Essays in Analytic Theology [1] 2020944362, 9780198866794


312 82 2MB

English Pages [224] Year 2021

Report DMCA / Copyright

DOWNLOAD PDF FILE

Recommend Papers

Essays in Analytic Theology [1]
 2020944362, 9780198866794

  • 0 0 0
  • Like this paper and download? You can publish your own PDF file online for free in a few minutes! Sign Up
File loading please wait...
Citation preview

OXFORD STUDIES IN ANALYTIC THEOLOGY Series Editors

MICHAEL C. REA and OLIVER D. CRISP

OXFORD STUDIES IN ANALYTIC THEOLOGY Analytic Theology utilizes the tools and methods of contemporary analytic philosophy for the purposes of constructive Christian theology, paying attention to the Christian tradition and development of doctrine. This innovative series of studies showcases high quality, cutting edge research in this area, in monographs and symposia.   : Metaphysics and the Tri-Personal God William Hasker

The Theological Project of Modernism Faith and the Conditions of Mineness Kevin W. Hector

The End of the Timeless God R. T. Mullins

Ritualized Faith Essays on the Philosophy of Liturgy Terence Cuneo

In Defense of Conciliar Christology A Philosophical Essay Timothy Pawl

Atonement Eleonore Stump

Humility and Human Flourishing A Study in Analytic Moral Theology Michael W. Austin

Humility, Pride, and Christian Virtue Theory Kent Dunnington

Essays in Analytic Theology Volume 1 MICHAEL C. REA

1

3

Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © Michael C. Rea 2020 The moral rights of the author have been asserted First Edition published in 2020 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: 2020944362 Set ISBN 978–0–19–886679–4 Volume 1 978–0–19–886680–0 Volume 2 978–0–19–886681–7 DOI: 10.1093/oso/9780198866800.003.0001 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.

Acknowledgements Except for Chapters 4 and 6 and the postscripts to Chapters 5 and 6, all of the essays in this volume have been previously published in English. Chapter 4 was previously published (only) in German translation. I have made no changes to the previously published material except to correct a few minor errors, add an occasional editor note, and make some formatting changes for the sake of uniformity. I am grateful to the following publishers for permission to reprint the material listed here. ‘Realism in Theology and Metaphysics’, pp. 323–44 in Conor Cunningham and Peter Candler (eds), Belief and Metaphysics (London: SCM Press, 2007). Used by permission of Hymns Ancient & Modern Ltd. ‘Theology without Idolatry or Violence’, Scottish Journal of Theology 68 (2015): 61–79. Copyright Cambridge University Press. Reprinted with permission. ‘Authority and Truth’, pp. 872–98 in D. A. Carson (ed.), The Enduring Authority of the Christian Scriptures (Grand Rapids, MI: Eerdmans, 2016). Reprinted by permission of the publisher. ‘Die Eigenschaften Gottes als Thema der analytischen Theologie’ (‘Divine Attributes as a Topic in Analytic Theology), in German; trans. into German by Martin Blay, Daniela Kaschke, and Thomas Schärtl), pp. 49–68 in Thomas Marschler and Thomas Schärtl (eds), Eigenschaften Gottes: Ein Gespräch zwischen systematischer Theologie und analytischer Philosophie (Münster: Aschendorff Verlag, 2016). ‘Gender as a Divine Attribute’, Religious Studies 52 (2016): 97–115. Copyright Cambridge University Press. Reprinted with permission. ‘The Trinity’, pp. 403–29 in The Oxford Handbook of Philosophical Theology, edited by Thomas P. Flint and Michael C. Rea (Oxford: Oxford University Press, 2009). Reprinted by permission of Oxford University Press. ‘Polytheism and Christian Belief ’, Journal of Theological Studies 57 (2006): 133–48. Reprinted by permission of Oxford University Press. ‘Relative Identity and the Doctrine of the Trinity’, Philosophia Christi 5 (2003): 431–46. Philosophia Christi is the journal of the Evangelical Philosophical Society (http://epsociety.org). ‘Material Constitution and the Trinity’ (with Jeff Brower), Faith and Philosophy 22 (2005): 487–505.

viii



Thanks are also due to Jeff Brower for his permission to reprint our coauthored paper, ‘Material Constitution and the Trinity’, to Oliver Crisp, Hud Hudson, and the anonymous referees for Oxford University Press for very helpful comments on the introductions and postscripts, and to Callie Phillips for preparing the index Finally, I would like to thank Oliver Crisp for encouraging me to publish these essays here and in the companion volume, for our ongoing collaboration on all things analytic-theological, and, most of all, for our many years of friendship. It is in gratitude for all of this that I dedicate the first volume to him. The second volume I dedicate to my youngest son, Matthias.

Introduction The activity of analytic theology—bringing the style, method, and literature of analytic philosophy to bear on theological topics—has been pursued for quite some time. One might reasonably see its origins and development in the work of analytically oriented philosophers like Alvin Plantinga, Richard Swinburne, Robert Adams, Marilyn McCord Adams, Nicholas Wolterstorff, Eleonore Stump, and others who, in the latter half of the twentieth century, played a major role in the revival of philosophy of religion and the growth of philosophical theology within academic philosophy. But the concept of analytic theology, the concept of a self-consciously interdisciplinary philosophical-theological activity of the sort just described that deserves both the label ‘analytic’ and the label ‘theology’, is of more recent origin. The idea of analytic theology grew out of conversations during the 2004–5 academic year between Oliver Crisp and myself about the puzzling fact that, despite the recent turn in mainstream philosophy of religion towards traditional topics in systematic theology (trinity, incarnation, and atonement most notably), there had been very little by way of genuine and productive interdisciplinary dialogue between philosophers of religion and their counterparts in theology. Philosophical theology as practised by analytic philosophers seemed not even to be recognized as theology; and, with few exceptions, academic theologians and analytic philosophers of religion seemed generally uninterested in exploring their intersecting research topics in dialogue with one another. Both states of affairs seemed problematic and, as we discussed the matter, we thought that perhaps a volume might be called for—a volume tendentiously entitled Analytic Theology— that would call attention to and begin some much needed conversation about the historical, methodological, and epistemological issues lurking in the background of this disciplinary divide. Analytic Theology: New Essays in the Philosophy of Theology was published in 2009. The same year marked the founding of a series of annual conferences in philosophical theology—the Logos Workshop in Philosophical Theology—held most years at the University of Notre Dame and dedicated to building bridges between philosophy and theology. Now, just over a decade later, the field of analytic theology is in full bloom, with a growing body of literature, a dedicated journal (The Journal of Analytic Theology), a monograph series with Oxford University Press (Oxford Studies in Analytic Theology), a research centre at the

Essays in Analytic Theology: Volume I. Michael C. Rea, Oxford University Press (2020). © Michael C. Rea. DOI: 10.1093/oso/9780198866800.003.0001

2

   :  

University of St Andrews (the Logos Institute for Analytic and Exegetical Theology), and much else in addition. My own work in analytic theology began in earnest just a few years prior to those early conversations with Oliver; and over the past fifteen years it has come to occupy a substantial portion of my research time. This book is the first of two volumes collecting together the most substantial work in analytic theology that I have done between 2003 and 2019. Broadly speaking, the chapters in this volume focus on the nature of God whereas those in the companion volume focus on humanity and the human condition. More specifically, the chapters in Part I deal with metatheological issues pertaining to discourse about God and the authority of scripture; the chapters in Part II focus on divine attributes; and the chapters in Part III discuss the doctrine of the trinity and related issues. The section headings of this introduction match the part divisions of the book, but it is not my aim here to summarize the chapters included in each section’s corresponding part. Instead, this introduction aims to supplement those papers with a more general discussion of some of my past and current thinking on the various loci covered by the chapters in the volume.¹

1. Metatheology Metatheological questions are questions about theology rather than questions about first-order theological topics. This is not necessarily to say that metatheology is a topic outside the discipline of theology; it is just to say that such questions are about the discipline itself rather than about the discipline’s primary subject matter, God. Metatheology includes, but is not limited to, questions about the nature of theology, about the evidential sources for theological claims, and about how we should interpret theological doctrines (e.g. as statements aiming to tell us the literal truth about God, or as mere ‘rules of discourse’, or in some other way).² Chapters 1, 2, and 3 deal with metatheological issues—theological realism and the authority and veracity of scripture (which is, in my view, the most important source of evidence for theological claims). Accordingly, I want to begin here with some remarks about the nature of analytic theology and its bearing on these topics.

¹ For some of the content of this introduction I have drawn on parts of two other essays not included here—one which I characterized at the time I wrote it as ‘a miniature sketch of a partial systematic theology’ (Rea 2017), and another that was contributed to an American Academy of Religion symposium on Analytic Theology, the volume that Oliver and I co-edited in 2009 (Rea 2013a). I am grateful to the publishers—Taylor and Francis, and Oxford University Press, respectively—for permission to reuse this material. ² On the idea of theological doctrines as rules of discourse, see Lindbeck 1984.



3

Just as there is no uniform party line on what constitutes theology, philosophy of religion, or analytic philosophy, so too one shouldn’t expect a uniform party line on the nature of analytic theology. But here is the line I have taken on that question. I start with a rough characterization of analytic philosophy. As I understand it, analytic philosophy is an approach to philosophical problems that is distinguished from other approaches by a particular rhetorical style, some common ambitions, an evolving technical vocabulary, and a tendency to pursue projects in dialogue with a certain evolving body of literature—one typically seen to trace its roots back to G. E. Moore, Bertrand Russell and others in the early twentieth century who were in philosophical dialogue with them. The ambitions seem generally to be these: (i) to identify the scope and limits of our powers to obtain knowledge of the world, and (ii) to provide such true explanatory theories as we can for non-scientific phenomena. The rhetorical style might roughly be characterized as paradigmatic, instances of which conform (more or less) to the following prescriptions: (1) Write as if philosophical positions and conclusions can be adequately formulated in sentences that can be formalized and logically manipulated.³ (2) Prioritize precision, clarity, and logical coherence. (3) Avoid substantive (not merely decorative) use of metaphor and other tropes whose semantic content outstrips their propositional content. (4) Work as much as possible with well-understood primitive concepts, and concepts that can be analysed in terms of those. (5) Treat conceptual analysis (insofar as it is possible) as a source of evidence. More might be added, of course. But my ‘official’ list stops at (5) because most of what else I would add would not really count as prescriptions that divide analytic from continental philosophers. Prescriptions (1) to (5) do mark a division, however. They are prescriptions that non-analytic philosophers typically reject or aim to violate, and for principled reasons. As I see it, analytic philosophy as such is not wedded to a particular theory of truth, nor is it committed to any particular epistemological theory. Contrary to what various critics of analytic philosophy have suggested, there are analytic philosophers aplenty who reject (for example) the correspondence theory of truth; there are also analytic philosophers who reject foundationalism. Analytic philosophers are not, as such, committed to belief in propositions (at least not ³ What exactly is it to write ‘as if ’ positions and conclusions can be formulated in the way described here? Sometimes writing this way may involve actually trying to produce such formulations; sometimes it may involve presupposing in what one says about the positions and conclusions one is discussing that such formulations are possible; and sometimes it may simply be a matter of omitting comments that suggest that certain views or conclusions cannot be formulated in ways that would allow us to derive consequences via familiar rules of logic.

4

   :  

where propositions are considered to be abstract entities that stand in the is expressed by relation to sentences). Nor are they committed to any brand of metaphysical realism or moral or metaphysical absolutism.⁴ In fact, so far as I can tell, there is no philosophical thesis that separates analytic philosophers as such from their rivals. To be sure, analytic philosophers typically write as if certain (mostly metaphilosophical) theses are true—in particular, whatever theses underlie the prescriptions sketched above. But it is easy enough to imagine an analytic philosopher objecting to any one of those presuppositions, and doing so more or less in the analytic style and in the service of what I have called the ambitions of the analytic philosophical tradition. Given this characterization of analytic philosophy, I take analytic theology to be, more or less, theology done with the ambitions of an analytic philosopher, in a style that conforms to the prescriptions that are distinctive of analytic philosophical discourse, and in dialogue with the literature of analytic philosophy. I say ‘more or less’ because I am not so much identifying necessary and sufficient conditions as characteristics that mark a family resemblance; and, to my mind, it is the ‘in dialogue with the literature of analytic philosophy’ characteristic that is typically the most salient marker of that resemblance. How, then, does analytic theology relate to analytic philosophy of religion? The answer depends on how one draws distinctions between philosophy and theology. One way of doing so is methodologically. One might say that theology overlaps philosophy topically in the domain known as ‘philosophy of religion’, but theology is methodologically constrained in a way that philosophy is not. The former, unlike the latter, is confessional—it is done in a way that is faithful to a particular tradition, or text, or set of doctrines, or some other authority besides ‘pure reason’. I do not myself endorse this way of drawing the distinction, but suppose, for the sake of argument, that one does. Nothing in the characterization thus far precludes the idea that theology can be done in accord with the style and ambitions of analytic philosophy and in dialogue with its literature; so the division between analytic theology and analytic philosophy of religion will simply reduce to the division between philosophy proper and theology proper. If, on the other hand, one thinks that philosophy and theology are topically individuated (with a great deal of overlap between them), then one might say that analytic theology and analytic philosophy of religion are overlapping subfields of philosophy and theology respectively, but the former includes work on theological topics that would

⁴ Some seem to think that the grand explanatory ambitions of analytic philosophy commit it to a brand of realism, or at least to ‘absolute metaphysical truth’. But this is manifestly false. If metaphysical realism is false, then that fact will be part of the grand explanation for which we all are striving. If there is no absolute truth (whatever exactly that means), then there would not be a unique grand explanatory theory, but analytic philosophy can proceed from different perspectives and starting points just as it always has. These two points seem not to be sufficiently appreciated by those who would criticize analytic philosophy.



5

typically be seen to fall outside the domain of philosophy (e.g. the development of a distinctively and self-consciously Reformed ecclesiology) whereas the latter includes work on religious topics (e.g. the nature of religion itself) that would typically be seen as having little to do with theology.⁵ As I noted earlier, the decade following the publication of Analytic Theology witnessed significant growth and activity in the field; but the idea of analytic theology has also faced both resistance and misunderstanding.⁶ This is not the place for an exhaustive survey of the objections to and misconceptions about analytic theology; but three in particular are worth mentioning here. First, analytic theology has been confused by some critics with what I have called its style. Second, it has been misunderstood as embodying a commitment to theological realism—a position I happily endorse, but do not take to be part and parcel of analytic theology. Third, it has been objected to on the grounds that it is overly rationalistic, elevating reason over scripture as the primary data source for scripture. I will discuss each in turn. In a recent essay, Martin Westerholm (2019) subjects the enterprise of analytic theology to sustained criticism, arguing ultimately that it ‘may be poorly suited to functioning as a free-standing enterprise in a way that would warrant the perpetuation of a distinctively analytic approach to theology as a whole . . . because it appears to rest on presuppositions that generate a drift into abstraction through which ideal objects of inquiry are substituted for real’ (232). There is a great deal to object to in Westerholm’s essay,⁷ but I will leave most of that aside for now and focus simply on the characterization of analytic theology that he offers en route to developing his objections. In laying out his characterization of analytic theology, Westerholm turns to my own work. He writes: The point [made in an earlier sentence about potential inconsistency in the enterprise of analytic theology] might be made by considering the account of analytic thought that is developed by Michael Rae [sic]. In a widely cited set of comments, Rae [sic] proposes that analytic theology is best taken to be characterized by its ‘rhetorical style’ because concrete definitions fail. [note omitted] He goes on to point to five features that mark this style; these features include ⁵ On the distinction between analytic theology and analytic philosophy of religion, see also BakerHytch 2016. ⁶ For recent critical discussion of analytic theology and replies to some of the criticisms, see (as a starting point) Bitar 2013; Couenhoven 2013; McCall 2015; McCall and Pawl 2018; Macdonald 2014; Oliver 2010; Rea 2013b; Sarrisky 2018; Vanhoozer 2017a, 2017b; Wessling 2017; Westerholm 2019; and Wood 2009, 2013a, 2013b, 2014. ⁷ Not least objectionable is the fact that Westerholm draws sweeping conclusions about the entire enterprise of analytic theology on the basis of features he claims to find in the work of just a couple of analytic philosophers. To my mind, this approach is a bit like drawing conclusions about the entire field of systematic theology on the basis of claims about the work of Barth and von Balthasaar. See Panchuk and Rea (2020) for further critique of Westerholm’s essay.

6

   :   emphasis on precision and clarity, on the priority of primitive concepts and of notions that can be formalized, and rejection of rhetorical and stylistic tropes whose semantic content outstrips their propositional content. This characterization has been formative for other work [note omitted]; but it appears in some ways to struggle to balance its accounts. It deflates the value of the rhetorical and presents clear definitions as the currency in which good work trades; but when it comes to discharge its debts in giving an account of itself, it suggests that is unable to pay in the currency of clear definition, and proposes instead to settle accounts in the deflated coin of the rhetorical . . . . We might note for now that critics have pointed out that inability to define oneself is a philosophical mark of shame, and the embarrassment would seem particularly acute where insistence on clear definition is a hallmark of the school in question. (232–3)

I note in passing that it is one thing to say that ‘inability to define oneself is a philosophical mark of shame’, but quite another thing to show that it is. Westerholm does not show that it is; and I do not believe that he can. As pretty much everybody recognizes, not nearly everything admits of precise definition; and fields and modes of inquiry are notoriously intractable in this regard. (Just witness the vast and ultimately inconclusive literature aiming to define the nature of science.) It is no part of analytic theology to suggest otherwise, and so it is no particular embarrassment for analytic theologians to find themselves in exactly the same boat as the practitioners of just about every other academic discipline when it comes to drawing neat boundaries between what they do and what other people do. Contrary to what Westerholm suggests, I think in fact that we should expect analytic theology to elude precise characterization; for, as some of my earlier remarks suggest, I think the most plausible characterizations of it (as with most fields and modes of inquiry) will deny that there are necessary and sufficient conditions for being a work of analytic theology and insist rather that something counts as such simply by virtue of a kind of family resemblance to paradigm cases. But the more important problem with the passage I have just quoted from Westerholm is that he gets my characterization of analytic theology absolutely wrong. Setting aside minor quibbles about his glosses on the five prescriptions I have used to identify the analytic style, it is incorrect to say, as he does, that my proposal is that ‘analytic theology is best taken to be characterized by its “rhetorical style” because concrete definitions fail’ (232). Of course, style is part of it; but, as I noted earlier, ambitions and philosophical-theological interlocutors are also—and equally—part of it. Moreover, I think that careful attention to these latter aspects of my characterization of analytic theology helps to defuse one of the key presuppositions that drives Westerholm’s essay—namely, that analytic theology somehow purports to be a free-standing enterprise that aims to encompass the whole of theology. Analytic theology was not born out of the idea that the only good theology is analytic in its approach. I cannot speak for others, but for my



7

own part I reject any such claim. In contrast to what Westerholm seems to assume, the impetus to do theology in the analytic mode comes (for many of us, anyway) simply from the idea that the style, methods, and literature of analytic philosophy offer resources that will help serve important theoretical aims. Not all worthwhile theoretical aims are best served by these resources (nor, I might add, does all good theology aim in the first instance to even to serve theoretical aims). Accordingly, it is no part of analytic theology as such to assume that these resources are exhaustive, that other ways of doing theology are without merit, that all non-analytic theological work can profitably be ignored, or anything else of the sort. Leaving this first misunderstanding of analytic theology aside, then, I now turn to the question whether analytic theology is committed to theological realism. As I understand it, theological realism has two components: first, the view that theological theories and doctrines have objective truth-values, and, second, the view that theological theories and doctrines are true only if the objects they apparently refer to genuinely exist, and the apparent kind-terms that they contain are genuine kinds with genuine instances.⁸ Can one pursue the ambitions of analytic theology in the style distinctive of that mode of theorizing without endorsing both components of theological realism? I think that the answer is clearly ‘yes’; but the reasons I would give for saying so depend to some extent on whether metatheological questions are properly thought to fall within the scope of analytic theology. Suppose they do. This is certainly a plausible supposition. Taking the contents of standard textbooks and course syllabi as guides, and noticing who mostly tends to participate in various other metadisciplinary debates, metametaphysics would seem to be part of the field of metaphysics, metaethics would seem to belong to the field of ethics, metaepistemology would seem to be part of epistemology, and so on. So it seems reasonable to think that metatheology belongs to the field of theology, even though its distinctive questions are about theology rather than about God. (Chapter 2 in this volume seems to me to illustrate this point: it strikes me as being a clear instance of analytic theology, despite some discussion of Levinas and other ‘continental’ and postmodern thinkers, while at the same time dealing mostly with metatheological issues.) In that case, it is easy to imagine an analytic theologian providing, in the requisite style, and in accord with the requisite ambitions, an argument against either component of theological realism. Thus, analytic theology as such is not committed to either component. But suppose metatheological questions do not fall within the scope of analytic theology. In that case, reasons for rejecting the first component of theological realism would not belong to analytic theology, since that is a purely metatheological thesis. Moreover, one who denies across the board the objective truth of

⁸ Cf. Chapter 1 in this volume, pp. 19–20.

8

   :  

theological theories would be committed to saying that none of the theories that analytic theology properly aims to produce are objectively true. Still, strictly speaking, the distinctive ambitions of analytic theology aim only at truth, not (necessarily) objectivity; and nothing in the distinctive style of analytic theology commits the analytic theologian to the pursuit of objectivity. More importantly, however, an analytic theologian might well reject theological realism not by denying the first component, but rather by denying the second—insisting (for example) that many of the objects putatively referred to in theological doctrines do not exist, or that many of the kinds referred to in such doctrines are not genuine. A theologian with broadly Bultmannian views about miracles, angels, demons, heaven, hell, and the like, or with John Hick’s views about the incarnation and about God more generally, would hardly count as a theological realist; but such a person might nonetheless arrive at these views in the course of doing analytic theology. So much for the idea that analytic theology is committed to theological realism. What of the third concern, that it is committed more to the authority of reason than to the authority of scripture as a source of evidence in theology? In response to this, I will simply lay out some of my own views about scripture and invite readers to consider whether these views either prioritize reason over scripture or put me at odds with the distinctives of analytic theology. My own view is that they obviously do not. One of the major distinctives of Protestantism is the ‘sola scriptura’ slogan, which has implications for how theology is to be done both individually and corporately. As I understand it, the slogan expresses at least three attributes that the Reformers held to be true of scripture: authority, clarity, and sufficiency. Concerning the authority of scripture, I take the traditional position to be that scripture is what we might call foundationally authoritative—i.e. more authoritative than any other source of information or advice—within the domain of all topics about which it aims to teach us something.⁹ (For convenience, let us refer to the topics in question together as matters of faith and practice.¹⁰) The claim that scripture is clear and sufficient amounts, roughly, to the claim that all doctrines and prescriptions necessary for salvation can easily be derived from scripture by persons concerned about the salvation of their souls without the help of the Church or Church tradition.¹¹ Together, these claims about authority, clarity,

⁹ For discussion of what it means to say that one source is ‘more authoritative’ than another, and for fuller discussion of what it means to say that scripture is authoritative, see Chapter 3 in this volume. ¹⁰ It is a matter of interpretive dispute—and hardly a trivial one!—exactly what topics fall within this domain. ¹¹ My gloss closely follows Bavinck 2003: 477, 488. Cf. Berkhof 1992: 167–8. Note that the clarity doctrine does not imply that it is easy to see that anything in particular is necessary for salvation—as if adherents of other religions are simply failing to understand scripture if they doubt (say) that faith in Christ is necessary for their own salvation.



9

and sufficiency provide what I take to be the core idea underlying the sola scriptura slogan. I affirm sola scriptura as I have just glossed it, and I want now to highlight three points in connection with it that pertain specifically to the question of how scripture, reason, and tradition ought to interact in our theologizing. First, sola scriptura carries no substantive interpretive commitments. It is consistent with the most wooden literalist approach to biblical texts; it is also consistent with rampant allegorical interpretations, and all manner of others. To this extent, it permits a great deal of theological diversity. Its import is simply to provide a loose but significant constraint on the development of theology. It implies that when we do theology, what we ultimately say must be consistent with our best judgment about what the text of scripture teaches. Proponents of sola scriptura cannot sensibly think ‘scripture teaches X, but it is more reasonable for me to believe not-X’; but they are free to use any and all tools at their disposal to determine for themselves what exactly it is that scripture teaches. Second, sola scriptura is plausible only on the assumption that scripture asserts and advises only what God, as divine author, asserts and advises. Absent that assumption, it assigns far too much authority to scripture alone. Surely if the assumption were false there would be no reason to regard scripture as a greater authority in the domain of faith and practice than every other human experience or testimonial report. For those who make the assumption, however, it is no light matter to pronounce either on what scripture teaches or on what topics fall within the domain of ‘matters of faith and practice’. For the doctrine implies that once we have reached a settled judgment about what the text of scripture teaches, we have in the content of that teaching reasons for belief and action that are at least as authoritative as reasons from any other source. Third, a consequence of my first two points is that proponents of sola scriptura have good reason to make careful and judicious use of all available tools for determining what the text of scripture might be saying. These tools include science, moral and other rational intuitions, the techniques of historical biblical criticism and literary analysis, and so on. Moreover, the assumption that scripture has a divine author licences a particular way of using these tools. We know in general that it is perfectly legitimate to interpret texts in light of what we reasonably believe about their authors. Historians of philosophy, for example, often allow their interpretations of great thinkers to be constrained by assumptions about the sorts of errors to which these thinkers may or may not be susceptible. If interpretation X implies that Aristotle was not very bright or wellinformed with respect to the science of his day, that by itself is a reason not to favour interpretation X. So likewise, it seems, with a divinely authored text. If our best science tells us that the sun, moon, and stars existed long before terrestrial plant life, that fact by itself constitutes good reason—as good as the science itself— to believe that a divine author would not teach anything to the contrary. If moral

10

   :  

intuition tells us that slavery is wrong, or that conquering armies should not seek to annihilate their enemies, or that men and women are equally suited for positions of ecclesial authority, these facts by themselves constitute good reason—as good as the intuitions involved—to believe that a divine author would not teach anything to the contrary. And these considerations will appropriately guide our interpretation of the relevant texts. Of course, the reasons just mentioned can be defeated. It is possible, for example, to acquire evidence that scripture really does contradict some of our moral views or some of our scientific views. But the only condition under which sola scriptura would bind someone to revise her intuitions or scientific beliefs in light of scripture (instead of revising her understanding of scripture in light of her intuitions or scientific beliefs) would be one in which her reasons for believing that scripture teaches something contrary to reason are evidentially stronger than the intuitions themselves.

