that a “candidate” time is a time t such that at no time before t has there occurred any state of affairs x such that x is incompatible with En’s freely performing A.15 The second is extending actualization-by-prediction to events such as ­particle-decays which God does not cause. Generalizing the definition of “candidate time” in an obvious way, I interpret the relevant sections of Almeida 2012 as indicating that he would accept the following conclusion: 13  This is stronger than the proposition that in every world it is true that there is at least one course of action such that if God were to take it then St would be actual. 14  My suggestions are prompted by remarks in Almeida  2012—specifically, the sentences following B3 on p. 116, and the second new paragraph on p. 117. 15  For example, if at time t1 God has unconditionally predicted or promised that x will not obtain then no time after t1 is a candidate time.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Unrestricted Actualization and Divine Providence  201 (3b)  It is true in every world that at all candidate times God can, by predicting that an uncaused state of affairs, event or action z will occur, bring about z without God’s causing z. Neither of the strengthened premises commits Almeida to the view that in the actual world there is some uncaused state of affairs, event or action z such that God does, by predicting that z will obtain, occur or be performed, infallibly bring about z without God’s causing z.16

2.  The Significance of U-actualization for the Theory of Goodness Almeida is concerned with the requirements and constraints that perfect goodness places on an agent who is both omnipotent and all-knowing. He thinks that his work on U-actualization is relevant in the following ways, amongst others: 1.  Suppose that God must actualize some world which has a quantitative overall value expressed by some minimum positive number at least as high as N. Let us say that such worlds are all and only the good enough worlds. Almeida argues that even if all the good enough worlds include uncaused events such as free creaturely actions, necessarily, God can actualize a good enough world. Suppose that necessarily, if God can actualize a good enough world then he does so. It follows that necessarily, God actualizes a good enough world. Almeida argues that this last proposition entails a contradiction, and that therefore it is false that necessarily, if God can actualize a good enough world then he does so.17 Combining this result with Almeida’s assertion that a perfect being must actualize some world or other,18 we obtain the conclusion that it is possible that God actualizes a world which is not good enough; and this result is independent of any specification of where the threshold constituting what is just barely good enough is located. 2.  Assume that necessarily, if there is a uniquely best world then it ­contains free human actions, and therefore human actions that are not strongly

16  Having made the point that (3b) should be understood as covering states of affairs and events as well as actions, from now on, for brevity’s sake, I will in most places speak of z as if it is an action, unless the context requires otherwise. 17  Almeida 2012, pp. 153–9. 18  Almeida 2008, p. 170, footnote 3.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

202  Bruce Langtry actualized. Assume also that it is not necessary that there are true CCFs suitable for making it the case that necessarily, God can weakly actualize the best world. It has hitherto seemed to follow that even if there is a uniquely best world, it is not necessary that God can actualize it. Almeida, however, declares the contrary, that (N) Given that there is such a world, it is necessary that God can actualize it. The only reason for doubt has been the suggestion that the best world would contain undetermined events including free human actions. Almeida, however, affirms (O) Necessarily, it is within God’s power to predict the undetermined events contained in the best world (if it exists), and thereby to actualize them. Nevertheless, Almeida argues, if there is a best world, it is not necessary that God does actualize it.19 3.  Almeida holds that a similar point applies to morally perfect worlds, i.e., worlds in which there are significantly free beings, each of whom performs morally significant actions and “goes morally right” with respect to each of those actions.20 He argues that necessarily, God can U-actualize every morally perfect world. This conclusion threatens the viability of the Free Will Defence developed by Plantinga.21 It is obvious that Almeida’s account of U-actualization is incompatible with Open Theism (which denies that God can have infallible foreknowledge of undetermined events). Furthermore, although the last sentence of my previous section implies that God’s having the option of U-actualizing events is compatible with Theological Determinism, the availability of this option would undermine some important motivations for theists’ adopting Theological Determinism. The three foregoing paragraphs suffice to indicate that Almeida’s account also has other significant philosophical implications which make it worth careful evaluation.

19  Notice that Almeida’s premise O does not suffice to license his conclusion N.  Suppose that (for at least one agent En and action A) the uniquely best world Wub contains the state of affairs Its being the case that En freely performs A and En’s action is not actualized by God; in the case, then it is impossible that God actualizes Wub, and therefore God cannot actualize it. The supposition is compatible with O: what follows from O is that if God were to predict the undetermined events contained in Wub then it would not be Wub that was actual. 20  “Significantly free being” and “morally significant action” and “an agent’s having significant freedom in performing an action” are defined in Almeida 2012, p. 87. 21  It should be clear from footnote 19, however, that Almeida’s argument requires the premise that no morally perfect world contains an agent En and an action A such that the state of affairs Its being the case that En freely performs A and En’s action is morally significant and En’s action is not actualized by God. My criticisms of (1b)–(3b) undermine this premise.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Unrestricted Actualization and Divine Providence  203

3.  Evaluation of Almeida’s Argument for (3b) After Almeida’s argument in support of his main claims about U-actualization has been modified in the ways proposed in Section 1 above, it looks like this: (1b)  Necessarily, at all candidate times God can strongly actualize the state of affairs T that includes the state of affairs of God’s having predicted or prophesied that a specific uncaused action z will be performed. [premise] (2b)  If, necessarily, God can do so, then, necessarily, at all candidate times God can, by predicting that z will be performed without God’s causing z, bring it about that z is performed. [premise] (3b)  Necessarily, at all candidate times God can, by predicting that z will be performed, bring it about that z is performed without God’s causing z. [from (1b), (2b)] Call this Almeida’s Argument Modified. Almeida does not quite advance it, but my foregoing remarks explain why I think that he would accept it as sound, and should do so if he continues to advance his original argument from (1) and (2) to (3).

3.1  Should we accept premise (2b)? Consider the propositions (i)  It is necessarily true of God that if he were to make an unqualified prediction that some human agent would freely perform some action z, then the agent would freely perform z. (ii)  It is necessarily true of God that if he were to make the unqualified prediction, then God would, merely by making the prediction, bring it about that the agent freely performed z. If (ii) is true then (ii) is necessarily true, and so (in classical logic) every proposition entails (ii). So let the sentence “p is eligible to establish q” express the proposition If one were to come to know that p is true in a way that did not depend on one’s already knowing that q is true, then one’s coming to know that p is true would provide one with an epistemically impeccable, deductively valid argument for the truth of q. (2b) is plausible only if it is plausible that (i) is eligible to establish (ii).

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

204  Bruce Langtry Almeida does not provide an argument for the view that (i) is eligible to establish (ii). It is worth considering, however, whether some remarks at 2012, p. 116, somehow indicate a candidate reason for thinking that (i) is eligible to establish (ii).22 Almeida says that in every world God can predict (without causing), and thereby U-actualize, every person’s always going right; he adds that the state of affairs (R) Every person’s always going right counterfactually depends on the state of affairs (P) God’s predicting that they will always go right but does not causally depend on it. Given the context, he is committing himself to the view that necessarily, if P is actual then R is actual and R counterfactually depends on P. By “R counterfactually depends on P” he must mean “P◻→R.” He cannot mean “non-P◻→non-R” or “both P ◻→R and non-P ◻→non-R,” since he should deny that necessarily, non-P◻→non-R.23 In general, is either y counterfactually depends on x or Necessarily, y counterfactually depends on x eligible to establish that x brings about y? Some specific pairs of propositions satisfying the schema x◻→y or the schema Necessarily, x◻→y are such that if the state of affairs identified in the antecedent were actual then it would bring about the state of affairs identified in the consequent. But others are definitely not—for example, Adam drinks tea today ◻→ Adam dies before turning 150; and Necessarily (Alice is 40 years old ◻→ Alice is more than 20 years old). Hence the remarks I summarized from Almeida 2012, p. 116, would require a lot of supplementation in order for Almeida to have a good reason for thinking that (i) is eligible to establish (ii).24 22 Cf. the remarks about restricted actualization in the sentences following B2 at Almeida 2012, p. 109. 23  Non-P is the state of affairs consisting in P’s not obtaining. Surely Almeida is not at this point in the discussion entitled to assume that necessarily God predicts every future undetermined event—after all, such an assumption would be question-begging in an argument whose premises are disputed by Theological Determinists and Open Theists. So he cannot exclude the conjecture that there is at least one world Wr in which God does not predict that everyone will always freely act rightly but nevertheless they all do. (Almeida is using “predict” in such a way that God foreknows that p does not entail God predicts that p.) Since at 2012, pp. 91–2, Almeida defends the view that for all propositions x and y, x&y entails x◻→y, he would infer that in Wr it is true that God does not predict that everyone will always freely act rightly ◻→ everyone always freely acts rightly, i.e., he would infer that in Wr non-P◻→ R This, together with the stipulated fact that, in Wr, P does not obtain, would commit him to concluding that in Wr it is not the case that non-P◻→ non-R.  So Almeida should accept that the situation in Wr constitutes a counterexample to the proposition Necessarily non-P◻→ non-R. 24  Remarks Almeida makes at 2016, pp. 445–8, suggests that he would endorse the following line of thought: In every world, God can strongly actualize the existence of a perfect predictor,

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Unrestricted Actualization and Divine Providence  205 Here is a prima facie strong argument, in the form of a direct objection to (ii), for doubting that Almeida can provide such a reason. No doubt Almeida would affirm that God could have made the following prediction a billion years ago: (D) There will occur free human actions bound up with the use of tanks and submarines in a major war between the British and German empires. Almeida would also affirm that if God had in fact made the prediction then God would have brought about the truth of (D). But for (D) to be true, there needed to come into existence human beings, the British and German empires, tanks, and submarines, and there also needed to occur such developments as industrialization (which did not form part of the prediction’s explicit content). It is not plausible to suppose, as Almeida is committed to supposing, that predicting the truth of (D) was all that God needed to do in order to bring all this about. If Almeida’s account of U-actualization is to contribute substantially to the theory of divine providence, then it has to connect, in the right kinds of way, God’s practical reason, his goals, what he can do, and what he does. If God has a reason all-things-considered to act with some specific state of affairs S as his goal, then he has to be able to contribute effectively to S’s being actual. But God’s merely doing something such that necessarily, if he were to do it then S would be actual does not suffice for God’s contributing effectively to S’s being actual, and so God needs some additional reason to act in the relevant way—e.g., that what he does will somehow make an important causal contribution to S’s being actual (even if it does not cause this). But it is hard to see that God’s merely predicting, a billion years ago, the truth of (D) could make an important causal contribution of the right kind. I conclude that it is very doubtful that (i) is eligible to establish (ii), and therefore that it is very doubtful that (2b) is true. an agent PP who God knows will make many predictions including the prediction that every free agent always goes right, and is such that, for any contingent q, (PP predicts that q) ◻→ q and (PP predicts that not-q) ◻→ not-q, even though PP does not cause it to be the case that q. Then in every world in which PP predicts that q, where q asserts the occurrence of a contingent uncaused event or state of affairs, God thereby brings it about that q, without causing it to be the case that q. This epicycle, however, does not help. Suppose that an engineer makes a faultdetector F, such that when F scans a solar panel, F’s indicator light flashes red ◻→ The solar panel is faulty, and F’s indicator light does not flash red ◻→ The solar panel is not faulty; the engineer does so knowing that F will be used to scan a specific batch of solar panels in the factory where he works. Furthermore, relative to the actual laws of nature, F is infallible. It does not follow that when F scans a solar panel from this batch and F’s indicator light flashes red then the engineer brings it about that this particular solar panel is faulty.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

206  Bruce Langtry

3.2  Should we accept premise (1b)? Let us, however, for the sake of discussion, accept (2b), despite its precarious status. Having done so, should we argue from (2b) and (1b) to (3b), or instead from (2b) and the negation of (3b) to the negation of (1b)? Alternatively, should we suspend judgment concerning each of (1b) and (3b)? Since the truth of (1b) is not obvious, and is denied by some theists, Almeida needs a positive argument in favor of (1b). He does not provide one.25 I suspect that if (1b) initially seems appealing, it is because of some line of thought which one has only partially articulated. For the rest of this chapter I propose to bring a few such considerations into the daylight, ­evaluate them, and thereby end any grip they may have in at least some stages of one’s thinking. The first thought is that God must be able to predict free creaturely actions in virtue of his infallible omniscience. Let us treat it as obvious that if God knows that En will freely perform A then God can predict that En will freely perform A. Nevertheless, it remains less obvious that if God can know that En will freely perform A then God can predict that En will freely perform A.26 So the first thought needs to be developed somehow into a formal argument. The attempt below comes in two stages: the first starts with (a1) and reaches (f1), and the second begins with (f1) and takes us down to (j1). (a1)  Necessarily, God exists and is at all times all-knowing.   [premise] (b1)  There are worlds in which specific human persons p1, p2, . . . freely perform actions of type A. [premise] 25  Is there a sign of such an argument at Almeida 2012, p. 117, when he says, “But since the set SU of CCFs whose antecedents God can strongly actualize is true in every possible world, we know that there are undetermined states of affairs that God can unrestrictedly actualize”? Almeida is unlikely to offer this proposition in support of (1b): the conclusion he draws from the line of overall line thought on pp. 116–18 (and then defends against objections) is merely that “the thesis that God can unrestrictedly actualize a morally perfect world is consistent with the Molinist position on the prevolitional truth of CCFs” (pp. 9, 102). Furthermore, if the (infelicitously worded) proposition The set SU of CCFs whose antecedents God can strongly ­actualize is true in every possible world were a premise of such an argument, then the supposed argument would depend on the assumption that every possible world contains true CCFs. This premise would be difficult to defend, and very few non-Molinists (and not all Molinists) would accept it; so Almeida can hardly think that relying on it would be widely persuasive. 26  After all, if I learn to speak French well, then I will be able understand conversations conducted in French; and I can learn to speak French well. Nevertheless I will not in fact do so. So I will not be able to understand conversations conducted in French. Thus in the supposed circumstances it is false that if I can learn to speak French well then I will be able to understand conversations conducted in French. (I do not claim that this point suffices to disprove the “less obvious” proposition.)

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Unrestricted Actualization and Divine Providence  207 (c1)  In all such worlds, God has infallible foreknowledge, at all times, that these agents freely perform these actions. [from a1, b1]

(d1)  For all worlds W and for all contingent untensed propositions q compatible with God at all times knows that q, if it is false in W that q, then it is true in W that q ◻→ God at all times knows that q. For all contingent tensed propositions of the form It will be the case that q, and for all worlds W, if it is false in W that q then in W q ◻→ (There is some time t such that at t it becomes the case that q and God knows at all times before t that at t it becomes the case that q). [from a1] (e1)  Even in worlds which do not contain p1, p2, . . . freely performing actions of type A, at all times God can at all candidate times infallibly know that they will do so.27 [from d1] (f1)  In every world W, and for all candidate times, if the human agents freely perform actions of type A at those times then God can infallibly foreknow that they do. [from (c1), (e1)] (g1)  Necessarily, (f1) is true if (3b) is true, and (f1) can be true only in virtue of the truth of (3b)28 [premise] (h1)  Hence (3b) is true [from (f1), (g1)] (i1)  Necessarily, (3b) is true only if (1b) is true [premise] (j1)  Therefore (1b) is true [from (h1), (i1)] Whatever its merits, (a1)–(i1) is best seen as in effect replacing Almeida’s Argument Modified—which proceeds from (1b) and (2b) to (3b)—with an argument whose second stage moves in the reverse direction: from (f1) and (g1) to (3b), and then from (3b) and (i1) to (1b). The key idea underlying the second stage is that God’s essential ability to know infallibly about future free creaturely actions (and other uncaused contingent items) is provided by and metaphysically depends upon God’s essential ability to predict free actions and thereby to bring it about that they obtain without God’s causing them. In that case, however, if (f1) is true 27  A scientific instrument may be capable of detecting events of a certain kind even though there are none—e.g., messages from inhabitants of other solar systems. If some world W1 does not contain the relevant free actions then in W1 God does not predict that they will obtain; nevertheless in W1 it is true, let us suppose, both that he can predict them and also that if he were to do so then some world other than W1 would be actual. 28  (g1) cannot be replaced by the weaker proposition Necessarily, (f1) is true if and only if (3b) is true, because (g1)’s role is to provide a reason for believing that (f1) entails (3b), assuming that there are no other reasons for believing that the entailment holds.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

208  Bruce Langtry then God’s essential ability to make the prediction is not provided by, is not explained by, and does not metaphysically depend upon, his essential ability to foreknow the actions. Hence if both (f1) and (g1) are true then God’s essential ability to make the prediction is metaphysically fundamental and unexplained. In that case, (g1) is surely self-undermining. For if it is legitimate to suppose that God’s ability to make the prediction is metaphysically fundamental and unexplained, why would it not be equally legitimate to suppose instead that God’s ability to foreknow free human actions is metaphysically fundamental and unexplained? Could a successful argument from God’s essential, infallible omniscience to the truth of (1b) be based on the premise that, necessarily, in all worlds, including worlds in which there are no free human actions, God can predict free human actions of type A in virtue of the fact that he can have direct, non-inferential, infallible awareness of them? In effect, this new argument proceeds through (a1)–(d1), defends (f1) from the objection that there is no way that (f1) could be true by identifying a way, and then moves immediately to (1b). But the transition from (f1) to (1b) surely requires justification by intermediate premises. Plainly, in any context—philosophical discussion with Open Theists, for example—in which one needs an argument for God’s being able to infallibly predict future free human actions, then there is at least as great a need for an argument in favor of God’s being able to have infallible foreknowledge of them, and if appeal is made to infallible direct awareness then there is at least as great a need for an argument in favor of God’s being able to be aware of future free actions directly, non-inferentially, and infallibly. Can (1b) be supported by some argument whose premises are compatible with the proposition that God’s ability to predict future uncaused items is not provided by, and does not metaphysically depend on, God’s foreknowing them? I will consider two candidates. Call the first of these the argument from divine practical rationality. (a2)  Necessarily, God can have a desire all-things-considered that finite human persons p1, p2, . . . freely perform actions of type A.29 [premise]

29  For brevity’s sake, (a2) is inexplicit about a qualification which should be understood as present, namely that it is possible that God has a desire that specific events will occur and will be uncaused.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Unrestricted Actualization and Divine Providence  209 (b2)  Necessarily, if God has such a desire all-things-considered then it gives God a good practical reason all-things-considered to bring it about that p1, p2, . . . will freely perform actions of type A. [premise] (c2)  Necessarily, if God has a good practical reason all-things-considered to bring it about that p1, p2, . . . will freely perform actions of type A then God can infallibly bring it about that they will. [premise] (d2)  Therefore, necessarily, God can infallibly bring it about that they will. [from (a2), (b2), (c2)] (e2)  Necessarily, God brings it about that p1, p2, . . . freely perform actions of type A only if both God predicts that they will freely perform actions of type A and also he brings it about that they do so by his prediction. [premise] (f2)  Necessarily, God can predict that specific human agents p1, p2, . . . freely perform actions of type A. [from (d2), (e2)] Let us consider the argument (a2)–(f2), on its merits as a candidate proof of (f2). Here is why, even if we have good reason to accept (e2), the argument fails. Firstly, the proposition whose conjunction with (c2) is eligible to establish (d2) is not (a2)&(b2). What is required is a proposition which says or is eligible to establish (NPR)  Necessarily, God has a good practical reason all-things-considered for bringing it about that human agents p1, p2, . . . will freely perform actions of type A. For suppose that NPR is false—i.e., that there are worlds in which God lacks a good practical reason all-things-considered to bring it about that p1, p2, . . . freely perform actions of type A.  In those worlds, (f2) implies, God still has the ability to make the prediction. Yet in those worlds, it cannot be a practical reason all-things-considered, or anything which entails that God has one, whose presence, together with the truth of (c2), is eligible to establish that God has the ability to make the prediction. But NPR and (c2) cannot both be true. For necessarily, God is not akratic. So necessarily, if God has a good practical reason, all-things-considered, to bring about an item x, and he can bring about x, then he does bring about x. Therefore NPR and (c2) together entail that necessarily, God brings it about that p1, p2, . . . will freely perform actions of type A. Suppose that one of the persons is Julius Caesar and that type A is cross the English Channel. It follows

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

210  Bruce Langtry that, necessarily, Julius Caesar exists and that he crosses the English Channel. Since this modal proposition is false, at least one of NPR and (c2) is false. Hence (d2) is left unsupported. I conclude that the argument from divine practical rationality is unsuccessful. Here is a second argument purporting to support (1b) using premises compatible with the proposition that God’s ability to predict future uncaused items is not provided by, and does not metaphysically depend on, God’s foreknowing them. Call it the argument from divine firm intentions. Let a divine intention that p be firm if and only if in any world in which God intends that p, it is certain both that God will not abandon this intention (before it is implemented) and also that God will take any steps he needs to take in order to ensure that p.30 Hence if God has any firm intentions, then they are all fulfilled. Most Theological Determinists and Molinists hold that necessarily all of God’s decisions are firm. Open Theists hold that, necessarily, God’s decisions to bring about specific free creaturely actions are non-firm: they are often abandoned or revised in the light of new incoming information, and God cannot ensure that they will be fulfilled. (a3)  Necessarily, God can firmly intend that finite human persons p1, p2, . . . freely perform actions of type A. [premise] (b3)  Necessarily, God can bring about whatever he firmly intends to bring about. [premise] (c3)  Necessarily, God can bring it about that p1, p2, . . . will freely perform actions of type A.[from (a3), (b3)] (d3)  Necessarily, if God can bring it about that p1, p2, . . . will freely perform actions of type A then God can predict that they will. [premise] (e3)  Necessarily God can predict that p1, p2, . . . will freely perform actions of type A. [from (c3), (d3)]

30  If it is certain that God will take any steps he needs to in order to ensure that p then it is certain that God can take steps which will guarantee that p. In the case of any specific actual divine firm intention, God’s ability to take such steps, and his immunity from akrasia, will be underwritten by his essential omnipotence, omniscience, and other excellences. The definition of “divine firm intention” does not imply that if p and God firmly intends that p then God brings it about that p; he might, for example, leave it to other agents or processes to do so, while being prepared to intervene if it turns out that they were not going to.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

Unrestricted Actualization and Divine Providence  211 It should be plain from my comments on the argument from divine p ­ ractical rationality that argument (a3)–(e4) cannot be employed in support of (1b) by anyone who is advancing Almeida’s Argument Modified. Furthermore, even assessed independently of its latter role, it is defective. To establish (c3), given (b3), one needs not (a3) but (a3*)  Necessarily, God firmly intends to bring it about that finite human persons p1, p2, . . . freely perform actions of type A. For suppose that (a3*) is false—i.e., that there are worlds in which God lacks the firm intention. In those worlds, (c3) implies, God still has the ability to bring it about that p1, p2, . . . will freely perform actions of type A; yet in those worlds, it cannot be the specified divine firm intention whose presence, together with the truth of (b3) guarantees that God has the ability to bring it about that p1, p2, . . . will freely perform actions of type A. (a3*), however, entails that in every world p1, p2, . . .freely perform actions of type A. But Almeida’s premise (1b), if it is true, applies to such cases as David Cameron’s resigning the Prime Ministership, even though it is obvious that David Cameron does not exist in every world, and that in many worlds in which he does exist, he does not resign the Prime Ministership. So (a3*) is unacceptable. Hence the argument from divine firm intentions fails. There is no need to investigate or discuss it further. I conclude that premise (1b) of Almeida’s Argument Modified, like ­premise (2b), is at best very insecure, and unable to serve in a strong ­argument for (3b). If Almeida’s Argument Modified is sound, then even if in the actual world there are no true counterfactuals of creaturely freedom, in the actual world God is able to engage in hands-on micromanagement of all that occurs, including causally undetermined actions, events, and states of affairs. In that case, God can act in such a way that all libertarian-free human actions are either intended by him or are foreseen consequences of what he intends. This is an example of the ingenious lateral thinking which Almeida’s contributions to philosophy of religion richly manifest. Unfortunately, this specific contribution is unsuccessful, because Almeida provides us with no reason to believe the central, novel claims that underlie it. This takes us back to the three main contemporary accounts of divine providence: Theological Determinism, Molinism, and Open Theism.

OUP CORRECTED PROOF – FINAL, 02/08/19, SPi

212  Bruce Langtry

References Almeida, Michael. 2008. The Metaphysics of Perfect Beings. London: Routledge. Almeida, Michael. 2012. Freedom, God, and Worlds. Oxford: Oxford University Press. Almeida, Michael. 2016. “Bringing About Perfect Worlds.” In K.  Timpe and D.  Speak (eds.), Free Will and Theism: Connections, Contingencies and Concerns. Oxford: Oxford University Press, pp. 195–213. Flint, Thomas  P. 2006. Divine Providence: The Molinist Account. Ithaca, NY: Cornell University Press. Plantinga, Alvin. 1974. The Nature of Necessity. Oxford: Clarendon Press.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

11

Unrestricted Actualization and Perfect Worlds A Reply to Langtry Michael Almeida

1. Introduction Alvin Plantinga maintains that every possible world actually exists.1 But he denies that more than one possible world obtains. In fact, at most one (and at least one) possible world obtains. What distinguishes the possible worlds that actually exist from the possible world that obtains, according to Plantinga, is that God actualizes the possible world that obtains. To actualize a possible world God must actualize a maximally consistent state of affairs. God actualizes a state of affairs—say, the state of affairs of Kawhi Leonard playing forward for the Spurs—only if God creates Kawhi or, as Plantinga would put it, God brings it about that Kawhi Leonard’s individual essence is instantiated, in a set of circumstances that includes there being a basketball team, there being an NBA, and so on. The details of divine creation are of course extraordinarily complex, but Plantinga distinguishes two general ways in which God can actualize or bring about a possible world. God can strongly actualize a possible world and God can weakly actualize a possible world. In the strong sense God can actualize only what he can cause to be actual; in that sense he cannot actualize any state of affairs including the existence of creatures who freely take some action or other. Plantinga’s incompatibilism requires that free actions have no divine causes. But in the weak sense God can actualize states of affairs that include the existence of creatures who freely take some action. God might know that were Curley left free with 1  It may surprise you to learn that the actual world is an abstract object, even granting that the actual universe is not abstract.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

214  Michael Almeida respect to accepting a bribe in certain circumstances, then Curley would freely reject it. If God creates Curley in such circumstances and leaves him free to act, then God weakly actualizes the state of affairs of Curley’s freely rejecting the bribe. Weak actualization is compatible with free choice and action. The problem with weak actualization, of course, is that God cannot weakly actualize any state of affairs that includes Curley freely rejecting a bribe unless Curley cooperates with God in rejecting the bribe. And Curley’s cooperation is entirely up to him. Plantinga has famously emphasized that, possibly, there is nothing at all God can do to ensure that Curley freely rejects the bribe. It is possible that, no matter what states of affairs God strongly actualizes, Curley simply would not cooperate in rejecting the bribe.2 Plantinga concludes that, possibly, God simply has to acquiesce in Curley’s moral recalcitrance. Possibly, God simply has to acquiesce in the moral recalcitrance of us all. That’s a fundamental part of his free will defense against the logical problem of evil. God can coexist with evil in worlds where we’re all unwilling to cooperate with a plan for perfect moral goodness. But the problem of Curley’s moral recalcitrance has a simple and straightforward solution. Contrary to Plantinga’s suggestion, God has available to him a way to actualize possible worlds that guarantees that Curley freely cooperates in rejecting the bribe. In addition to strong actualization and weak actualization there is unrestricted actualization.3 In previous work, I made the following observation: It is also true that, necessarily, God can strongly actualize the state of affairs T that includes the state of affairs of God’s having predicted . . . that En will perform A.4 But if, necessarily, God can predict that En performs A then it is true in every world that God can bring it about that En performs 2  Matters are of course worse. It is possible that, no matter what states of affairs God strongly actualized that included some free moral agent, that moral agent would not cooperate in acting morally. This is essentially Plantinga’s problem of transworld depravity. 3 I introduce unrestricted actualization in Freedom, God, and Worlds (Oxford: Oxford University Press, 2012). It is also discussed in my “The Logical Problem of Evil Redux,” in Peter  A.  French and Howard  K.  Wettstein (eds.), The Concept of Evil, Midwest Studies in Philosophy, Volume XXXVI (Hoboken, NJ: Wiley, 2012), pp. 163–78, and “Bringing about Perfect Worlds,” in K.  Timpe and D.  Speak (eds.), Free Will and Theism: Connections, Contingencies and Concerns (Oxford: Oxford University Press, 2016), pp. 195–213. 4  The original version includes predicting or prophesying that E will perform A. I want to focus exclusively on predicting.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Unrestricted Actualization and Perfect Worlds  215 A without causing En to perform A.  Call that unrestricted actualization. Unrestricted actualization ensures that God can strongly actualize a state of affairs T such that necessarily, T only if God actualizes a morally perfect world.5 And God can actualize T in every possible world unrestrictedly.6 Unrestricted actualization (u-actualization) occurs when God strongly actualizes a state of affairs T—where T is a largest state of affairs that God can cause to obtain—that includes God’s predicting that some moral agent(s) will freely perform some action(s). Predictions are not causes. The famous perfect predictor in Newcomb’s Problem might predict that you will two box. If he does, then it is guaranteed that you will two box.7 But the prediction does not cause you to two box, and the prediction is perfectly consistent with libertarian views on free will.

2.  Langtry’s Objections to U-Actualization Bruce Langtry includes among the significant implications of unrestricted actualization that God can unrestrictedly actualize a good enough possible world.8 That is of course true. For any minimum standard of value, God can ensure that a possible world that meets that minimum is u-actualized. It is a more important implication of unrestricted actualization that, necessarily, God can unrestrictedly actualize a morally perfect world. Even if there are true counterfactuals of creaturely freedom (CCF) that might preclude God from weakly actualizing a morally perfect world, there’s no distribution of true CCF’s that can preclude God from unrestrictedly actualizing a morally perfect world. The consequence is that the logical problem of evil remains unresolved.9

5  A morally perfect world is one in which every moral agent always does what is right. 6  See Almeida, Freedom, God, and Worlds, p. 110 ff. 7  If the predictor is indeed perfect, and not merely near perfect, then his prediction that you two box entails that you two box. See Don Hubin and Glenn Ross, “Newcomb’s Perfect Predictor,” Noûs 19(3) (1985): 439–46. Other important features of Newcomb problems with perfect predictors include a failure of independence. This makes one boxing with perfect predictors guaranteed to pay off and two boxing guaranteed to be costly. 8  But it is false that (i) necessarily, God actualizes a good enough possible world. (i) is false because it is impossible that, necessarily, God actualizes a good enough world. See Almeida, Freedom, God, and Worlds, p. 153 ff. 9 Almeida, Freedom, God, and Worlds, p. 135 ff. But see also my “The Logical Problem of Evil Redux,” and “Bringing about Perfect Worlds.”