2. The Attributes of God The following passage from the Belgic Confession, one of the doctrinal standards of the Christian Reformed Church, provides what I take to be a decent initial list of the essential attributes of God: Article 1: We all believe in our hearts and confess with our mouths that there is a single and simple spiritual being, whom we call God—eternal, incomprehensible, invisible, unchangeable, infinite, almighty; completely wise, just, and good, and the overflowing source of all good. (Christian Reformed Church 1988: 78)

A decent initial list, but not a perfect or complete one. For example, the attributes of incomprehensibility, simplicity, unchangeability, and infinity are so difficult to understand that ascribing them to God is apt to mislead without extended comment (which I shall not provide here). I think that the attributions express truths; but I do not, for example, think that divine simplicity implies that there are no distinctions to be made within the Godhead or that incomprehensibility implies that God cannot be understood or talked about except via analogy or metaphor, or that divine unchangeability implies that it is false to say that God became incarnate, and so on. More importantly, the quoted passage leaves out some attributions that I would want to include (most of which the Confession itself includes, at least implicitly, elsewhere in its text). For example, I would say that God is necessarily existent, essentially triune, and omniscient; God is loving and merciful, and capable of sorrow and anger; God is a perfect person,¹² and the ¹² The Christian tradition maintains that God exists in or as three persons, but it also resoundingly affirms that God is personal and that God is perfect as a personal being. Not every way of



11

creator and sustainer of the concrete contingent universe. None of these additional attributes, however, are mentioned in the quoted passage. For some of these attributions, there is clear scriptural warrant. For others, however, there is not. What, then, justifies their presence in standard confessions, creeds, and other formal statements of Christian belief? A traditional but controversial answer is that the attributions not clearly derivable from other parts of scripture can nonetheless be derived from the scriptural claim that God is perfect. This answer has methodological implications that deserve further comment. In accord with many others in the Christian tradition, I think that our grasp of perfection can serve as a reliable guide to fleshing out our understanding of the divine attributes. (This idea makes substantive appearance in Chapters 4 and 5 of this volume.) It is not an infallible guide, for there is no good reason to think that any of us has a perfect grasp of it. Nor is it entirely clear exactly how it is a guide. Jeff Speaks (2018) has compellingly argued that the idea that substantive divine attributes are derivable from the claim that God is a perfect being, the greatest possible being, or something similar is fraught with problems; and I am not sure how or whether the problems can be overcome. Perhaps instead, then, we should think that the claim that God is perfect merely imposes constraints on our theorizing about the divine attributes; or perhaps we should think (as I said at the beginning of this paragraph) that our grasp of perfection simply helps us to flesh out our understanding of divine attributes that we arrive at via special revelation or some other route.¹³ In any case, Speaks’s arguments do not undercut the view that, to the extent that we have warrant for thinking that a perfect being would have some property p, we also have warrant for the claim that God has p. The question is simply how we can get such warrant, and (depending on the answer to that question) what further use the claim that God is perfect might be for the task of theology. In addition to whatever help it might give us in arriving at or understanding the divine attributes, our grasp of perfection also serves as a defeasible guide to interpreting scripture. For example, scripture describes God as our heavenly father, and most of the pronouns and other images used to refer to and characterize God are masculine. Now we face an interpretive choice. Ought we to infer that God is masculine, and prefers to be characterized as masculine? My own view, which I defend in Chapter 5, is that the answer is ‘no’: a perfect being would either transcend gender or belong to all genders equally, and would furthermore have no

understanding the trinity can comfortably accommodate the unqualified claim that God is a person; but (as we shall see) mine can. ¹³ I am also inclined to think that a full response to Speaks’s worries will require a revisionary understanding of ‘perfection’—one characterized not in terms of maximal greatness but more in terms of worship-worthiness. I develop this idea in fuller detail in Rea 2019.

12

   :  

preference in favour of a metaphysically misleading mode of characterization. This is one way, then, in which intuitions about perfection serve as an interpretive guide. In saying all of this I presuppose that at least some of our concepts apply univocally to God, and express truths about what God is like. I take perfection to be one of these concepts. If that presupposition were false, then the belief that God is perfect and perfect beings are F would not necessarily support the claim that God is F. This for the same reason that God is a father, and fathers are male does not support the thesis that God is male: analogies break down, so one must take special care in drawing inferences from analogical claims. But, of course, in saying that perfection is predicated univocally of God, I put myself at odds with views that motivate extremely apophatic and so-called ‘therapeutic’ approaches to theology.¹⁴ Chapter 6 represents my best attempt to make sense of and do justice to an extreme form of apophaticism—one according to which God is in some meaningful sense beyond being. As I explain in the postscript to the chapter, I ultimately decided that I could not endorse the view that I defended there; but some of the key ideas made their way into the chapters on divine transcendence in The Hiddenness of God (Rea 2018). In that book, I endorsed a moderate conception of divine transcendence according to which predications of intrinsic, substantive divine attributes involving non-revealed concepts (i.e. concepts whose content is not fully given in divine revelation¹⁵) are at best analogical.¹⁶ One might reasonably wonder whether the claim that God is perfect is such a predication and, if so, what that means for inferences from that claim to other divine attributes. This is a complicated issue, and my thinking on the topic has shifted in recent years. The short answer is that I think that the bare claim that God is perfect (as contrasted, say, with claims like ‘God is perfectly loving’) does not attribute an intrinsic property to God;¹⁷ so I am under no pressure to give up my view that perfection is univocally predicated of God. Likewise, then, I do not think that the doctrine of divine transcendence poses any threat to the kinds of inferences that are central to perfect being theology.

3. The Trinity The chapters in Part III of this book focus on the doctrine of the trinity. According to this doctrine, there is exactly one God, but three divine persons—Father, Son,

¹⁴ Cf. Hector 2011 for discussion. ¹⁵ The content of a concept is ‘fully given in divine revelation’ just if the complete content of that concept is part of or derivable from the content of divine revelation. ¹⁶ Cf. Rea 2018. ¹⁷ For more detail, see Rea 2019.



13

and Holy Spirit. A little more precisely, the doctrine includes each of the following claims:¹⁸ (T1)

There is exactly one God, the Father almighty.

(T2)

Father, Son, and Holy Spirit are not identical.

(T3)

Father, Son, and Holy Spirit are consubstantial.

To say that two things are consubstantial is to say that they share a common nature—i.e. they are members of exactly the same kind. Saying that two or more divine beings are consubstantial, then, implies that they are identical with respect to their divinity—they are not divine in different ways, neither is more or less divine than the other, and if one is a God then the other is a God too.¹⁹ It would be quite an understatement to say that this is a puzzling doctrine. At first glance (and, many would say, even after a much closer look) it appears to be incoherent. There are various ways of trying to demonstrate the incoherence. The one I prefer proceeds as follows: Suppose T1 is true. Then the Father is a God. But, given what I have just said about consubstantiality, T2 and T3 say that the Son and the Spirit are distinct from the Father (and from one another) but exactly the same kind of thing as the Father. So if the Father is a God, then the Son is a God, the Spirit is a God, and each is distinct from the other two. But then it follows that there are three Gods, contrary to T1. So the doctrine is incoherent. Resolving the contradiction means giving up a premise or saying that one of the inferences is invalid. Chapters 7, 8, 9 and 10 talk at length about what not to say in response to this problem (if one cares about creedal orthodoxy), about the nature of monotheism (which is a crucial matter to sort out if one is to understand how the doctrine of the trinity could be consistent with monotheism), and about the solution I favour. In short, the solution is to reject the inference from (T4) to (T5): (T4) The Father is a God, the Son is a God, and the Spirit is a God; and each is distinct from the other two. (T5)

Therefore: There are three Gods.

The challenge is to explain how this can sensibly be done. The model I favour begins with the Aristotelian idea that every material object is a compound of matter and form. The form might be thought of as a complex

¹⁸ This is not the only way of formulating the doctrine. But I choose this formulation because it is faithful to the creeds, suffices as well as others to raise the problem I wish to discuss, and emphasizes one central tenet of the doctrine—T3—that is all too often omitted in the contemporary literature. On the importance of T3, see Rea 2009 or, at length, Ayres 2004. ¹⁹ For purposes here I treat ‘God’ as a kind term rather than a name, obviously in keeping with its use in T1.

14

   :  

organizational property—not a mere shape, but something much richer. For Aristotle, the form of a thing is its nature. Thus, on this sort of view, St Peter would be a compound of some matter and the form humanity; St Paul would be a compound of the same form but different matter. Sharing the same form is what it means for Peter and Paul to be consubstantial. Now imagine a case in which some matter has two forms. Suppose, for example, that being a statue and being a pillar are forms; and suppose an artistic building contractor fashions a lump of marble that exemplifies both. The contractor has made a statue. She has also made a pillar. Furthermore, the two compounds are genuinely distinct: e.g. the pillar could survive erosion that would obliterate the statue. But surely we don’t want to say that two material objects—a statue and a pillar—occupy exactly the same place at the same time. What then might we say about this situation? What Aristotle would have said is that the statue and the pillar are the same material object, but not the same thing, or even the same compound. This sounds odd. How can two things or two compounds count as one material object? Answer: All there is to being a material object is being some matter that exemplifies at least one form. So we count one material object wherever we find some matter that exemplifies at least one form. To say that the statue and the pillar are the same material object, then, is to say no more or less than that the two things share all of the same matter in common. If this view is correct, then the following will be true: The statue is a material object, the pillar is a material object, the statue is distinct from the pillar but each is the same material object as the other; so exactly one material object (not two) fills the region occupied by the statue. Now let us return to the trinity. God is not material, of course; but we might still suppose that each divine person has constituents that play the same roles that matter and form play in material objects.²⁰ If we do, then we can say about the divine persons something like what we said about the statue and the pillar. Suppose that the divine nature plays the role of matter in the divine persons; and suppose that three separate properties (let’s just label them ‘F’, ‘S’, and ‘H’) play the role of form. Then we can say that all there is to being ‘a God’ is being a compound of the (one and only) divine nature and some person-making property (like ‘F’). Furthermore, to say that Father, Son, and Spirit are the same God is just to say that Father, Son, and Spirit share the same ‘matter’—i.e. the same divine nature. Father, Son, and Spirit are, on this view, genuinely distinct compounds

²⁰ In fact, I think some of the most important theologians who hammered out the NicenoConstantinopolitan formulation of the doctrine of the trinity did think of God in this way. Cf. Rea 2009 for discussion and references.



15

and genuinely distinct persons; but, precisely by virtue of sharing the same divine nature, they count as one and the same God.²¹ If all of this is right, then (as in the statue/pillar example) we can say the following about the divine persons: The Father is a God, the Son is a God, and the Holy Spirit is a God, but each is the same God as the others; so, since there are no other Gods, there is exactly one God. The inference from T4 to T5 is therefore blocked. Furthermore, we can say without qualification that God is a person, because on any way of resolving the ambiguity of ‘God’, ‘God is a person’ comes out true. We can even say unqualifiedly that God is triune, so long as we understand triunity as the attribute (possessed by each divine person) of sharing one’s ‘matter’ with exactly two other divine persons.

References Ayres, Lewis. 2004. Nicaea and Its Legacy: An Approach to Fourth-Century Trinitarian Theology. Oxford: Oxford University Press. Baker-Hytch, Max. 2016. ‘Analytic Theology and Analytic Philosophy of Religion’. Journal of Analytic Theology 4: 347–61. Bavinck, Hermann. 2003. Reformed Dogmatics, Vol. 1: Prolegomena. Grand Rapids, MI: Baker Academic. Berkhof, Louis. 1992. Systematic Theology. Grand Rapids, MI: Wm. B. Eerdmans Publishing Co. Bitar, Ray Paul. 2013. ‘The Wisdom of Clarity and Coherence in Analytic Theology’. Journal of the American Academy of Religion 81: 578–85. Christian Reformed Church. 1988. Ecumenical Creeds and Reformed Confessions. Grand Rapids, MI: CRC Publications. Couenhoven, Jesse. 2013. ‘Fodge-Ogs and HedgeOxes’. Journal of the American Academy of Religion 81: 586–91. Hector, Kevin. 2011. Theology without Metaphysics: God, Language and the Spirit of Recognition. Cambridge: Cambridge University Press. Lindbeck, George A. 1984. The Nature of Doctrine: Religion and Theology in a Postliberal Age. Louisville, KY: Westminster John Knox Press. McCall, Thomas H. 2015. An Invitation to Analytic Christian Theology. Downers Grove, IL: IVP Academic. McCall, Thomas H. and Timothy Pawl. 2018. ‘Why a Catholic Theologian Can Be an Analytic Theologian by Tom McCall and Timothy Pawl’. BLOGOS: http://blogos.

²¹ Why do Peter and Paul not count as two persons but one human being? Because, unlike the divine nature, human nature does not play the role of matter.

16

   :  

wp.st-andrews.ac.uk/2018/02/15/why-a-catholic-theologian-can-be-an-analytic-theo logian/. Macdonald, Paul A., Jr. 2014. ‘Analytic Theology: A Summary, Evaluation, and Defense’. Modern Theology 30: 32–65. Oliver, Simon. 2010. ‘Review of Crisp and Rea, eds., Analytic Theology’. International Journal of Systematic Theology 12: 464–75. Rea, Michael. 2009. ‘The Trinity’. In The Oxford Handbook of Philosophical Theology, 403–29. Oxford: Oxford University Press. Rea, Michael. 2013a. ‘Analytic Theology: Précis’. Journal of the American Academy of Religion 81: 573–7. Rea, Michael. 2013b. ‘Analytic Theology Roundtable: Replies to Bitar, Couenhoven, and Wood’. Journal of the American Academy of Religion 81: 614–19. https://doi. org/10.1093/jaarel/lft044. Rea, Michael. 2017. ‘(Reformed) Protestantism’. In Inter-Christian Philosophical Dialogues, edited by Graham Oppy and Nick Trakakis, 4: 67–88. London: Routledge. Rea, Michael. 2018. The Hiddenness of God. Oxford: Oxford University Press. Rea, Michael. 2019. ‘Deflating Perfect Being Theology’. Unpublished ms. Sarrisky, Darren. 2018. ‘Biblical Interpretation and Analytic Reflection’. Journal of Analytic Theology 6: 162–82. Speaks, Jeff. 2018. The Greatest Possible Being. Oxford: Oxford University Press. Vanhoozer, Kevin J. 2017a. ‘Analytic Theology as Sapiential Theology: A Response to Jordan Wessling’. Open Theology 3 (1): 539–45. https://doi.org/10.1515/opth-20170041. Vanhoozer, Kevin. 2017b. ‘Analytics, Poetics, and the Mission of Dogmatic Discourse’. In The Task of Dogmatics: Explorations in Theological Method, edited by Oliver D. Crisp and Fred Sanders, 23–48. Grand Rapids, MI: Zondervan. Wessling, Jordan. 2017. ‘Analytic Theology as Sapiential Theology: Reflections on a Concern Raised by Kevin J. Vanhoozer’. Open Theology 3: 380–96. https://doi.org/ 10.1515/opth-2017-0030. Westerholm, Martin. 2019. ‘Analytic Theology and Contemporary Inquiry’. International Journal of Philosophy and Theology 80: 230–54. Wood, William. 2009. ‘On the New Analytic Theology: The Road Less Travelled’. Journal of the American Academy of Religion 77: 941–60. Wood, William. 2013a. ‘Analytic Theology and the Academic Study of Religion’. Paper presented at the Logos Workshop in Philosophical Theology, University of Notre Dame. Wood, William. 2013b. ‘Philosophical Theology in the Religious Studies Academy: Some Questions for Analytic Theologians’. Journal of the American Academy of Religion 81: 592–600. Wood, William. 2014. ‘Analytic Theology as a Way of Life’. Journal of Analytic Theology 2: 43–60.

1 Realism in Theology and Metaphysics Since the early 2000s, increasing attention has been paid in two separate disciplines to questions about realism and ontological commitment. The disciplines are analytic metaphysics on the one hand, and theology on the other. In this chapter, I shall discuss two arguments for the conclusion that realism in theology and metaphysics—that is, a realist treatment of doctrines in theology and metaphysics—is untenable.¹ ‘Realism’ is variously defined in the literature. For purposes here, I shall adopt the following characterizations: • where ‘x’ is a singular term, realism about x is the view that there is a y such that x = y • where ‘F’ is a putative kind-term, realism about Fs is the view that there are Fs and that F is a genuine kind-term • where ‘T’ refers to the linguistic expression of some claim, theory, or doctrine, to interpret or treat T realistically is (a) to interpret T as having an objective truth-value (and so to interpret it as something other than a mere evocative metaphor or expression of tastes, attitudes, or values); and (b) to interpret T in such a way that it has realist truth-conditions—ie., it is true only if realism about the xs and Fs putatively referred to in the theory is true. • where ‘D’ refers to a discipline (like metaphysics or theology), realism in D is or involves interpreting the canonical statements of theories or doctrines in D realistically. Thus, one way to be an anti-realist about God, say, is to affirm explicitly that there is no such being as God; but another way to be an anti-realist about God is to say, for example, that ‘God exists’ expresses a truth, but that the truth it expresses isn’t that there is an x such that x = God. Likewise, one way to be an anti-realist about beliefs, say, is to affirm explicitly that there are no such things as beliefs; but another way to be an anti-realist about beliefs is to offer paraphrases of belief-talk

¹ By ‘theology’ in the present context I have primarily in mind those sub-disciplines of theology that go by such labels as ‘systematic theology’, ‘dogmatic theology’, ‘philosophical theology’, and the like. The arguments of this chapter do not (to my mind, anyway) have any obvious bearing on (say) historical and biblical theology or the various kinds of biblical criticism that are practised in contemporary theology departments.

Essays in Analytic Theology: Volume I. Michael C. Rea, Oxford University Press (2020). © Michael C. Rea. DOI: 10.1093/oso/9780198866800.003.0002

20

   :  

according to which ‘there are beliefs’ expresses a truth, but the term ‘belief ’ doesn’t pick out a genuine kind of mental state. Furthermore, in light of the above characterizations, theists and atheists alike can interpret the same theological claims realistically. Indeed, their disagreement will most perspicuously be expressed as a disagreement over the truth value of the claim ‘God exists’ realistically interpreted. One motivation for doubting that we should interpret doctrines in metaphysics or theology realistically is the vague worry that practitioners of both disciplines are spinning out theories with no reliable way of determining which of the competing theories is true. The worry is that the practitioners of each discipline are simply talking past one another, that their ‘debates’ lack substance, and that their theories don’t tell us anything of interest about the world or its inhabitants. In short, theorizing in both disciplines is but idle word play; and so it is doubtful that the theories in either discipline have objective truth values or truth values with realist truth conditions. Those caught in the grip of this worry then face the question of what to do with metaphysics and theology. In the case of metaphysics, the verdict is often that we should simply view it as a game and either stop playing it or else leave it to weekends and spend our day jobs on more serious activities—like, perhaps, philosophy of science. Theology is more complicated because many of the objectors still want to maintain that there is some value in religion, and they recognize that the sentences typically taken to express the core doctrines of religions like Christianity still have some value even if they can’t be taken with literal seriousness. Indeed, whereas the objectors to metaphysical realism tend also to be objectors to metaphysics in general, the objectors to theological realism often style themselves as people interested in saving religion from the pernicious influence of modernism, fundamentalism, ontotheology, or other villains. Still, for many of us, theology is of far lesser interest and import if the anti-realist verdict is allowed to stand. If, in the end, the theories produced by theology are not fitting objects for belief, it is hard to see why we should take the discipline very seriously. In this chapter, I want to examine two ways of making the vague worry more precise. In God and Realism, Peter Byrne (2003) offers an argument against realism in theology that is readily modified to cut against realism in metaphysics as well. And in The Empirical Stance, Bas van Fraassen (2002) offers an argument against the very practice of analytic metaphysics that is both readily seen as an argument against realism in metaphysics and easily adapted into an argument against realism in theology. In what follows, I will examine these two arguments and defend four conclusions: first, that Byrne’s argument is answerable; second, that van Fraassen’s argument is unanswerable if we adopt what he calls the ‘empirical stance’; third, that there is (and can be) absolutely no reason why metaphysicians or theologians ought to adopt the empirical stance; and, finally,

    

21

that for those who don’t adopt the empirical stance, van Fraassen’s objections can be answered in precisely the same way as we answer Byrne’s. The chapter has three sections. In the first, I briefly present and respond to Byrne’s argument against theological realism. In the second, I present van Fraassen’s argument against analytic metaphysics and I show how, if sound, it constitutes a reason to reject both metaphysical and theological realism. Finally, I show how van Fraassen can be answered. Obviously what I am doing here falls far short of a full-blown defence of realism in either metaphysics or theology. But the objections raised by van Fraassen and Byrne are tokens of a type of objection that I think is rather widely endorsed among those who are suspicious of these two brands of realism. Thus, responding to those objections constitutes an important first step in the direction of a defence.

1. Byrne’s Argument Peter Byrne sums up his argument against theological realism as follows:² (1) All disciplines of thought that can be interpreted realistically show the accumulation of reliable belief. (2) Theology does not show the accumulation of reliable belief. (3) Therefore, theology cannot be interpreted realistically. Byrne declares that this argument is ‘simple’ (2003: 162) and ‘decisive’ (2003: 161). As a matter of fact, however, it is no simple matter at all to figure out precisely what Byrne means by terms like ‘interpreted realistically’ or ‘show the accumulation of reliable belief ’; nor is it a simple matter to figure out why exactly he thinks that the two premises of the argument are true. Since time will not permit the sort of detailed exegetical discussion it would take to sort out the terminological issues, I will simply offer glosses that I think are faithful to what Byrne was aiming at. I will then try to reconstruct as best I can his defence of the premises. Readers who think that the resulting product is not something Byrne would be happy with are welcome to take the argument of the present section as one of my own invention (albeit inspired by the work of Byrne and others) and offered primarily as a prelude to the discussion of van Fraassen. Byrne seems to think that to interpret a discipline of thought realistically is just to see it as the sort of discipline whose methods of enquiry are successfully aimed at truth, whose theories are grounded in and responsive to evidence, and whose

² Byrne 2003: 162.

22

   :  

conclusions are intended to tell us the literal, objective truth about the world.³ Thus, those disciplines which we can interpret realistically in Byrne’s sense are presumably just those disciplines whose theories we can sensibly interpret realistically in my sense. Byrne also seems to think that a discipline shows the accumulation of reliable belief just in case it generates an increasing number of statements that we can rationally expect not to be contradicted by future well-established theories in the discipline.⁴ Reliable beliefs in a discipline D are just those beliefs that can be expected to remain permanently sanctioned by D’s theoretical apparatus.⁵ To say that a belief is reliable, then, is not to say that it is likely to be true (though it might in fact turn out that the reliable beliefs of a discipline are just the ones that are likely to be true). Rather, it is just to say that it is unlikely to be overturned by future evidence or theoretical developments. Given all of this, Byrne’s argument might be restated as follows: Consider some discipline D. We can take D’s theories as worthy of belief and as aiming to tell us the literal truth about the world only if the practice of D over time generates an increasing number of statements that we can rationally expect not to be contradicted by future well-established theories in D. But we don’t find such an increase of ‘reliable belief ’ in theology. Thus, we should not treat theological theories as worthy of belief or as aiming to tell us the literal truth about the world. And, we might add, what goes for theology also goes for metaphysics: we don’t find the accumulation of reliable belief in that discipline either. Thus, we should not be realists about theories in metaphysics either. So much for the argument. Now, what shall we think of the premises? Let us begin by observing that neither of the premises is obviously true. A relatively narrow discipline that hits on the truth right at the outset will show no accumulation of belief at all; but that by itself is not obviously a reason to doubt that it is to be interpreted realistically. Thus, there is prima facie reason to think that premise

³ Cf. Byrne 2003: ch. 1, passim and, especially, pp. 155–9. On p. 159, Byrne offers what might appear as an outright definition of what it is to interpret a discipline of thought realistically. He says: ‘We have reached the conclusion that to interpret a discipline of thought realistically is to see its evolving conclusions as the outcome of real-world influences.’ But, of course, this offers us nothing by way of precision; for, after all, superstitions, prejudices, fears and ambitions, peer pressure and other sociological influences, and so on are all ‘real-world influences’. Every discipline—from biology and chemistry on the one hand to astrology and iridology on the other—is such that its ‘evolving conclusions’ are the outcomes of ‘real-world influences’. But, of course, this can’t be what Byrne has in mind. To find out what he has in mind, however, we have to look elsewhere and then offer a gloss; and my own view is that if we do this, and if we do it in the most charitable way possible, we arrive at something like the gloss that I have just offered. ⁴ Cf., especially, Byrne 2003: 159–61. ⁵ Note, however, that this definition of reliable belief leaves open the possibility that reliable beliefs in one discipline might be contradicted by reliable beliefs in another discipline. If we were looking for sufficient conditions for the realistic interpretation of a discipline, we would want to rule this out. But since Byrne is concerned to show that theology fails to meet a necessary condition for being interpreted realistically, I doubt that this problem will cause much trouble for present purposes.

    

23

(1) is false. Moreover, many branches of theology within Christendom seem clearly to have shown the accumulation of reliable belief (as defined above) over the centuries. In the Catholic Church, for example, the Nicene Creed and the Canons and Decrees of the Council of Trent—to name just two of a variety of doctrinal standards within Catholicism—are not at all likely to be contradicted by future developments in (official) Catholic theology. The Nicene Creed and the Canons and Decrees of the Council of Trent are explications of and elaborations on doctrines that the Catholic Church claims to have found in scripture. They were needed precisely because their contents were not explicitly part of Christian belief prior to their formulation—so, in other words, they constitute genuine theoretical developments rather than being, like the scriptures, mere sources for theological reflection. Though plenty of Roman Catholics, including Roman Catholic theologians, disagree with them in part or in their entirety, the Catholic Church is set up in such a way that we can be virtually certain that they will not be contradicted by future established theories in official Catholic theology. Likewise, and for similar reasons, it is highly unlikely that either the Nicene Creed or the Westminster Confession will be contradicted by future developments in (traditional, orthodox) Presbyterian theology—here not because the Presbyterian Church is set up so as to guarantee that those doctrinal standards will be preserved, but rather because theology as it is practised by those with a traditional, orthodox bent is not at all revisionary in the way that certain other brands of theology might be.⁶ And, of course, similar things might be said for various other denominational theologies. Thus, there is prima facie reason to reject premise (2) as well.⁷ What then are Byrne’s arguments for these premises? Premise (2), oddly enough, is offered without any argument at all. Byrne simply declares that it is obviously true, and then follows that declaration with remarks that effectively just restate and elaborate on it. Thus, he writes: Consider this question: do we know anything more about God than we did at the dawn of Christian theology nearly 2,000 years ago? Answer: No. During that ⁶ Moreover, even the revisionists in the Presbyterian camp will likely agree on permanence of conditional claims to the effect that, given an appropriately strong view of the inspiration and infallibility of the Bible, the doctrines expressed in the Nicene Creed, the Westminster Confession, and various other doctrinal standards are true. ⁷ Is it really fair, though, to treat official Catholic theology, or traditional, orthodox Presbyterian theology as disciplines in their own right, rather than as branches of a single discipline—theology? It is hard to see why not; but, in the end, nothing hinges on treating them as such. For surely Byrne would not countenance this sort of reply to his argument: ‘Granted, we cannot interpret theology realistically. But that doesn’t matter; for all I claim is that we can interpret the distinct theory-building enterprise of Catholic theology realistically.’ But so far as I can tell, the only argument he has against this reply is an adapted version of the argument currently under discussion: i.e. a theory-building enterprise can be interpreted realistically only if it shows the accumulation of reliable belief; but in these various theorybuilding enterprises there has been no accumulation of reliable belief. If this is the argument he would use, then my reply is as above: these theory-building enterprises have shown the accumulation of reliable belief after all.

24

   :   period many theological theories have come and gone in Christian thought, but there has been no accumulation of insight and discovery whatsoever. The stock of reliable beliefs about the Christian God, about its attributes and plans, has not increased one iota . . . . Theology has not possessed intellectual traditions and modes of discovery [analogous to those in science] to enable its practitioners to be open to influences from divine reality and its practitioners have not been put in cognitive contact with divine reality. The academic discipline of theology is simply not productive of reliable beliefs about God—or about anything else for that matter. It cannot be understood realistically. QED. (2003: 162)

But why should we believe any of this? The stock of reliable beliefs about the Christian God has not increased one iota? Again, it is hard to take this claim at all seriously in light of what we know of the histories of Catholic theology, traditional Presbyterian theology, and any of a number of other denominational theologies within Christendom. The ‘QED’ at the end of the paragraph seems, to put it mildly, a bit premature. Premise (2), then, is a natural target for resistance. But for present purposes I want to waive worries about premise (2) and focus instead on premise (1). Here Byrne does want to offer argument; though what the argument amounts to, exactly, is rather hard to tell. What he says explicitly in favour of (1) is just this: ‘Premise (1) has been established through consideration of the example of science’ (162). What we find, however, upon reviewing his consideration of the example of science is that, really, he has defended not (1) but (1a): (1a) Disciplines that show the accumulation of reliable belief are to be interpreted realistically. And what he offers in support of (1a) is just a version of the familiar ‘no miracles’ argument for scientific realism. In his words: The story of science is a human story, but one which is comprehensible only if we assume that human theory and practice are being in part, at least, shaped by what the world is really like. If there is a progressive, cumulative structure to the development of science, this strongly suggests real-world cognitive contact and influence; otherwise the accumulation of reliable belief would be the merest accident. (2003: 156)

A generalization on this argument yields (1a); but it yields nothing close to (1). Nevertheless, there is an argument for (1) lurking in the neighbourhood. Suppose we endorse the following premises: (1b) For any discipline D, there is no initial presumption that D is to be interpreted realistically.