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

216  Michael Almeida But Langtry offers some interesting reasons to doubt that God has the ability to unrestrictedly actualize possible worlds. Langtry notes that I do not construct an argument for the claim that, necessarily, God can predict that some undetermined state of affairs z obtains or for the claim that, ­necessarily, God can bring it about that z obtains. Langtry is exactly right. I take it to be a direct consequence of God’s omnipotence that, necessarily, God can predict that z. And I take it to be a consequence of God’s predicting that z that God brings it about that z. As Langtry suggests, these claims do need further defense. From the brief passage cited above, Langtry kindly extracts the following argument in defense of the view that God’s predicting that z brings it about that z: (1b)  Necessarily, at all candidate times God can strongly actualize the state of affairs T that includes the state of affairs of God’s having predicted or prophesied that a specific uncaused action z will be performed. [premise] (2b)  If, necessarily, God can do so, then, necessarily, at all candidate times God can, by predicting that z will be performed without God’s causing z, bring it about that z is performed. [premise] (3b)  Necessarily, at all candidate times God can, by predicting that z will be performed, bring it about that z is performed without God’s causing z. [from (3a), (3b)] Langtry’s main objection to unrestricted actualization is that the argument from (1b) and (2b) to (3b) is unsound. He doubts the truth of both (1b) and (2b), and concludes that we therefore have no argument for (3b). According to Langtry, it is false that, necessarily, God can bring about z, where z is some uncaused state of affairs. He offers an interesting ­counterexample to (2b) that aims to establish that not everything that God predicts is something that God brings about. I consider the counterexample in Section 2.1. Langtry argues further that there are no good arguments that necessarily, God can predict that z. It is not true in every world that God has the ability to predict that z. He generously offers a few candidate arguments for (1b) but he rejects each of them. I consider Langtry’s objections to (1b) in Section 2.2. My aim in Sections 2.1 and 2.2 is to show that the counterexample to (2b) and the arguments against (1b) present no serious obstacle to unrestricted actualization.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Unrestricted Actualization and Perfect Worlds  217

2.1  Langtry’s Objection to (2b) The main argument offered against premise (2b) above is a counterexample according to which, Langtry argues, God predicts that z but God does not bring about that z: No doubt Almeida would affirm that God could have made the following prediction a billion years ago: (D) There will occur free human actions bound up with the use of tanks and submarines in a major war between the British and German empires. Almeida would also affirm that if God had in fact made the prediction then God would have brought about the truth of (D). But for (D) to be true, there needed to come into existence human beings, the British and German empires, tanks, and submarines, and there also needed to occur such developments as industrialization (which did not form part of the prediction’s explicit content). It is not plausible to suppose, as Almeida is committed to supposing, that predicting the truth of (D) was all that God needed to do in order to bring all this about.10

How does the counterexample cast doubt on the view that, necessarily, were God to predict that D, then God would bring it about that D? We should note right off that God’s essential perfect rationality ensures that he cannot make an inaccurate prediction. That much, I think we agree, is impossible. The accuracy of God’s predictions exceeds even the accuracy of Newcomb’s contingently perfect predictor. So, the counterexample could not show that God predicts D and D does not occur. Where then is the problem? Consider again Langtry’s last sentence in the passage quoted above: It is not plausible to suppose, as Almeida is committed to supposing, that predicting the truth of (D) was all that God needed to do in order to bring all this about.

Langtry’s main worry seems to be that, in the situation above, God does not exercise the right sort of providential control in the actualization of D. God did not cause D to be the case and God did not weakly actualize D. God unrestrictedly actualized D, and that does not provide God with the right sort of providential control over D. Since God did not exercise the right sort of providential control in the actualization of D, God cannot properly be said to have brought it about that D. Thus, Langtry writes, 10  See this volume, Bruce Langtry, “Unrestricted Actualization and Divine Providence,” p. 205.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

218  Michael Almeida If Almeida’s account of U-actualization is to contribute substantially to the theory of divine providence, then it has to connect, in the right kinds of way, God’s practical reasons, his goals, what he can do, and what he does. If God has a reason all-things-considered to act with some specific state of affairs S as his goal, then he has to be able contribute effectively to S’s being actual. But God’s merely doing something such that necessarily, if he were to do it then S would be actual, does not suffice for God’s contributing effectively to S’s being actual, and so God needs some additional reason to act in the relevant way—e.g., that what he does will somehow make an important causal contribution to S’s being actual (even if it does not cause this). But it is hard to see that God’s merely predicting, a billion years ago, the truth of (D) could make an important causal contribution of the right kind.11

Now, there is no doubt that God’s unrestricted actualization of D does not cause or causally contribute to the actualization of D.12 That’s part of the point of unrestricted actualization! It enables God to bring about states of affairs without having to cause those states of affairs. But it’s evident that the consequences Langtry discusses here are just an artifact of the particular example he has offered. Certainly, if God makes a passive voice prediction concerning some outcome—the prediction that there will occur free human actions bound up with the use of tanks and submarines in two major wars between the British and German empires— then God does not exhibit much control over who might be involved in those free actions, what they might do, when it might occur, etc. God does not exercise much providential control in making that prediction, but that’s only because Langtry has God making a vague prediction. But obviously there’s no need to make a vague prediction. God could have made a fully determinate prediction instead. The degree to which God makes his predictions determinate is the degree to which God approaches meticulous providence. ­ eterminate If God strongly actualizes T in world w, and T includes God’s d prediction concerning exactly how the wars will transpire, who will be involved, what they will do and so on, and God’s determinate prediction concerning everything else that will indeterministically occur in this 11  Langtry, this volume, p. 205. 12  If propositions are not themselves states of affairs, then we should replace my talk of actualizing D with talk of actualization the states of affairs S that would make D true were it to obtain.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Unrestricted Actualization and Perfect Worlds  219 particular world, then God exhibits about as much providential control over the events in w as anyone could want. Unrestricted actualization is perfectly consistent with meticulous providence. It’s perhaps another question whether what God does in w is properly called “bringing about” the uncaused states of affairs in w. But it doesn’t matter much how that question is settled. It remains true that unrestricted actualization allows God to exercise meticulous providence without causing, or causally contributing, to the actualization of any state of affairs.

2.2  Langtry’s Objections to (1b) Langtry provides what he takes to be the best arguments that might be offered in favor of (1b). All of these arguments are unsuccessful, according to Langtry, so we have no good reason to believe (1b). Recall that (1b) states the following. (1b)  Necessarily, at all candidate times God can strongly actualize the state of affairs T that includes the state of affairs of God’s having predicted or prophesied that a specific uncaused action z will be performed.  [premise] The first proposed argument for (1b) proceeds from the fact that necessarily God foreknows z to the fact that necessarily God can predict z. The argument effectively is that God’s ability to foreknow that an undetermined state of affairs z is best explained by God’s ability to predict z. Langtry’s complaint about this argument to the best explanation of God’s foreknowledge is that it would leave God’s ability to predict z unexplained: The key idea underlying the second stage is that God’s essential ability to know infallibly about future free creaturely actions (and other uncaused contingent items) is provided by and metaphysically depends upon God’s essential ability to predict free actions and thereby to bring it about that they obtain without God’s causing them. In that case, however . . . God’s essential ability to make the prediction is not provided by, is not explained by, and does not metaphysically depend upon his essential ability to foreknow the actions . . . [But in that case,] God’s essential ability to make the prediction is metaphysically fundamental and unexplained.13

13  Langtry, this volume, p. 207.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

220  Michael Almeida As we have noted, the target of this objection is an argument to the best explanation. The argument is that the best explanation for God’s ability to foreknow z is that God can predict z. Since God does foreknow z, God can predict z. Therefore, (1b) is true. How could it be a criticism of this argument that we do not have an explanation of God’s ability to predict that z? Suppose for the moment that we do not have an explanation of God’s ability to predict that z. We can agree that an explanation of God’s ability to predict that z would be a good thing to have. But how does currently lacking such an explanation vitiate the argument to the best explanation offered for (1b)? That argument does give us some reason to believe (1b), whether or not we have the additional explanations. Nevertheless Langtry rejects the initial argument for (1b) and proposes an alternative argument. He suggests that considerations of divine rationality might provide a better reason to believe (1b). Consider the small argument from (1) to (2): (1)  Necessarily, God can have a good practical reason all-thingsconsidered to predict that z. (2)  Necessarily, God is able to predict that z.14 The argument from (1) to (2) seems perfectly reasonable and easy to make valid. But Langtry’s complaint is that, given only (1), God is able to predict that z in worlds where God has no practical reason to predict that z. He complains that God is able to make a prediction in those worlds despite the fact that God doesn’t have a practical reason to make a prediction. According to Langtry, God’s prediction in that world would lack an explanation: In those worlds . . . God still has the ability to make the prediction. Yet in those worlds, it cannot be a practical reason all-things-considered, or anything which entails that God has one, whose presence, together with the truth of (c2), is eligible to establish that God has the ability to make the prediction.15

14  Langtry offers a fairly complicated argument that we cannot get from (1) to (2) (see this volume, p.  206). For the reasons that follow, I think the argument depends on a scope confusion. 15  Langtry, this volume, p. 207.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Unrestricted Actualization and Perfect Worlds  221 It is certainly true, according to the small argument, that God has the ability to make a prediction in worlds where he does not have a practical reason to make a prediction. But this is a concern only if we misread the modal claim—that is, only if we misread the claim that God has the ability to make a prediction and God has no practical reason to make a prediction. The modal claim has two readings—one is true and the other is false. On the wide scope reading the claim is of course false: (3)  ◊(God makes the prediction that z & God has no practical reason to predict that z). What is possible, according to (3), is that both God makes a prediction and he has no reason to do so. Let’s agree that there are no possible worlds like that. So, in that sense, it is impossible that God is able to make a prediction in worlds where he lacks a reason to do so. But, on the narrow scope reading of the modal claim, it is true: (4)  ◊(God makes the prediction that z) & God has no practical reason to predict that z. What (4) tells us is that there are worlds in which both it is possible for God to make a prediction and he has no reason to do so. There isn’t any problem with worlds like that. If we suppose at least a weak principle of sufficient reason, then there are no worlds in which God predicts that z and there is no practical reason for God’s making that prediction. So, let’s also agree that (5) is true, where ◻ symbolizes metaphysical necessity: (5) 

◻(God predicts that z → God has a practical reason to predict that z).

Assuming we agree on (5), Langtry insists that there are no worlds in which God is able to make a prediction and he has no practical reason to do so. What’s his worry? I think it is the following. Assume there is a possible world w in which God is able to make a prediction and God has no practical reason to do so. Langtry’s concern is that God could not make a prediction in w. This is because it would be true in w that if God were to predict that z, then God would have made a prediction and he would have had no practical reason for doing so. In short, (3) would be true. And we both agreed that (3) is false. But this reasoning confuses (4) and (3). w is a possible world in which God does not have a practical reason to predict that z. Is it true in w that

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

222  Michael Almeida God is able to predict that z despite the fact that God has no practical reason in w to predict that z. Is that possible? The answer is of course yes. What is true in w is that God does not have a practical reason to predict that z, but were God to predict that z, then it would be true that God had a practical reason for doing so. That is to say, it is true in w that, were God to predict that z, then the prediction would not take place in w wherein God has no practical reason to predict that z. Instead, the prediction would take place in some world w’ diverse from w in which it is true that God does have a practical reason to predict that z. And in general, following (5), in every possible world in which God makes a prediction it is true that God has a good practical reason to make the prediction. The availability of a good practical reason to predict that z explains why God is able to predict that z. Let us suppose that it is true in world w that God can predict that z and that God has no practical reason to predict that z. Here’s another common mistake. If God were to predict that z, then there would be some temporal sequence in which God first makes the prediction that z and then—subsequent to the prediction—God has a practical reason to predict z. That misunderstands what is true under the counterfactual supposition that God predicts that z. Under that supposition, it is not even true that were God to predict that z, then God’s practical reason for z would not be temporally prior to God’s prediction. The practical reasons might well be temporally prior to the prediction and in a perfect position to explain why God predicted that z. Both (6) and (7) are also true: (6)  ◻(God predicts that z → God knew prior to his prediction that z will occur) (7)  ◻(God predicts that z → God had a practical reason, prior to his prediction, to predict that z)

Not only does God know, prior to his prediction, what will occur, he also has practical reasons, prior to his prediction, to predict that z. So we have an explanation for why God can make predictions and why he does make predictions. God makes predictions for reasons of practical rationality. And God can (wide scope) make predictions only if God has practical reasons to do so. But is there something else that lacks an explanation? Do we need further explanation for God’s ability to make a prediction that z that goes beyond the explanation we have offered? Is there some reason to believe that God would lack the ability to make a prediction in some world or other? What

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Unrestricted Actualization and Perfect Worlds  223 could prevent him from doing so? Would it involve a contradiction to make such a prediction? I think that if further explanation of God’s ability to make a prediction is necessary, then the only explanation remaining is God’s omnipotence. Indeed, I’m certain that it is his omnipotence, and not his practical reason, that explains his ability to make a prediction in any possible world. Given God’s essential omnipotence, God can make any prediction the making of which does not entail a contradiction.16 It is of course extremely rare that counterfactual suppositions—even in controversial cases—generate contradictions. Lewis’s case of Tim the time traveler is illuminating: . . . what if Tim had killed Grandfather? In that case, some of the story I told would not have been true. Perhaps Tim might have been the timetraveling grandson of someone else. Perhaps he might have been the grandson of a man killed in 1921 and miraculously resurrected . . . . It is hard to say what is the least revision of Tim’s story to make it true that Tim kills Grandfather, but certainly the contradictory story in which the killing both does and doesn’t occur is not the least revision. Hence it is false (according to the unrevised story) that if Tim had killed Grandfather then contradictions would have been true.17

Can Tim kill his grandfather? Does his doing so involve a contradiction? Certainly in the closest worlds in which Tim does fire the gun, there are no true contradictions! Contradiction worlds are definitely not among the closest. A consistent story would rather have Tim unrelated to the person that he has killed. Now consider the case of God making a prediction. No contradiction would be true under the counterfactual supposition in w, or any other world, that God predicts that z. It wouldn’t then be true, for instance, that God makes a prediction that z and z does not occur, which is impossible. It wouldn’t be true that God makes a prediction that z and God has no reason for doing so. That too is impossible. What is the least revision possible in God’s story to make it true that God predicts that z? It is certainly not the revision in which a contradiction is true. I say that the least revision is one in which God had a practical reason for predicting that z prior to making 16 We can strengthen this to any prediction that does not entail a metaphysical impossibility. 17  See David Lewis, “The Paradoxes of Time Travel,” in his Philosophical Papers II (Oxford: Oxford University Press, 1986), pp. 67–80.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

224  Michael Almeida the prediction. I say that because (6) and (7) are both true for beings like God. I say that God is able to make the prediction in w since making the prediction does not yield any contradiction at all. God’s ability to make the prediction is simply a consequence of God’s omnipotence. Finally, I note that unrestricted actualization does not entail that God can make any prediction at any time in any possible world. All unrestricted actualization entails is that God can make any prediction—any prediction the making of which does not yield a contradiction—in any possible world before he creates any concrete objects or any free moral agents. And there are no good arguments that divine omnipotence does not afford God the ability to make such predictions.18 18  An earlier version of this paper was presented at the August 2015 Rutgers Workshop on God and Value. I would like to thank the John Templeton Foundation for sponsoring that workshop as well as the other participants for their feedback.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

12

Prediction and Providence Rejoinder to Almeida’s Reply Bruce Langtry

I thank Michael Almeida for his careful attention to my paper “Unrestricted Actualization and Divine Providence” (UADP) in his response “Unrestricted Actualization and Perfect Worlds: A Reply to Langtry,” in this volume.1 After preliminary explanations, Almeida begins the main part of the Reply with a defense of (2b): If, necessarily, at all candidate times God can predict that a specific uncaused action z will be performed then, necessarily, at all candidate times God can, by predicting that z will be performed without God’s causing z, bring it about that z is performed.2 He then defends (1b): Necessarily, at all candidate times God can predict that a specific uncaused action z will be performed.

1.  Almeida’s Reply to my Objection to (2b) Almeida’s (2012, chapter  4) discussion of U-actualization is for the most part in terms of schemata—for example, what God can predict is that En freely performs A, and what this implies is that God can bring it about that En performs A without causing En to perform A.  I assumed that Almeida had in mind examples like God can predict that Elizabeth will drink coffee at 5pm on 1st January 2022. This proposition embodies indeterminacy. (For example, concerning “coffee”: white or black? if black then a short black or a long black? and so on.) But if God’s prediction is to be fulfilled, something 1  “Unrestricted Actualization and Divine Providence” (UADP) is Chapter 10 in this volume, pp. 195–212, and “Unrestricted Actualization and Perfect Worlds: A Reply to Langtry” is Chapter 11, pp. 213–224. Unless otherwise indicated, the quoted material in this chapter is from these texts. 2  I had stipulated that a “candidate” time is one earlier than any state of affairs which entails that the agent does not perform the action.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

226  Bruce Langtry other than this very prediction has to settle the details not included in the content specified above. I imagined, nevertheless, that Almeida would affirm (α)  If, necessarily, God can predict that Elizabeth will freely drink coffee at 5pm on 1st January 2022, then God can, merely by making this prediction, bring it about that Elizabeth will drink coffee at 5pm on 1st January 2022.3 Almeida declares, in his Reply, that God could have made a fully determinate prediction instead of the somewhat unspecific one in my example, (D) There will occur free human actions bound up with the use of tanks and submarines in a major war between the British and German empires. Even if Almeida has good arguments for (1b) that will also support God’s ability to make a fully determinate prediction, they will surely not invalidate the points I made, using (D) as an example, while addressing what he says about U-actualization in his book Freedom, God and Worlds. For there he showed no signs of limiting the scope of his main claims via some general requirement of complete specificity in God’s predictions. Nor does his Reply spell out the content of such a requirement. His response to my objection to (2b) involves replacing (2b) by a premise that reflects his current view that God can exercise meticulous providence by making a fully determinate prediction specifying the exact way the war will come about, which human individuals will be involved, what they will do and so on, and also specifying “everything else that will indeterministically occur in this particular world.”4 Plainly, Almeida will be willing to affirm (β)  If, necessarily, at some candidate time t1 God can make a fully determinate prediction of the entire history of the universe up to and including some later time t2, and can include in this prediction En’s freely performing A at t 2 , then God can, by doing so, exercise meticulous providence with respect to En’s freely performing A at t2. (β) Is compatible with premise (2), the original second premise of the argument I ascribed to Almeida on the basis of the passage from his (2012) 3  Probably a great many people will be drinking coffee at 5pm on 1st January 2022. So it is evident that drink coffee at 5pm on 1st January 2022 is an action-type rather than an actiontoken (i.e., a particular). 4  By “indeterministically occur,” Almeida must mean “will not be determined, directly or indirectly, by the state of the universe at the time God makes the prediction.”

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Prediction and Providence  227 quoted in UADP (this volume, pp. 196–7). Almeida would, or at least should, now say that (β) corresponds more closely than (2b) to how his original second premise should be reformulated in the face of arguments I offered for replacing (2) by (2b). There would need to be a corresponding reformulation of the conclusion of his main argument concerning U-actualization. I will argue that Almeida’s defense against my objection involving (D) does not succeed in defending either (2b) or (β). Suppose that one billion years ago God wants it to be the case that (D) is true. Can God be certain of attaining this goal? Almeida’s thought seems to be that God has open to him the option of making a prediction along the following lines: (i)  It specifies exactly how the relevant war or wars will transpire, which individual humans will be involved, and what they will do. (ii)  Since the fulfillment of clause (i) involves there coming into existence the British and German empires, tanks, and submarines, largely as a result of free human actions, the prediction also specifies exactly how these items will come into existence. (iii)  It also specifies how there will come to obtain, again largely as a result of free human actions, various states of affairs presupposed in the account provided in clause (ii), including the existence of relevant financial institutions and of heavy industries and industrialization involving advanced engineering. And so on, for further clauses (iv), (v), etc. For each clause n, a clause n+1 will be required, as long as clause n predicts or presupposes the occurrence of undetermined events and states of affairs. Almeida thinks that God’s making a prediction along the foregoing lines would yield meticulous providence, complete control or at least a very high degree of control, with respect to all the predicted items. If there are undetermined actual events not covered by this one, then to exercise meticulous providence with respect to the entire history of the universe God must make, before the first undetermined events occur, other predictions, whose conjunction with the foregoing multi-clause prediction specifies all the undetermined events in the universe. Suppose that God can indeed make a fully determinate prediction that is complete in the foregoing way, and that its content includes (D). He can also make a similar prediction whose content includes a small or large variation (D*) on (D). Necessarily, if God were to do so then all the predicted states of

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

228  Bruce Langtry affairs, events, and actions would in due course become actual. Can we properly conclude that necessarily God could exercise complete control or at least a very high degree of control over the entire course of subsequent events, including the undetermined events bound up with (D)? In Freedom, God and Worlds, and in his Reply, Almeida provides no positive arguments for a Yes answer. The need for him to say a good deal more is plain from the fact that in general it is not the case that if an agent performs an action x, and her doing so entails the occurrence of a state of affairs or event y, then by performing x the agent exercises control over whether y occurs. For example, if Beatrice drives a tractor, she does not thereby exercise control over whether there are tractors. Perhaps Almeida has in mind an account of control which includes this part: (δ)  A person P controls whether event x occurs if it is true both that (a) if P were to firmly intend that x occur then x would occur, and if P were to firmly intend that x not occur then x would not occur, and also that (b) if x were to occur then P would have firmly intended that x occur, and if x were not to occur then P would have firmly intended that x not occur. But (a) and (b) are not sufficient. For consider some non-human created person Cheryl, of whom epiphenomenalism is true. Unknown to Cheryl, for all physical events x of which (a) and (b) are true (such as her arm’s rising), entirely non-mental states of her brain, her body, and the environment bring it about that (a) and (b) are both true. They do so by the way they bring about, on the one hand, Cheryl’s intentions and other mental states, and also, on the other hand, those physical events she intends to occur and the non-occurrence of those physical events she intends to not occur. Cheryl’s intentions, desires, and other mental states are all ineffectual. In virtue of this fact, Cheryl cannot properly be said to be controlling anything physical. Exercising complete, fine-grained control over some domain does not require that in this domain the agent actually bring about everything that happens and prevent anything else from happening. It does require, amongst other things, that (c)  For all x within the domain, if x occurs then the agent has the power to bring about x, and if x does not occur then the agent has the power to prevent x. If God can make a complete, fully determinate prediction, then he can do something such that necessarily, if he were to do it then (D) would be true.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Prediction and Providence  229 But this is not all that is required for him to exercise complete, fine-grained control over whether (D) is true. He must also have the power to bring it about, infallibly, that the various clauses of his prediction are true. In the Reply, Almeida accepts that (1b), (2b), and (3b) fairly express central parts of the (unrevised) theory advanced in Freedom, God and Worlds: God’s ability to make the prediction suffices to confer on God the ability to bring about, infallibly, the predicted states of affairs. Nevertheless, in his Reply, Almeida steps back from this aspect of the 2012 version of his theory, saying now that it does not matter how we settle the question whether making the prediction is properly called “bringing about” the uncaused states of affairs. But it does matter, for reasons indicated above. Control over some domain is not delivered merely by the agent’s ability to do something such that, if she were to do it then her intentions with respect to that domain would be fulfilled. Even if we were to suppose, for example, that Cheryl is necessarily 100 percent accurate in all her predictions about the physical world, we would still have no argument supporting the legitimacy of our inferring that even though Cheryl’s mental states are all epiphenomenal, she has control over her arms.5

2.  Almeida’s Reply to my Objection to (1b) 2.1  The argument from God’s essential comprehensive knowledge In his Reply, Almeida interprets (a1)–(j1) as an argument to the best explanation, and my criticism as the point that we do not have an explanation of God’s ability to predict undetermined events such as free human actions. He responds: But how does currently lacking such an explanation vitiate the argument to the best explanation offered for (1b)? That argument does give us 5  Almeida (2016, pp. 442, 445) holds that God, who is himself a perfect predictor, can create a perfect predictor, a being who is, of de re necessity, 100 percent accurate in his predictions. Although Almeida says that prefect predictors need not be essentially perfect predictors, in footnote 7 of the Reply (this volume, p. 215), when he is speaking of Newcomb’s problem, he says that “if the predictor is indeed perfect, and not merely near perfect, then his prediction that you two box entails that you two box.” Almeida’s (2016) discussion of created perfect predictors is largely confined to one who can predict that every free created agent will always “go right.” Plainly, however, he would affirm that God can create a perfect predictor whose infallible predictive powers were much wider in scope (though finite), even though it would have very little capacity or opportunity to make a causal contribution to events in the lives of any other created beings.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

230  Bruce Langtry some reason to believe (1b), whether or not we have the additional explanations.

(a1)–(j1), however, was not intended as an argument to the best explanation. In its crucial premise (g1), it employs the claim that (f1) can be true only in virtue of the truth of (3b), rather than the claim that (f1) can be explained only by the truth of (3b). My objection to the argument for (1b) was not that it leaves us without an explanation of God’s ability to predict specific free actions and other uncaused events. My objection was that we should not accept (g1). I expressed my reason for doubting (g1) this way: If it is legitimate to suppose that God’s ability to make the prediction is metaphysically fundamental and unexplained, why would it not be equally legitimate to suppose instead that God’s ability to foreknow free human actions is metaphysically fundamental and unexplained?

The question’s significance may be plainer if we consider the propositions: (ε)  Necessarily, God’s essential ability to infallibly foreknow all future free actions is metaphysically based on and explained by God’s essential ability to infallibly predict all future free actions and thereby to bring them about. (ζ)  Necessarily, God’s essential ability to infallibly predict all future free actions is metaphysically based on and explained by God’s essential ability to infallibly foreknow all future free actions. (η)  Necessarily, God’s having each of the foregoing abilities is metaphysically fundamental and unexplained. Why, unless we are for some strong reason committed to upholding (a1)–(j1) as a good argument for (1b), should we believe (ε) and disbelieve both (ζ) and (η)? If we lack a good answer, then we have little or no reason to believe (g1). None has been supplied.

2.2  The argument from divine practical rationality I rejected the argument on the grounds that premises (a2), (b2), and (c2) do not jointly entail (d2), from whose conjunction with (e2) the conclusion (f2) is inferred. Let R-worlds be worlds in which God has no practical reason to predict that persons p1, p2, . . . freely perform A. My main point was

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Prediction and Providence  231 that (a2)&(b2)&(c2) is compatible with there being R-worlds, and in R-worlds “it cannot be a practical reason all-things-considered, or anything which entails that God has one, whose presence, together with the truth of (c2), is eligible to establish that God has the ability to make the prediction.” Almeida discusses the argument (a2)–(f2) via what we can regard as a “proxy” inference, from (1) Necessarily, God can have a good practical reason all-things-considered to predict that z to (2) Necessarily, God is able to predict that z. The idea is that if my objection to (a2)–(f2) were to succeed, it would also refute the inference from (1) to (2). I will address, in reverse order, the three parts of Almeida’s response. Here is the third: I think that if further explanation of God’s ability to utter a prediction is necessary, then the only explanation remaining is God’s omnipotence. Indeed, I’m certain that it is his omnipotence, and not his practical reason, that explains his ability to make a prediction in any possible world. Given God’s essential omnipotence, God can make any prediction the making of which does not entail a contradiction . . . There are no good arguments that divine omnipotence does not afford God the ability to make such predictions.6

Let (κ) be the proposition God can make any prediction the making of which does not entail a contradiction; and let us assume that entails a contradiction is equivalent to is metaphysically impossible. (κ) is an insecure assumption for use in defense of the availability of U-actualization. One problem with (κ) concerns accidental necessities. For example, there are worlds in which Emily’s first child is born in 2022. So God predicts in 2018 that Emily’s first child will be born in 2022 does not entail a contradiction. But since in fact by November 2013 Emily had already given birth to two children, God cannot in 2018 make the prediction that Emily’s first child will be born in 2022. Although the proposition By November 2013 Emily had already given birth to two children and God predicts in 2018 that Emily’s first child will be born in 2022 entails a contradiction, this point does not save (κ). Almeida can deal with the problem by replacing (κ) by (λ) God can at any candidate time t make any prediction such that God’s making this prediction at t does not entail a contradiction, where a “candidate” time relative to an event or state of affairs x is one earlier than any event or state of affairs y such that y, of de re necessity, excludes x. (In the example, x is God’s 6  Almeida had earlier briefly foreshadowed these claims by saying, “I take it to be a direct consequence of God’s omnipotence that, necessarily, God can predict that z.”

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

232  Bruce Langtry predicting in 2018 that Emily’s first child will be born in 2022, and y is Emily’s having, by November 2013, already given birth to two children.) (λ) surely stands or falls with the broader proposition (μ) God can at any candidate time do anything such that his doing it at that time does not entail a contradiction. Let H1 be the proposition God infallibly brought about En’s freely performing A and God’s doing so was a basic action. (Even though various other things had to happen if En was to freely perform A, H1 implies that God directly brought about En’s free action—God did not bring it about by doing anything else, and hence did not bring it about by employing means. The underlying idea is that unless there are basic actions an infinite regress arises.) Given the controversial assumption that God’s infallibly bringing about a creaturely action is incompatible with the action’s being free, divine omnipotence seems to underwrite a strong prima facie case for H1’s being possible. Is this case defeated by the presence of some hidden contradiction? If not, then, given (μ), H1 licenses the conclusion (H2) God can infallibly bring about En’s freely performing A without doing so by predicting or prophesying that En will so act (or by announcing afterwards that En has so acted).7 Since Almeida’s accepting H2 would sap most of the m ­ otivation for his invoking U-actualization, no doubt he would declare that (H2) does somewhere involve a contradiction. Perhaps it does. But does God’s essential omnipotence provide a stronger argument for (λ) than it provides for H2? Does invoking omnipotence yield a better explanation of how (μ) can be true (if (μ) is true) than it provides of how H2 can be true (if H2 is true)? Let us turn now to the second part of Almeida’s response to my criticisms of the argument from divine practical rationality. Let z be Persons p1, p2, . . . freely perform A and let w be an R-world—i.e., a world in which God has no practical reason to predict that persons p1, p2, . . . freely perform A. Almeida argues: What is true in w is that God does not have a practical reason to predict that z, but were God to predict that z, then it would be true that God had a practical reason for doing so. That is to say, it is true in w that, were God to predict that z, then the prediction would not take place in w wherein God has no practical reason to predict that z. Instead, the prediction would take place in some world w’ diverse from w in which it is true that God does have a practical reason to predict that z. And in general, . . . in every possible world in which God makes a prediction it is true that God has a 7  The clause in parentheses reflects what Almeida (2012) says about restricted actualization on pp. 8, 109–10, and elsewhere in the book.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Prediction and Providence  233 good practical reason to make the prediction. The availability of a good practical reason to predict that z explains why God is able to predict that z.