    

25

(1c) There can be no evidence supporting a realistic interpretation of a discipline apart from the accumulation of reliable belief. (1d) Absent an initial presumption for interpreting a discipline D realistically, and absent evidence that D is to be interpreted realistically, D cannot rationally be interpreted realistically. If we do endorse (1b)–(1d), and if we are persuaded by Byrne’s argument for (1a), then we have a ready argument for (1). The idea, then, is something like this: For any discipline D, the practitioners of D aren’t entitled simply to adopt, without argument, a realist interpretation of D. Rather, if we want to interpret D realistically, we need to do so on the basis of evidence—evidence that D is really putting us in touch with the truth about things. But what evidence could we possibly acquire? In the case of science, we have (allegedly) the accumulation of reliable belief. And it is, one might think, very hard to explain how we could have that if science weren’t putting us in touch with the truth about things. But absent the accumulation of reliable belief, what other evidence could we have for interpreting science realistically? What other phenomenon would be best explained by the supposition that science, or any other discipline, is putting us in touch with truth? Apparently none. Thus, the only disciplines that we are entitled to interpret realistically are those that show the accumulation of reliable belief—which is just to say that premise (1) is true. The trouble with this line of reasoning is just that, if it were sound, we would face the threat of global scepticism. Let D be the discipline of Detecting Reliable Beliefs (DRB). (If you prefer, you could treat it as the discipline for detecting success, and then fill in your favorite criterion for success. But since we’re talking about Byrne, we’ll focus on his.) Practitioners of DRB—all of us, I suppose, to some extent or another—are engaged in the enterprise of trying to find out which, if any, of their beliefs count as reliable. Moreover, if Byrne is correct about the criteria for interpreting a discipline realistically, realism about any other discipline is predicated in part on a realist interpretation of DRB. That is, unless we assume that the claims of DRB have objective truth values (and, indeed, that they tell us the objective truth about things), we will not be entitled to believe that any discipline has shown the accumulation of reliable belief.⁸ And if we are not ⁸ Suppose you think that some claim of DRB has a truth value, but that the truth value is not objective. Thus, suppose you think something like this: ‘It is true, but only true-for-me, that B1–Bn are reliable beliefs in discipline D.’ Given our understanding of reliable belief, this would seem to be equivalent to the view that you, but not necessarily anyone else, can rationally expect that B1–Bn will be permanently sanctioned parts of D’s theoretical apparatus. But isn’t this claim selfundermining? Note that the claim isn’t equivalent to the (perhaps perfectly sensible) claim that you have evidence E that (for all you know) nobody else has, and that given this, it is objectively rational for you (but not necessarily for anyone lacking E) to believe that B1–Bn will be permanently sanctioned parts of D’s theoretical apparatus. Rather, if it is really only true-for-you that B1–Bn are reliable beliefs in D, the idea is that even people in your same epistemic position might not rationally be able to expect

26

   :  

entitled to believe that any discipline has shown the accumulation of reliable belief, then (by Byrne’s lights), we cannot interpret any discipline realistically. So can we interpret the theories of DRB realistically? Well, following Byrne’s reasoning, in order to assess this question we should ask: Has DRB itself shown the accumulation of reliable belief? There are two ways of trying to answer this question. Practitioners of DRB might assume from the outset (until given reason to do otherwise) that the methods they employ in practising DRB are successfully aimed at truth; and, on the basis of this assumption, they will likely say ‘yes, DRB has shown the accumulation of reliable belief ’. Note, however, that these practitioners of DRB are interpreting DRB realistically, and they are doing so not on the basis of its success, but rather in advance of any awareness of its success. Indeed, their assessment of DRB’s success depends on their realist interpretation of DRB. On the other hand, practitioners of DRB might refrain from interpreting DRB realistically until they have independent evidence that DRB has shown the accumulation of reliable belief. But this will be a long wait; for, after all, any mode of detecting whether DRB has shown the accumulation of reliable belief will itself fall under the scope of DRB. The upshot, then, is this: If you can’t interpret a discipline realistically until it has shown the accumulation of reliable belief, then you won’t ever be able to interpret DRB realistically. And if you can’t interpret DRB realistically, then you can’t interpret any discipline realistically. But Byrne acknowledges that some disciplines can be interpreted realistically. Thus, he must concede that at least some disciplines can be interpreted realistically even if they haven’t shown the accumulation of reliable belief. In the case of DRB, it seems, in fact, that there is a rational initial presumption that DRB is to be interpreted realistically, contrary to (1b) above. And, for all we know, there might be other disciplines for which there is evidence, but of an entirely different sort, that supports a realist interpretation of the discipline—contrary to (1c). Absent (1b) and (1c), however, there is no clear argument for (1). And so Byrne’s argument fails—even ignoring worries about (2). One might object that there is no such discipline as DRB; thus, my argument against Byrne—which apparently rests on the supposition that there is such a discipline—fails. I admit that it seems odd to characterize DRB as a discipline. After all, it is hard to imagine awarding university degrees in DRB, or applying for National Endowment for the Humanities or National Science Foundation funding to pursue DRB. But, really, the objection does not rest in any important way upon the supposition that the criteria for discipline-individuation allow us to treat DRB as a discipline. For whatever else DRB happens to be, it is, at the very least, a that B1–Bn will be permanently sanctioned by D. But this fact by itself counts as good reason to question whether practitioners of D will continue to sanction B1–Bn. Thus, in affirming that it is only true for you that B1–Bn can rationally be expected to be permanently sanctioned by D, you acquire a defeater for the belief that B1–Bn will always be sanctioned by D; and so it becomes irrational for you to expect that B1–Bn will be permanently sanctioned by D.

    

27

theory-building activity. And what the argument just presented shows is that at least some theory-building activities must be interpreted realistically in the absence of the accumulation of reliable belief. But if that is right, then the door is open for thinking that other theory-building activities—including, for all we know, full-blown disciplines (whatever exactly a ‘discipline’ amounts to in Byrne’s usage) can be interpreted realistically in the absence of the accumulation of reliable belief. In closing this section, I would like to note a connection between my argument here and some conclusions that I have defended elsewhere. In World Without Design: The Ontological Consequences of Naturalism (Rea 2002), I argued that, assuming we want to form beliefs on the basis of evidence, we cannot avoid treating at least some sources of evidence as basic sources, where a source is treated as basic just in case it is trusted as reliable in the absence of evidence for its reliability. The reason is simple: in order to acquire or appreciate evidence for the claim that a source S is reliable, one must either invoke evidence from some distinct source S*, or one must invoke evidence from S itself. Since we do not have infinitely many sources of evidence, it follows that at least one of our sources is such that some, if not all, of the evidence we have in support of its reliability comes from the source itself. But in such a case, our disposition to trust the source precedes our having evidence for its reliability. Thus, our trust in the source does not depend on the evidence—even if, in the end, our belief that the source is reliable does depend (circularly) on the evidence. So, assuming it is rational for us to trust any of our sources of evidence as reliable, it must be rational to treat at least one of them as basic. But if all that is right, then (ceteris paribus) it will be rational to interpret realistically any discipline that can be practised just by relying on basic sources, and furthermore, it will be rational to do so in advance of any evidence of success. Suppose, for example, we (rationally) treat source S as basic. Then, from our point of view, source S is a reliable source of evidence. Other things being equal, then, it will be rational for us to form beliefs on the basis of evidence garnered from source S. In other words, propositions supported by evidence from S will be fitting objects of belief for us. But then, of course, it follows that (ceteris paribus—i.e., absent counterevidence from other sources, strange circumstances, and so on) the conclusions of any discipline that can be practised just by relying on S will be fitting objects for belief. In other words, it will be rational for us to believe those conclusions as expressions of the literal, objective truth about the world. Thus, by definition, it will be rational for us to interpret the discipline realistically. Of course, one might try to oppose realism in theology or metaphysics by arguing that these disciplines inevitably rely on evidence from sources that are not rationally treated as basic, or by arguing that they are in some other way defective; but those would be very different arguments, and ones that have yet to appear in the literature in any sort of developed form.

28

   :  

2. Van Fraassen I now want to turn to a different sort of challenge: an objection to the very enterprise of metaphysics that, if sound, carries over to theology as well. The objection is raised by Bas van Fraassen in the opening chapter of The Empirical Stance. According to van Fraassen, analytic ontology—which, for purposes here, I am taking to be identical to the discipline of metaphysics—aims to answer questions that science doesn’t ask, and to do so in the same way that science answers its questions. What does this mean exactly? Primarily, it means that metaphysicians posit things to do explanatory work, and then they try to justify the acceptance of their explanatory posits by appeal to the sorts of virtues that justify the acceptance of scientific explanations—not virtues like predictive success and increased ability to control nature, however; rather, virtues like explanatory power, simplicity, elegance, conservatism, and the like. (For convenience, I shall refer to the latter sorts of virtues as ‘explanatory virtues’. In doing so, however, I do not mean to make any presuppositions about whether or to what extent the virtues that I am calling ‘explanatory’ can be distinguished from putatively nonexplanatory virtues like empirical adequacy; nor do I mean to make any presuppositions about whether the allegedly different explanatory virtues can be distinguished from one another.) Thus construed, analytic ontology is open to two objections. The first objection is that relevant differences between science and ontology cast doubt on the justification for accepting the ontologist’s explanatory posits. Van Fraassen lays out the objection as follows: You will have understood me correctly if you now see science and analytic ontology caught in a Pascalian wager . . . . If the wager is on a choice of theories or hypotheses, then from a God’s-eye view, success consists in selecting the true and failure in choosing the false . . . . As in all success and failure, however, although there is value in winning as such, there are also collateral value and damage that win and loss bring along with them . . . . In science, the stakes are great for all of us: safety, food, shelter, communication, all the preconditions for life in peace and justice that a successful science can enhance. The risk of acquiring some false beliefs matters little in comparison. Most important for us, here, the acquisition of false beliefs by itself, apart from their practical empirical consequences, is no great matter in its contrast class of practical risk and gain. That is very far from how it is in metaphysics. There the gains to be contemplated are those of having true beliefs . . . and of being in a position to explain . . . . The risk is precisely that of acquiring false beliefs . . . . Where is the metaphysician who shows us how likely it is that inference to the best explanation in ontology

    

29

will lead to true conclusions? Why is he or she missing? Where is the metaphysician who makes the case that the gain of explanatory power outweighs the risk of ending up with a tissue of falsehoods? (2002: 15–16; emphasis mine)

The argument, in sum, seems just to be this: False belief, as such, is to be avoided; and there is no evidence that explanatory virtues lead us to true belief. Thus, we are justified (if at all, one might add) in forming beliefs on the basis of bestexplanation arguments only if the gain from doing so outweighs the risk. In science, we might have a case for the conclusion that gain outweighs risk; but not in metaphysics.⁹ Thus, absent some argument for the conclusion that inference to the best explanation in metaphysics is likely to lead us to the truth (and given that there is something rather absurd about constructing metaphysical theories while at the same time withholding belief in them and suspending judgment about their explanatory status), one ought not to engage in metaphysics at all. Of course, one might note that there are arguments in the literature for the conclusion that explanatory virtues are truth-indicative. But such arguments typically focus on the status of explanatory virtues as criteria for theory choice in science rather than metaphysics; and, more importantly, they typically reason from the premise that this or that feature of science is hard to explain apart from the assumption that choosing theories on the basis of certain explanatory virtues is a reliable way of reaching the truth to the conclusion that, therefore, the assumption in question is true. But according to van Fraassen, the demand for explanations is precisely what good empiricists aim to resist.¹⁰ In other words: If one is already committed to the metaphysical enterprise, one will take seriously the demand for explanation, and, in doing so, one may well be led to the conclusion that choosing theories on the basis of their explanatory power (among other virtues) is a reliable way of getting to the truth. But if the question is why we should be committed to the metaphysical enterprise in the first place, then arguments that presuppose a need for explaining things will be impotent. Thus, the empiricist will be completely unmoved by the usual arguments for the claim that explanatory virtues are truth-indicative. But, one might wonder, why should metaphysicians care whether empiricists are moved by replies to their objections? As it turns out, I think that they should not care; and I think that, in the end, van Fraassen’s objections against metaphysics are, at best, grumblings that express the empiricist distaste for metaphysics but that will not and should not convince the unconvinced that the distaste ought to ⁹ Here I take it that van Fraassen is suspending, for the sake of argument, his view—defended at length in The Scientific Image (van Fraassen 1980)—that explanation is not the aim of science, and that believing that scientific theories provide true explanations is to take an objectionably metaphysical view of science. ¹⁰ On this, see (for example) van Fraassen’s reply to Richard Boyd in van Fraassen 1985.

30

   :  

be shared. Showing this, however, will take a bit of argument—argument that is best left until after van Fraassen’s second objection has been presented. The second objection is that the procedure of explanation via theoretical posit results in the creation of ‘simulacra’ which then replace the real things that we aim to be theorizing about and thereby make our theoretical activity into an idle exercise in wordplay. So, for example, van Fraassen argues that when philosophers ask the question ‘Does the world exist’, what they inevitably do is to make the question rigorous with technical definitions of ‘world’ and related terms that map on to some but not nearly all uses of the term ‘world’ and then stipulate that the world exists if, and only if, the world as they have defined it exists. He notes that one might just as well introduce a new technical term—‘Sworld’, for example— and then ask whether the Sworld exists. The trouble, of course, is that we wouldn’t really care about the answer—unless we had reason to think that ‘the world exists’ means the same thing as ‘the Sworld exists’. But therein lies the rub; for, after all, it doesn’t, exactly. As van Fraassen puts it: “Sworld” is intelligibly related to “world,” taking over a carefully selected family of uses, regimenting them, and is then used to make new, logically contingent, fully intelligible assertions. If we are careful not to let other usages of “world” creep back into our professional discourse, then “the world exists” is a perfectly good way of saying “the Sworld exists.” The unfortunate negative verdict forced on us by this . . . line of reasoning, which grants sufficiency to such lenient standards [of meaningfulness], is that it is very easy, all too easy, to make sense. We can sit in our closets and in a perfectly meaningful way, kneading and manipulating the language, create new theories of everything and thereby important contributions to ontology. In other words, to put it a little more bluntly, this “word play” we [are engaged in] is merely idle world play; although shown to be meaningful, it is merely idle world play nevertheless. (2002: 27)

In sum, then, ontologists who try to answer the question ‘Does the world exist?’ in the way just described inevitably replace talk of the world with talk of a simulacrum—the Sworld. But the simulacrum isn’t what we care about when we ask the question the ontologist is trying to answer; and so the ontologist is engaged in a project that is somewhat removed from our real interests and concerns. So go van Fraassen’s objections against metaphysics. I’ll consider responses in a moment; but first I want to comment briefly on how these objections might carry over to theology. It is clear that van Fraassen thinks that they do—or, at any rate, that they carry over to some kinds of theology—for some of the examples he uses are examples drawn from the philosophy of religion. I take it that, as applied to theology, the objections are just these: (a) there are no empirically detectable

    

31

payoffs for the procedure of explanation by theoretical posit in theology, and (b) the God that is talked about in that part of theology that resembles or overlaps with analytic philosophy of religion is a simulacrum—either different from the real God, or else simply stipulated to be identical to the real God. Obviously not all of theology is indicted by these objections. But at least two kinds probably will be: so-called perfect being theology, an approach adopted by many medieval and contemporary analytic philosophers that attributes properties to God (such as simplicity and changelessness) on the basis of intuitions about perfection; and systematic and philosophical theology, at least insofar as these activities involve a certain amount of explanatory postulation in the effort to build detailed theories out of the data provided by divine revelation and religious experience. And, again, the objection will be that perfect being theology results in the creation of a simulacrum (one which privileges for theological purposes the characterization of God as perfect over characterizations of God as our parent, our employer, our shepherd, and so on), and that the explanations offered by systematic and philosophical theology have no payoffs that outweigh the risks associated with treating explanatory power as a theoretical virtue. But now what shall we think of these objections? Let us begin with the ‘simulacrum’ objection. And here, I think, it will be helpful to begin by considering a rather simple-minded response to van Fraassen. The objection, again, is that metaphysicians aren’t talking about things that we care about: they talk about the Sworld rather than the world; they talk about the God of the philosophers rather than about God; and so on. But, one might wonder, how do we know that the Sworld isn’t the world? How do we know that the God of the philosophers isn’t God? How, in other words, do we know that these things are just simulacra? It seems, in fact, that we can’t know unless we already have a metaphysical story to tell about the nature of the world or about the nature of God. But, of course, to have a metaphysical story to tell, we’d have to do some metaphysics. But the response isn’t quite right. It’s not true that the only way to know that you’ve constructed a simulacrum is by comparing the object of your discourse with the real thing. The other way to know that you’ve constructed a simulacrum is by knowing that constructing simulacra is pretty much all you can do. And here, I think, we start to get at the real objection in van Fraassen’s text. As I see it, the concern is just this: Metaphysicians are in the business of offering explanations by postulate. We postulate definitions, entities, properties, and the like; and we use our postulates to build theories that explain the world to us. Our postulates are by their very nature props and models—‘simulacra’, if you will, that may or may not manage to represent the things they are about in a full and accurate way. And— this is the concern—precisely because we have no evidence that explanation by postulate is a reliable way of reaching the truth, we have no way of knowing the extent to which simulacra represent the things that they are about. To the extent,

32

   :  

then, that we try to force our talk about (say) God or the world to conform to our idea of the God of the philosophers or of the Sworld, we change the subject from something we know and recognize to something that, for all we know, may be (and probably is) at best a shadow of the thing we actually care about.

3. Two Stances Suppose we grant that the, or a, defining characteristic of analytic metaphysics (and of certain theological enterprises as well) is a willingness to engage in explanation by theoretical postulate. How shall we address the sceptical worry raised at the end of the previous section—the worry, in short, that we have no evidence that the methods of metaphysics are reliable and therefore we have no reason to think that the simulacra we construct bear any important relation to the things we care about? The answer, so I shall argue, just depends on which of two ‘stances’ we adopt: the empirical stance or the metaphysical stance. According to van Fraassen, empiricism is a philosophical position that involves, in addition to a familiar sort of respect for science, empirical investigation, undogmatic theorizing, and the like, the following two values, tendencies, or attitudes: (a) a rejection of demands for explanation at certain crucial points, and (b) a strong dissatisfaction with explanations (even if called for) that proceed by postulation. It is (a) and (b), he thinks, that separates empiricists from metaphysicians. Moreover, empiricism, on his view, is not a belief or a philosophical thesis. Rather, it is a stance: an ‘attitude, commitment, approach, a cluster of such—possibly including some propositional attitudes such as beliefs as well’ (2002: 48–9). To adopt empiricism, or the empirical stance, then, is to adopt the characteristic attitudes, commitments, and so on of the empiricist tradition—including (a) and (b). And we might go on to say that, in contrast with the empirical stance there is also a metaphysical stance—a stance characterized in part by a willingness to embrace demands for explanation and to be satisfied with explanations that proceed by postulation. So far as I can tell, those who adopt the empirical stance can have no answer to van Fraassen’s objections against metaphysics. We have already conceded for the sake of argument that metaphysics involves offering explanations by postulation; and, so far as I can tell, apart from the sorts of explanatory arguments already in the literature, there is no argument forthcoming for the conclusion that selecting theories on the basis of explanatory virtues is, in general, a reliable way of reaching

    

33

the truth. And, of course, explanatory arguments will be of no use in persuading an empiricist to take metaphysics seriously; for those are precisely the sorts of arguments of which empiricists are most suspicious. From within the empirical stance, then, it is hard to find any resources for answering van Fraassen’s objections. On the other hand, the objections are readily answerable from within the metaphysical stance. Indeed, those who adopt the metaphysical stance may simply adapt the reply I offered to Byrne. We know that some sources of evidence have to be taken as basic. Even the empiricist must acknowledge as much for, lest she fall into global scepticism, she will be forced to take at least sense perception and logical reasoning as basic sources. But then why not take the methods and sources of evidence employed by metaphysicians as basic sources as well? There are, to be sure, arguments in the literature for the conclusion that some of these sources— intuition, for example—are unreliable or unworthy of being treated as basic sources. But, notoriously, these arguments are not decisive—in no small part because (as is often pointed out) it seems impossible to run the arguments without presupposing the reliability of intuition. In any case, however, they are not the arguments that van Fraassen has offered; and to the extent that his objections depend on them, they are all the weaker for that dependency. My suggestion, then, is that van Fraassen’s scepticism about metaphysics ought to be treated in precisely the way that Byrne’s theological scepticism ought to be treated: both are rightly ignored. Matters would be different, of course, if there could be some argument for the conclusion that it is more rational to adopt the empirical stance than to adopt the metaphysical stance. Is there such an argument? More pertinently, is there any such argument that those who have already adopted the metaphysical stance will be rationally bound to accept? It is hard to see how there could be. For, after all, the two stances are distinguished in part by different views about what constitutes good argument and rational theorizing; and, unless the standards and sources characteristic of the metaphysical stance are self-defeating, there is no reason to think that those adopting the metaphysical stance ought to be persuaded by empiricist arguments for the conclusion that it is irrational to adopt the metaphysical stance. And, so far as I am aware, there is no argument forthcoming for the conclusion that the standards and sources characteristic of the metaphysical stance are self-defeating. At any rate, no such argument has been offered by Byrne or van Fraassen. Interestingly, van Fraassen does seem to have a pragmatic argument for adopting the empirical stance. The pragmatic argument is just his first objection against metaphysics: namely, that the payoff we gain from believing the theories we choose on the basis of explanatory virtues is too small to offset the risk of false belief. In fact, however, the pragmatic argument seems unsound. Granted, food,

34

   :  

shelter, and safety are not at stake in metaphysics. But other things are.¹¹ Metaphysics impinges on morality; it also impinges on our very conception of ourselves. Are we free? If not, does it follow that we are not morally responsible? Can a person existing now be the same person as one existing a thousand years from now? If not, as Derek Parfit seems to suppose, is there any point in taking steps—as both scientists and religious believers do in various different ways—to try to significantly prolong our lives? The list of questions might well go on; and to the extent that our metaphysical views do impinge upon our moral lives and upon our self-conception, they will impact our intellectual and emotional lives, our social interactions, and a variety of other aspects of life. Indeed, many religious believers have thought that one’s metaphysical beliefs—particularly one’s theological beliefs—make the difference between eternity in heaven and eternity in hell. For those who think this, the risk of false belief and reward of true belief rise exponentially. So, contrary to what van Fraassen argues, false belief in metaphysics can be a serious risk; true belief can be an important reward. And the reward of getting it right might well justify taking the risk. Van Fraassen will almost certainly be unmoved by the response just offered on behalf of metaphysics. But really that doesn’t matter. For plenty of people will be moved. That is, plenty of people will agree that the potential reward of true belief in metaphysics does outweigh the risk. And for those people, there will be ample pragmatic justification for adopting the metaphysical stance.

4. Conclusion In this chapter, I have examined two objections against realism in theology and metaphysics, and I have concluded that practitioners of metaphysics and theology ought simply to ignore these objections. In a very important sense, both objections are simply instances of ‘preaching to the choir’. Those who are already sceptical of theology or metaphysics or both will find in the objections plenty to agree with. But they should not convince the unconvinced.¹²

¹¹ Some of the points here, particularly the point about free will, are borrowed from Alicia Finch’s discussion of van Fraassen’s objection in her review of The Empirical Stance (Finch 2003). ¹² Versions of this chapter were read at the 2006 ‘Belief and Metaphysics’ Conference, sponsored by the University of Nottingham’s Centre for Theology and Philosophy, the 2007 Pacific Regional Meeting of the Society of Christian Philosophers, and the 2007 Midwest Regional Meeting of the Evangelical Philosophical Society. I am grateful to audiences on those occasions for helpful discussion. I would also like to thank the participants in my ‘Metaphysics and Christian Theology’ seminar (especially Alex Arnold, Andrew Bailey, Adam Green, Nate King, Jenny Martin, Luke Potter, and Chris Tucker) and participants in the University of Notre Dame Center for Philosophy of Religion reading group (in particular, Robert Audi, Alicia Finch, Tom Flint, Don Howard, Alan Padgett, J. Brian Pitts, Kevin Sharpe, and Nick Trakakis) for very helpful discussion of the ideas in this chapter and for comments on an earlier draft.

    

35

References Byrne, Peter. 2003. God and Realism. Aldershot: Ashgate. Finch, Alicia. 2003. ‘Review of The Empirical Stance’. American Catholic Philosophical Quarterly 77 (2): 302–7. Fraassen, Bas van. 1980. The Scientific Image. Oxford: Clarendon Press. Fraassen, Bas van. 1985. ‘Empiricism in the Philosophy of Science’. In Images of Science: Essays on Realism and Empiricism, with a Reply from Bas C. van Fraassen, edited by Paul M. Churchland and Clifford A. Hooker, 245–308. Chicago, IL: University of Chicago Press. Fraassen, Bas van. 2002. The Empirical Stance. New Haven, CT: Yale University Press. Rea, Michael. 2002. World without Design: The Ontological Consequences of Naturalism. Oxford: Clarendon Press.

2 Theology without Idolatry or Violence Since the 1960s, metaphysics has flourished in Anglo-American philosophy. Far from wanting to avoid metaphysics, philosophers have embraced it in droves. There have been critics, to be sure; but the criticisms have received answers and the enterprise has carried on. Matters have been different outside the so-called ‘analytic’ philosophical tradition, and particularly so in theology throughout the past century. Witness, for example, Kevin Hector’s recent book, Theology without Metaphysics (Hector 2011), which takes as a starting point the idea that metaphysics is a thing to be avoided, a succubus from whose embrace we must struggle to extricate ourselves. In the opening chapter, Hector writes: Modern thought has engaged in a recurrent rebellion against metaphysics . . . . This recurrent rebellion against metaphysics indicates that although we moderns may want to avoid metaphysics, we have a hard time doing so. It would appear, in other words, that metaphysics is a kind of temptation: we want to resist it, but find it difficult to do so. (2011: 2; emphasis in original)

As it arises in theology, the temptation towards metaphysics is supposed to have its origin in our natural propensity to speak positively and substantively about God. Cataphatic theology, so the reasoning goes, is inherently metaphysical. So our propensity to engage in it constitutes a temptation towards metaphysics. In turn, the concern about metaphysics is that it results in both idolatry and violence—idolatry because it shifts our attention away from God and onto a simulacrum of our own creation, and violence because it denies the otherness of God and forces God into creaturely categories. Would-be theologians are thus offered a dilemma: apophatic theology on the one hand, idolatrous and violent theology on the other. In Theology without Metaphysics, however, Hector seeks a middle ground. He offers a broadly Wittgensteinian theory about the nature and deployment of human concepts and predicates with the goal of showing how both can be applied to God in a non-metaphysical way. In this way, he hopes to show that cataphatic theology is not inherently metaphysical, and that one can therefore engage in it without falling into idolatry or violence. In what follows, I will argue that Hector has not succeeded in forging a path between the horns of the aforementioned dilemma. I shall begin by attempting to Essays in Analytic Theology: Volume I. Michael C. Rea, Oxford University Press (2020). © Michael C. Rea. DOI: 10.1093/oso/9780198866800.003.0003

    

37

identify the proper target of the ‘idolatry’ and ‘violence’ objections and to provide a clear statement of each. I will then explain Hector’s proposal and show how it is supposed to provide a non-metaphysical way of doing substantive, cataphatic theology. In the third section, I will highlight five difficulties that beset Hector’s view.