My main question about R-worlds did not concern what in an R-world could confer on God the ability to predict that p1, p2, . . . freely perform A, or what could explain his having the ability. Instead, it concerned what truth about R-worlds could be eligible to establish that God has the ability. Almeida’s answer identifies such a truth, namely: (θ) For each R-world w, it is the case in w that if God were to predict that z then God would have a good practical reason for making the prediction. He seems to be arguing: If (θ) is true then for each R-world w it is true in w that there is some world w’ suitably related to w in which God predicts that z and has a good practical reason for making the prediction.8 That this practical reason is present in w’ is a truth about w which is eligible to establish that in w God can make the prediction. Therefore for each R-world w, it is the case in w that God can predict that z. But consider the following argument. In the actual world Alpha, Meg does not have a good practical reason to climb Mt. Everest. Yet it is true in Alpha that if she were to climb Mt. Everest then she would have a good practical reason for doing so. Hence there is some world w”, suitably related to Alpha, in which Meg climbs Mt. Everest and has a good practical reason for doing so. That this practical reason is present in w” is a truth about Alpha which is eligible to establish that in Alpha, the actual world, Meg can climb Mt. Everest. Therefore Meg can climb Mt. Everest. Take my word for it that the first and second premises of the argument concerning Meg are true yet the conclusion is false. Since the argument is valid, I conclude that the third premise (about what follows from the presence in w” of the practical reason) is false. Almeida’s foregoing argument is defective in a similar way. The first part of Almeida’s response to my criticism of the argument from divine practical rationality centers on his claim that I committed a modal fallacy. Here is how he introduces his reason for saying so: Langtry’s complaint is that, given only (1), God is able predict that z in worlds where God has no practical reason to predict that z. He complains that God is able to make a prediction in those worlds despite the fact that

8  “Suitably related” is to be filled out, on demand, with conditions supplied by possibleworld semantics for counterfactuals (such as (θ)).

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

234  Bruce Langtry God doesn’t have a practical reason to make a prediction. According to Langtry, God’s prediction in that world would lack an explanation.

Almeida distinguishes between two schemata, (3) and (4): (3)  ◊(God makes the prediction that z & God has no practical reason to predict that z). (4)  ◊(God makes the prediction that z) & God has no practical reason to predict that z.9 The latter, Almeida believes, is true and the former is false. Immediately after stating (4), he paraphrases it as follows: “What (4) tells us is that there are worlds in which both it is possible for God to make a prediction and he has no reason to do so.” But this is not what (4) tells us. There is only one possibility operator in (4), and the location of the parentheses indicates that in (4) its scope includes only the sentence “God makes the prediction that z.” What (4) tells us instead is that, firstly, there are worlds in which both it is possible for God to predict that z, and secondly, God has no reason to predict that z. That is, the second part of what (4) tells us is that in the actual world God has no reason to predict that z. Obviously, Almeida is unwarranted in saying so. Perhaps what he intended to say was that there are specific propositions which, when substituted for “z,” turn (4) into a truth. (Obviously other substitutions might be made, such that the second clause of (4) would turn into a falsehood and therefore (4) would turn into a falsehood.) Perhaps, alternatively, what Almeida had in mind was not (4) but instead (4*), which evidently he would affirm: (4*) ◊(God can make the prediction that z & God has no practical reason to predict that z). He continues: Assume there is a possible world w in which God is able to make a prediction and God has no practical reason to do so. Langtry’s concern is that God could not make a prediction in w. This is because it would be true in w that if God were to predict that z, then God would have made a prediction and he would have had no practical reason for doing so. In short,

9  I stipulated that z was to be Persons p1, p2, . . . freely perform A’; the italicized sentence is itself a schema, which can be turned into a proposition by replacing “Persons p1, p2, . . . ” by (say) “Elizabeth and Eve” and “freely perform A” by “freely drink coffee at 5pm on 1st January 2022.”

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Prediction and Providence  235 (3) would be true. And we both agreed that (3) is false. But this reasoning confuses (4) and (3) . . . .

It is Almeida’s reasoning which is confused at this stage. I did not commit myself to denying (4*) and to accepting that, necessarily, God can predict an undetermined action only if God has a practical reason to predict it. My self-assigned task in my discussion of (1b) was not to disprove (1b) but to point out that there do not appear to be good arguments for its truth. In terms of the proxy argument Almeida introduces, what I tried to do, and all I needed to do, was show that that it is doubtful that (1) is eligible to establish (2). I focused on R-worlds, and asked, “What truth or truths about R-worlds are eligible to establish that (2) is true?” Almeida has offered answers in his Reply, and I have evaluated them above. In the absence of better answers, it is indeed doubtful that (1) is eligible to establish (2). My line of thought did not involve the modal fallacy Almeida attributes to me. I conclude that my main claims in UADP survive Almeida’s critique of them in the Reply.

References Almeida, Michael. 2012. Freedom, God, and Worlds. Oxford: Oxford University Press. Almeida, Michael. 2016. “Bringing About Perfect Worlds.” In K.  Timpe and D.  Speak (eds.), Free Will and Theism: Connections, Contingencies and Concerns. Oxford: Oxford University Press, pp. 195–213.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

13

A Naturalistic Intrinsic Value Theodicy Scott A. Davison

1. Introduction Recent work in axiology points in the direction of a theodicy that echoes ancient and medieval approaches but is rarely pursued by contemporary philosophers. I’m inclined to think that many parts of it are true, and that the approach as a whole is promising, and so I will explore it briefly in this chapter. At several junctures, I will need to make assumptions in order to develop the basic picture, but different assumptions at those junctures are clearly worth exploring also, so this essay should be regarded as a partly speculative venture. A theodicy is designed to explain what justifies God in creating this world, given its apparent shortcomings. We should distinguish reasons that require something from reasons that merely justify it (see Gert 2012, pp. 19–39 and the discussion in Murphy 2014, section 6). To be promising, a theodicy should provide at least a justifying reason for God to create our world; this is what my theodicy is designed to do. Of course, God might have more than one justifying reason to create our world, so I will not argue here that my proposed theodicy is the only viable one. A successful theodicy will not appeal to things that are known to be false or extremely implausible, because good explanations do not appeal to such things. The history of theistic thought includes a number of prominent attempts to identify a successful theodicy, some more successful than ­others. Thanks to advances in the natural sciences, of course, our understanding of the world has changed quite a bit since theists started constructing theodicies many years ago. This means that the task of providing a successful theodicy today faces rather different constraints than it did just a century ago. The theodicy I will present here is naturalistic in the following sense: Except for mentioning God’s reasons for creating the natural world that we

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

A Naturalistic Intrinsic Value Theodicy  237 observe, my theodicy does not appeal to the existence of good things or events or processes that cannot be studied using the natural sciences. More specifically, unlike most of the theodicies that are typically discussed in the literature, mine will not involve any claims about human survival of death, the existence of a soul, libertarian human freedom, or divine intervention, miraculous or otherwise. I will appeal to various philosophical claims about values and reasons, of course, despite the fact that these cannot be studied using the natural sciences, but this is an unavoidable part of carrying out any task of justification. This approach has the advantage of providing a response to the problem of evil that even those with strongly naturalist leanings could accept. Along the way, I will also indicate here and there some strands in various theistic religious traditions that could be seen as supporting my approach. Of course, since it is a theodicy, my explanation involves the idea that God created the world, so one might wonder just how naturalistic it can be. I will not discuss here exactly how God is supposed to have created the world. On some accounts, the creation of the world is clearly inconsistent with what we know about the origins of the world from current natural science, but on other accounts, this is not the case. I will bracket the question of creation off to one side, asking instead whether there is anything about the universe as we find it that would provide God with a justifying reason for creating it. When I talk about the creation of the world, I will restrict myself to the physical universe as we currently understand it, without committing myself either way to the existence of any other kinds of things. Although I will talk as though God exists for the sake of convenience, my arguments do not assume this, and I am not trying to provide evidence for that conclusion. Strictly speaking, my conclusion is hypothetical: If God exists, then the following explanation provides God with a justifying reason for creating our world. The theodicy I will explore includes the following claims: Everything that exists is intrinsically valuable to some degree. The universe as a whole is a thing of immense intrinsic value. The immense intrinsic value of the universe as a whole provides God with a justifying reason for creating it. The evil in the world is offset by the intrinsic values of the creatures affected together with the intrinsic value of the world that comes from its regularity.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

238  Scott A. Davison

2.  Creaturely Freedom Before turning to a more detailed elaboration of this picture, it will be instructive to consider a popular group of quasi-naturalistic theodicies (or defenses: see Plantinga 1974a, p. 10) that appeal to libertarian human freedom. Following the ancient Stoics, the influential Christian theologian/­ philosopher Saint Augustine of Hippo argued that created persons who could act freely were better than those who could not, and blamed the existence of evil on the free activity of such persons, which not even an omnipotent God could determine (see Augustine  1993). More sophisticated versions of this approach have been developed in our time by prominent Christian philosophers (e.g., Plantinga 1974a, 1974b; Swinburne 1979; and van Inwagen 1988, 2006). The appeal to creaturely freedom in these approaches requires the libertarian conception of freedom, according to which one’s free choices cannot be determined by things over which one has no control (as Augustine saw clearly: see Augustine 1993, pp. 69, 72–3). Libertarians differ with respect to how precisely we should characterize freedom, but since nothing of substance turns on this question in what follows, I will simplify things and continue to talk about the libertarian view of freedom as if there were no disagreement here. The appeal to creaturely freedom may explain how some possible world might contain a great deal of evil despite having been created by God, but it is doubtful that such an approach could explain the quality and quantity of evil in our world. Instead of defending a detailed argument for this conclusion, which has received a great deal of attention in the literature, I will simply list here a number of worries that point in this direction. First, even if human beings are free in the relevant sense, they seem to exercise this capacity rarely (see van Inwagen 1989, 1994), whereas they seem to be the cause of evil rather frequently. Second, even if human beings act freely in the libertarian sense, it does not follow that on those occasions, they are fully morally responsible for their actions, let alone the consequences of those actions, and it is very hard to find cases in which human beings are fully morally responsible for what they do (see, e.g., Levy 2011). Third, there is the fact that the vast majority of evil caused by human beings seems to be caused not by free choices but by perfectly natural ­psychological processes, processes that probably served the early members of our species well but now sometimes cause us a great deal of trouble (see Baumeister 1999).

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

A Naturalistic Intrinsic Value Theodicy  239 Fourth, one might assume that if we could show that human beings were morally responsible for all evil, then God would not be morally responsible for any of it, but it is possible for more than one person to be fully morally responsible for the very same thing (see Zimmerman 1985). Fifth, there is also the so-called problem of natural evil, evil clearly not caused by moral agents, including much of the vast problem of animal suffering. (Recently I  discovered Peter van Inwagen’s “anti-irregularity defense” regarding the suffering of beasts (van Inwagen 2006, pp. 114–29), which is very similar to my approach to the entire problem of evil in several key respects; I will say more about this below.) Sixth, it is important to point out that the scope of moral evaluation is much wider than the sphere of intentional control selected for special attention by the theodicies that appeal to libertarian freedom (see Adams 1985 for a defense of this point from within the JudeoChristian theological tradition). Finally, there is the fact that human beings may not possess libertarian freedom at all. These worries do not amount to a knock-down argument against theodicies based on human libertarian freedom, but they are sufficiently serious to give us pause, and to lead us to look in a different direction, especially if we are interested in what I have been calling a naturalistic theodicy.

3.  The Distribution of Intrinsic Value Leibniz famously argued that our world is the best possible world. This was not because he thought that every event that occurs in the world was the best possible event, or because every creature it contains was the best ­possible creature. Rather, he thought our world was the best possible world either because (a) it contained a maximization of mirroring of divine goodness or because (b) it contained the best ratio of simple laws to complex and varied phenomena. (In different places, he seems to endorse each of these points: see Murray and Greenberg 2013.) Here is a description of the first possibility; I will return to the second one later: In general, Leibniz holds that God creates the world in order to share his goodness with created things in the most perfect manner possible [Gr 355–6]. In light of the fact that created beings, in virtue of their limitations, can mirror the divine goodness only in limited respects, God creates a variety of things, each of which has an essence that reflects a different facet of divine perfection in its own unique way. Since this is God’s ­purpose in creating the world, it would be reasonable to think that

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

240  Scott A. Davison maximizing the mirroring of divine goodness in creation is the goal that God seeks to achieve. And this in fact is one of the standards Leibniz seems to endorse. We might call this the “maximization of essence” standard. Leibniz seems convinced that the actual world meets this standard and that creatures are to be found that mirror the divine perfections in all the sorts of ways that creatures can do this. Thus, there are creatures with bodies and creatures without, creatures with freedom and intelligence and creatures without, creatures with sentience and creatures without, etc. (Murray and Greenberg 2013)

In a related vein, Robert Adams argues that God is the best candidate for playing the role of the Good itself, and provides a corresponding account of the nature of created excellence: There are features by virtue of which things resemble God, and features that could serve as reasons for God’s love. It is features that have both qualifications that will constitute excellence.  (Adams 1999, p. 36)

Later, he says that “if there is a sacredness of finite persons, it is a derivative or secondary sacredness that belongs to them as images of the transcendent Good” (Adams 1999, p. 107). But this means that for both Leibniz and Adams, the excellence or perfection of creatures depends upon standing in a particular relationship to God, so it is not something that we can identify independently from God. It would be preferable to have an account of the value that did not have this feature, because the appeal of a naturalistic approach to theodicy (as I have described it) lies in the way it points to features of our world that make no essential reference to supernatural beings. (For more on Adams’s account in connection with the concept of intrinsic value, see Davison 2012, chapter 7; for a response and further discussion, see Wielenberg 2012.) A more promising approach, I think, involves an appeal to the idea that our world contains a tremendous amount of intrinsic value, the existence of which does not require a relationship to God. Before describing this approach in detail, it is worth noting that contemporary discussions of the problem of evil in Anglo-American analytic philosophy started with papers by J.  L.  Mackie (1955) and H.  J.  McCloskey (1960), at a time when the ­logical positivist’s dismissive, deflationary approach to ethics (and value theory in general) exerted considerable influence. Plantinga’s groundbreaking version of the free will defense (1974a,  1974b) was designed to  answer this so-called logical problem of evil, and incorporated a