1. Objectionable Metaphysics Despite his provocative title, Hector is explicit about the fact that he does not mean to argue that theology can or should be done without metaphysics of any sort. Rather, his aim is simply show how theology can be done without falling into metaphysics of a particular objectionable sort. Just what sort is that? There are several ways in which one might distinguish metaphysics from other forms of theorising. One might characterise it by its subject matter, its methods, its presuppositions, or some combination of these. Hector and his interlocutors (to some extent following Heidegger) seem to treat objectionable metaphysics as a mode of inquiry defined primarily by its guiding assumptions. Hector identifies objectionable metaphysics with what he calls essentialistcorrespondentist metaphysics (ECM for short). ECM, according to Hector, is . . . the attempt to secure human knowledge by identifying the fundamental reality of objects—their being as such—with our ideas about them . . . . [W]hat sets [it] apart [from other forms of metaphysics] is precisely an understanding of the being of beings—their essence—as that which must correspond to the ideas of a human knower. (2011: 8)

Note that it is one thing to say that the essences of material objects do or must correspond to the ideas of human knowers, and quite another to identify those essences with human ideas. Since the ‘correspondence’ requirement is plausibly entailed by the ‘identity’ requirement, it is perhaps best to suppose that it is the correspondence requirement at which he intends to take aim. Doing so provides Hector with a more expansive target. Thus, we might say that objectionable metaphysics is just ECM, construed as the view that the being, or essence, of any object that falls within the purview of our theorising corresponds (in some sense) to human ideas. I am not convinced, however, that ECM is the proper target of the idolatry and violence objections as Hector characterises them; nor am I convinced that it is the characterisation of metaphysics that those who raise such objections themselves have in mind. So I would like to suggest an alternative characterisation. For Marion and Heidegger, both of whom loom large in Hector’s text, metaphysics involves putting the ‘being of beings’—i.e. the being of things in the world,

38

   :  

as opposed to God, or the ultimate ground of being—on a par with Being itself, and assuming that the latter grounds but can be accounted for or explained in terms of the former.¹ The idea that Being itself can be explained in terms of the ‘being of beings’ amounts, in practice, to the supposition that human concepts— concepts shaped by the experience of beings in the empirical world—can be used to characterise God.² In a similar vein, Levinas, who does not feature prominently in Hector’s discussion, but who is nevertheless strongly associated with the idea that metaphysics is violent,³ takes metaphysics to be the ‘promotion of the Same before the other, the reduction of the other to the Same’.⁴ The Same, for Levinas, is the thinking subject; so the idea here is that metaphysics presumes (objectionably) that the objects of our study are relevantly just like us—that they can be classified according to our own category system, understood in terms of our own familiar concepts acquired by way of our own perspectivally conditioned experience of the world. My own inclination, then, would be to characterise ‘objectionable metaphysics’ as follows. To engage in objectionable metaphysics is to conduct inquiry in a kind of ‘self-centred’ mode, one which tacitly privileges one’s own conceptual scheme and cognitive capacities as the standard by which the world is to be understood. It is, in particular, to approach a phenomenon that we wish to study or understand as if the following three claims are true of it. First, it is a being or object like me and like other things with which I am acquainted. Second, it is similar enough to me and these other things to be understood in terms of the same fundamental categories (univocally applied). Third, to whatever extent I fail to understand it, the failure is not due to defects or limitations in such concepts as I have or could acquire by further experience but rather to my failure to have a sufficiently wide range of experiences to have the concepts I need in order adequately to grasp the phenomenon I am studying. (I will generally speak of these three claims as ¹ There are various technical senses of ‘world’ in Heidegger, but I do not mean to invoke any of those here (cf. Heidegger 1999, part 2). As to why metaphysics involves putting the being of beings on a par with Being and assuming that the latter can be explained in terms of the former, the reason in short is as follows: Metaphysics is concerned both with the ‘being of beings’, but also with their ground—the highest being, Being, which Heidegger identifies with the ‘god of philosophy’ (Heidegger 2002: 70–2). But insofar as metaphysics seeks to ‘represent beings as such’, it does so ‘with an eye to their most universal traits’ and ‘only with an eye to that aspect of them that has already manifested itself in being’ (Heidegger 1998: 287, 288). Here, then, we may identify a guiding assumption: to investigate something in metaphysical mode is to do so under the supposition that its very essence can be understood in terms of universal characteristics that have been made manifest in beings—i.e. mundane things. But the ultimate ground, Being, is not a mundane thing; and so the supposition that it can be understood via concepts and categories crafted for understanding beings is suspect. Thus one finds Heidegger speaking of overcoming metaphysics, a goal that is accomplished just when one manages to ‘think the truth of Being’ (1998: 279). Cf. also Marion 1994. ² On the relation between Being and God, see (e.g.) Heidegger 2002, 70–2; Heidegger 1998; and Marion 1994. For both, Being understood metaphysically—the god of philosophy, or of ontotheology— is but an idol. But it is an idol often enough confused with God. ³ But see note 5 below. ⁴ Kosky 2001: 9; cf. also Levinas 2005: 43, 45–6.

    

39

‘presuppositions’ involved in objectionable metaphysics, though I do not really mean to suggest that they are always or even typically explicitly or self-consciously assumed.) The violence and idolatry objections, as I understand them, apply most saliently to an approach towards God which treats God as just another object of inquiry to which human concepts can univocally be applied. Let us begin with the idolatry objection. Idolatry for Marion is not the creation or worship of religious idols. It is a broader concept. To fall into idolatry is to direct one’s thought towards a simulacrum of the phenomenon about which one aims to think or theorise rather than towards the phenomenon itself. This happens when one fails to allow things to ‘give themselves’ in their own way, or to appear on their own terms, presuming (tacitly, via our intention to locate them within our own conceptual schemes) that they will appear only in ways that conform to our concepts. Hence Tamsin Jones’s characterisation of idolatry in Marion: [Idolatry] is the constraining of any phenomenon within limits alien to the way it gives itself or shows itself. Defining the phenomenon according to one’s own subjective conceptual limitations is . . . idolatrous. (2011: 9)

Accordingly, Hector construes idolatry (of the relevant sort) to involve the subjection of God to human conditions for the experience of the divine; and he takes Marion to think that concepts are themselves a kind of ‘human condition’ (2011: 16). But if that is right, then it looks as if idolatry will be a danger anytime we take it for granted that God can be understood within the confines of human conceptual schemes. The proper target of the idolatry objection, then, is not ECM, but rather objectionable metaphysics as I have construed it—most saliently (though not exclusively), the view that concepts apply to God. Now to the violence objection. Hector characterises it in two separate places: once in discussing the work of Heidegger, and again in discussing the work of Caputo. As Hector understands Heidegger, the concern is that metaphysics (ECM in particular) ends up equating an object’s fundamental reality with that which fits within the bounds of [human] categories. The danger is obvious: if one thinks that one’s preconceived ideas correspond to an object’s fundamental reality, one may be tempted to force the object to fit one’s conception of it, whether because one fails to see anything beyond one’s conception or, worse, because one tries to make it conform to that conception. (2011: 11)

Later, characterising Caputo, he writes: Language is violent, according to Caputo, in as much as it seeks to fit objects within its horizon, to pin them down, and to hold them within its grasp. This

40

   :   being the case, Caputo reasons that ‘there really is nothing we can say about God that is not violent in the sense that it does not cast God in certain terms, that it does not subject God to a certain horizontality, and so set up something anterior to God, with a kind of ontological violence.’ (2011: 20–1)

(The notion of ‘ontological violence’ in play is left unexplained in both Hector’s text and Caputo’s, but traces to Levinas. More on this below.) Thus, the core of the violence objection, as Hector sees it, is this: applying concepts to God is violent because it ‘cuts God down’ to creaturely size, force-fitting God into ‘antecedently defined’ human categories (2011: 49). The objection thus construed seems to depend on an understanding of concepts according to which nothing can strictly and literally satisfy a concept F unless it exactly resembles, with respect to its F-ness, some other creature that strictly and literally satisfies F (cf. Hector 2011: 49–50). One way, but not the only way, of motivating this idea is to suppose that concepts are just human ideas which arise out of experience, and that the extensions of concepts are classes of objects which exactly resemble in some particular respect other objects which lie within our experience. If this sort of view is right, then to say (for example) that God is wise is to say that God is at best paradigmatically wise in the creaturely way. It is surely hyperbolic to say that this violently ‘cuts God down’ to creaturely size; but it is easy enough to see that it would at least result in a distorted vision of God. Thus far the violence objection, as Hector characterises it. But I am not sure that this fully captures the fundamental worry that the objectors have in mind. Here I think it is illuminating to consider the way in which Levinas (who is not among Hector’s primary interlocutors) associates metaphysics with violence. For Levinas, as we have seen, metaphysics is ‘the promotion of the Same before the other, the reduction of the other to the Same’.⁵ It is, in other words, a tacit denial of the otherness of the other, a failure to allow the other to ‘appear’ on his/her/its own terms. In less colourful and evocative language: one is engaged in metaphysics to the extent that one takes as a methodological starting point the idea that the others one encounters are similar enough to oneself that they can be accurately understood and characterised in terms of one’s own concepts and categories, without regard for the possibility that they might in fact transcend those categories or somehow otherwise elude characterisation within one’s own familiar conceptual scheme. In doing this, one (conceptually speaking) forces the other into a pre-cut mould; and this is where the association with violence comes in. Violence, for Levinas, ‘does not consist so much in injuring and annihilating persons as in interrupting their continuity, making them play roles in which they ⁵ For purposes here, I conflate metaphysics and ontology. The distinction matters to Levinas (2005: 42ff.); but what I am calling objectionable metaphysics is appropriately assimilated to what Levinas calls ontology (and to what I think even he would describe as objectionable metaphysics).

    

41

no longer recognize themselves, making them betray not only commitments but their own substance, making them carry out actions that will destroy every possibility for action’ (2005: 21). One might, of course, challenge this characterisation of violence; but, taking it as read, it becomes easy to see what the connection between violence and metaphysics is supposed to be. In metaphysics, one denies the alterity of the other; one ‘tries to integrate the other into [one’s] project of existing as a function, means, or meaning’ (Burggraeve 1999: 30). In so doing, one thereby risks ‘reduc[ing] the other to his countenance’, which, in turn, risks making the other play a role in which she no longer recognises herself. In this way, the ‘reduction of the other to the Same’ amounts to an exercise of power over the other. As Burggraeve puts it: In [reducing the other to myself] I approach the other not according to his otherness itself, but from a horizon or another totality . . . I look the individuality of the other, so to speak, up and down, forming a conception of him not as thisindividual-here-and-now but only according to the generality of a type, an a priori idea, or an essence . . . . The ‘comprehending’ I, or ego, negates the irreducible uniqueness of the other and tries to conceive of him in the same way as he does the world. Comprehensive knowledge is thus also no innocent phenomenon but a violent phenomenon of power. By my ‘penetrating insight’ I gain not only access to the other, but also power over him. (1999: 36)

These remarks apply all the more strongly when the relevant other is God, who is supposed to be radically other. Caputo seems to have something quite similar in mind. Consider, for example, the following excerpt from the same essay from which Hector draws: You see the idolatrous functioning of the metaphysical concept: the concept seizes God round about, measures the divine by humanly comprehensible standards, holds the look of the mind’s eye captive, and cuts off the infinite incomprehensible depths of God. Lacking infinite depths, the metaphysical look is accordingly not sent off into the distance but is reflected back onto itself. A metaphysical concept of God, let us say that of the causa sui, is an image of the metaphysician. It is not inspired but constructed, not infinite but finite, not an excess but an incision into the divine. (Caputo 1992:132)

The concern about idolatry is in the foreground; but Caputo’s broader concern is violence, and the key ideas in this passage resonate strongly with those in the quotations from Levinas and Burggraeve in the previous paragraph. Note, however, that the differences between these various ways of casting the objection pertain not to the target (which is, again, simply the presupposition that God can be understood in terms of human concepts) but rather to the reasons for

42

   :  

thinking that metaphysics treats the objects of our theorising violently. Moreover, it appears that the Levinasian reason for thinking that metaphysics is violent is quite similar to Marion’s reason for thinking that it is idolatrous. In both cases, the fundamental problem is that the metaphysical presupposition occludes the (perhaps quite radical) otherness of our objects of study. So the violence and idolatry objections, as I understand them, are mainly directed not at ECM but at objectionable metaphysics as I have characterised it; and they apply most saliently to the supposition that human concepts apply (univocally) to God. Assuming this is what puts us at risk of shifting our attention away from God and onto a simulacrum, it is what most clearly manifests our tacit denial of the radical otherness of God, and it is what constitutes our attempt to ‘force’ God to fit human categories. Responding to the objections, then, will involve establishing each of the following theses: (1) Assuming that our concepts apply to God does not risk shifting our attention to a simulacrum. (2) Assuming that our concepts apply to God does not deny the radical otherness of God. (3) Assuming that our concepts apply to God does not ‘force’ God to fit into human concepts and categories. Although Hector does not cast his project in these terms, his philosophy of language is crafted so as to guarantee each of these theses. In the next two sections, I shall summarize the main components of his view, and then highlight what I take to be its central problems.

2. Hector’s Proposal Hector’s main goal is to show that language is non-metaphysical in the following sense: concept application does not force objects into pre-determined creaturely categories, nor does it presuppose that concepts express or correspond to the essences of the things that satisfy them. Key to accomplishing this goal is the deployment of a robust theory of concept use, meaning, reference, and truth evaluation. Here I shall focus primarily on what Hector has to say about concept use, but I shall begin by laying out the main components of the overall theory. Hector summarizes his philosophy of language helpfully as follows: I propose a[n] . . . account of language, according to which (a) to use a concept is to intend one’s usage as going on in the same way as certain precedents and to claim this same precedent status for one’s own usage; (b) the meaning of a

    

43

concept is a product of the normative trajectory implicit in a series of such precedents, which entails that a concept’s meaning changes, if only slightly, each time a candidate use is recognized; (c) to refer to an object is to link up with a chain of precedents that carries on the normative commitment implicit in an initial picking-out, in such a way that one inherits (and renders further inheritable) that commitment; and (d) to judge some proposition to be true is to see it as going on in the same way as one’s other commitments and to use it to judge still other propositions. One can thus arrive at an account of concepts that do not “contain,” of meaning without “meanings,” of reference without “presence,” and of truth without “correspondentism.” We arrive, in other words, at an account according to which language might be fit for God-talk. (2011: 38)

Note that, thus far, we have only the claim that language might be fit for God-talk. To show that it is fit for God-talk—that theology can be done without idolatry or violence—Hector adds a further component to his theory. This further component is a story about how ordinary concepts come to apply to God; and central to this story is a view about the work of the Holy Spirit. Drawing on the work of Schleiermacher, Hector develops a view according to which the Spirit of Christ works so as to shape the meanings of our concepts in a way that makes them applicable to God. The details of this theory are complex; but for present purposes the following rough summary should suffice. According to Hector, Jesus taught his disciples how to follow him; he also recognised them as competent at doing so and as competent judges of what counts as following him. They, in turn, did likewise for others, and so on down through the history of the Church. This is important because, by virtue of his own divinity, Christ himself faced no problem in applying concepts to God; he was able to ensure that the ‘normative trajectories’ implicit in his own use of the concepts he applied to God had God as their ‘fulfilment’. (Whatever else it might imply, having God as the ‘normative fulfilment’ of a concept guarantees that the concept is applicable to God.) Furthermore, Hector thinks, learning to follow Christ is, in part, a matter of learning what it is to ‘go on in the same way as Christ’ in one’s own God-talk. Learning to ‘go on in the same way’ as someone in one’s own use of concepts, in turn, involves becoming attuned to the normative trajectories implicit in the ways the other person uses her concepts. It is this sort of attunement that, with the aid of the Holy Spirit, has been passed down by and through the Church so that we twenty-first-century theologians can go on in the same way as Christ and can reliably recognise others as doing so. Thus, Christ’s own ways of using concepts are among the normatively authoritative precedent uses for our own, and they are linked to our own God-talk via a chain of further precedents, all of which go on in the same way as Christ’s. So, in short, we can apply concepts to God because Christ’s concepts applied to God, and at least some of the concepts that we apply to God are continuous with

44

   :  

Christ’s. This does not tell us just how many of our concepts apply to God; and, as we shall see, real sceptical monsters lurk in the neighbourhood. But I shall leave that concern to be developed later. The question now is why exactly this view allows us to say that concepts apply to God without falling prey to the idolatry and violence objections as they were characterised earlier. Hector takes the answer to lie in part (b) of the executive summary of his position: the meaning of a concept is a product of the normative trajectory implicit in a series of such precedents, which entails that a concept’s meaning changes, if only slightly, each time a candidate use is recognized.

One might question the entailment claim: is it really impossible for there to be, over time, a perfectly stable normative trajectory in the usage of some concept? Has, for example, the meaning of the concept of pi changed in (say) the last twenty minutes as the concept has been introduced or employed in a myriad elementary school classrooms? But never mind this for now. What is important is just the claim that concept meaning does change over time with use, and that meaning is a function of the concept’s inherently unstable and therefore not entirely predictable normative trajectory. The result is that concepts are open-ended and vaguely bounded in their extensions. In other (perhaps more familiar) terms: concepts are vulnerable to semantic drift, so much so that it is indeterminate to what objects they might eventually apply. From this, then, it is supposed to follow that concepts do not necessarily correspond to fixed, unchanging essences of things. Nor do they necessarily correspond to or constitute ‘predetermined’ or even well-defined categories. According to Hector, . . . if a concept is the product of a series of precedent uses, and if this series changes every time a new use is recognized as carrying it on, it follows that concepts are continually being reconstituted. This means . . . that it makes no sense to talk about the meaning of a concept being fixed by its application to creaturely reality . . . (2011: 97)

This is partly because it makes little sense to talk about ‘the meaning’ of a concept, and partly because concept meanings are not fixed. Furthermore, it follows on this view that applying a concept to God does not imply that God exactly resembles any creaturely paradigm; for, on this picture, the extension of a concept is not determined by anything like exact resemblance to paradigms. Thus, according to Hector, applying concepts to God no longer threatens to ‘cut God down to size’, forcing God into merely creaturely categories. Nor is it plausible to think of concept application on this view as contributing essences to things. After all, if

    

45

what it means to say (for example) that God is wise might change over time, it is hard to see why one should think that calling God wise makes God into something which is essentially (and thus unchangeably) characterisable by the present content of the concept wisdom. Likewise, Hector might say, once we appreciate the fact that concepts work in this way, concept application does not even threaten to distort our vision of God; nor does it threaten to ‘make [God] to play a role in which [God] no longer recognizes [God-self]’; nor does it ‘reduce [God] to [God’s] countenance’. For it will be built into the very idea of applying concepts to God that much is being left out and that the sands are always vulnerable to some shifting. Thus the violence objection is addressed. Nor does it ‘subject God to human conditions’. According to Hector, the norms by which one assesses one’s [theological] concept use are not external to God—and neither are the concepts themselves. On the present account, one learns how to use theological concepts by submitting one’s performances to Christ, and one is able to submit them to Christ through the power of the Spirit . . . . One conceptualizes God through God, which is to say that one conceptualizes God by grace alone. (2011: 96)

This, then, does away with the concern about idolatry.

3. Problems As noted earlier, there are further components to Hector’s overall theory that merit close attention and serious discussion. But I think that the material thus far presented provides a clear enough picture of where the view is going for us to see some of the problems it will face. In this section, I shall discuss five problems, each of which strikes me as rather serious. Let us begin by noting that one possibility not at all ruled out by Hector’s theory is that there are, after all, sharply bounded (non-vague) properties which ‘correspond to’ or serve as the content of our concepts, and that the instability of our concepts over time has simply to do with the fact that concepts take different properties as their precise contents at different times, depending on vagaries of usage.⁶ Indeed, we might even suppose that there are determinable properties (like being wise) that always serve as the ‘normative fulfilment’ of the concepts (like ‘wisdom’) to which they correspond, but that the precise content of the concept at

⁶ This general picture is built into the epistemic theory of vagueness, according to which the vagueness of a term or concept (like bald) is simply a result of our ignorance of the sharp boundaries on the property to which it refers. For a thorough and classic defence of epistemicism, see Williamson 1996.

46

   :  

any given time is some determinate of the determinable (e.g. Socratic wisdom, or creaturely wisdom, or some other more specific sort of wisdom). On this picture, everything that Hector says about how concept meaning depends on ‘going on in the same way’ as precedent uses, and about the instability of concept meaning, might still be true; but it will nonetheless be a picture according to which concepts, despite their instability, may still correspond to fixed properties, or categories. Now consider this fact in light of Hector’s stated goal of ‘free[ing] us from the metaphysical picture of concepts, according to which one’s use of a concept corresponds to an essence-like idea or “meaning” and so [also] free[ing] us from the sense that concept use fits objects into pre-determined categories’ (2011: 48). What is a predetermined category? One answer is that it is nothing more or less than a sharply bounded property that objects might or might not objectively, mind-independently exemplify. But if this is right, then by not ruling out the possibility described in the previous paragraph, Hector’s view fails to free us from the problematic ‘metaphysical picture of concepts’. Concepts will still correspond to determinate properties, and so, in applying them to objects, we will still be sorting things into predetermined categories. Note, too, that they will still be characterisable as familiar, human categories; for the properties to which our concepts correspond would, in order to secure the correspondence, have to be humanly graspable. Moreover, if predetermined categories are nothing other than mindindependent, non-vague properties, Hector’s talk of ‘freeing us’ from the metaphysical picture of concepts does not make sense. For to say that objects objectively and mind-independently exemplify certain properties is, among other things, to say that human conceptual activity is not part of what makes it the case that they exemplify those properties. Thus, simply by virtue of believing in mindindependent property exemplification we are already freed from a picture according to which human concept usage fits objects into predetermined categories. Objects either fit into those categories or not under their own steam, as it were. Human concept usage does not generate the facts about how objects fit into categories; it simply tries to represent those facts. Another alternative—and the more plausible one, in light of what has just been said—is that the predetermined categories that Hector has in mind are minddependent essences. Suppose, for example, that being a statue is not an objective property of a gold statue, but rather depends in some way upon our thinking of it as a statue rather than (say) as a mere lump of gold. Hector’s view, of course, implies that the concept ‘statue’ might shift in meaning over time; so the view allows for the fact that what counts now as a statue might not count in the future (or might not have counted in the past) as such. But it doesn’t follow from this that applying the concept now to this particular piece of gold doesn’t fit the piece of gold into a predetermined category. Indeed, on this picture, our thinking of the piece of gold as a statue (rather than a mere piece of gold) is part of what makes it a

    

47

statue, and it is part of what gives the statue persistence conditions like being unable to survive being recast in the shape of a cube that are not otherwise had by a mere piece of gold. So here too it seems that Hector’s theory of concepts fails to achieve his goals. Second problem: Suppose we grant that Hector’s theory succeeds at showing us how concepts (in general) can apply to God. Still, it is not clear that the theory shows us how ordinary concepts can apply to God. In showing how we can use his model to illuminate theological concept use, Hector writes: . . . one’s concepts are applied correctly to God . . . if they are faithful to God’s revelation in Christ, and they count as such if they go on in the same way as precedent applications which carry on the normative trajectory that stretches back to Christ . . . . To count as using (Christian) theological concepts . . . one’s use must be recognizable as trying to carry on the normative trajectory implicit in this chain [that stretches back to Christ]. (2011: 95)

This is all fair enough; but the thing to notice is that the criterion for the use of Christian theological concepts is not a criterion for the use of ordinary concepts. Thus, unless it can be shown that, as a matter of contingent historical fact the normative trajectory implicit in the use of some ordinary concept C is recognisably identical to the normative trajectory implicit in some chain of precedent uses that stretches back to Christ, we have no reason for thinking that C is a ‘Christian theological concept’, and so no reason for thinking that C properly applies to (the Christian) God.⁷ To illustrate, consider the concept ‘wisdom’ as it is used in the predication ‘God is wise’. Suppose we grant that Christians who say that God is wise are recognisably trying to carry on a normative trajectory implicit in a chain of precedent uses of the term ‘wise’ that stretches back to Christ. Suppose further that we grant that, by virtue of this, ‘wisdom’ applies to God. Still, it does not follow that any ordinary concept applies to God; for it does not follow that the theological concept ‘wisdom’ is identical to (or even intelligibly related to) the ordinary concept that goes by the same name. To show that wisdomtheological = wisdomordinary, we would have to show what Hector has not: namely, that the normative trajectory implicit in nontheological uses of the concept ‘wisdom’ is recognisably the same as the normative trajectory implicit in the chain stretching back to Christ of precedent uses of the theological concept. I do not know how one could show this; but without showing this, it seems that Hector’s philosophy of language inadvertently gives victory to ⁷ Note that the point here does not depend on interpreting ‘going on in the same way’ as equivalent to ‘going on in the same identical way’. Hector has informed me (personal conversation) that he intends ‘same’ in this context to be understood more with the sense of ‘similar’ than with the sense of ‘identical’. Nevertheless, even granting this, it seems that the non-theological concept of wisdom is identical to the theological concept of wisdom only if the normative trajectory of the one is identical to that of the other.

48

   :  

the apophatic theologians. Ordinary concepts do not apply to God. Rather, we can speak of God only in a special theological language, employing concepts that, for all we can tell, may not even be intelligibly related to our ordinary ones.⁸ Third problem: Thus far I have been assuming that Hector’s theory of concept usage is viable, at least in broad outline. But this too is questionable. For Hector, concept usage depends upon judgements to the effect that one concept is the same as another; it also depends upon the existence and recognition as such of precedent uses. Thus: To use a concept is to recognize the concept use of others as having a certain normative authority over one, and to claim this same status for one’s own use. To count as using a particular concept, one’s usage must be recognizable as using the same concept as others, which is to say that it must go on in the same way as uses that are recognized as precedential—that is, as uses of the concept in question. (2011: 62)

But it is easy to see why the requirement that there be recognised precedent uses will pose a problem. Assuming one has to use a concept in order to introduce it, the requirement implies that concepts can never be introduced. It also implies (absurdly) that behind every instance of concept use is an infinite series of precedent uses (since any precedent use will count as the use of a concept only if it goes on in the same way as some precedent). Of course, one can fix this problem by stipulating that one can also use concepts without precedent at the moment of introduction. But then one faces a dilemma: either concept introduction is always violent (and theological concept introduction always idolatrous) or else Hector’s elaborate story about how concept-use depends on normativetrajectories implicit in chains of precedent uses is ultimately irrelevant to thwarting the idolatry and violence objections. But what of the claim that concept-usage depends upon judgements to the effect that one concept is the same as another? Here we face a dilemma. Either the judgements must all be explicit, or not. Suppose they must all be explicit. Then we face a vicious regress. Why? Because making the judgement ‘A is the same concept as B’ involves applying the concept ‘sameness’ to A and B. In order to apply that concept to A and B, however, I have to judge that my concept ‘sameness’ is the same concept as one involved in precedent applications of the term ‘sameness’. To make this judgement, however, I have to use the (or, better, a) sameness concept again; and to do that, I have to judge it to be the same as precedent uses, thus using another sameness concept, and so on.

⁸ Note that semantic drift together with the bifurcation of Christendom into different theological communities means that we might not even be using the same theological concepts as other Christians.

    

49

Better, then, to say that the judgements need not be explicit. And, indeed, this is what Hector does seem to say (2011: 64ff.). Hector speaks of the ‘implicit recognition’ that one instance of concept usage goes on in the same way as others, and he also seems to want to allow that one can ‘implicitly try’ to go on in the same way as others (which is good, because otherwise trying to go on in the same way as others would generate the same regress of concept application) (2011: 81–2). But now we face in spades the problem of determining when a person (or even we ourselves) intends to go on in the same way as someone else. Suppose I say that God is wise, but I am neither explicitly (i.e. consciously) judging myself to go on in the same way as Christ nor explicitly intending to go on in the same way as Christ. We might ask what makes it the case that I am doing these things. But even if we waive that worry, Hector’s conditions for concept-usage require not only that I be doing them, but also that I be recognisably doing them. But if I have not even acknowledged to myself, so to speak, the relevant judgements or intentions, how could others recognise them in me? The cases where I am implicitly trying to go on in the same way as Christ might be indistinguishable in practice from the cases where I am implicitly trying to go on in the same way as someone else who is working with a subtly but genuinely different concept of wisdom. Fourth problem: According to Hector, a candidate performance counts as going on in the same way as precedent performances if and only if it is recognisable as such by those who know how to undertake such performances (2011: 68). But we might well ask what ‘recognisable’ means. If the idea is that one performance goes on in the same way as another if and only if someone who knows how to undertake such performances is capable of forming the belief that it goes on in the same way, then concepts are a lot more open-ended than Hector has led us to believe. Indeed, it would seem that anything goes; for, after all, given world enough and time, together with the fact that those who know how to undertake various kinds of performances are always capable of forming deviant or mistaken beliefs about what goes on in the same way as precedent performances, it would seem that any concept is capable of morphing in virtually unlimited ways. On the other hand, the idea might simply be that there are fixed (and recognition-independent) standards for ‘going on in the same way’ as precedent performances, and that satisfying these standards somehow guarantees that people who know how to undertake the relevant performances can see whether the standards are satisfied. If this is right, however, then every concept will have fixed boundaries after all. For present uses of a concept count as such, on this view, only if they go on in the same way as all of their precedents. But if concepts have fixed boundaries, then they start to look like fixed, predetermined categories after all. Finally, a fifth problem: Consider the concept of using (or applying) a concept. Marion and Caputo have a concept (call it ‘C1’) of using or applying a concept according to which applying a concept to God is idolatrous and violent. Hector has a concept (call it ‘C2’) of using or applying a concept according to which it is

50

   :  

not the case that using or applying a concept to God is idolatrous or violent. However, if Hector is to be taken as responding to Marion and Caputo rather than simply talking past them, C2 must be identical to C1. Or, at the very least, the two concepts must overlap significantly in their meaning. But in order for that to be the case, Hector must be ‘going on in the same way’ as Marion and Caputo; and it is not immediately evident that he is. Admittedly, neither is it evident that he is not. The term ‘going on in the same way as’ is rather vague. He is clearly not going on in a way identical to that in which Marion and Caputo are using the term. But, as discussed in note 7, ‘going on in the same way’ in fact requires only that he be going on in a sufficiently similar way. How similar is sufficient? It is hard to say; but there is at least some evidence that Hector’s way of going on is not sufficiently similar to Marion’s and Caputo’s. For one thing, we know that his concept of a concept is different from theirs. Hector’s idea of concepts, elaborated so thoroughly in his book, is intended as a rival to the idea of concepts that gives rise to the idolatry and violence objections. Insofar as Hector has also developed his own theory of concept usage (also intended to contribute to defusing the objections), we have good reason to think that his concept of using a concept—and so his concept of applying a concept—is different from theirs. If that is right, then we should probably deny that Hector intends to go on in the same (relevantly similar) way as Marion and Caputo. Indeed, it seems clear that he means to urge all of us to go on in a relevantly different way from them in our own usage of words like ‘concept’ and ‘apply a concept’. If I am right about this, then, by the terms of his own theory, Hector is not so much responding to Marion and Caputo as he is simply talking past them.

4. Conclusion I have argued in this chapter that, despite its ingenuity, Hector’s theory of concepts and concept application is not adequate to the task of rescuing cataphatic theology from the idolatry and violence objections. Assuming those objections do not fail for other reasons, the dilemma thus remains: apophatic theology on the one hand, or idolatrous and violent theology on the other. Despite what I have argued in this chapter, however, my broader sympathies do lie with Hector and other cataphatic theologians. I agree that the dilemma is a false one; I believe that cataphatic theology can be done without idolatry or violence. This is not the place to develop my own response to those objections in detail. However, I would like to close with a few very brief remarks reporting my own view of where the objections go wrong. Both objections turn on the idea that applying concepts to worldly phenomena somehow distorts our vision, so much so that it is no longer the phenomena themselves that we have in view but rather idols, and so much so that we do a kind

    

51

of violence to the phenomena by reducing them to what is familiar and humanly graspable, and by privileging our own particular perspective on them, rather than allowing them to appear on their own terms and attending to them in ways that respect their alterity.⁹ The idea seems to be that, in providing what we take to be an objective theoretical description of something in terms that are familiar and intelligible, we will somehow inevitably take ourselves (incorrectly) to have a full, complete, and perfectly accurate grasp of something that in fact can never be fully, completely, or perfectly accurately grasped; or, worse, we will inevitably make it the case that the thing is as we conceive of it rather than simply allowing it to be what it is in and of itself. The latter, of course, is a danger only if the anti-realists are correct and human conceptual activity plays a role in constructing the world.¹⁰ To the extent that the idolatry and violence objections depend on this sort of anti-realism, I think that that is where they go wrong. The answer, then, is not to reconceive the nature of human concepts and human concept application as Hector tries to do, but rather to reject the underlying picture of how concepts relate to objects and their properties. As to the former—the idea that concept application somehow presupposes that our grasp of worldly phenomena is more complete and accurate than it ever could be—I think that the answer is again not to reconceive the very nature of human conceptualisation but rather to cultivate an appreciation for the richness of the properties to which many of our concepts refer. Heidegger says (with many others in his wake) that ‘the difficulty lies in language’ (2002: 73). My own view, on the other hand, is that such problems as there are in this neighbourhood lie more in our own contingent attitudes towards the objects about which we theorise and the things that we say about those objects. To say that God is wise, for example, is (as I see it) to say no more or less than that God has a certain rather familiar property—wisdom—a property also had by many human beings. In saying this, we might presume many further things: that wisdom is a universal, God is a particular, and God exemplifies wisdom; that divine wisdom has exactly the same qualitative character as human wisdom; that our grasp of wisdom as we experience it in humans is so thorough as to allow us to deduce many other things about God from the simple thesis that God is wise; that wisdom is part of God’s essence, and that in grasping wisdom (and maybe a few other familiar properties) we have a thorough and secure understanding of the totality of the divine essence, and so on. We might presume these things, but we need not do so. All of these claims and more are up for grabs as we theorise further about what we mean when we say that God is wise, about the extent to which that claim can be thought of as literal or ⁹ On the privileging of our perspective, cf. Heidegger 1999: 63–4. ¹⁰ It is perhaps worth noting that Heidegger, at any rate, seems to reject this brand of anti-realism (1999: 63).

52

   :  

metaphorical, and so on. Divine wisdom must bear some intelligible relation to human wisdom in order for the predication to be apt; but we might well be ignorant of exactly what that relation amounts to. So long as we steadfastly maintain this understanding of the limits on what it means to say things like ‘God is wise’, it is hard to see how the idolatry and violence objections can have any real purchase. So, in the end, I think that we can have theology without idolatry or violence. My difference with Hector simply concerns the reason why.¹¹

References Burggraeve, Roger. 1999. ‘Violence and the Vulnerable Face of the Other: The Vision of Emmanuel Levinas on Moral Evil and Our Responsibility’. Journal of Social Philosophy 30: 29–45. Caputo, John D. 1992. ‘How to Avoid Speaking of God: The Violence of Natural Theology’. In Prospects for Natural Theology, edited by Eugene Thomas Long, 128–50. Washington, DC: Catholic University of America Press. Hector, Kevin. 2011. Theology without Metaphysics: God, Language and the Spirit of Recognition. Cambridge: Cambridge University Press. Heidegger, Martin. 1998. ‘Introduction to “What Is Metaphysics?” ’ In Pathmarks, edited by William McNeill. Cambridge: Cambridge University Press. Heidegger, Martin. 1999. Ontology : The Hermeneutics of Facticity. Bloomington: Indiana University Press. Heidegger, Martin. 2002. Identity and Difference. Translated by Joan Stambaugh. Chicago, IL: University of Chicago Press. Jones, Tamsin. 2011. A Genealogy of Marion’s Philosophy of Religion: Apparent Darkness. Bloomington: Indiana University Press. Kosky, Jeffrey L. 2001. Levinas and the Philosophy of Religion. Bloomington: Indiana University Press. Levinas, Emmanuel. 2005. Totality and Infinity: An Essay on Exteriority. Translated by Alphonso Lingis. Pittsburgh, PA: Duquesne University Press. Marion, Jean-Luc. 1994. ‘Metaphysics and Phenomenology: A Relief for Theology’. Translated by Thomas A. Carlson. Critical Inquiry 20: 572–91. Williamson, Timothy. 1996. Vagueness. New York: Routledge.

¹¹ I am grateful to Kevin Hector for helpful conversations about his book and about some of the objections presented in this chapter. Special thanks also to Matthew Halteman, Tamsin Jones, and Andrea White for their comments on portions of an earlier draft, for patiently talking with me about Heidegger, Levinas, and Marion, and for directing me to several very helpful texts.

3 Authority and Truth Discussions of the authority of scripture are commonly intertwined with discussions of the truthfulness of scripture. A casual survey of the literature on biblical authority might well convey the impression that questions about the nature and scope of biblical authority just are questions about the nature and scope of biblical truthfulness.¹ In fact, however, these issues are very different—as we can see quite easily by attention to the fact that questions about the nature and scope of parental authority (say) are very different from questions about the nature and scope of parental truthfulness. Taking a low view of a parent’s truthfulness might imply taking a low view of her authority in some particular domain. But the precise connection between authority and truthfulness in this case is hardly immediately obvious. Saying anything helpful about the connection would require us to get clear on what exactly we mean by calling a parent authoritative (no doubt distinguishing, along the way, various different ways in which a parent might be authoritative), what we mean by calling her truthful, and the like. So too, I think, with the relation between biblical authority and biblical truthfulness. This chapter is divided into three parts. In Section 1, ‘Authority’, I try to get clear about what we might mean in calling a text authoritative. In Section 2, ‘Truth’, I draw distinctions between different things that we might mean by saying that a text is truthful. My goal in both of these sections is to arrive at some general conclusions about texts, rather than specific conclusions about the Bible. Consequently, I try to refrain from making assumptions about (e.g.) biblical interpretation or about the truth of particular biblical texts. Indeed, for much of the discussion, the Bible is not even directly in view. In Section 3, ‘Authority and Truth’, I draw out some of the implications of the discussions in the first two sections for the question of how textual authority and textual truth are connected to one another. I also comment on the significance of these conclusions for discussions about the relation between biblical authority and biblical inerrancy.

¹ To see just how the impression might be conveyed, see (for example) Achtemeier (1999: 148–50), Moo 1997, and Rogers and McKim 1979. Both the title of Moo’s edited volume and the title of the book by Rogers and McKim convey the impression that each book is primarily about biblical authority; and yet the clear focus of each is, instead, the question of biblical truthfulness. In similar vein, Achtemeier moves very quickly within the space of two paragraphs from saying that ‘the Bible is authoritative . . . because it is true’ to saying that ‘the divine authority of the Bible consists in propositional truth’ (1999: 149).

Essays in Analytic Theology: Volume I. Michael C. Rea, Oxford University Press (2020). © Michael C. Rea. DOI: 10.1093/oso/9780198866800.003.0004

54

   :  

Many people seem to think that if a text is authoritative, then what it says must be true. We shall see, however, that this is not quite correct. To say simply that a text is authoritative is, in fact, to say nothing at all definitive about whether it is, or even could be, true. Nor does it guarantee freedom from falsehoods. There are, however, connections to be drawn between certain kinds of authority and reliability, or likelihood of being true. Many also seem to think that if a text is truly authoritative over a person, then she must regard it as perfectly reliable, or inerrant. (Thus, for example, one sometimes hears arguments to the effect that if one thinks that the Bible makes mistakes, then one thereby treats human sources of evidence—reason, sense perception, scientific investigation, testimony from others, etc.—as authoritative over the Bible rather than the other way around.) We shall see, however, that there is no clear way to derive the claim that a text is perfectly reliable from the claim that it is authoritative over a person unless we also make substantive philosophical assumptions about the very nature of the text’s author. What this means, then, is that a high view of biblical truthfulness—the view that the Bible is perfectly reliable, or close to it—stands or falls not with the claim that the Bible is authoritative (which a text may or may not be, regardless of how truthful it is), but rather with philosophical-theological claims about the nature of God and the divine authorship of scripture.

1. Authority Throughout this chapter, as I have indicated, I shall be concerned first and foremost with questions about the authority and truthfulness of texts. This, of course, raises the question of what counts as a text. Books, articles, contracts, and written signs are paradigm examples. But where do we draw our boundaries? Are films texts? Photographs? Symphonies? Paintings? Dance performances? Ordinary artefacts or natural objects that signify or can be taken to signify something to somebody? Though I can see the attraction of drawing the lines liberally enough to allow a very wide range of communicative acts and objects to merit the label ‘text’, for purposes here I shall simply follow Kevin Vanhoozer in defining a text as a ‘communicative action fixed by writing’.² We treat a wide variety of texts as authoritative, and we do so in different ways. In Scrabble®, the latest edition of the Official Scrabble® Dictionary is authoritative with respect to questions about which sequences of letters form admissible words and which do not. In a physics class, the assigned text is generally authoritative with respect to questions about physics. Homer’s Iliad is authoritative with respect to questions about certain matters of Greek mythology. An uncontested will is

² Vanhoozer 1998: 229. See also pp. 103–13 for a helpful survey of alternative characterizations.

  

55

authoritative with respect to questions about how a person’s assets are to be distributed after her death. The United States Constitution is authoritative with respect to questions about the permissibility of a wide variety of executive, legislative, and judicial acts, election practices, and the like. Reflection on examples like these helps to shed light on what it might mean to say that a text is authoritative. Note first that it would make little sense to say (e.g.) that the Iliad, or Jack’s last will and testament, or the Scrabble® Dictionary is authoritative, simpliciter. Each of these texts has a domain within which it is authoritative. One might think that to say that a text is authoritative simpliciter, where not even the claim’s context specifies a domain, is just to say that the text is authoritative in every domain within which a text might possibly be authoritative. If so, however, then no text— and certainly not the Bible—is authoritative simpliciter. For Jill’s last will and testament is (we may suppose) the only text that is authoritative in the domain of questions about how precisely Jill’s possessions are to be distributed after her death, and that text is authoritative in virtually no other domain. In light of this, one initial conclusion we might draw is that no ascription of authority to a text is complete without the specification of the domain within which it has authority. We may also conclude that unqualified ascriptions of authority are charitably interpreted as saying something other than that the text has authority in every domain within which a text might have authority. Note, second, that there are different kinds of authority. The authority of a physics book is primarily theoretical: it is ‘belief-guiding’ rather than actionguiding. The Official Scrabble® Dictionary and a last will and testament are partly belief-guiding, but primarily action-guiding, albeit in different ways. Thus, their authority is primarily practical. Homer’s Iliad is theoretically authoritative for us about (among other things) the contours of Greek mythology and ancient legends about the Trojan War. For the ancient Greeks it was theoretically authoritative in a different way: it provided authoritative information about (at least) the personalities and dispositions of the beings who (so it was thought) inhabited Olympus. In other words, for us the text is an authoritative source of information about mythology; for them, it was (to some degree) an authoritative source of information about the gods. It was also practically authoritative, insofar as it helped to define Greek ideals about courage and other virtues, for example.³ Saying that a text is authoritative, then, is twice incomplete: not only do we need to specify the domain within which the text is authoritative, but we need to specify the kind of authority that the text has. In the previous paragraph I indicated that there is a distinction to be drawn between practical and theoretical authority. It will be helpful in what follows to

³ Cf. MacIntyre 1998, esp. chs 1–2.

56

   :  

understand that distinction a little more fully. Roughly speaking, practical authorities provide decisive reasons for action in the domains over which they are authoritative (tennis matches, state law, morality, etc.); theoretical authorities provide decisive reasons for belief in the domains over which they are authoritative (chemistry, mathematics, etc.).⁴ Authorities are sources of information or directives. An authority might be a communicator (e.g. a speaker or institution); it might be the product of a communicative act (e.g. a text or a gesture); or it might even be something like a cognitive faculty (e.g. reason, or sensory experience).⁵ A decisive reason for action is, among other things, a reason to ignore in one’s deliberations alternative courses of action.⁶ A decisive reason for believing a proposition is, among other things, a reason to disbelieve conflicting propositions and perhaps also a reason to discontinue further investigation into the matter. Note, too, that reasons here are justifiers, not (necessarily) motivators. Having decisive reason to tell the truth, for example, is different from being motivated to tell the truth.⁷ Likewise, having decisive reason to believe that you are an inferior chess player is different from being motivated to believe this. (It might seem odd to speak of motivation and stubbornness with respect to belief-formation, but that is only because many of us think of belief-formation as an involuntary matter. I don’t want to take a position on that issue here.) This is, of course, not a full account of the distinction. It tells us only what is necessary for practical and theoretical authority, not what is sufficient. We learn, for example, that in order for A to have practical authority over B, then A’s directives must supply B with decisive reasons for action. But we don’t learn anything to the effect that whenever a certain range of conditions are satisfied, A has practical authority over B. This much is fine for our purposes, however. We have no need right now for a full analysis of either kind of authority. Still, we do need to attend to at least one complication. The complication has to do with cases in which authority is defeated. For instance: Dad tells Bart to go mow the lawn. Unbeknownst to Dad, Mom has told Bart to clean his room. Both Mom and Dad are authorities for Bart, but, given that mowing and cleaning preclude ⁴ Here I follow Mark Murphy, who writes: . . . if A is a genuine practical authority over B in some domain, then in that domain A’s telling B to ϕ is a decisive reason for B to ϕ . . . . If A has theoretical authority over B in some domain, then A must be a speaker and B must be one who can believe things for reasons; and if A tells B that it is the case that p, then A’s telling B that it is the case that p is a reason to B to believe that p, a reason that is decisive from B’s point of view. (2006: 413) Whereas Murphy focuses on the authority of speakers, however, we are also, and primarily, concerned with the authority of texts. ⁵ I assume that it’s not just metaphor to speak, for example, of ‘the authority of reason’. I also assume that not just any object is a source of information or directives. Experience of a sofa, for example, is a source of information about the sofa; but (in the usual case) a sofa is not itself a source of information. ⁶ Cf. Murphy 2006: 413. On this point, Murphy follows Raz 1979. ⁷ Cf. Hampton 1998: 85–92.

  

57

one another, Bart does not have decisive reason to take either course of action. Claire is told by her doctor that a trans-oceanic flight is safe at this stage in her pregnancy; but she is told by another doctor that it is not. Assuming she has no reason to distrust either doctor, both may be theoretical authorities for her with respect to this issue. But she does not have decisive reason for believing that the trans-oceanic flight is safe. The two examples just mentioned—cases in which we have conflicting authorities—show that it isn’t strictly true to say without qualification that authorities provide decisive reasons for belief or action. Problems also arise in cases where there is no overt conflict among authorities. Consider again the case of Claire, and her question about the trans-oceanic flight. Although doctors generally function as authorities for us in matters of health, we know enough about the prevalence of disagreement among doctors to continue investigation when we receive answers from our doctors that, for one reason or another, we are tempted to doubt. Thus, if Claire were sufficiently worried about the safety of the flight, reassurance from her doctor might fail to settle the matter for her even in the absence of an explicitly conflicting report from another doctor. Even while granting that her doctor is an authority on this matter, she might reasonably take herself to lack decisive reason for believing that it is safe to fly until she has consulted several doctors. To handle these sorts of cases, we must add ‘no defeater’ conditions to our account of the necessary conditions on practical and theoretical authority. Authorities are suppliers of reasons (more on this below), and reasons have the power to confer justification on beliefs or actions. In epistemology, a reason for a belief is said to be defeated whenever its power to confer justification is neutralized by some other belief or experience. The neutralizing belief or experience is then called a defeater for the reason. The defeater might be a reason to believe contrary to the original belief (i.e. it might be a rebutting defeater). Or it might be a reason to think that one’s original set of reasons does not really support the belief, or indicate that it is true (i.e. it might be an undercutting defeater).⁸ Similar might be said in the case of reasons for action. An example will help to illustrate the distinction between rebutting and undercutting defeaters. Suppose Betty believes on the basis of Barney’s testimony that Barney is a brilliant neurosurgeon. If she later learns that Barney has a pathological habit of lying about his profession in order to attract women, she will have an undercutting defeater for her reason (Barney’s testimony) for believing that he is a neurosurgeon. If, on the other hand, she learns that Barney flunked out of

⁸ For more precise characterizations, see Bergmann 2006, ch. 6; Plantinga 2000, ch. 11; and Pollock and Cruz 1999: 195–7. Bergmann and Plantinga treat beliefs as the objects of defeat whereas Pollock and Cruz treat reasons as the objects of defeat. As Bergmann (2006: 159–60) explains, however, this is mostly just a terminological difference.

58

   :  

college, has never been to medical school, and spends every day of the week working at the local rock quarry, she will have a rebutting defeater—i.e. she will have reason to believe that Barney is not a brilliant neurosurgeon. In the problem cases above, what we seem to encounter are situations wherein authority is defeated. Importantly, these are not cases in which an erstwhile authority has simply ceased to be one—as if, say, Claire’s doctor was once an authority in the domain of medicine but has now entirely lost whatever authority she once had. Nor does it seem exactly right to say that the reasons supplied by the authority are defeated. The fact that Claire’s doctor says that it is safe to take the flight is still a serious reason for thinking that it is safe to fly. (So it is not in that respect like the case of poseur-Barney’s testimony.) It’s just that the doctor’s testimony no longer supplies decisive reason. In other words: relative to the present context and the particular issue at hand the doctor is no longer authoritative. Likewise in the other cases. Putting it just this way, however, suggests wrongly that authority might be an entirely situation-relative matter, and it still gives us no clear information about the circumstances under which authority is defeated. It will help, then, to seek a bit of further clarity. In order to understand the phenomenon of authority-defeat and the circumstances under which it occurs, we need first to understand three other concepts: the concept of one reason being prior to another, the concept of someone or something being a source or supplier of reasons, and the concept of a source of reasons being more authoritative than another. I’ll take these in order.⁹ Suppose Fred has a reason—self-interested fear of consequences—to lie to his boss, and a different reason—the fact that lying is immoral—to tell the truth. Suppose further that he has no other reasons relevant to the decision whether to lie. One of these reasons has priority over the other if, and only if, it is more (practically) rational to act on the basis of one rather than the other.¹⁰ Likewise with reasons for belief: one has priority over another if, and only if, it is more

⁹ I am particularly grateful to Matthew Lee for pressing objections that helped me to see that clarification of these three concepts was necessary in order to understand authority-defeat. I am also grateful to him for helping me to make some of the relevant clarifications. ¹⁰ You might be tempted to object that whether it is more rational to act on the basis of reason R1 rather than on the basis of reason R2 depends on the circumstances. But, as I see it, any circumstances that would be relevant to how one should act would simply be additional reasons. Suppose, for the sake of argument (and contrary to philosophers like Thomas Aquinas) that there is an exception to the moral prohibition on lying: it is always wrong (and therefore irrational) to lie, except when doing so will protect innocent life, in which case it might be more rational to lie than to tell the truth. If this is true, then whether it is more rational for Fred to lie or tell the truth to his boss depends on whether the consequences of telling the truth will include loss of innocent life. But this doesn’t at all show that circumstances make a difference to the question whether it is more rational for Fred to act on the basis of the reason supplied by morality or the reason supplied by his self-interested fear. Suppose innocent life will be lost if Fred tells the truth. In that case (assuming, again, that it might be rational to lie in order to protect innocent life), it might be more rational for Fred to act in accord with the reason supplied by his self-interested fear; but there is still no reason to think that it is more rational for him to act on the basis of that fear.

  

59

(epistemically) rational to believe on the basis of one rather than on the basis of the other.¹¹ A source or supplier of reasons is, intuitively, an individual or a cognitive faculty or a text or some other entity from which we are able to acquire cognitive input that functions for us as a reason for belief or action. But it is a rather tricky matter to identify once and for all the list of things from which we are able to acquire cognitive input on a topic. There is a perfectly good sense, for example, in which God is a source of input on any topic whatsoever. God knows everything, after all; and God can reveal to us anything that can be understood by our finite minds. But there is also a perfectly good sense in which God is not a source of cognitive input on most topics. For example, one can’t simply consult God to find out how one’s deceased relatives want their worldly possessions distributed. One can’t expect God to adjudicate between theories in fundamental physics. Though God has that information, there is absolutely no reason to think that God is willing to distribute it on demand (or even in response to a lot of effort). It is this sense of being a source of reasons—the sense in which God is not a source of reasons in every domain—that I have in mind when I talk about sources of reasons in what follows. Roughly (and I don’t think we can do much better than ‘roughly’) a source of reasons is an entity from which it is physically possible without miracles or special divine intervention to acquire reasons for belief or action.¹² We can now say what it is for a source of reasons to be more authoritative than another. Authorities are dispute-settlers. They are, in other words, sources of reasons that have priority over reasons that we get from non-authoritative sources. But, of course, the reasons that come from one authority might well trump the reasons that come from another. Schoolteachers have authority over children: they supply reasons for the children to act that trump a lot of other reasons they might have for acting (reasons like the desire to throw chalk, the belief that it would be amusing to make Billy cry, etc.). But, dysfunctional cases aside, parents are more authoritative than schoolteachers: the reasons for action supplied by a parent have priority over reasons for action supplied by schoolteachers. For example, other things being equal, the fact that a fifth grader’s schoolteacher has assigned watching The Matrix as homework is reason for the fifth grader to watch that ¹¹ The distinction between practical and epistemic rationality may be understood as the distinction between what it is rational to do given the goal of furthering one’s overall best interests and what it is rational to believe given the goal of believing in accord with the truth (cf. Rea 2001: 139–44). ¹² Note, too, that being a source or supplier of reasons is—for purposes here—not the same as being a source or supplier of information about reasons that one has, as it were, from other sources. A parent might inform her child that it is wrong to lie; but in so doing she is not, in the sense I have in mind, becoming the source or supplier of a reason to tell the truth. When a parent tells her child that it is wrong to lie, she is simply informing the child of the existence of a reason that is available and supplied, as it were, from another source. The reasons of which parents are sources are reasons that we might express with phrases like because mom told me not to lie (a reason for telling the truth) or because mom told me that it is wrong to lie (a reason for believing that it is wrong to lie, and so a reason for believing that one has, from another source, a reason to tell the truth).