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

A Naturalistic Intrinsic Value Theodicy  241 contemporary version of Augustine’s notion of libertarian freedom. But it did not incorporate anything like Augustine’s rich value theory with regard to the entire created order, a value theory that included the thesis of the convertibility of being and goodness that would come to characterize medieval views of the goodness of creation (see Wouter and Aertsen 2016). It seems to me that we are still trying to overcome setbacks in ethics and value theory stemming from the errors of logical positivism. Fortunately, the concept of intrinsic value has enjoyed a kind of revival recently (see, for instance, Chisholm 1986; Lemos 1994; and Zimmerman 2001). Recently, I have developed and defended a contemporary analogue of the medieval thesis of the convertibility of being and goodness, recommending a kind of pluralism about the ontology of intrinsic value (Davison 2012). According to this approach, the bearers of intrinsic value need not be restricted to a single ontological category of things (see Davison  2012, chapter  3). Currently this is not the majority view, since most value theorists view the bearers of intrinsic value as restricted to a single ontological category, such as states of affairs (e.g., Lemos 1994) or concrete states (e.g., Zimmerman 2001), but this pluralistic position has rarely been taken s­ eriously in the literature and some have expressed sympathy for it (e.g., Rønnow-Rasmussen 2012). I will assume something like my view in this chapter; the next few paragraphs contain a quick summary of this pluralist position. According to the picture I favor, something is intrinsically valuable to a certain degree if and only if it possesses intrinsically natural properties that would lead it to be valued to that degree for its own sake by a properly functioning, fully informed valuer. Here I follow the general idea behind Thomas Scanlon’s so-called “buck passing” approach to value, according to which natural, non-normative properties serve as the grounds or basis of value (see Scanlon 1998, pp. 96–8). The notion of a properly functioning, fully informed valuer is an ideal that human valuers only approximate. On my view, our empirical knowledge of the properties possessed intrinsically by various things, aided by the natural sciences, informs our defeasible judgments about their intrinsic value. My approach to intrinsic value, involving as it does the notion of an ideal valuer, is both naturalistic and non-anthropocentric. Nothing in this account implies, for instance, that human beings possess the highest degree of intrinsic value among living things; for all we know, dolphins may in fact enjoy this honor. (For more on the determination of degrees of intrinsic value, see Davison 2012, chapter 5). In fact, for all we know, there may exist alien forms of life on other planets that possess a much higher degree of

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

242  Scott A. Davison intrinsic value than we do. Also, there could come to be future creatures on Earth who are like this; we cannot say in advance—it is an empirical matter. Some have claimed that statements of the form “X is intrinsically ­valuable” are necessarily true if true at all; this is often called “universality,” “incorruptibility,” or “essentiality” (see Lemos 1994; Feldman 1998; and Zimmerman 2001, respectively). By contrast, since I hold that concrete particular objects can be bearers of intrinsic value, and concrete particular objects can change with regard to the properties that they possess i­ntrinsically, I hold that the analogue of this form of statement (namely, one that accounts for degrees: “X is intrinsically valuable to degree D”) is sometimes only contingently true, if true at all. I argue in various ways that we should recognize concrete particular things as bearers of intrinsic value (Davison  2012, chapter  3); here I will mention only one of those arguments, without defending it. Just as our ­paradigm cases of existing objects are middle-sized artifacts and animals, our paradigm cases of things of value are middle-sized artifacts and animals. Although human beings are not ideal valuers, we are valuers nonetheless, so our valuing activity provides some (defeasible) evidence about the structure of value. (This much Mill got right in his famous argument for the Greatest Happiness Principle: see Mill 1969, p. 61.) The idea that concrete particular things are bearers of intrinsic value seems also to find support in some strands of the traditional theisms of Judaism, Christianity, and Islam. According to those strands, God finds their existence good upon creation (see Genesis, chapters 1 and 2), and God loves particular creatures. As Jesus of Nazareth said, Are not two sparrows sold for a penny? Yet not one of them will fall to the ground outside your Father’s care. And even the very hairs of your head are all numbered. So don’t be afraid; you are worth more than many sparrows.  (Matthew 10:29–31, New International Version)

Although this need not be interpreted as a statement about intrinsic value, it could be, since it is plausible to suppose that loving X involves finding X to be intrinsically valuable (although the converse does not hold: see Brentlinger 1970). I propose a replacement for the so-called isolation test for intrinsic value (see, e.g., Carter  1979; Korsgaard  1983; Tolhurst  1983; Lemos  1994; and Zimmerman  2001), which I call the Annihilation Test: How would a properly functioning, fully informed valuer regard the annihilation of X

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

A Naturalistic Intrinsic Value Theodicy  243 when considering this and nothing else? Appealing to the Annihilation Test, I develop the “Anything is Better than Nothing” argument, which implies that even a grain of sand has some degree of intrinsic value. After considering alternative pictures, including a number of attempts to draw a sharp distinction between that which possesses intrinsic value and that which does not (these are attempts to answer what I call “the Cutoff Question”), I conclude that it is most reasonable to suppose that everything that exists possesses some degree of intrinsic value, however small. I will not repeat all of my arguments for this conclusion here, but suppose that I am right about this. If every concrete particular thing has some degree of intrinsic value, then we have the beginnings of a naturalistic, intrinsic value-based theodicy. To see why, consider Nicholas Everitt’s argument from scale against the existence of God (Everitt 2004, chapter 11). Everitt argues that if God exists, and humanity is the focus of all creation, and the point of everything is to give us libertarian freedom for some kind of test, then the universe should be much smaller and younger than it is, and our species should not have arrived by means of natural selection, since evolution is an awfully inefficient and ghastly means to adopt in order to lead the world to the desired arrangement. Instead, we find that for more than 99.999 percent of the history of the universe, there have been no human beings, and most life forms are microscopic; the most distant star is about 200,000,000,000,000,000 times as far away from us as the sun is, and there are roughly 100,000,000,000,000,000,000,000 stars in the universe (Everitt  2004, chapter 11, especially pp. 216–17). Everitt wonders why “the universe as it is revealed to us by modern science is hugely unlike the sort of universe which the traditional theist would lead us to expect” (italics his), and so argues that in the end, the scale of the universe makes theism improbable (Everitt 2004, p. 216; see also p. 217 ff.). But if every concrete particular thing is intrinsically valuable, then we would expect God, an ideal valuer, to find valuable for its own sake, and to varying degrees, everything. So we might actually expect a world created by God to contain all kinds of interesting things, spread across vast stretches of space and time, which is exactly what we do find. Our universe is literally teeming with intrinsic value, and contains an enormous amount of it. Perhaps also it was always a mistake to think of human beings as the single focal point of all creation, contrary to certain strands in various theistic religious traditions, so that Everitt’s predictions about how a world created by God would look are simply misguided. In the book of Job, which has the

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

244  Scott A. Davison structure of a debate, God’s speech from the whirlwind certainly seems to suggest that the focal point of creation is the entire world taken together, including non-human creatures, entire ecosystems, and even inanimate structures. So if every concrete particular thing is intrinsically valuable, we might expect God to create a universe like this one in some respects. We might also expect existing concrete particulars to stand in interesting relations to one another, which brings us back to Leibniz’ second idea and to the notions of parts and wholes.

4.  Organic Unity and the World as a Whole If we recognize concrete particular things as bearers of intrinsic value, then we should recognize the possibility of organic unities. Taken together, the parts of a strand of my DNA, for instance, have less intrinsic value than the whole that they compose. According to my account of intrinsic value, this is because the strand as a whole possesses properties intrinsically that no mere collection of such molecules does, properties that serve as the basis of dispositions and powers that make things better or worse intrinsically (see Davison 2012, chapter 5). But even if we set aside my account of why this is so, it still seems intuitively plausible—emergent properties, after all, enjoy a high degree of philosophical respectability. It would be natural, at this point, to observe that some cases of organic unity seem to be bad. For instance, the virus T4 that infects E. coli bacteria has a kind of organic unity in exactly the same way as a strand of my DNA. John O’Neill argues that a virus like T4 should have no place at all in our ethical deliberations (O’Neill  1992, p. 117), but here it seems that O’Neill and others confuse intrinsic value with instrumental value for human beings (see Davison 2012, pp. 54–5). Intrinsic value need not involve things that are conducive to human flourishing, and so not all organic unities need be good for us, even if they are intrinsically good to some degree. Contemporary physicists appear to have demonstrated experimentally a puzzling phenomenon known as quantum entanglement, which seems to be a case of action at a distance: If two particles interact with one another at a time, then changes in one particle of the pair at a later time will produce, instantaneously and across any distance, a corresponding change in the other particle of the pair (see Kaiser 2014 for a helpful summary). Because of this kind of relationship and others, we might view the entire concrete

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

A Naturalistic Intrinsic Value Theodicy  245 physical universe, the organic unity of all things related to one another in space and time, as a single object of immense intrinsic value. One recent approach in this vein that has received substantial attention was suggested by James Hartle and Stephen Hawking, who have applied the idea of a wave function of a particle to the universe as a whole, creating what is known as “quantum cosmology” (see Hartle and Hawking 1983, and the discussion of theism in Smith 1994). Could the intrinsic value of our world as a whole, whether or not it counts as a single object, give God a justifying reason for creating it, despite its shortcomings? In order to approach this question carefully, I will make several assumptions. First, let’s assume that God knows, in complete detail, everything that would happen in the entire history of the universe “before” it is created. This is a controversial assumption, of course, and although I love the debates surrounding it, I will not discuss them here. (This assumption, of course, makes the task of theodicy more difficult than any other assumption, but this project is a bold one.) Second, given my desire to stay within certain naturalistic bounds, let’s suppose that God creates the world and sustains it in being, but does not intervene in the natural order. I know that this is a matter of considerable controversy among theists, and also involves a controversial way of speaking, and that articulating God’s relationship to the created world is a complicated affair that traditionally involves many components (including creation, conservation, and cooperation with secondary causes). But I have nothing new to say about this set of issues. Later I will return to the question of whether or not God should intervene in the natural order, but for now I will simply assume that God does not. In some ways, my assumption about God’s relationship to the created world in this paper resembles the deistic view that God creates the universe and lets it run by itself, like a clock that has been wound up and set upon a shelf. Van Inwagen takes a different approach to this question, arguing that for all we know, God does intervene in the world regularly, but God must draw an arbitrary line among possible horrors concerning which to prevent and which to permit in order to achieve certain ends (van Inwagen 2006, pp. 105–11). He also says that for all we know, “Being massively irregular is a defect in a world, a defect at least as great as the defect of containing patterns of suffering morally equivalent to those found in the actual world” (van Inwagen 2006, p. 114); I will return to his view below. So given my assumptions, I want to know whether the intrinsic value of the physical universe provides a justifying reason for God to create it all at

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

246  Scott A. Davison once, so to speak. We know from proponents of the so-called “Fine Tuning Argument” for God’s existence (which I am not endorsing here) that there seem to be around 20–30 physical constants such that if one of them were ever so slightly higher or lower in value (and all of the other constants remained the same), then no stable physical objects would be physically possible (see, e.g., Koperski 2015). For the sake of convenience, I will call a universe “flat” in which no stable physical objects are physically possible. In a flat universe, of course, not only would the middle-sized artifacts and animals of which we are so fond fail to exist, but none of the relations between them would exist, either. Such a universe might contain a very large number of fundamental particles, and it might be vast, but what it possessed in quantity, we might say, it would lack in quality. Now I suppose God might find flat universes intrinsically valuable to some degree, but it seems obvious that such universes could not be nearly as intrinsically valuable as ours. In creating this world, God would have selected the fundamental ­physical constants in order to permit the emergence of a universe of immense intrinsic value. In so choosing, God would be expressing a value preference in favor of a universe like ours rather than a flat one, in accordance with Leibniz’s view that . . . the goodness of a world is measured by the ratio between the variety of phenomena that a world contains and the simplicity of the laws that govern that world . . . . [T]the perfection of a world that maximizes the variety of phenomena it contains is enhanced by the simplicity of its laws, since this displays the intelligence of the creator who created it. (Murray and Greenberg 2013)

This universe not only contains a very large number of concrete particular things, each of which is intrinsically valuable, but these things also stand in interesting relations to one another, constituting together a single thing of incredible value. Although the universe also contains some terrible things, which I will discuss shortly, on the whole, it is something of immense intrinsic value, and this is sufficient, all by itself, to justify its creation by God. This is the core claim of my naturalistic intrinsic value theodicy.

5.  Objections: Value, Existence, Death, Pain One objection to my approach here involves skepticism about the value judgments that it involves. In discussing an imaginary person Frank, who

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

A Naturalistic Intrinsic Value Theodicy  247 forms various judgments about the relative value of large-scale states of affairs, van Inwagen says the following: But why should Frank suppose his inclinations—why should anyone suppose that anyone’s inclinations—to make judgments about the relative value of various states of affairs are reliable guides to the true relative values of states of affairs of cosmic magnitude that have no connection with the business of human life? One’s intuitions about value are either a gift from God or a product of evolution or socially inculcated or stem from some combination of these sources. Why should we suppose that any of these sources would provide us with the means to make correct value-judgments concerning matters that have nothing to do with the practical concerns of everyday life?  (van Inwagen 2006, pp. 121–2).

Although this argument raises many issues that I cannot address fully here, I should think that Frank (and those like him) are probably engaged in a comparative assessment of various claims, and that his judgments are based on the outcome of the kind of process of reflective equilibrium that typically justifies philosophical opinions generally. (For a helpful and insightful account of the persuasive elaboration of philosophical pictures, see Gutting 2009.) Such a process presumably grounds van Inwagen’s own judgments concerning similar questions of value. For instance, in connection with his own expanded free will defense, he seems inclined to believe that an eternity of loving friendship between God and human persons “outweighs the horrors” of the current period of estrangement (van Inwagen 2006, p. 90). He also suggests that it is plausible to suppose that “a world that was as the Earth was just before the appearance of human beings would contain a much larger amount of intrinsic good, and would, in fact, contain a better balance of good over evil, than a world in which there were no organisms higher than worms” (van Inwagen 2006, p. 120). And he finds very compelling the claim that the actual total suffering of human beings throughout history is much worse than the actual total suffering of beasts (van Inwagen  2006, p. 127) (a claim which I myself find highly implausible, incidentally). A different objection has to do with what actually exists. If one holds that concrete states are the only bearers of intrinsic value (as Michael  J. Zimmerman does, for example), then it is much easier to deny the existence of organic unities altogether (as he does, in fact: see Zimmerman  2001, chapter 5, especially section 3). I am happy to admit that concrete states can be bearers of intrinsic value, as long as they are not regarded as the only

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

248  Scott A. Davison bearers. But here is an interesting way to press a related worry against the intrinsic value theodicy I have proposed: In order to be intrinsically valuable, something must exist. But unless and until we settle some pressing questions about mereology, we won’t know which concrete particular things actually exist, and so we won’t be able to say that all concrete particular things have intrinsic value to some degree. One might worry, for instance, that given some mereological views, there exist odd, gerrymandered objects (such as the fusion of my pet rabbit, your left ear, and Dean Zimmerman) that have a very high degree of intrinsic value, especially given the account of degree of intrinsic value that I have defended elsewhere. (This worry is expressed with respect to my view in Bader 2014.) This argument strikes me as a kind of Parmenidean rationalism, a denial of common sense for the sake of theoretical consistency, of the sort seen clearly in the wonderful paradoxes of Zeno of Elea (for example). With Aristotle, I take as paradigm cases of existence those middle-sized artifacts and animals with which we are acquainted. So I assume that any view that implies that these things do not exist must be mistaken, even if I cannot produce a coherent mereological account according to which all of those objects (and no strange, gerrymandered objects not recognized by common sense) qualify as existing things. Attempts to provide principled answers to this sort of mereological challenge (such as, e.g., Van Inwagen  1995 and Merricks 2001) strike me as fascinating, important, and heroic. (I also love debates about the existence and nature of holes, and wonder about their proper resolution.) But I find myself having firmly entrenched beliefs about the existence of the paradigm cases, the middle-sized artifacts and ­animals mentioned above, even though I have no resolution of these questions in hand, and I find it hard to imagine that I am unjustified in holding these beliefs. Suppose, though, that I am wrong about this. Suppose that God, being omniscient, knows the true mereology, and that according to it, the middlesized artifacts and animals that we take to be paradigmatic examples of existing objects do not, in fact, exist—including our own selves. What then? Would my proposed intrinsic value theodicy collapse? I don’t think so. Even if these things would not exist, strictly speaking, still there would be repeating, ever-changing patterns in the flux corresponding to those things that I and others mistakenly take to be middle-sized artifacts and animals, and God could find such repeating patterns intrinsically valuable. So even if only one concrete particular thing exists (namely, the entire physical universe), strictly speaking, the nature of that one thing,

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

A Naturalistic Intrinsic Value Theodicy  249 especially as compared to alternative flat universes, would give God a ­justifying reason for creating it. Couldn’t God have created a world containing more and better things, as some have argued in response to my argument for the conclusion that ­everything has some degree of intrinsic value (see, e.g., Mikkelson 2014)? This question really contains two: A question about more things, and a question about better things. Consider first the first question. I suppose God could have created more things, and given the axiological approach that I have assumed, more things probably would have made the universe better. But the question here is whether or not the intrinsic value of our universe, given considerations of both quantity and quality, is good enough to provide a justifying reason for God to create it. Many have tried to show that God’s moral perfection requires that God create the best available world, but these arguments are strained at best (for the tip of the proverbial iceberg here, see Adams  1972; Howard-Snyder and Howard Snyder  1994, 1996; Rowe 2006; and Wierenga 2007). The second question is whether or not God could have created better creatures. I cannot address every kind of evil here in this short essay, so I will consider death and physical pain, with the understanding that further discussion would be necessary to address successfully other kinds of evil. If God loves the concrete particular creatures that exist in our world, as I have suggested, then why do they suffer and die in horrible ways? First, let’s consider death. Suppose that there is no afterlife, that bodily death means the end of every living creature. (Although I am interested in debates about the possibility of personal survival and the empirical evidence for it, given my purposes here, I will simply set these questions aside, along with all theodicies that presuppose personal survival, including those developed by Marilyn McCord Adams, John Hick, Richard Swinburne, Peter van Inwagen, Eleonore Stump, etc.) Based on our current understanding of physics, it is hard to see how a world could contain the forces and tendencies that permit life to exist without also containing disorder and the loss of organic unity, which is the ultimate cause of death. (I will say more about what is possible for God to create below.) This doesn’t mean that death is a good thing, or that God would be indifferent to it; perhaps no sparrow falls to the ground without God’s care, but the sparrows continue to fall, and so do we, because of the sorts of creatures we are. Second, let’s consider pain. Living creatures must obtain reliable information from the environment in order to respond to opportunities and threats. Among protozoa, for instance, flagellates detect the level of salinity

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

250  Scott A. Davison in their environment and propel themselves away from high concentrations that are dangerous to their membranes. But there is no reason to think that they experience high saline concentrations as painful. By contrast, you detect the temperature in your environment, and assuming that you are a properly functioning person (and not someone who suffers from congenital analgesia), you move yourself away from extremes in temperatures that are dangerous to you. If you cannot move away, though, you experience such extremes as painful. Could there be properly functioning living creatures that were just as intrinsically valuable as you are but who would feel no pain as a result of detecting harmful conditions in the environment? We often assume we know which worlds are possible based upon our ability to imagine them, but this is not a reliable guide, in general. It is one thing to imagine some state of affairs, such as an alternative to a given example of horrific pain; it is quite another to imagine a complete possible world in which that alternative plays the role we imagine. Here it is helpful to consider van Inwagen’s “anti-irregularity defense” concerning the suffering of non-human, sentient creatures (“beasts”). According to this defense, four propositions are true, for all we know, including the following one: “Every world God could have made that contains higher-level sentient creatures either contains patterns of suffering morally equivalent to those of the actual world, or else is massively irregular” (van Inwagen 2006, p. 114). Van Inwagen argues persuasively that it is not at all clear how one might design an alternative universe—our only example of a universe is this one, after all, and one would need to specify a great deal of detail about how everything would work, including the physical constants, the process of evolution, and the mechanism that would take the place of pain. For all we know, there is no viable alternative to this arrangement (see van Inwagen 2006, pp. 117–19), and hence no viable alternative to a universe that contains both pain and death. We can say more about pain, though. Most discussions of the problem of evil simply assume without argument that pain is intrinsically bad (e.g., Draper 1989), but this is not obviously true. It is very hard to find an argument for this view, let alone a persuasive one, despite the fact that it is so widely held. Part of the problem here is that pain is a very complicated phenomenon, philosophically speaking. (Here I will mention briefly some of the arguments I have developed elsewhere (see Davison  2012, chapter  5, section 5), but see also Zimmerman  2001, chapter  6, sections 2–4; Aydede 2013; Klein 2015; and the discussion of God and requiring reasons in Murphy 2014.) I may always have a (defeasible) reason to prevent pain,

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

A Naturalistic Intrinsic Value Theodicy  251 in myself or another person, as Thomas Nagel claims (Nagel 1986, p. 161), but this by itself does not imply that pain is intrinsically bad. Pain is typically instrumentally bad. When I had an 8mm kidney stone lodged in between my kidney and bladder a few years ago, which I could not pass, I experienced sharp pain resulting from hydraulic pressure in the kidney. A nurse in the emergency room told me that she had experienced both this pain and the pain of childbirth, and in her opinion, the kidney stone pain was worse. (I remember noting to myself at the time that this judgment indicated a test of comparative painfulness of which John Stuart Mill would surely approve.) One of the worst things about those episodes was that they prevented me from doing anything meaningful for hours—I could only spin slowly on the ground on my left side like a breakdancer in slow motion, groan, and count in my head. I could pursue none of my goals and I could discharge none of my responsibilities, so the pain caused by the kidney stone was clearly instrumentally bad for me. Of course, the kidney pain was also instrumentally good for me, since it forced me to seek medical attention to alleviate the condition that caused the pain (and I was in a position to do so, fortunately). It is commonplace among those who discuss theodicy to attempt to locate, for each evil thing, some good thing to which it is essentially connected, so that one could have the good thing without having the evil thing also, and where the goodness of the good thing somehow outweighs the evil of the evil thing. I will refer to this strategy as the attempt to argue that all evil is “offset” by some good or goods. There seem to be two things that are candidates for offsetting physical pain. The first is the intrinsic value of the creatures who suffer. As Augustine noted, The pain that beasts feel reveals a power that is amazing and praiseworthy in its own way, because it shows that even the souls of beasts have a strong drive toward unity in governing and animating their bodies. For what is pain but a sense of resistance to division and corruption? (Augustine 1993, p. 117)

The very existence of pain requires a high degree of organic unity—there exist no free-floating pains (and no instantaneous ones, either). In order for pain to be experienced, there must exist a subject, a whole that possesses properties intrinsically that no mere collection of its parts does. These properties serve as the basis of dispositions and powers that make experience in general possible and ground a high degree of intrinsic value. In this way, the very existence of pain always implies the concomitant existence of

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

252  Scott A. Davison something that is highly intrinsically valuable. (I can’t resist pointing out that the contemporary focus on pain is probably another symptom of the enduring influence of logical positivism in the area of value theory, but I will not dwell on this point.) If we assume that suffering from pain is an event, a state of a concrete object, or a state of affairs, my suggestion that the intrinsic value of those creatures who suffer helps to offset the evil of their pain requires that we compare somehow the relative intrinsic value of things from different ontological categories. Michael J. Zimmerman argues that we should stick to a single ontological category for the bearers of intrinsic value in order to permit a formula for the computation of the total intrinsic value involved in complex states of affairs (Zimmerman 2001, p. 46; see also the discussion of this question in Davison 2012, chapter 3); this is part of his argument for reducing all intrinsic value to states of affairs. But I think that the following example shows that relative judgments across categories are possible without any kind of precise computation. Suppose that a mad scientist forces you to choose between the following unpleasant options: Either you must administer a full minute of mild but painful electrical shocks to a perfect stranger who is unlucky enough to be strapped into a shocking machine, or you must push a button that will result in the complete annihilation of the creature locked in an annihilation machine across the room. Depending upon the nature of the creature in the box, we might have different views about what you should choose, assuming that your choice is to be determined entirely by considerations of intrinsic value and nothing else. If the creature in the box is a slug, you might not hesitate at all to annihilate it. But what if the creature is a squirrel? Or a dog? Or another human being? At some point, you will decide that the intrinsic value of the concrete particular creature at risk outweighs the badness of the pain involved. Even if pain is intrinsically bad to some degree, as many people assume, it could be the case that the particular pains we find in our world are offset, at least in part, by the intrinsic values of the creatures that experience them. (Some cases of pain are worse than others, of course, and we might have  different judgments about the degrees of intrinsic value that various creatures possess, so this is a complicated matter.) Ah, one might say, but God is not in a position like the one imagined, where a mad scientist forces one to choose between inflicting pain on someone and saving something else. Since God is almighty, God could intervene miraculously to prevent pain, at least where it serves no good purpose for

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

A Naturalistic Intrinsic Value Theodicy  253 the ones who experience it. This is an important objection. There are at least two different responses that one might have to it. The first response is to say that the kind of miraculous intervention imagined here would result in a very disjointed world, one that lacked the organic unity (and hence some of the intrinsic value) of a more regular world (once again, see van Inwagen 2006, pp. 117–19). Together with the intrinsic value of the creatures who suffer pain, this consideration may provide an explanation of how the actual pain in the world is offset by intrinsic value. Someone might protest, at this point, that if this explanation is successful, then God would be justified in creating a world containing millions of creatures who suffer intense and uninterrupted pain forever, since they would be beings of intrinsic value and such a world could be perfectly regular and so on. But is such a world really possible? And why would God choose to create it? A related question (posed by an anonymous reviewer) wonders why scientists should not create a chicken (for example) that suffers continuous pain, since the chicken has intrinsic value (whether or not it suffers). There are several things that could be said in response to this objection. First, we should not expect the justifying reasons that theodicies recommend with respect to God’s creation of the world to provide justifying reasons for human action in general. Also, I have not assumed here that the moral rightness or wrongness of an action (for humans or for God) is determined completely by the net increase or decrease of intrinsic value in the world. In fact, I am inclined to reject such a consequentialism (see Davison 2012, chapter 7). There is more to be said about this objection, but I will not pursue it further. The second response to the objection that God could intervene miraculously to prevent pain is very different. Suppose that God could make the world better by intervening miraculously to prevent pain. Does it follow, though, that that nature of the world as it is, without divine intervention, is insufficiently good to justify God’s creation? It does not. This is a crucial point, I believe: Just because God could have created a better world, it does not follow that this world is not good enough. And this world is very, very good, if what I have said about intrinsic value so far is correct. At this point, one might respond by saying that such an attempted justification is clearly inadequate to square the horrors of the human pain we find in the actual world with the idea of God as perfectly good and loving, so much so that one might wonder what it means to call such a God “good” or “loving.” The naturalistic value theodicy seems to point to a God who is

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

254  Scott A. Davison distant and remote and concerned with the big picture and the long term, rather like a university administrator who is indifferent to the fates of particular faculty because of a focus on the health of the institution as a whole—or worse. I am sympathetic to this objection. I recognize that this is a significant worry about many approaches to the problem of evil, but it is especially significant here. I do not have a convincing reply, except to point out that all other approaches to theodicy seem to face a similar problem in one form or another—why does God not prevent particular horrors? Some will say that this shows that no theodicies are promising, and I have a great deal of sympathy for that reaction also, despite the focus of this essay; others will say that perhaps we have made mistaken assumptions about divine goodness and love (and still others will say that we are no longer talking about God if we make that move). Despite the fact that this remains a pressing objection against the naturalistic intrinsic value theodicy I have outlined here, still I recommend it for further consideration.

6. Conclusion Earlier attempts to address the problem of evil tended to focus on moral evil (caused by free creatures) rather than natural evil, but there is a growing recognition that the problem of natural evil is significant and requires attention. For instance, in his free will defense, Plantinga follows St. Augustine in proposing a possible reduction of natural evil to moral evil, suggesting that natural evil could result from the actions of non-human, malevolent spirits (Plantinga  1974a, p. 10). A few decades later, van Inwagen developed a detailed defense addressing the suffering of beasts, but still found it ­important to develop his “expanded free will defense” to address fully the question of moral evil (van Inwagen  2006, pp. 114–27). I suggest that we have not yet gone far enough in this direction, that the appeal to free will is not very helpful for theodicy, and that perhaps the appeal to intrinsic value can do all of the work by itself. By way of summary, the theodicy I have explored here involves the following claims: Everything that exists is intrinsically valuable to some degree. The universe as a whole is a thing of immense intrinsic value. The immense intrinsic value of the universe as a whole provides God with a justifying reason for creating it.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

A Naturalistic Intrinsic Value Theodicy  255 The evil in the world is offset by the intrinsic values of the creatures affected together with the intrinsic value of the world that comes from its regularity. At the beginning of this chapter, I said that a theodicy is designed to explain what justifies God in creating this world, given its shortcomings. Following Plantinga (Plantinga 1974b, pp. 166–7), some would distinguish a theodicy (designed to identify God’s actual reasons for creating this world) from a defense (designed to identify possible reasons for God to create this world); if the explanation I have explored here fails as a theodicy, perhaps it would be successful instead as a defense. Many questions remain, and what I have said here is very preliminary and subject to many questions, but further exploration along these lines seems to me to be warranted.1

References Adams, Robert Merrihew (1972). “Must God Create the Best?” The Philosophical Review 81(3), pp. 317–32. Adams, Robert Merrihew (1985). “Involuntary Sins,” The Philosophical Review 94(1), pp. 3–31. Adams, Robert Merrihew (1999). Finite and Infinite Goods (Oxford: Oxford University Press). Aydede, Murat (2013). “Pain,” The Stanford Encyclopedia of Philosophy (Spring edition), ed. Edward  N.  Zalta. http://plato.stanford.edu/archives/spr2013/ entries/pain/. Bader, Ralf (2014). “Review of Scott  A.  Davison, On the Intrinsic Value of Everything (2012),” Journal of Moral Philosophy 11(3), pp. 353–6. Baumeister, Roy (1999). Evil: Inside Human Violence and Cruelty (New York: Holt Paperbacks). Brentlinger, John (1970). “The Nature of Love,” in The Symposium of Plato, trans. Suzy  Q.  Gooden, ed. John Brentlinger (Amherst: University of Massachusetts Press), pp. 113–129. [Reprinted in Alan Soble (ed.), Eros, Agape, and Philia (St. Paul, MN: Paragon House, 1989), pp. 136–48.]