60

   :  

movie. But that reason will be trumped by a parent’s general prohibition on watching movies with an ‘R’ rating. Likewise, Scientific American provides, in the domain covered by the natural sciences, reasons for belief that trump a lot of other reasons we might have for belief in that domain (reasons coming from untutored empirical intuition, for example, or testimony by non-experts). But the word of a Nobel prizewinning physicist speaking in earnest in her area of expertise trumps whatever conflicting reasons we might get from Scientific American. Thus, a bit more formally: T ‘M A T’ R A is more authoritative than B (for a person C, in a domain D) if, and only if, A and B are both sources of reasons for belief or action for C in D, and the reasons supplied by A have priority for C in D over the reasons supplied by B. We may now return to the notion of authority-defeat. Note first that this whole messy issue arose because we found cases wherein (i) we still want to say that someone—Claire’s doctor, for example—is an authority in a domain, but (ii) the authority in question has supplied reasons for belief or action that are not decisive in that domain. What ought to be clear by now is that these are cases where, for one reason or another, there are conflicting reasons supplied by other sources that are on a par with or have priority over the reasons supplied by the authoritative individual. But how can this occur, if an authority is, by definition, a supplier of decisive reasons? The most natural response, I think, is to say (as indicated above) that being an authority is not an absolute matter: whether something counts as an authority depends a lot on context. This much is easily verified. For example, around my dinner table, I can say truthfully to one of my daughters that my 13-year-old son is an authority on how to play the French horn. I can say this truthfully in that context because he is, for us around the family table, more authoritative in that domain than any other source to which we have access. But, of course, if I were to go to the members of the music department at Notre Dame and announce that they should take heed, for my son Aaron is an authority on how to play the French horn, I would be saying something false. But in saying that being an authority is not an absolute matter, we are in no way committed to the view that all of the facts about authority are situation relative. We can (and should) say that the absolute facts about authority are facts expressed by claims of the form ‘X is more authoritative than Y (for individual S in domain D)’. To be an authority in some domain, then, is just to be, in that domain, a source of reasons that is (objectively, absolutely) more authoritative for some contextually salient group of people than other contextually salient sources of reasons. In the case of Claire’s doctor, we can still truthfully say that she is an authority in the domain of medicine because, in context, that boils down to something like the claim that Claire’s doctor is more authoritative for non-doctors than sources of evidence that

  

61

are neither doctors nor medical texts written by doctors, and so on. But in the particular context of our example, the authority of Claire’s doctor is defeated by the fact that she is not more authoritative in the domain of medicine than the other doctors who are giving conflicting opinions. In light of all of this, we can now identify the following conditions under which authority-defeat occurs: A-D A’s authority (for a person B, in a domain D) is defeated at a time if, and only if (i) there are contexts relative to which it would be true to say that A is an authority for B in D, but (ii) at the time of defeat it is not rational for B to treat some statement or directive from A in D as decisive reason for belief or action. Condition (ii) is satisfied in cases like our earlier examples involving conflicting directives from parents and conflicting information from medical authorities. It is also satisfied in cases where B acquires evidence independent of conflicts with other authorities for thinking that an erstwhile authority is no longer authoritative. Thus, for example, if B learns that A is morally corrupt in relevant ways, at least some of A’s commands will no longer be decisive reasons for action; if B learns that A lacks expertise that B once thought she had, then there will be at least some contexts in which A’s say-so is no longer decisive reason for belief; and so on. Recall that I introduced the concept of authority-defeat because I needed it in order to flesh out my accounts of theoretical and practical authority. Now, I have already said that being an authority (and so, likewise, having authority) in a domain D is a context-relative matter: it is a matter of being, for some salient group of people, more authoritative than other salient sources. Employing this insight, as well as my characterization of authority-defeat, I can now say more clearly what it is for someone or something to have theoretical or practical authority over an individual in a domain: H T/P A Where A is a source of reasons and B is someone who can believe or act on the basis of reasons, A has theoretical authority over B in D only if A’s affirming something in D is, in the absence of authority-defeaters, decisive reason for B to believe what A affirms. Likewise, A has practical authority over B in D only if A’s telling B to do something that falls within D is, in the absence of authority-defeaters, decisive reason for B to do that thing. We have now taken note of two features of authority: first, that it is domain relative; second, that it comes (broadly speaking) in at least two varieties. Thus, in

62

   :  

order to be clear about what we mean in calling a text authoritative, we must both specify a domain and specify the type of authority we have in mind. Before closing, there are two further bits of terminology that I’d like to introduce. First, we should note the distinction between de facto and de jure authority. Second, it will be helpful to have in what follows a concept of foundational authority. The difference between de facto and de jure authorities corresponds roughly to the distinction between what we happen to treat as sources of decisive reasons and what are in fact sources of decisive reasons.¹³ (Note that treating something as a source of decisive reasons doesn’t necessarily involve believing it to be a source of such reasons. Parents are generally treated by small children as sources of decisive reasons even though small children typically lack the belief—by virtue of lacking the relevant concepts—that their parents are sources of such reasons.) Even atheists will agree that, for example, biblical assertions and injunctions are de facto authoritative for many people. It is undeniable that many people treat the Bible’s say-so as decisive reason for belief or action. But there will be real controversy over the question whether biblical assertions and injunctions are de jure authoritative. For there is real controversy over the question whether the Bible’s say-so provides anyone with genuinely decisive reasons for belief or action. Thus far, I have been treating an attribution of ‘genuine authority’ as equivalent to an attribution of de jure authority, and I will continue to do so in what follows. Thus, A’s taking a text as authoritative—its being a de facto authority for A—isn’t, in my terminology, sufficient for its being genuinely authoritative over A. Accordingly, unless otherwise noted, in the remainder of this chapter all talk of authority should be construed as talk of de jure authority. The notion of foundational authority may be characterized as follows: F A A is foundationally (theoretically/practically) authoritative over B in D if, and only if, A has (theoretical/practical) authority over B in D and there is no source of (epistemic/practical) reasons for B in D that is more authoritative than A. Given this definition, it should be easy to see that there can be more than one foundational authority for a given person in a given domain. The reason is simply that more than one source can be such that there is no source more authoritative than it. Given the characterizations of authority-defeaters above, it should also be clear that foundational authorities can defeat one another. Foundational ¹³ To a certain extent, the distinction in view here could also be captured by the terms ‘subjective’ and ‘objective’ (replacing ‘de facto’ and ‘de jure’ respectively). But there are problems with these terms as well. For example, it seems rather odd to say that a text might be subjectively, but not objectively, authoritative for a nation, or for all of humanity at a time, whereas it sounds less odd to say that the same text is de facto authoritative but not de jure authoritative for a nation, or for all of humanity at a time.

  

63

authorities can also suffer authority-defeat if other evidence comes to light— perhaps even from less authoritative sources—that they are no longer to be treated as suppliers of decisive reasons for belief or action.¹⁴ So, it would seem that a very rough and ready answer to the question ‘What does it mean to call a text authoritative?’ is something like this: To call a text (genuinely, de jure) authoritative is to say that, within some domain and for some individual or individuals, the text supplies reasons for belief or action (or both) that are, absent defeaters, decisive. Again, it doesn’t follow from this that the supplied reasons motivate the relevant beliefs or actions. Nor does it even follow that the individual or group in question will recognize them as decisive. (De jure authorities might fail to be de facto authorities—as often happens, we might think, when people casually and habitually act contrary to state laws or common moral intuitions.) But they will, nonetheless, decisively justify the relevant beliefs or actions. This characterization of textual authority comports well with a variety of formal, informal, and partial characterizations of biblical authority that one finds in the literature. Thus, for example, in discussing Jesus’ view of the authority of the Old Testament, Edward J. Young writes: [Jesus] believed that both as a unit and in its several parts [the Old Testament] was finally and absolutely authoritative. To it appeal might be made as to the ultimate authority. Its voice was final. When the Scriptures spoke, man must obey.¹⁵

Here the idea seems to be that scripture is a de jure foundational supplier of decisive reasons for action. Robert Gnuse takes a similar view, quoting with approval Robert Bryant (1968: 156): The basic definition of authority is to speak of that “which is acknowledged as rightly and worthily commanding loyalty and obedience.”¹⁶

Young, Bryant, and Gnuse are clearly (in these passages, anyway) focused on the practical authority of scripture. But the literature is also replete with characterizations of scripture as having what I am here calling theoretical authority. Thus, for example, Clark Pinnock assimilates the notion of scriptural authority to the

¹⁴ In the domain of theoretical authority, what I have elsewhere called basic sources of evidence will be in some ways akin to foundational authorities (cf. Rea 2001: 2). To treat something as a basic source of evidence is to trust it as reliable even in the absence of evidence for its reliability. Our basic sources of evidence are de facto authorities for us and, in the typical case, they are treated as foundational (in the sense just described) as well. Of course, since we can treat hopelessly unreliable sources as basic, something’s being a basic source of evidence for us doesn’t guarantee that it is a genuine authority, much less a foundational authority. Still, treating something as a basic source of evidence will typically (though not necessarily) go hand in hand with believing it to be a foundational authority in some relatively wide domain. ¹⁵ Young 1967: 59. ¹⁶ Gnuse 1985: 2.

64

   :  

notion of ‘primary evidence’.¹⁷ Similarly, John Woodbridge notes that one meaning for the Latin word auctoritas (from which our term ‘authority’ is derived) is ‘the things which serve for the verification or establishing of a fact’.¹⁸ The clear suggestion in the text following this remark is that this meaning is part of what we invoke when we call scripture authoritative. On the other hand, there are other characterizations of biblical authority with which my own is inconsistent—such as, for example, characterizations according to which authority is equivalent to canonicity.¹⁹ Being canonical, after all, has much less to do with being a de jure supplier of decisive reasons for belief and action and a lot more to do with being regarded by the Church as being divinely inspired or as containing a record of divine revelation. It should also now be clear that the bare claim that the Bible is authoritative does not tell us very much. We can also see that many of the questions about biblical authority that have vexed theologians over the centuries (but perhaps especially in the past two centuries) are more perspicuously characterized as, or at least dependent on, questions about the domain over which the Bible exercises authority (history? science? just ‘faith and practice’? or what?) and about the type of authority (de jure, de facto, foundational, non-foundational, theoretical, practical, etc.) that the Bible exercises. Does one challenge biblical authority if one thinks that one can use reason or science to discover factual errors in the biblical text? Does one challenge biblical authority if one tries to demythologize the Bible? Does one challenge biblical authority if one thinks that the Bible reflects an outmoded system of moral values that aren’t binding on enlightened contemporary thinkers? In each case, the answer depends on what one takes the domain of biblical authority to be, and on whether one thinks that the Bible has practical authority, theoretical authority, or both. I turn now to a discussion of what it means to say that a text is truthful, after which I draw some conclusions about the relations between authority and truth.

2. Truth Philosophers have worried quite a bit over the question of what it means to say that a sentence, or a belief, or a proposition is true. Our question here is orders of magnitude more complicated: we are asking what it means to say that a text is

¹⁷ Pinnock 2006: 95. ¹⁸ [Editorial note: In the original version of this chapter, I cited Woodbridge 2016. However, this was apparently an error—neither the claim nor the quoted material appears in that paper. I have found the quotation and relevant additional material in Woodbridge and James 2013, ch. 9, sec. IV; but I do not believe that is a source I consulted in writing this paper. I expect, therefore, that the claim and the quotation came from John Woodbridge in correspondence; but I cannot be sure.] ¹⁹ See, e.g. Stonehouse 1967: 92–3.

  

65

true—any text, including texts that go on for hundreds of pages, texts that consist of nothing more than pictures, texts that include literary tropes like metaphor and hyperbole, and so on. I cannot hope to do this question (or even the simpler first one) real justice here. But I can, I think, hope to say enough about truth to draw some interesting conclusions about the relationship between authority and truth. Let me begin by saying a few words about the nature of truth. In the philosophical literature, the major theories about the nature of truth are primarily theories about what it is for a sentence, belief, or proposition to be true. They are not theories about truth in pictures, truth in metaphor, truth in literature, or anything of the sort. For this reason, an extended foray into the overgrown jungle of theories about truth is unlikely to be productive for present purposes. Those theories are simply too narrow; they will not help us (much) with questions about what it means to say that a metaphor, or a parable, or a psalm, or an epistle is true. Still, it will be helpful to be aware of two of the main divisions within that literature. The first is that between realist and anti-realist conceptions of truth. To understand this division, and its import, we must first say what we mean by ‘realism’. Realism has been characterized in many different ways. For purposes here, I’ll adopt the following characterizations:²⁰ R (1) Where ‘x’ is a singular term, realism about x is the view that some particular thing is identical to x. (2) Where ‘F’ is the name of a kind (e.g. ‘unicorn’), property (e.g. ‘moral wrongness’), or relation (e.g. ‘causation’), realism about F (or about Fs) is the view that the following conditions hold: (i) Something is (an) F or stands in the F-relation to something else. (ii) If some particular thing is an F or stands in the F-relation to some other things, that fact does not depend upon the experiences that other creatures might have of those things, nor does it depend upon what other creatures think about them or about the conditions under which being an F or standing in the F-relation is exemplified. Given these characterizations, one way to be an anti-realist about God (say) is to affirm that there is no such being as God; but another way is to say, for example, that ‘God exists’ expresses a truth but that the truth it expresses isn’t that there is something identical to God, but rather (say) the claim or ‘conviction that free and loving persons-in-community have a substantial metaphysical foundation, that

²⁰ [Editorial note: In the original version of this chapter, note 20 inaccurately reported a difference between the present formulation of realism and the one I offered in Chapter 1 of this volume. Rather than reproduce the error here, I have decided instead just to omit it and to explain the omission.]

66

   :  

there are cosmic forces working toward this sort of humanization’.²¹ Likewise, one way to be an anti-realist about statues is to say that there are none; but another way is to say that nothing counts as a statue unless it is believed to be one by creaturely observers. One way to be an anti-realist about beliefs, or about minds, is to believe that whether a person has a belief or a mind depends in some way upon how individuals other than that person think about or experience her. And so on. Accordingly, anti-realism about truth will be the thesis that (i) nothing is true or (ii) something’s being true depends in some way upon people’s beliefs about the conditions under which something is true, or (iii) something’s being true is at least partly constituted by facts about actual or hypothetical creaturely mental states.²² For example, the view that p is true if, and only if, p is uncontroversial would count as a particularly radical version of anti-realism about truth. So too would the view that p is true if, and only if, p would be uncontroversial in a community of ideal rational agents. It is sometimes thought that anti-realism about truth goes hand in hand with some sort of metaphysical anti-realism. But this is not correct. Metaphysical antirealism is likewise characterized in many different ways. Sometimes it is treated as a view about the aims of metaphysics (roughly: metaphysical theories aren’t aimed at truth). Sometimes it is treated as a view about commonsense and scientific kinds (roughly: anti-realism about all or most commonsense and scientific kinds is true). Sometimes it is treated as equivalent to the thesis that metaphysical theories aren’t objectively true. And so on. We can capture the most important kinds of metaphysical anti-realism, I think, by saying that metaphysical anti-realism is the thesis that theories in metaphysics are not to be ‘interpreted realistically’, where interpreting a theory realistically is to be understood as follows: I  T R To interpret a theory realistically is (a) to take it as having an objective truth-value (i.e. to take it as something other than a mere evocative metaphor or expression of tastes, attitudes, or values), and (b) to interpret it as being true only if realism about the objects and properties (supposedly) referred to in the theory is true.²³ To be sure, some ways of being an anti-realist about truth will imply metaphysical anti-realism. For example, the view that nothing is true (because, e.g. the whole concept of truth is defective) counts as a version of anti-realism about truth, and it

²¹ Kaufmann 1981:49, quoted in Plantinga 2000: 41. ²² We might note that beliefs count as true and they are mental states, but it doesn’t follow that their being true depends on mental states in the relevant sense. To say that the truth of my belief that p is ‘mind dependent’ in the relevant sense is to say either that its being true depends on what other people think about the conditions under which a belief is true, or that the fact that my belief is true is partly constituted by facts about other people’s mental states. ²³ Cf. Chapter 1 of this volume, pp. 19–20.

  

67

straightforwardly implies that no theory in metaphysics or any other discipline is to be interpreted realistically. But anti-realism about truth is also separable from metaphysical anti-realism. Suppose, for example, you think that p is true if and only if p is what would be believed by rational inquirers at the end of an ideal process of inquiry. This view implies anti-realism about truth. But one can hold this view while at the same time holding (for example) that God is the only rational inquirer capable of engaging in an ideal process of inquiry, that God believes that there are objective truths in metaphysics, and that God is also a realist about various commonsense and scientific kinds (e.g. the kinds ‘person’ and ‘human being’). Thus, even though the view in question implies anti-realism about truth, it does not imply that truth in metaphysics is always subjective, or that no theory in metaphysics is to be interpreted realistically, or that anti-realism about commonsense and scientific kinds is true.²⁴ So it doesn’t imply metaphysical anti-realism. What goes for the relation between realism about truth and metaphysical realism goes also for the relation between realism about truth and theological realism. Theological realism is, in effect, a species of metaphysical realism: it is the thesis that theories in theology (or in some particular domain of theology—e.g. Christian theology, or distinctively Presbyterian theology) are to be realistically interpreted (in the sense described above). And just as one might endorse antirealism about truth without endorsing metaphysical anti-realism, so too one might endorse anti-realism about truth without endorsing theological antirealism. As noted above, the division between realism and anti-realism about truth is just the first of the divisions to which we must attend. The second is that between epistemic and non-epistemic conceptions of truth. Epistemic conceptions of truth maintain that there is a necessary connection between what is true and what would be believed under certain specified conditions by a certain kind of rational agent or agents. Epistemic conceptions of truth abound in the literature. To take just a few examples: C. S. Peirce says that ‘The opinion which is fated to be ultimately agreed to by all who investigate, is what we mean by the truth . . . ’;²⁵ William James claims that ‘[t]rue ideas are those that we can validate, corroborate, and verify’;²⁶ and Brand Blanshard holds that ‘truth consists in coherence [of our ideas]’.²⁷ Non-epistemic conceptions of truth, on the other hand, maintain that there is no necessary connection between truth and what would be believed by rational creatures. The main import of this second division within the literature on truth for our purposes is just this: Whatever other links there are between authority and truth, epistemic conceptions of truth introduce additional and very direct connections. ²⁴ On this, cf. Alston 1996 and Devitt 1991. ²⁷ Blanshard 1940: 269.

²⁵ Peirce 1878: 139.

²⁶ James 1907: 142.

68

   :  

For example, undefeated theoretical authorities validate, corroborate, and verify our ideas, and, by definition, their assertions cohere with the assertions of all authorities to which they are not prior. Thus, it looks as if on the theories offered by James and Blanshard, undefeated authoritative assertions are automatically true. Be that as it may, I propose to set aside consideration of epistemic accounts of truth in what follows. This is not because they are unworthy of attention, but rather because I have elsewhere already given them the attention that I think they deserve. Elsewhere, I have argued—developing a line of reasoning found in the work of Alvin Plantinga²⁸—that epistemic accounts of truth imply something very much like theism.²⁹ Naturally, I am a fan of the implied consequence; but it is hard to take seriously a theory of truth that is so heavily laden with ontological commitments. Better, I think, to look elsewhere for a theory of truth. Once we do so, however, we find that differences among theories of truth do not make for substantive differences in what we say about the relationship between authority and truth. And so we save ourselves the trouble of proceeding piecemeal, asking with respect to each theory of truth what the relationship between authority and truth is if that theory is true. The third and final division to which I want to attend lies within the realist camp: the division between correspondence theories of truth and all the rest. The main thing that I want to note about this division is simply that it exists: that is, some realist theories are correspondence theories, but not all of them are. It is sometimes thought that realism about truth goes hand in hand with a correspondence theory of truth. But this is a mistake. The confusion is especially important in the present context because it seems to be one to which Christian theologians, and especially evangelicals, are particularly prone. One often finds theologians taking a stand for the correspondence theory of truth, apparently in part because of the misguided belief that one must do so in order to preserve the (absolute) truth of core doctrines of the Christian faith. The basic idea seems to be that giving up the correspondence theory means giving up realism about truth which, in turn, means giving up realism in theology. We have already seen that the second link in this chain of reasoning is weak; but it is important to see that the first is as well. The correspondence theory of truth comes in different varieties, all of which maintain that being true consists in the obtaining of a relation—the correspondence relation, whatever exactly that is—between truth-bearers (propositions, beliefs, or sentences usually) and something in the world (states of affairs, or facts, for example). But one need not believe in the existence of any sort of correspondence relation, nor need one even believe that truth has a substantive,

²⁸ Plantinga 1982.

²⁹ Rea 2000.

  

69

analysable nature, in order to be a realist about truth. A coherence theory of truth, for example, maintains (roughly) that being true is a matter of cohering with certain other beliefs or propositions. But it does not follow from this doctrine that being true is ‘mind dependent’ in the sense specified by condition (ii) of the schematic characterization of ‘realism about F’.³⁰ Similarly, William Alston has defended at length the thesis that realism about truth is consistent with minimalism about truth—the thesis, roughly, that there is nothing more to be said about the nature of truth beyond the so-called T-schema (‘X is true if, and only if, p’, where particular instances are obtained by replacing ‘p’ with a declarative sentence and ‘X’ with an expression referring to that sentence).³¹ So, again, realism about truth does not require the correspondence theory. So much for remarks about the nature of truth. I now turn to a more direct consideration of the main question of this section—the question of what it means to attribute truth (or falsity, for that matter) to a text. First, some easy cases. If the text in question expresses a single (perhaps conjunctive) proposition, then the text is true if and only if the proposition it expresses is true. If, on the other hand, the text is a sentence like ‘Stop!’ or ‘Enter at your own risk’, then attributing truth to the text is just a category error. But, of course, we are not here interested in the easy cases. What interests us here ultimately is the Bible—a text composed of smaller book-length works of various genres, which includes within it sentences of various grammatical types, as well as various literary tropes like metaphor, allegory, hyperbole, and the like. What could it mean to say that a text like that is true? Suppose someone were to tell you that Tolkien’s Lord of the Rings is true—not that there are hobbits and elves and a place called Mordor, of course, but that, despite the nonexistence of these sorts of things and despite the fact that the events narrated therein never took place, the text is nonetheless true. Or suppose someone were to hand you a book of sonnets and poetic lamentations, and suppose they were to tell you that the poems were true. Could anyone make sense of these claims? What I want to suggest is that we can make sense of claims like this in much the same way in which we make sense of similar claims about interesting metaphors. The philosophical literature on metaphor is vast and complicated, but we don’t need a full-blown theory of metaphor in order to appreciate roughly what it might mean to call a metaphor ‘true’. Attention to two main points of agreement in the literature ought to suffice for present purposes. Note first that there is a distinction to be drawn between the semantic content of a text and what the speaker or author means by the text (or, in other words, what the text ‘pragmatically conveys’). I don’t have a theory about the relationship ³⁰ Though, of course, there are anti-realist ways of developing the coherence theory. ³¹ Alston 1996. For discussion of the T-schema and its relation to Tarski’s ‘Form T’, see pp. 30ff.

70

   :  

between semantics, pragmatics, and ‘what is said’ by a text,³² but we don’t need any such theory in order to appreciate the basic point that what authors of metaphors typically mean by their words is something other than the bare semantic content of the metaphor. There is controversy in the literature over whether live metaphors typically (or ever) mean anything determinate. That is, there is controversy over whether what is conveyed by a metaphor is a definite proposition, or whether metaphors merely invite or stimulate certain kinds of reflection.³³ But there is general agreement that if metaphors convey anything determinate, what they convey goes beyond their semantic content.³⁴ Accordingly (and this is my second point), so-called ‘metaphorical truth’, whatever it is, will attach to what is intentionally conveyed rather than to the semantic content. This, of course, is why metaphors that are semantically false— e.g. ‘Juliet is the sun’—can nevertheless be assessed as true. If the author of a metaphor intentionally conveys propositional content by way of the metaphor, then the metaphor (as such) will be true if and only if the relevant proposition is true. On the other hand, if the author of the metaphor fails to convey propositional content (perhaps because metaphors as such generally lack such content), then either the author fails to convey any ‘metaphorical truth’ or the metaphor’s ‘truth’ consists in its being insightful or revealing or otherwise interestingly illuminating by virtue of the sorts of reflections it stimulates.³⁵ Return, now, to the cases we began with—cases wherein someone tells you, for example, that The Lord of the Rings is true, or that a book of poetry is true. It seems that we can understand these sorts of truth claims in roughly the way just described. Either we attribute to the speaker—the one who is telling us that the text is true—the view that the text intentionally conveys a message that is somehow different from the semantic content of the text, and that the message in question is genuinely true; or we attribute to the speaker the view that the metaphor intentionally invites lines of reflection that naturally lead to genuine insights, fruitful lines of research, and so on. If we buy into the view that the text conveys a true message, we employ our skills at literary analysis to try to extract the message and then to assess it as true or false. If we think that the text merely invites lines of reflection that lead to genuine insights, then we pursue the relevant lines of reflection. This, then, is one way in which attributions of truth to a text require clarification. For all but the most simple texts, however, there is a further way in which

³² For a start into this literature, with particular attention to metaphor, see Camp 2006a. ³³ For example, compare, on the one hand, Davidson 1979 and Cooper 1986 with, on the other hand, Black 1954/5, 1977; Soskice 1985, and, more recently, Stern 2000, 2006 and Camp 2005, 2006a, 2006b. ³⁴ See, e.g. Camp 2006a, 2006b and Stern 2000. ³⁵ On this, see Cooper 1986, ch. 4.

  

71

such attributions must be clarified. Consider, for starters, Benjamin Franklin’s poem, ‘Death is a Fisherman’: Death is a fisherman, the world we see His fish-pond is, and we the fishes be; His net some general sickness; howe’er he Is not so kind as other fishers be; For if they take one of the smaller fry, They throw him in again, he shall not die: But death is sure to kill all he can get, And all is fish with him that comes to net.

Someone might well attribute truth to this poem; and, indeed, one might well suppose that the poem conveys propositional truth (as opposed to merely stimulating interesting lines of reflection). But note that one can do so while disagreeing with central features of the allegory. So, for example, it seems that the following speech represents one perfectly sensible, though hardly inevitable, response to the poem: The poem is quite true—Death is merciless and takes all who fall into his net; and Death seems like a fisherman in other respects as well (Death is predatory and so on). But it just doesn’t seem right to say that ‘some general sickness’ is his net. Death catches us in many ways. Does the remark that ‘it doesn’t seem right to say that “some general sickness” is his net’ negate the remark at the beginning of the speech—namely, that the poem is ‘quite true’? No. For, clearly enough, an attribution of truth to a text need not mean that every single message conveyed by the text or by some part of the text is true. In fact, I suspect that it is relatively rare for attributions of truth to a complex text to mean anything more than that the central, or most important message conveyed by the text is true. (Obviously for a long or complex text, the central or most important message may be a longish conjunction of smaller messages.³⁶) Suppose you read an intellectual biography of Sartre that devotes hundreds of pages to a description of his upbringing, early intellectual development, and

³⁶ In some cases it will be a bit of a stretch to speak of ‘the central message’, even if we insist that we are thinking of it as a long conjunction of smaller such messages. You might, for example, say of a newspaper that it tells the truth. But it sounds odd to say that the paper has a central message—even one that is just the conjunction of the central messages of all of the stories. Point taken (and I thank Paul Draper for raising the point); but, despite the oddity, I am inclined to think that if one were to assert of a newspaper that it is true, or tells the truth, what one would probably mean is just that the ‘central message’ (in the ‘longish conjunction’ sense of the term) is true.

72

   :  

philosophical career but, as regards his romantic life, offers only one sentence: ‘Sartre enjoyed a life-long romance with Simone de Beauvoir, to whom he was ever faithful.’ Given that Sartre was notorious for his many mistresses, we would be reluctant to say of a biography like this that it is true, even if it were absolutely inerrant with respect to all of the other details of Sartre’s life. The reason is that it gets wrong a fact that many of us will regard as pretty important. On the other hand, we probably would assess it as true if it were accurate on this and other salient details but included mistakes on such matters as the number of students attending some famous lecture of Sartre’s, the precise date on which Sartre first met Camus, or the number of pages in the first edition of Being and Nothingness (so long as it was not mistaken by much). If asked how we could possibly say that the text as a whole is true when it contains manifest falsehoods, our reply would likely just be that the falsehoods simply aren’t central—i.e. they are not part of ‘the most important message’ of the text. Moreover, it seems clear that which details count as part of the central message of a text (and so which details the text must get right in order to count as true) may shift from context to context. A time traveler who took Newton’s Principia back to Egypt of 3000  would not be lying if she said that everything in the Principia is true. But she would be lying (or confused) if she brought the text into a twenty-first-century graduatelevel physics course and announced that everything in the text is true. The reason is that the ways in which the Principia goes wrong are not salient in a context where Newton’s work is being introduced to Egyptians of the fourth millennium , but they are highly salient in a twenty-first-century graduatelevel physics course. We may now formulate a recipe for assessing attributions of truth to biblical texts or to the Bible as a whole. For any such attribution, we must determine whether the claim is (1) that the (central, or most important) semantic content of the text is true, (2) that the text intentionally conveys a genuinely true message that both is the central or most important message conveyed by the text and also somehow differs from or goes beyond its semantic content, (3) that the text invites reflection that will naturally lead to genuine insight, or (4) some combination of these three. We must then employ our skills at literary analysis to uncover the central or most important message (which, again, will likely be a conjunction of smaller central and important messages), pursue the relevant lines of reflection, and so on. This means, of course, that there is real complexity involved in evaluating attributions of truth to texts—so much so that it is not clear that one says anything deeply informative when one says of a long and complicated text simply that it is true. Before we can even begin to evaluate such a claim, we need a lot more information—information about what is central to the text, about the proper interpretation of the literary tropes within the text, information about genre and much more. But I think that, rather than be troubled by such a conclusion, we ought to welcome and embrace it. No one is helped in their

  

73

understanding or assessment of the Bible to be told, without further clarification, that the Bible, as a whole, is true. The really crucial questions, it seems, are questions about the parts.