1  I wish to thank Paul Draper, John Schellenberg, Andrei Buckareff, Mark Murphy, Thomas Flint, Michael Bergmann, participants in the 2015 summer Value and Religion Workshop at Rutgers University, participants in the Notre Dame Center for Philosophy of Religion discussion group, and an anonymous referee for helpful comments concerning earlier versions of this paper. I would also like to thank the John Templeton Foundation for sponsoring the aforementioned workshop at Rutgers.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

256  Scott A. Davison Carter, Robert Edgar (1979). “Comparative Value Theory,” Journal of Value Inquiry 13 (Spring), pp. 33–56. Chisholm, Roderick  M. (1986). Brentano and Intrinsic Value (New York: Cambridge University Press). Davison, Scott  A. (2012). On the Intrinsic Value of Everything (New York: Continuum Press). Draper, Paul (1989).“Pain and Pleasure: An Evidential Problem for Theists.” Noûs 23(3), pp. 331–50. Everitt, Nicholas (2004). The Non-Existence of God (New York: Routledge). Feldman, Fred (1998). “Hyperventilating about Intrinsic Value,” Journal of Ethics 2, pp. 339–54. Gert, Joshua (2012). “Moral Worth, Supererogation, and the Justifying/ Requiring Distinction,” The Philosophical Review 121(4), pp. 611–18. Gutting, Gary (2009). What Philosophers Know: Case Studies in Recent Analytic Philosophy (Cambridge: Cambridge University Press). Hartle, James and Stephen W. Hawking (1983). “Wave Function of the Universe,” Physical Review D 28, pp. 2960–75. Howard-Snyder, Daniel and Frances Howard-Snyder (1994). “How an Unsurpassable Being Can Create a Surpassable World,” Faith and Philosophy 11(2) (April), pp. 260–8. Howard-Snyder, Daniel and Frances Howard-Snyder (1996). “The Real Problem of No Best World,” Faith and Philosophy 13(3) (July), pp. 422–5. Kaiser, David (2014). “Is Quantum Entanglement Real?” New York Times, November 14. Klein, Colin (2015). What the Body Commands: The Imperative Theory of Pain (Cambridge, MA: MIT Press). Koperski, Jeffrey (2015). The Physics of Theism (Chichester: Wiley-Blackwell). Korsgaard, Christine  M. (1983). “Two Distinctions in Goodness,” The Philosophical Review 92(2) (April), pp. 169–95. Lemos, Noah (1994). Intrinsic Value (Cambridge: Cambridge University Press). Levy, Neil (2011). Hard Luck: How Luck Undermines Free Will and Moral Responsibility (Oxford: Oxford University Press). McCloskey, H. J. (1960). “God and Evil,” Philosophical Quarterly 10, pp. 97–114. Mackie, John L. (1955). “Evil and Omnipotence,” Mind 64, pp. 200–12. Merricks, Trenton (2001). Objects and Persons (Oxford: Oxford University Press). Mikkelson, Gregory  M. (2014). “Review of Scott  A.  Davison, On the Intrinsic Value of Everything (2012)”, Environmental Ethics 36(3), pp. 381–2.

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

A Naturalistic Intrinsic Value Theodicy  257 Mill, John Stuart (1969). “Utilitarianism,” in James M. Smith and Ernest Sosa (eds.), Mill’s Utilitarianism: Text and Criticism (Belmont, CA: Wadsworth). Murphy, Mark (2014). “Toward God’s Own Ethics,” in Challenges to Religious and Moral Belief, ed. Michael Bergmann and Patrick Kain (Oxford: Oxford University Press), pp. 154–71. Murray, Michael and Sean Greenberg (2013). “Leibniz on the Problem of Evil,” The Stanford Encyclopedia of Philosophy (Spring edition), ed. Edward N. Zalta. http://plato.stanford.edu/archives/spr2013/entries/leibniz-evil/. Nagel, Thomas (1986). The View from Nowhere (Oxford: Oxford University Press). O’Neill, John (1992). “The Varieties of Intrinsic Value,” The Monist 75(2) (April), pp. 119–37. [Reprinted in Heimir Geirsson and Michael Losonsky (eds.), Beginning Metaphysics (Malden, MA: Blackwell Publishers, 1998), pp. 105–22.] Plantinga, Alvin (1974a). The Nature of Necessity (Oxford: Oxford University Press). Plantinga, Alvin (1974b). God, Freedom, and Evil (Grand Rapids, MI: William B. Eerdmans). Rønnow-Rasmussen, Toni (2012). “Review of Scott Davison, On the Intrinsic Value of Everything (New York: Continuum Press, 2012),” in Notre Dame Philosophical Reviews, http://ndpr.nd.edu/news/on-the-intrinsic-value-of-everything-2/. Rowe, William L. (2006). Can God Be Free? (Oxford: Oxford University Press). St. Augustine (1993). On Free Choice of the Will, translated with an introduction by Thomas Williams (Indianapolis, IN: Hackett Publishing Company). Scanlon, Thomas (1998). What We Owe to Each Other (Cambridge, MA: Harvard University Press). Smith, Quentin (1994). “Stephen Hawking’s Cosmology and Theism,” Analysis 54(4), pp. 236–43. Swinburne, Richard (1979). The Existence of God (Oxford: Clarendon Press). Tolhurst, William (1983). “On the Nature of Intrinsic Value,” Philosophical Studies 43 (May), pp. 383–96. van Inwagen, Peter (1988). “The Magnitude, Duration, and Distribution of Evil: A Theodicy,” Philosophical Topics 16(2), pp. 161–87. van Inwagen, Peter (1989). “When is the Will Free?” Philosophical Perspectives 3, Philosophy of Mind and Action Theory, pp. 399–422. van Inwagen, Peter (1994). “When the Will Is Not Free,” Philosophical Studies 75, pp. 95–113. van Inwagen, Peter (1995). Material Beings (Ithaca, NY: Cornell University Press).

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

258  Scott A. Davison van Inwagen, Peter (2006). The Problem of Evil (Oxford: Clarendon Press). Wielenberg, Erik (2012). “Review of Scott A. Davison, On the Intrinsic Value of Everything (2012)”, Ethics 123(1) (October), pp. 141–6. Wierenga, Edward (2007). “Perfect Goodness and Divine Freedom” [a review of William Rowe, Can God Be Free? (Oxford: Oxford University Press, 2006)], Philosophical Books 48(3) (July), pp. 207–16. Wouter, Goris and Jan Aertsen (2016). “Medieval Theories of Transcendentals,” The Stanford Encyclopedia of Philosophy (Winter edition), ed. Edward N. Zalta. https://plato.stanford.edu/archives/win2016/entries/transcendentalsmedieval/. Zimmerman, Michael J. (1985). “Sharing Responsibility,” American Philosophical Quarterly 22 (April), pp. 115–22. Zimmerman, Michael  J. (2001). The Nature of Intrinsic Value (Lanham, MD: Rowman & Littlefield).

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Index Aas, Sean  5, 6, 16, 17 Adams, Marilyn  67, 85, 94, 249 Adams, Robert M.  13, 16, 163, 239, 240, 249, 255 Aertsen, Jan  241, 258 Alexander, Scott  68, 94 Almeida, Michael  166, 182, 195–207, 211, 212, 215, 217, 218, 225–9, 231–5 Alston, William  34, 50 Amundson, Ron  6, 16 Aquinas, Thomas  14, 16, 24, 111, 114, 148, 155 Arbour, B.  35, 52 Aristotle  14, 16, 132, 150, 248 Asch, Adrienne  5, 17 Audi, Robert  166, 182 Augustine  15, 238, 241, 251, 254, 257 Aydede, Murat  250, 255 Azadegan, E.  31, 50 Bader, Ralf  248, 255 Barnes, Elizabeth  2, 4–6, 13, 16 Baumeister, Roy  238, 255 Bedau, Mark  6, 16 Bennett, Karen  150 Bergmann, Michael  28, 257 Blustein, Jeffrey  5, 17 Bonaventure 155 Boorse, Christopher  5, 6, 16 Brentano, Franz  182 Brentlinger, John  242, 255 Cantor, Georg  106, 111, 129–34, 162 Cardinal du Four, Vital  106, 151 Carlson, Eric  185, 194 Carter, Robert E.  242, 256 Chang, Ruth  18, 19, 23, 26 Chisholm, Robert  34, 50, 167–71, 174, 175, 178, 181–5, 188, 190, 192–4, 241, 256 Chrzan, K.  41, 47, 48, 50 Climenhaga, Nevin  86, 94

Conway, John  134 Cooper, John M.  148 Cover, J. A.  28 Cross, Richard  5, 16 Crummett, Dustin  35, 36, 46, 47, 50, 83, 94 Dancy, Jonathan  172, 173, 182, 185, 194 Darwin, Charles  25 Dasgupta, Shamik  112, 113, 164 Davis, Richard  31, 32, 50 Davis, S. T.  31, 50 Davison, Scott A.  240–2, 244, 250, 252, 253, 255–8 Dougherty, Trent  33, 50 Dragos, Chris  31, 32, 49, 51 Draper, Paul  24, 250, 256 Dumsday, Travis  31, 50 Ehrlich, Philip  134, 164 Ellerby, Nathan  31, 52 Everitt, Nicholas  243, 256 Ewing, Alfred C.  166, 182 Farrer, Austin  67 Feldman, Fred  242, 256 Fine, Kit  143 Finnis, John  148, 155 Fischer, John M.  62 Flint, Thomas P.  196, 212 Forster, E. M.  117 Francis, Leslie  6, 16 Franks, W. P.  31, 32, 50 French, Peter A.  206, 214 Gelinas, Luke  43, 50 Gellman, Jerome  35, 50 Gert, Joshua  236, 256 Granado, Diego  120 Gulesarian, Theodore  15, 17 Gutting, Gary  247, 256

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

260 Index Hájek, Alan  161, 162 Harding, Brian  158 Hartle, James  245, 256 Hasker, William  31, 33–53, 56 Hawking, Stephen W.  245, 256, 257 Hawthorne, Nathaniel  158 Herman, Barbara  79, 94 Hick, John  67, 249 Himma, K.  45, 51 Holmes, Avram J.  6, 17 Hooker, Brad  77, 81, 94, 95 Howard–Snyder, Daniel  51, 53, 249, 256 Howard–Snyder, Frances  41, 51, 249, 256 Howard, Dana  5, 6, 16, 17 Hubin, Don  215 Johnston, Mark  6, 113, 132 Jordan, Jeffrey  32, 35, 51, 67 Kagan, Shelly  75, 94, 183, 194 Kahane, Guy  13, 17, 31, 32, 51 Kaiser, David  244, 256 Kant, Immanuel  78–81, 94, 160 Karris, Robert J.  155 Keller, James  45, 47, 48, 51 Kierland, Brian  70, 86, 94 Klein, Colin  250, 256 Koons, Robert C.  6, 17 Koperski, Jeffrey  246, 256 Korsgaard, Christine  78, 81, 95, 183, 194, 242, 256 Kraay, Klaas J.  31, 32, 51, 52, 54, 56–63, 65, 86, 95 Langtry, Bruce  213, 215–21, 225, 233, 234 Leibniz, Gottfried  120, 239, 240, 244, 246, 257 Lemos, Noah  182–94, 241, 242, 256 Lerner, Adam  2, 17 Leslie, Sarah–Jane  2, 17 Levy, Neil  238, 256 Lewis, David  223 Lin, Martin  113 Linford, Daniel  31, 51 Linsky, Bernard  150 Lombard, Peter  155 Lougheed, Kirk  31, 49, 51, 52 Luck, Morgan  31, 52 Mackie, John L.  240, 256 Maimonides 114

Maitzen, Stephen  35, 52, 59 Mann, William E.  106 Mavrodes, George  55 Mawson, T.  31, 35, 52 McAllister, Dawson  145, 146 McBrayer, Justin  33, 52 McCloskey, H. J.  240, 256 McConnell, F. J.  76, 95 Mclean, G. R.  31, 52 McMahan, Jeff  68, 95 Megill, Jason  31, 51 Meinong, Alexius  147 Merricks, Trenton  248, 256 Mikkelson, Gregory M.  249, 256 Mill, John S.  242, 251, 257 Millikan, Ruth G.  6, 17 Moltmann, Friederike  147 Moore, G. E.  166–8, 172–6, 182–4, 188, 194 Morriston, Wes  24 Moser, P.  31, 52 Mugg, J.  31, 32, 52 Murphy, Mark C.  15, 17, 22, 85, 87, 95, 236, 250, 257 Nagel, Thomas  162, 251, 257 O’Connor, David  32, 47, 52 O’Neill, John  244, 257 Olson, Jonas  175, 182, 187, 194 Parfit, Derek  78, 79, 81, 89, 95 Patrick, Lauren  6, 17 Penner, M. A.  31, 35, 52 Peterson, Michael  35, 52, 67 Plantinga, Alvin  17, 40, 52, 56, 146, 147, 196, 202, 212–14, 238, 240, 254, 255, 257 Plato  7, 17 Plotinus  107, 108 Pollock, John  92, 95 Pruss, Alexander R.  14, 113, 119 Putnam, David  5, 17 Rabinowicz, Wlodek  183, 194 Rawls, John  77, 95 Reichenbach, B.  47, 52 Rhoda  34, 45, 52 Rocca, Michael D.  113 Rønnow–Rasmussen, Toni  183, 194, 241, 257 Rosen, Gideon  95 Ross, Glenn  215

OUP CORRECTED PROOF – FINAL, 06/08/19, SPi

Index  261 Rowe, William L.  32–4, 41–3, 46, 47, 52, 53, 56, 62, 65, 86, 95, 249, 257 Rowlands, Mark  80, 95 Ruiz de Montoya, Diego  120 Sandel, Michael  70, 95 Savulescu, Julian  13, 17 Scanlon, Thomas M.  77, 79, 80, 91, 95, 241, 257 Scheffler, Samuel  94, 95 Schellenberg, John L.  24 Silvers  2, 17 Singer, Peter  89, 95 Smith, Quentin  245, 257 Speak, Daniel  212, 214, 235 Spinoza  107, 113 Sterba, James  67, 94 Stump, Eleonore  24, 60, 249 Sullivan, Meghan  157 Swenson, Philip  7, 44, 59, 70, 75, 86, 94, 194 Swinburne, Richard  19–21, 23, 26, 27, 29, 238, 249, 257 Timpe, Kevin  212, 214, 235 Tognazzini, Neal  62 Tolhurst, William  242, 257 Urs von Balthasar, Hans  158

van Inwagen, Peter  34, 41, 53, 62, 238, 239, 245, 247–50, 253, 254, 257, 258 Wainwright, William J.  140 Wallace, R. J.  72, 95 Walzer, Michael  76, 95 Wasserman, David  5, 17 Weber, Bruce  7, 17 Wettstein, Howard K.  214 Wielenberg, Erik  240, 258 Wierenga, Edward  249, 258 Williamson, Timothy  111, 150 Wilson, Luke  20 Wisdom, John  166, 182 Wolf, Susan  79, 95 Wood, Allen  79, 95 Wouter, Goris  241, 258 Wright, Larry  6, 17 Wykstra, Stephen  32, 53 Zagzebski, Linda  140, 141, 150 Zalta, Edward  17, 50, 150, 255, 257, 258 Zermelo, Ernst  106 Zimmerman, Dean  44, 59, 248 Zimmerman, Michael  177–82, 187, 188, 190–4, 239, 241, 242, 247, 250, 252, 258