3. Authority and Truth We are now in a position to draw some conclusions about the relationship between attributions of authority and attributions of truth. First, it should be clear from the foregoing that to say without qualification that a text is authoritative is to say nothing definitive about whether it is true. The reason is that the text might be merely practically authoritative, and a merely practically authoritative text need not semantically express or pragmatically convey any propositional message. It need not even stimulate insightful lines of reflection. It might merely issue directives. (The Ten Commandments, for example, cannot sensibly be said to be true in any sense that we are interested in here.) This is not to say, however, that truth is entirely irrelevant to practical authority. For example, it would be hard to take seriously a text’s claim even to practical authority in the domain of morality if the text in question were filled with supposedly factual assertions to the effect that members of a particular race or gender are morally worthless, or if it went on for pages extolling the health benefits of cannibalism, or if it contained enough other misguided assertions as to lead one seriously to doubt the wisdom or sanity of its author. On the other hand, a text might have a great deal of merely practical authority and still be mistaken about a wide variety of facts, so long as its being so mistaken is consistent with the view that its author possesses sufficient wisdom and knowledge within the relevant domain to merit obedience. The point, then, is just that practical authority on its own is no guarantee of general truthfulness, both because it is possible that a merely practical authority make no factual assertions, but also because it is possible that the factual assertions such an authority does make are entirely irrelevant to its status as a practical authority. Second, it should also be clear that if a text has theoretical authority over an individual in some domain, then the text’s assertions within that domain must be reliable enough to warrant belief in the absence of defeaters. (For purposes here, ‘what the text says’ is just the conjunction of whatever it semantically asserts with whatever its author intentionally conveys;³⁷ and ‘reliability’ is to be understood in ³⁷ Note, too, that what a text semantically asserts is not necessarily the same as what it semantically expresses. The proposition that Juliet is the sun, for example, is identical to the proposition that Juliet is the gaseous object around which Earth, Mars, Venus, and various other planets are in orbit. Romeo’s metaphorical remark that Juliet is the sun semantically expresses this proposition—i.e. the proposition is its semantic content. But, obviously, his remark does not assert that proposition; it asserts a different proposition which is pragmatically conveyed by the words, in context, ‘Juliet is the sun’.

74

   :  

terms of likelihood of truth.³⁸) How reliable is ‘reliable enough’? That is hard to say; but we seem to understand the concept well enough to employ it in everyday judgments. Tabloids are not reliable enough within the domain of ‘world news’ to warrant belief. If you believe that Elvis Presley is still alive because you read in The National Enquirer that he was recently seen at a mall in Kentucky, your belief is not warranted. The Wall Street Journal, on the other hand, is reliable enough to warrant belief in this domain—and this despite the fact that it is hardly perfectly reliable. Note, in this connection, that it would be wholly implausible to say, in general, that if a text has theoretical authority over someone within some domain, then whatever it says to that person within the domain must be perfectly reliable. Theoretical authority implies reliability; it does not imply infallibility. Third, we may conclude that if a text has foundational theoretical authority over S in D, then it must be at least as reliable as any other authority for S in D. For suppose this were not the case. Suppose, for example, you have two books (T1 and T2) about some historical event, one of which (T1) has foundational theoretical authority in that domain and the other of which (T2) doesn’t. (Perhaps the first book contains the only surviving eyewitness account of the event, and there is no forensic evidence available about the event apart from the testimony of that source.) It follows from our earlier characterization of foundational authority that if T1 and T2 were to conflict in their assertions about the event, it would be more rational to believe in accord with T1 than with T2. But it is hard to see how this could be the case if T1 were less reliable than T2 in D. Thus, at a minimum, T1 must be at least as reliable in D as T2. Thus, if one understands attributions of authority to the Bible to be attributions of foundational practical authority or foundational theoretical authority in some domain, it follows that the Bible must be at least truthful enough to be as reliable as any other authority in that domain. Thus far, my conclusions about the general connections between authority and truth—and so my conclusions about the connection between biblical authority ³⁸ At the conference connected with the volume in which this chapter appeared, it was pointed out that some have argued that ‘the biblical concept of truth’ is very close to the concept of reliability. Supposedly, then, according to ‘the biblical concept’, saying that something is true is roughly equivalent to saying that it is trustworthy, able to be relied upon, etc. If this view about truth is correct, then here we have a very close connection indeed between a kind of authority and truth simpliciter; and if it is indeed the biblical concept, then this might mean that Christians have good reason to endorse it. However, I am not persuaded by the arguments that purport to establish that this is the biblical concept. One problem is that those who give such arguments are not always careful to distinguish epistemic from non-epistemic conceptions of truth. So, for example, Roger Nicole (1983) sometimes characterizes ‘the biblical concept of truth’ in a way that makes it sound like an epistemic conception, and sometimes characterizes it in a way that makes it sound more like a correspondence theoretic concept. Moreover, Thiselton (2006) explicitly claims that multiple conceptions may be found in scripture. These two facts cast some doubt, I think, on the idea that there is one single ‘biblical concept of truth’. Indeed, one might reasonably be sceptical that the Bible offers a robust enough set of linguistic data to determine any such thing as the biblical concept of truth. Even if it is true that ‘the biblical concept of truth’ is connected with reliability in the way described above, however, it must further be noted that this is a different notion of reliability from the one I am invoking here—which is, again, a notion that is partly definable in terms of truth.

  

75

and biblical truthfulness—have been rather milquetoast. In short: Attributions of authority, depending on what exactly they amount to, might require a reasonably high degree of reliability in some interesting domain, but (so far) nothing like reliability across the board, and nothing even close to inerrancy. But religious believers are often inclined to make assumptions about God and God’s relation to the biblical text that, given what else I have said here, would justify stronger conclusions. Importantly, however, many of these assumptions are negotiable. Consider, for example, the following schema: (α) G is the author of T, and, necessarily, for any text τ authored by G and for any individual S other than G, τ has foundational authority over S in D. Many Christians are willing to assume that α is true in the case where ‘G’ is replaced with ‘God’, ‘T’ is replaced with ‘the Bible’, and ‘D’ is replaced with the name of some fairly large domain like ‘the domain of truths about God, morality, the human condition, and salvation’ or, much more generally, ‘the domain defined by the text itself ’—i.e. the domain which is just the conjunction of every proposition that is semantically asserted or intentionally conveyed by the text. And if some instance of α is true, it follows that whatever T says within D is true and whatever directives T issues within D constitute (for everyone) decisive reasons for action. In other words, T is perfectly reliable. Here is the proof: Consider a possible world W1 in which G produces a text T1 that is authoritative over some individual S (≠ G) but less than perfectly reliable in D. Then there is a possible world W2 that differs from W1 in no relevant respect apart from these facts: (a) W2 contains, in addition to T1, a text T2 that is both authoritative over S and perfectly reliable in D; and (b) S knows that T2 is more reliable than T1.³⁹ In W2, then, T2 is clearly more authoritative than T1: reasons supplied by T2 have priority over reasons supplied by T1. But if that is so, then it is possible that there be someone (other than G) over which T1 does not have foundational authority, contrary to α. Now let us consider the assumption, β, generated from α by substituting ‘God’ for ‘G’, ‘the Bible’ for ‘T’, and ‘the domain defined by the text itself ’ for ‘D’: (β) God is the author of the Bible, and, necessarily, for any text τ authored by God and for any individual S other than God, τ has foundational authority over S in the domain defined by the text itself. My fourth—and certainly more substantive—conclusion, then, is that β implies that the Bible is perfectly reliable within the domain defined by the text itself. ³⁹ I assume that texts must be of finite length. Thus, there will be no text whose content is an infinitely long conjunction that express in full detail the entire truth about a possible world.

76

   :  

Thus, if β is true, then every proposition that the Bible semantically asserts or intentionally conveys must be true, and all of its directives constitute decisive reason for action. Of course, β is not an implication of the bare claim that the Bible is authoritative, nor is it an implication of the bare claim that the Bible is true, nor is it even clearly implied by any passage in scripture. It is, it would seem, just a philosophical-theological doctrine. It is an interesting question whether its second conjunct in particular could be shown to follow from any of the central tenets of classical theism, or of some plausible alternative version of theism such as open theism or process theism. But this is a question that will not be taken up here. The fourth conclusion is naturally construed as drawing a connection between a certain view about the nature of the authority of scripture—the view captured by β—and the inerrancy of scripture. For a lot of evangelicals, the go-to document for an ‘official’ statement of the doctrine of inerrancy is the Chicago Statement on Inerrancy. That document does a lot more than merely define the doctrine; but according to the portion of the document that looks most like a definition, ‘Scripture is without error or fault in all its teaching, no less in what it states about God’s acts in creation, about the events of world history, and about its own literary origins under God, than in its witness to God’s saving grace in individual lives.’⁴⁰ This is pretty close to the claim that the Bible is perfectly reliable in whatever it semantically asserts or intentionally conveys, and also perfectly reliable in whatever directives it issues. If there is a difference at all, it will lie in the difference between the sum total of ‘what the Bible teaches’ on the one hand and, on the other hand, the sum total of whatever the Bible says to us, along with whatever directives it issues. The Bible teaches, for example, that Jesus told his apostles to go unto all nations, baptizing them in the name of the Father, Son, and Holy Spirit. I am not so sure, however, that it teaches that we are to do the same. If anything, by including the teaching that Jesus said this, the divine author of the Bible intentionally conveys to believers the directive that they are to do the same. But that is no more one of the Bible’s teachings than a command to ‘Stop!’ is one of a policeman’s teachings. Still, since I am inclined to think that believers in inerrancy would not want to exempt intentionally conveyed directives from the scope of what they take to be ‘without fault’ in scripture, I am inclined to think that this fourth conclusion here does indeed draw a connection between a certain view of biblical authority and the doctrine of biblical inerrancy. Fifth, it is not clear that any (non-question-begging) assumption weaker than β will forge the same link between authority and inerrancy. We have already seen that a text can be foundationally authoritative for an individual in a domain without being perfectly reliable in that domain. β manages to connect authority with perfect reliability because it implies that there is an author who is such that

⁴⁰ Geisler 1980: 494.

  

77

any text of which s/he is the author would have to be at least as reliable in some domain as a text that is perfectly reliable in that domain. But that implication holds only when the author is necessarily such as to be the author only of texts that have foundational authority of some kind. So it seems that those interested in maintaining a connection between scriptural authority and scriptural inerrancy will be best served by devoting their philosophical-theological energies to a defense of β. Moreover, it is worth noting that the inerrancy doctrine itself has relatively few implications about what we actually ought to believe in light of scripture. It is, after all, compatible with widely varying views about what (if any) propositional messages are asserted or conveyed by biblical texts. Suppose that what the biblical text says is just what the divine author of the text semantically asserts or intentionally conveys through the text. Then what the biblical text says depends heavily upon what God aims to do with it. If we think that God intends to teach us the sober historical facts about the genesis of life on Earth, the origins of Israel, and the conquest of Canaan, then we will have reason to take most of the declarative sentences in the Old Testament as genuinely asserting their semantic contents. If, on the other hand, we think that God’s authorship of the Old Testament amounts primarily to his having appropriated a variety of myths and hyperbolic tales to make rather general points about divine sovereignty, divine faithfulness, and the like, then we will have little reason at all to think that the same Old Testament texts say much of anything at all about the origins of life or of Israel, or about the conquest of Canaan. The claim that the Bible is perfectly reliable in whatever it says will imply that whatever it says is true; but it implies nothing at all about what the Bible actually says. As I see it, then—and this is the sixth and final conclusion—our views about the nature and scope of biblical authority shed, all by themselves, relatively little light on the most interesting questions about the truthfulness of problematic passages in scripture. Consequently, it is a mistake to treat the topic of biblical authority as somehow lying at the heart of debates about the reliability and inerrancy of scripture. Far more pertinent to these latter debates are questions about the nature of God and divine authorship: In what sense is God an (or the) author of scripture? What are God’s aims in scripture? What might be God’s aims in this or that part of scripture? Is God the sort of author about whom β is true? These are the questions that promise to shed the most light on the topics that really worry us. Of course, some of them—especially the last one—will involve us in questions about authority; but none of them are fundamentally about the nature of authority, and all of them seem to be questions different from those that have occupied so much of the literature.⁴¹ ⁴¹ This chapter has benefitted from discussion with my colleagues in this project (June 2010) and in the weekly discussion group of the Center for Philosophy of Religion at the University of Notre Dame in January 2011. For helpful comments I am particularly grateful to Billy Abraham, Alex Arnold,

78

   :  

References Achtemeier, Paul. 1999. Inspiration and Authority. Peabody, MA: Hendrickson Publishers. Alston, William P. 1996. A Realist Conception of Truth. Ithaca, NY: Cornell University Press. Bergmann, Michael. 2006. Justification without Awareness. Oxford: Oxford University Press. Black, Max. 1954/5. ‘Metaphor’. Proceedings of the Aristotelian Society, New Series 55: 273–94. Black, Max. 1977. ‘More about Metaphor’. Dialectica 31: 431–57. Blanshard, Brand. 1940. The Nature of Thought. New York: The MacMillan Company. Bryant, Robert. 1968. The Bible’s Authority Today. Minneapolis, MN: Augsburg. Camp, Elisabeth. 2005. ‘Critical Study of Josef Stern’s Metaphor in Context’. Noûs 39: 715–31. Camp, Elisabeth. 2006a. ‘Contextualism, Metaphor, and What is Said’. Mind and Language 21: 280–309. Camp, Elisabeth. 2006b. ‘Metaphor and that Certain “Je Ne Sais Quoi” ’. Philosophical Studies 129: 1–25. Cooper, David. 1986. Metaphor. Oxford: Oxford University Press. Davidson, Donald. 1979. ‘What Metaphors Mean’. In On Metaphor, edited by Sheldon Sacks, 29–46. Chicago, IL: University of Chicago Press. Devitt, Michael. 1991. Realism and Truth, 2nd edn. Oxford: Basil Blackwell. Geisler, Norman (ed.). 1980. Inerrancy. Grand Rapids, MI: Zondervan. Gnuse, Robert. 1985. The Authority of the Bible: Theories of Inspiration, Revelation and the Canon of Scripture. New York: Paulist Press. Hampton, Jean. 1998. The Authority of Reason. Cambridge: Cambridge University Press. James, William. 1907. ‘Pragmatism’s Conception of Truth’. Journal of Philosophy and Scientific Methods 4: 141–55. Kaufmann, Gordon. 1981. The Theological Imagination: Constructing the Concept of God. Philadelphia, PA: Westminster Press.

Robert Audi, Jim Beilby, Michael Bergmann, Henri Blocher, Christina Brinks Rea, Jeff Brower, Don Carson, Graham Cole, Dan Doriani, Paul Draper, Simon Gathercole, Paul Helm, Michael Hickson, Peter Jensen, Matthew Lee, Richard Lints, Phil Long, Tom McCall, Sam Newlands, Ryan Nichols, Todd Ryan, Amy Seymour, Jeff Snapper, Kevin Vanhoozer, Jerry Walls, and John Woodbridge. I have also benefitted from helpful conversations with and unpublished work by Luke Potter on matters related to the discussion of metaphor in Section 2 of this chapter.

  

79

MacIntyre, Alasdair. 1998. A Short History of Ethics: A History of Moral Philosophy from the Homeric Age to the Twentieth Century. Notre Dame, IN: University of Notre Dame Press. Moo, Douglas (ed.). 1997. Biblical Authority: Conservative Perspectives. Grand Rapids, MI: Kregel. Murphy, Mark. 2006. ‘Authority’. In The Encyclopedia of Philosophy, edited by Donald Borchert, 412–18, 2nd edn. Detroit, MI: MacMillan Reference. Nicole, Roger. 1983. ‘The Biblical Concept of Truth’. In Scripture and Truth, edited by D. A. Carson and John Woodbridge, 287–98. Leicester: InterVarsity Press. Peirce, C. S. 1878. ‘How to Make Our Ideas Clear’. Reprinted in The Essential Peirce, edited by N. Houser and C. Kloesel, 124–41. Bloomington: Indiana University Press. Pinnock, Clark. 2006. The Scripture Principle: Reclaiming the Full Authority of the Bible, 2nd edn. Grand Rapids, MI: Baker Academic. Plantinga, Alvin. 1982. ‘How to be an Anti-Realist’. Proceedings and Addresses of the American Philosophical Association, September. Plantinga, Alvin. 2000. Warranted Christian Belief. New York: Oxford University Press. Pollock, John L. and Joseph Cruz. 1999. Contemporary Theories of Knowledge. Lanham, MD: Rowman & Littlefield. Raz, Joseph. 1979. The Authority of Law. Oxford: Oxford University Press. Rea, Michael. 2000. ‘Theism and Epistemic Truth-Equivalences’. Noûs 34: 291–301. Rea, Michael. 2001. World without Design: The Ontological Consequences of Naturalism. Oxford: Oxford University Press. Rogers, Jack and Donald McKim. 1979. The Authority and Interpretation of the Bible: An Historical Approach. San Francisco, CA: Harper & Row. Soskice, Janet Martin. 1985. Metaphor and Religious Language. Oxford: Oxford University Press. Stern, Josef. 2000. Metaphor in Context. Cambridge, MA: MIT Press. Stern, Josef. 2006. ‘Metaphor, Literal, Literalism’. Mind and Language 21: 243–79. Stonehouse, N. B. 1967. ‘The Authority of the New Testament’. In The Infallible Word: A Symposium by the Members of the Faculty of Westminster Theological Seminary, edited by N. B. Stonehouse and Paul Wooley, 92–140, 2nd edn. Philadelphia, PA: P&R Publishing. Thiselton, Anthony C. 2006. ‘Does Lexicographical Research Yield “Hebrew” and “Greek” Concepts of Truth? (1978), and How Does This Research Relate to Notions of Truth Today (new summary)’. In Thiselton on Hermeneutics: Collected Works and New Essays, 267–86. Grand Rapids, MI: Eerdmans. Vanhoozer, Kevin. 1998. Is There Meaning in this Text? The Bible, the Reader, and the Morality of Literary Knowledge. Grand Rapids, MI: Zondervan.

80

   :  

Woodbridge, John. 2016. ‘German Pietism and Scriptural Authority: The Question of Biblical Inerrancy’. In The Enduring Authority of the Christian Scriptures, edited by D. A. Carson, 137–70. Grand Rapids, MI: Eerdmans. Woodbridge, John and Frederick James, III. 2013. Church History, Volume Two: From Pre-Reformation to the Present Day: The Rise and Growth of the Church in Its Cultural, Intellectual, and Political Context. Grand Rapids, MI: Zondervan. Young, Edward J. 1967. ‘The Authority of the Old Testament’. In The Infallible Word: A Symposium by the Members of the Faculty of Westminster Theological Seminary, edited by N. B. Stonehouse and Paul Wooley, 55–91, 2nd edn. Philadelphia, PA: P&R Publishing.

4 Divine Attributes as a Topic in Analytic Theology The attributes of God have been among the staple topics in philosophy of religion since its revival in the 1960s. There is now a vast literature on the nature of omnipotence, omniscience, timelessness, simplicity, aseity, and a variety of other attributes typically thought to characterize the very nature of God. The conversations about divine attributes taking place in the philosophical literature today are heavily informed by the thought of medieval philosopher-theologians, Early Modern philosophers, and contemporary (analytic) metaphysicians. There has been remarkably less engagement, however—indeed, virtually no engagement— with the work of twentieth- and twenty-first-century theologians. More exactly, there has been little or no such engagement in the literature in analytic philosophy of religion. This state of affairs would make sense if, say, contemporary theologians had simply given up talking about the divine attributes. But they haven’t. On the contrary, theologians like Barth, Tillich, Pannenberg, Jenson, Daly, McFague, Hampson, and many others have had quite a lot to say on the topic of divine attributes; yet their voices have been strikingly absent (or at least marginalized) from the conversation in analytic philosophy of religion. Why so? Whence the divergence of these two disciplines on such a rich and conceptually central topic? Some of the reasons, of course, are sociological. Set those aside. What I want to focus on are some of the distinguishing features of the way in which the topic of divine attributes is approached in analytic philosophy of religion as contrasted with the way(s) in which many contemporary theologians are inclined to approach it. Naturally, I shall be painting with broad strokes; there will be many exceptions to the generalizations that I make about both philosophy and theology. Nevertheless, I think that some useful generalizations can fairly be made, and that these go some distance both towards explaining the divergence between the two disciplines and towards showing what must be done to bring them back into conversation with one another. The analytic philosophical literature on the divine attributes is, by and large, paradigmatic of what is now known as analytic theology (Crisp and Rea 2011). Analytic theology differs from other forms of theology primarily in its methodology: its ambitions, its style, its conversation partners, and so on. This, furthermore, is where the most interesting differences between analytic philosophical discussions of the divine attributes and contemporary theological discussions of Essays in Analytic Theology: Volume I. Michael C. Rea, Oxford University Press (2020). © Michael C. Rea. DOI: 10.1093/oso/9780198866800.003.0004

84

   :  

that topic are to be found. Indeed, my main positive thesis in this chapter is that the most distinctive features of the approach to divine attributes that one finds in the analytic philosophical literature, those features most commonly highlighted by the many and various critics of that approach, are simply instances of more general distinctives of analytic theology. What we have, then, in the analytic philosophical literature on divine attributes is simply an instance of a particular type of theology—one which can be given a fairly general characterization and assessed on its merits in comparison and contrast with other types of theology. This chapter unfolds as follows. In Section 1, I briefly recap my preferred characterization of analytic theology. In Section 2, I try to locate it as a type of theology within the five-category taxonomy offered by Hans Frei (1994). In Section 3, I discuss and defend the main positive thesis described in the previous paragraph. In so doing, I highlight what I take to be the most distinctive features of an analytic philosophical approach to the divine attributes. In Section 4, I comment on the relationships between analytic theology, ontotheology, and feminist theology. In so doing, I note some commonly criticized features of analytic theology that I think are wholly inessential to it. The end result, I hope, will be a clearer picture both of the nature of analytic theology in general and of the distinctive character of an analytic approach to the topic of divine attributes.

1. It will be helpful to begin by saying something about the nature of analytic philosophy. In the introduction to Analytic Theology: New Essays in the Philosophy of Theology, I said that, roughly speaking, analytic philosophy is an approach to philosophical topics characterized by a certain rhetorical style, a set of common ambitions, an evolving technical vocabulary, and a tendency to pursue one’s projects in conversation with a particular evolving body of literature. I identified two ambitions as characteristic of analytic philosophy: (i) to identify the scope and limits of our powers to obtain knowledge of the world, and (ii) to provide such true explanatory theories as we can for non-scientific phenomena. I characterized the rhetorical style by saying that paradigmatic instances of it conform (more or less) to the following five prescriptions: (1) Write as if philosophical positions and conclusions can be adequately formulated in sentences that can be formalized and logically manipulated. (2) Prioritize precision, clarity, and logical coherence. (3) Avoid substantive (non-decorative) use of metaphor and other tropes whose semantic content outstrips their propositional content. (4) Work as much as possible with well-understood primitive concepts, and concepts that can be analysed in terms of those. (5) Treat conceptual analysis (insofar as it is possible) as a source of evidence.

       

85

I focused on these five prescriptions because, unlike other prescriptions in accord with which analytic philosophers tend to write, these five are prescriptions that those working outside the tradition of analytic philosophy typically aim to violate, often for principled reasons. Analytic theology, then, is theology done with the ambitions characteristic of analytic philosophy in a style that conforms to the prescriptions that are distinctive of analytic philosophical discourse; and it is theology that is informed by the same evolving body of literature that the analytic philosophers are themselves trying to engage. In analytic philosophy of religion, the topic of divine attributes has been approached in precisely this way. For the most part, the attribute of perfection has been held up as the attribute around which all of our other thought ought to be structured. Philosophers in the analytic tradition (following upon the heels of philosopher-theologians working under the influence of Plato and Aristotle) have felt comfortable making robust theoretical claims about the nature of God on the basis of intuitions about the correct analysis of perfection. They have also felt comfortable making theoretical claims on the basis of intuitions concepts like perfect power, perfect freedom, perfect goodness, and so on. In the contemporary literature, discussions of the divine attributes have drawn heavily on relevant other literature in contemporary analytic metaphysics, action theory, moral theory, and the like. The goal has generally been to arrive at maximally clear and precise accounts of the nature of perfections like omnipotence, omniscience, eternity, and so on, often with an eye to resolving apparent tensions between the attributes or solving conceptual puzzles about them. In short, then, one looking for paradigmatic examples of analytic theology need look no further than the contemporary analytic philosophical literature on divine attributes. That said, one might wonder how analytic theology relates to analytic philosophy of religion. Are they the same discipline under different names, or are there real distinctions to be drawn? One might also wonder how analytic theology relates to various other types of theology. Should Barthians view analytic theology with suspicion or hostility? Is it of a piece with Gordon Kaufman’s ‘ordinary language’ approach to theology? Is it a form of ontotheology? Is it inherently opposed to feminist theology (and perhaps in the same ways in which analytic philosophy of religion is alleged to be)? The characterization I offered in the introduction to Analytic Theology does not explicitly address any of these questions. But we can make strides towards answering some of them by examining the typology of theologies offered by Hans Frei in his posthumously published Types of Theology.¹ Doing so will also provide us with

¹ Frei 1994. I make use of Frei’s taxonomy because doing so provides a convenient way of framing the discussion I want to have about analytic theology’s relationship to certain other ways of doing theology. I don’t intend here to endorse Frei’s taxonomy; nor do I endorse his readings of the various figures he takes as paradigm instances of the different types of theology that he identifies.

86

   :  

resources for explaining what is distinctive—and what is not—about analytic philosophical approaches to the specific topic of divine attributes.

2. Frei begins by identifying two different but not mutually exclusive conceptions of what theology is, or (as I would prefer to put it) of the task that is most fundamental to theology. He then distinguishes five different types of theology on the basis of their different attitudes towards these two ways of thinking about theology. In so doing, he also distinguishes several different views about the relationship between philosophy and theology, which discussion bears obvious relevance to the task of the present chapter. On one conception, (Christian) theology is an ‘academic discipline’; its fundamental task is the global, holistic task of locating particular doctrinal claims and perhaps the entire body of Christian discourse within a larger body of scientific and philosophical theory.² The business of theology, on this conception, is to determine, at least partly in light of what we know from other disciplines, the nature and function of Christian doctrines and, where appropriate, the meaning and truth value of particular doctrinal claims. On the other conception, theology is ‘Christian self-description’; its fundamental task is the local, internal task of systematizing Christian doctrine, elucidating the relationships between particular doctrinal claims, and, where appropriate, forging or explaining logical and normative connections between doctrine and experience or between doctrine and religious practice. Although I (following Frei) have presented these two conceptions as embodying distinct visions of the fundamental task of theology, it is obvious enough that one might regard both tasks as equally fundamental to theology. Alternatively, one might take one task as fundamental and the other as important but subordinated to the first; or one might insist that one task is fundamental to theology and the other does not belong to theology properly speaking at all. Within Frei’s taxonomy of theologies, the first two types take the holistic task to be fundamental and the internal task to be subordinate; the third places both on a par with one another; the fourth reverses the priority relations, while preserving the idea that both have a place within theology; and the fifth denies the holistic task any importance to theology whatsoever. Since I think that it is evident from my earlier descriptions that the holistic task will be of some importance to the analytic theologian, I shall ignore Type 5 in the discussion that follows. The remainder I shall discuss in turn. ² For purposes of Frei’s discussion—and so here too—the term ‘theology’ just refers to Christian theology.

       

87

In the first type of theology, the holistic task takes ‘complete priority’—so much so that the internal task is subsumed under the holistic as just one part of the overall endeavour. Gordon Kaufman is Frei’s paradigm for this type of theology. According to Kaufman, says Frei, ‘the task of the theologian is to search out the rules governing the use of the word or concept “God” as the organizing focus for a whole vocabulary’ (1994: 28). Executing the internal task, on this conception, would involve examining the way in which key theological terms are used within the Christian community; executing the holistic task would involve examining the way in which those same terms are used in the broader culture. One’s overall theory about the nature, function, meaning, and truth conditions of particular theological claims would then be crafted to accommodate all of these data. It is in just this way that the internal task is subsumed under the holistic task. On Kaufman’s view, it is a mistake to see theology as ‘essentially a work of the church’, and it is a mistake to think that ‘theologians can or must accept without question the “faith” of the church (as found in scripture or dogma)’ (Kaufman 1995: 3). The business of theology is to understand the nature and function of the concept of God; and in doing so it ‘works largely with public, not private or parochial materials’ and ‘is not restricted either to the language and traditions of a particular esoteric community (the church) or to the particular experience of unusual individuals’ (1995: 11). On this way of thinking about theology, one could not sensibly develop a doctrine of God simply by (for example) looking at what scripture and tradition have said about God (e.g. that God is omnipotent, omniscient, a se, and so on) and then trying to explicate those claims in a way that is clear, precise, systematic, and relevantly informed by our best theories in metaphysics, ethics, the natural sciences, and so on. To do that would be closer to working only with ‘private or parochial materials’. Rather, one would have to look at the concept of God as it appears across traditions and in ordinary discourse, and one would want to avoid giving priority to the use of that concept in any one particular tradition over its use in other traditions. As Kaufman sees it, the task of understanding the nature and function of the concept of God is fundamentally a philosophical project rather than, say, a project in empirical psychology or linguistics. Thus, for Kaufman, theology seems to be just philosophical reflection on theological vocabulary and concepts. He characterizes it as the ‘disciplined effort to see what we are trying to do and say with these complexes of meaning [terms like “holy”, “divine”, “sacred”, “transcendent”, and so on] so as to enable us to say and do them better’ (1995: 9). Thus construed, theology has a ‘revisionary’ aspect to it.³ It is not merely a descriptive project; rather, it is constructive, and not in a way that is constrained by the dogmatic commitments of any particular body of religious doctrine.

³ Cf. the distinction between ‘descriptive’ and ‘revisionary’ metaphysics in Strawson 1959.

88

   :  

Kaufman furthermore thinks that once we have understood the nature and function of the concept of God, we can see that it cannot sensibly be treated as the concept of an ‘object’ of the sort studied by the sciences: ‘a reality existing over against and independently of the investigator, a reality to be perceived, studied, analyzed, theorized about, understood’ (1995: 25). But I think that this view is a result of his employment of his own theological method rather than a part of the method itself. Consequently, I am reluctant to say that it is any part of Type 1 theology as such to deny that theology investigates an independent reality that can be studied, analysed, theorized about, and understood. There are a few similarities between Kaufmanian theology and analytic theology; but the differences, I think, are of greater importance. The most salient similarity is that Kaufmanian theology and analytic theology both treat philosophy and theology as methodologically continuous with one another. However, in contrast to Kaufmanian theology, I do not think that it is any part of analytic theology to insist that the primary business of theology is to understand the nature and function of theological concepts. Nor do I think that in trying to understand theological concepts (like the concept of God) theologians must refrain from prioritizing the claims of one tradition over the claims of others as ‘data’ for their theorizing. Analytic theology can be done without assuming that the claims of any one faith tradition take priority over the claims of others. But it can equally be done under the assumption that the ‘faith of the church’ as revealed in scripture and interpreted within some particular tradition of faith is to be taken as read. Thus, whereas one would naturally expect Kaufmanian theology to be highly revisionary from the point of view of any given faith tradition, one would not necessarily expect an analytic theological project to be so revisionary. Frei’s paradigm for the second type of theology is the approach of David Tracy as described in his Blessed Rage for Order (Tracy 1975). According to Frei, Tracy, like Kaufmann, sees theology as primarily an academic discipline whose fundamental task is to locate Christian theological claims within a larger body of theory; but for Tracy, unlike Kaufman, the larger body of theory is not what would result from tradition-transcendent philosophical reflection on the nature and function of the concept of God, but rather from an attempt to ‘critically correlate’ the results of two separate kinds of investigation into what he sees as the two sources for theology. In Tracy’s own words: [T]here are two sources for theology (common human experience and language, and Christian texts); those two sources are to be investigated by a hermeneutic phenomenology of the religious dimension in common human experience and language and by historical and hermeneutical investigation of the meanings referred to by Christian texts; the results of these investigations should be correlated to determine their significant similarities and differences and their truth-values. (1975: 53)

       

89

As indicated at the end of the quoted passage, however, mere correlation of claims is not all that the theologian is ultimately interested in; for even if the completion of the correlative project results in our knowing ‘the basic religious referents of Christian texts and the fundamental meaningfulness of religious and theistic categories for common experience’ (54), it will not tell us anything about the truth-values of our theological claims. For that, he says, we need metaphysics. In particular, we need to investigate the ‘ground in our common experience for having any notions of religion and God at all [and] how these notions may be conceptually explicated to avoid both vagueness and incoherence’ (55). Moreover, on his view, the task of the theologian is complete only when this ‘explicitly metaphysical’ task has been carried out (56). Tracy and Kaufman differ over what it would look like to execute the holistic task. Furthermore, on Tracy’s approach but not Kaufman’s, a specific philosophical framework—the methods and presuppositions of phenomenology and contemporary hermeneutic theory, and the methods and results of ‘transcendental’ or ‘metaphysical’ reflection—is set up to serve as a kind of foundation for Christian theology. These differences aside, however, there seems to be little else separating Type 1 from Type 2. In both types, the internal and holistic projects converge; in both types the holistic project is the one primarily in view; both types are revisionary in the sense that they do not take theology to be constrained by particular scriptural or dogmatic claims; and in both types theology and philosophy are methodologically continuous.⁴ By contrast with the first two types, Type 3 does not prioritize the holistic task over the internal task. Rather, in this type of theology, the two tasks are treated as being of equal importance. For Frei, the paradigm for this type of theology is Schleiermacher, though he introduces his discussion of Type 3 by saying that ‘we have it only in ambiguous form because its greatest representative [Schleiermacher] may not fit it’ (1994: 34). There is no question in his mind that the internal task is central to theology for Schleiermacher; the question rather concerns the centrality (and nature) of the holistic task. For Schleiermacher, all dogmatic claims are ultimately just expressions of ‘pious emotions or states of consciousness’ (Dole 2011: 140); but some are more transparently so than others. Executing the internal task is a matter of reducing the dogmatic claims that are not transparent expressions of pious emotions and states of consciousness to ones that are.⁵ Executing the holistic task, on the other hand, is (Frei seems to think) a matter of theorizing about the nature of Christian dogmatic claims and locating ⁴ Tracy’s own brand of theology is perhaps less revisionary just in that the correlative task that he proposes is more likely to issue in revisionary interpretations of particular scriptural and dogmatic claims rather than outright rejections of them; but, as with Kaufman’s theological anti-realism, this seems to me to be more an outgrowth of the particular applications of the two methodologies than a difference between the methods themselves. ⁵ For my understanding of Schleiermacher on these matters, I am primarily indebted to Dole 2011.

90

   :  

them in relation to other forms of expression or self-consciousness. Both, on this reading, are tasks for theologians; both are important; and both are relatively autonomous from one another. There are some important contrasts to be drawn between theology as Schleiermacher conceives it and analytic theology.⁶ But for purposes here I shall leave those aside since, as I have already mentioned, even Frei is ambivalent about taking Schleiermacher as a representative of Type 3 theology. Rather, what I want to note is simply that one need not think of the holistic and internal tasks as autonomous in the way that Schleiermacher allegedly does in order to fall within this type; nor, obviously, need one construe the tasks in the particular ways that Schleiermacher does. Both facts are important because, though I see little resemblance between analytic theology and theology as Schleiermacher conceives it, I think that (with one crucial caveat in place) Frei’s Type 3 is the best location for analytic theology within his taxonomy. The caveat is as follows: Rather than saying that the theological approaches within Type 3 give equal priority to the internal task and the holistic task, I would prefer to say simply that these approaches refrain from prioritizing one task over the other. They are neutral with respect to the priority relations between the two tasks. In other words, Type 3 approaches as such carry no commitment to the priority relations between the internal task and the external task; so an individual theologian can count as a practitioner of a particular Type 3 approach regardless of his or her own private views about relative importance of those two tasks. With this caveat in place, I think that Type 3 is the natural home for analytic theology. For the analytic theologian, executing the internal task will involve exploring the logical relations among distinctively Christian claims, systematizing them in relation to one another, and trying to express them with greater clarity and precision. Executing the holistic task will involve expanding, explaining, and evaluating distinctively Christian doctrines in light of theories in contemporary analytic metaphysics, ethics, philosophy of science, and so on. These tasks can be executed in tandem; moreover, it seems to be no part of analytic theology as such to treat either task as more fundamental than the other. To be sure, particular analytic theologians might emphasize one task to the relative exclusion of the other; but from the point of view of the methodology itself, no priority relationship is specified. Likewise, as we have already mentioned, it is no part of analytic theology to prioritize or refrain from prioritizing one source of ‘data’ over other sources. In particular, analytic theology as such does not prioritize reason over revelation (or vice versa); nor does it necessarily prioritize (or de-prioritize) the claims of one religious tradition relative to others. Particular analytic theologians might conduct

⁶ Cf. Dole 2011.

       

91

their business under varying different assumptions about what takes priority over what; but the methodology itself makes no such prioritizations. Finally, and importantly, it is no part of analytic theology as such to try to provide a ‘philosophical foundation’ for theology. That is, in contrast to David Tracy (as Frei characterizes his views), the analytic theologian as such is not antecedently committed to a set of philosophical principles that are supposed to ground her theological claims, or serve as a lens through which those claims are to be viewed and understood. Of course, particular analytic theologians might bring their own particular philosophical commitments to the table as they go about their theologizing; but those commitments will belong just to the theologians themselves and not to their methodology. On this latter point—i.e. the point that theology stands in no need of philosophical grounding—the analytic theologian can be in full agreement with Karl Barth, whose approach to theology is Frei’s paradigm instance of Type 4. On the question of how the internal and holistic tasks are to be prioritized, however, the analytic theologian is as much at odds with Barth as with Kaufman and Tracy. I hope to make it clear, however, that the opposition between analytic theology and Barthian theology (or, for that matter, Kaufmanian theology) on this score is much shallower than it at first appears. Whereas Kaufman gives complete priority to the holistic task, Barth, on Frei’s reading, gives complete priority to the internal task. Whereas Kaufman denies that theology is essentially a work of the Church, Barth affirms this unequivocally. As Frei puts it: Theology as specific and critical Christian self-description and self-examination by the Church of its language takes absolute priority over theology as an academic discipline. Philosophy as conceptual system describing and referring to ‘reality’ is not a basis on which to build theology, and even philosophy as a set of formal, universal rules or critiera for what may count as coherent and true in Christian discourse as in every other kind of conceptual practice is not basic to or foundational of Christian theology. (1994: 40)

Interestingly, theology on this conception does have a revisionary aspect; but, as Barth sees it, rather than revise her theological beliefs in light of her theorizing about non-theological matters, the theologian ought to be prepared to revise her views on non-theological topics in light of her theorizing about theological matters. Thus, for example, instead of allowing one’s views about fatherhood to colour one’s views about God (construed as our heavenly Father), one ought to let the Christian view of God colour one’s views about what it means to be a good and loving father. As I said in discussing Type 3, it is no part of analytic theology as such to endorse this Barthian view about the priority of the internal task over the holistic

92

   :  

task. But neither is it part of the methodology to deny the Barthian view. Rather, it can be pursued equally under that view or its denial. In this way, then, I think that the opposition between analytic theology and the Barthian approach to theology is merely superficial.

3. In light of the foregoing, it is easy enough to assemble a sketch demonstrating what it means to approach the topic of divine attributes in analytic theological mode. Positive analytic theological treatment of this topic will be a robust theorybuilding enterprise that aims for clear, precise characterizations of each of the divine attributes, seeks to understand the divine attributes in terms of familiar categories, and takes very seriously the idea that we can learn about the divine attributes (and therefore about God) via analysis of the concepts involved in our best descriptions of the divine attributes. Importantly, such an approach may or may not take claims from scripture or tradition as regulative for the endeavour; and it may or may not proceed under the influence of some particular metaphysical or moral theory. A Barthian analytic theologian might, for example, develop a metaphysic of freedom in light of systematic-theological claims about divine freedom, or she might develop a theory about love (in general) on the basis of what can be known about divine love. As I have already mentioned, the divine attributes in analytic philosophy of religion have most often been discussed under the assumption that the attribute of perfection is the one around which all of our other thought ought to be structured. Interestingly, it has also focused much more predominately upon what Barth calls ‘the perfections of the divine freedom’ (especially aseity, omnipotence, eternity, and unity), almost to the exclusion of ‘the perfections of the divine loving’ (patience, mercy, grace, wisdom). This is partly because a great deal of analytic philosophy of religion has been done under classical theistic suppositions about the nature of God, and classical theism seems to place greater theoretical emphasis on the former attributes. But it is also, I think, partly because properties like aseity and omnipotence have generally been regarded as more amenable to analytic philosophical treatment than properties like (perfect) wisdom or mercy. There is another point of contrast with Barth worth highlighting. Barth, as we have seen, advocates thinking about who God is and allowing one’s views about God and scriptural revelation to condition our understanding of ordinary properties, like fatherhood, immutability, love, and so on. In other words, on Barth’s approach we learn from scripture that God is loving, or unchanging; and then we seek to understand being loving and being unchanging in light of what else we know about God, rather than the other way around. But in analytic philosophy of

       

93

religion the order of understanding typically goes the other way around: we reflect on what we know about power and perfection in order to understand divine omnipotence; we reflect on what we know about love and moral goodness in order to understand divine omnibenevolence; we use our beliefs about (ideal) human fathers in order to arrive at theses about how our heavenly Father can be expected to interact with us; and so on. However, if I am right in thinking that analytic philosophical treatment of the divine attributes is co-extensive with analytic theological treatment of the same topic, and if what I have said in Sections 1 and 2 about the nature of analytic theology is correct, then all of these apparent distinctives of the analytic approach to divine attributes are of mere sociological significance. There is no reason internal to analytic theology as such for thinking that perfection has to be seen as the controlling attribute; there is no reason for thinking that the perfections of the divine loving should be de-prioritized in an analytic-theological doctrine of God; and there is no reason for thinking that the ‘order of understanding’ should proceed from ordinary predications to divine predications rather than the other way around.

4. I want to close now by making a few remarks about the way in which analytic theology relates to ontotheology and feminist theology. In a recent essay on the relationship between analytic theology and the academic study of religion, William Wood speculates that the most salient concern about analytic theology is its alleged involvement with ontotheology: I think the real worry is . . . that analytic theology is necessarily guilty of ontotheology. This is a difficult objection, not least because it isn’t entirely clear what ontotheology even is or what it would be to be guilty of it, but I think the most salient form of the charge goes something like this: If the goal of analytic theology is to use the tools of human reason to come to know God better, then surely that kind of inquiry presupposes that God can be known, which in turn presupposes that God is an object in the world, fundamentally like us but bigger and more powerful, etc. A related form of the objection would argue: If the goal of analytic theology is to give philosophical explanations of Christian doctrines, then that presupposes that doctrines are like scientific theories, fit to be reductively analyzed, rather than mysteries, or proper objects of faith. (Wood 2013: 23)

Kevin Vanhoozer raises similar worries. He asks, ‘Can one by doing metaphysics find out God?’ In reply, he says:

94

   :   Not if by metaphysics one means speculation that begins “from below”, with human experience, and seeks through a process of incremental and inferential reasoning to arrive at conclusions about what God ‘above’ must be like. The problem with “totalizing” metaphysics is the underlying assumption that there is one set of categories, accessible to unaided human reason, which applies both to the world and to God, created and uncreated reality. This inevitably leads to ontotheology, a unified system of thought that employs concepts such as Supreme Being or Unmoved Mover as conceptual stopgaps to prevent infinite metaphysical regress. [note omitted] Call it “bad” metaphysics: bad, because it imposes a system of categories on God without attending to God’s own selfcommunication. (Vanhoozer 2010: 8)

Vanhoozer furthermore seems to think that ontotheology is inextricably entwined with perfect being theology which he takes, in turn, to be the dominant mode of thinking about God and God’s attributes in analytic philosophy of religion (2010: 94–102). Hence the indictment of the latter: it is up to its neck in ‘bad metaphysics’. This particular indictment of ontotheology harks back to Heidegger. Metaphysics, Heidegger observes, is concerned with being. It is concerned with the being of beings (i.e. ordinary things in the world) and also with their ground, Being itself, the highest being, which Heidegger identifies with the ‘god of philosophy’ (Heidegger 2002: 70–2). Its concern with the being of beings is the ontological aspect of metaphysics; its concern with Being is its theological aspect; hence the term ontotheology for metaphysics thus construed. But the problem with (ontotheological) metaphysics is just this. Insofar as it seeks to ‘represent beings as such’, it does so ‘with an eye to their most universal traits’ and ‘only with an eye to that aspect of them that has already manifested itself in being’ (Heidegger 1998: 287, 288). Here, then, we may identify a guiding assumption: to investigate something in metaphysical mode is to do so under the supposition that its very essence can be understood in terms of universal characteristics that have been made manifest in beings—i.e. mundane things. But the ultimate ground, Being, is not a mundane thing; and so the supposition (central to ontotheology) that it can be understood via concepts and categories crafted for understanding beings is suspect. Thus one finds Heidegger speaking of overcoming metaphysics, a goal that is accomplished just when one manages to ‘think the truth of Being’ (1998: 279). Thus, too, one finds the followers of Heidegger associating ontotheological metaphysics with idolatry. (Cf. Marion 1994; for discussion of the ‘idolatry’ objection against metaphysics, see Hector 2011 and Rea 2015.) I think that it is an interesting question whether ontotheology is really as problematic as Heidegger and those in his wake make it out to be. Vanhoozer asserts that metaphysics from below inevitably results in ontotheology (though we should note that he seems to be using the term in a way somewhat different from Heidegger, associating it with the results of a certain kind of metaphysics rather

       

95

than with the activity itself); and he implies that it inevitably fails to attend to God’s own self-communication. But why should we think this? And why should we think that it is a problem to suppose that concepts and categories crafted for understanding mundane things also apply to the supra-mundane? The answers to these questions are not entirely clear. But defending ontotheology is not my project here. Rather, all I want to say here is that analytic theology as such need not involve ontotheology as it has been characterized above. Nothing in the analytic approach to theology requires one to think that divine attributes are mere idealizations of familiar creaturely attributes; nothing in the analytic approach even requires one to think that God’s being is on a par with creaturely being. (In the contemporary literature in analytic metaphysics there is much debate over the question whether ‘being’ is univocal.) The Heideggerian emphasis on divine transcendence is rather at odds with the method of perfect being theology which (as I have already indicated) has tended to dominate the analytic philosophical literature on divine attributes. But there is no reason in principle why analytic theologians could not structure their thinking about God in some other way—perhaps taking divine holiness, or transcendence, rather than perfection as the structuring attribute. Analytic philosophy of religion has also come in for a great deal of criticism as a ‘masculinist’ enterprise, one which tends to conceive of God in ways fundamentally at odds with feminist values. There are various ways in which this is alleged to be so; but with regard to the topic of divine attributes in particular, the objection is usually that the God of analytic philosophy of religion (and of classical monotheism more generally) is not only commonly referred to by masculine pronouns and designations like ‘father’, but ‘he’ is also typically conceptualized in a way that idealizes stereotypically male traits or interests rather than stereotypically female ones. Following upon some illustrative quotations from analytic philosophers like Richard Swinburne, Richard Gale, and A. E. Taylor, Daphne Hampson offers the following characteristic statement of the problem: God is conceptualized as ‘free’, ‘able to do anything’, ‘autonomous’, as acting ‘according to no discernible rule or law’, ‘the source of everything other than itself ’, not needing to ‘take into account anything else whatsoever’, an ‘object of unqualified adoration or worship’. What a nightmare! God would seem to be the reflection of the male’s wildest dreams. Even more interesting perhaps than that ‘he’ is powerful . . . is that God is self-sufficient, alone, not in any way constrained, and not needing to take into account what is other than ‘himself ’. Indeed, classically God has been held to have aseity, that is to be a se, complete in God’s self . . . . God . . . represents that which is perfect and powerful, while humanity is by contrast weak and sinful. Consequently humanity . . . is often conceptualized as ‘female’ in relation to a God for whom male metaphors are employed. Monotheism is thus instrumental in consolidating a gendered conception of reality. (Hampson 1996: 124–5)

96

   :  

Similar statements specifically indicting analytic philosophy of religion can be found in the work of Grace Jantzen, Pamela Anderson, and others. Let us, for the sake of argument, stipulate that Hampson and others are correct in criticizing the classical theistic vision of God as problematically gendering God as male and privileging distinctively masculine ideals and interests over feminine ones. Still, as with the critique of ontotheology, what we have here is a critique that applies at most to a view about God and the attributes of God that have been common among those taking an analytic philosophical approach to the divine attribute but that are by no means essential to that approach. Indeed, these presuppositions are not even essential to the method of perfect being theology, though they have been entirely common among those employing that methodology. One who shares Hampson’s understanding of the inherent masculinity of certain classical-theistic divine attributes might simply reject the idea that a perfect being would be omnipotent, a se, unconstrained, and so on. Alternatively, she might employ the method of perfect being theology to defend the conclusion that a perfect being would either lack gender or would belong equally to all genders.⁷ Thus, like the critics of ontotheology, feminist critics of analytic theology are best seen as criticizing views that are not essential to analytic theology as such but rather are simply widely held among those who tend to practise analytic theology.

5. Conclusion Towards the beginning of this chapter, I noted the lack of engagement between analytic philosophical discussions of the divine attributes and the diverse body of work in contemporary theology on the very same topic. I then asked why we find such a sharp divergence between the two disciplines. I hope that by now we are starting to see at least the beginnings of an answer. Analytic theologians as such bring to the table certain theoretical ambitions, a particular rhetorical style, and a disposition to try to understand theological topics in light of the terminology and results of contemporary analytic metaphysics, epistemology, ethics, and so on. They have also, however, tended to bring certain extraneous philosophical assumptions to the table: assumptions characteristic of classical theism, for example, or of perfect being theology. The assumptions are not essential to the method; but they have served as obstacles to conversation with theologians who themselves bring to the table a fairly robust set of contrary philosophical presuppositions. What I hope to have gone some distance towards showing in this chapter, however, is that the opposition between analytic theology

⁷ Cf. Rea 2016.

       

97

and other types of theology is much shallower than it initially appears. I do not claim that there is no opposition whatsoever; but it is easy enough to imagine analytic theologians discussing topics like the divine attributes in a way that is quite congenial to Kaufmanians, Barthians, feminist theologians, and a variety of others.

References Crisp, Oliver D., and Michael C. Rea (eds). 2011. Analytic Theology: New Essays in the Philosophy of Theology. Oxford; New York: Oxford University Press. Dole, Andrew. 2011. ‘Schleiermacher’s Theological Anti-Realism’. In Analytic Theology: New Essays in the Philosophy of Theology, edited by Oliver D. Crisp and Michael C. Rea, 136–54. Oxford; New York: Oxford University Press. Frei, Hans W. 1994. Types of Christian Theology. Edited by George Hunsinger and William C. Placher. Reissue edn. New Haven, CT: Yale University Press. Hampson, Daphne. 1996. After Christianity, 1st North American edn. Valley Forge, PA: Trinity Press International. Hector, Kevin. 2011. Theology without Metaphysics: God, Language and the Spirit of Recognition. Cambridge: Cambridge University Press. Heidegger, Martin. 1998. ‘Introduction to “What Is Metaphysics”?’ In Pathmarks, edited by William McNeill. Cambridge: Cambridge University Press. Heidegger, Martin. 2002. Identity and Difference. Translated by Joan Stambaugh. Chicago, IL: University of Chicago Press. Kaufman, Gordon D. 1995. An Essay on Theological Method, 3rd edn. Atlanta, GA: Scholars Press. Marion, Jean-Luc. 1994. ‘Metaphysics and Phenomenology: A Relief for Theology’. Translated by Thomas A. Carlson. Critical Inquiry 20: 572–91. Rea, Michael. 2015. ‘Theology without Idolatry or Violence’. Scottish Journal of Theology 68: 61–79. Rea, Michael. 2016. ‘Gender as a Divine Attribute’. Religious Studies 52: 97–115. Strawson, P. F. 1959. Individuals: An Essay in Descriptive Metaphysics, 1st edn. London: Methuen & Co. Tracy, David. 1975. Blessed Rage for Order: The New Pluralism in Theology. New York: Seabury. Vanhoozer, Kevin. 2010. Remythologizing Theology: Divine Action, Passion, and Authorship. Cambridge: Cambridge University Press. Wood, William. 2013. ‘Analytic Theology and the Academic Study of Religion’. Paper presented at the Logos Workshop in Philosophical Theology, University of Notre Dame.

5 Gender as a Divine Attribute Traditional Christian theology maintains that God created human beings, male and female, in the divine image. But the traditional pattern of characterizing God in scripture, authoritative confessions and creeds, and much of Christian theology throughout the centuries seems predicated on the assumption that God is masculine. God is traditionally referred to as ‘he’ and as ‘our heavenly Father’, for example; and even the figurative characterizations in scripture and elsewhere (e.g. as the faithful husband of wayward Israel) are predominantly masculine. By virtue of the relationship between masculinity and maleness, this pattern might seem to suggest that males bear the image of God to a greater degree than females, that they are more fit to represent God, and that whatever traits contribute to making someone masculine are somehow more divine than whatever traits contribute to making someone feminine. So many traditionalists have thought; so many feminists have objected. There is a persuasive case to be made for the conclusion that the traditional pattern of characterizing God in predominantly masculine terms has been harmful to women.¹ To the extent that it encourages a false view of their own superiority, and enables various kinds of oppressive behaviour, there is also a case to be made for the conclusion that it has been harmful to men.² There is, furthermore, good reason to think that this harmful state of affairs is at least partly a result of the fact that women’s voices have largely been excluded from the theological conversations that have shaped the tradition.³ In the face of such considerations, it is an urgent question whether the traditional pattern of characterizing God in predominantly masculine terms is theologically mandatory. That is to say, it is an urgent question whether we are somehow morally or rationally (in the epistemic sense) required to uphold and persist in the traditional pattern, or whether doing so is somehow liturgically important. If the answer is ‘no’, then the traditional pattern is theologically optional. Some, to be sure, have suggested that the harms done by the traditional pattern are overstated.⁴ But even if that were true, it is undeniable that many people believe that women have been and continued to be harmed by the traditional pattern, and that Christendom’s failure to adopt more gender-inclusive ways of speaking about God has presented such people with serious, ongoing obstacles to faith and to ¹ On this, see Daly 1986, as well as Carr 1988, Johnson 1993: 22–8 and other references therein. ² Cf. Carr 1988. ³ Cf. Coakley 2013, ch. 1. ⁴ Cf. Jenson 1992.

Essays in Analytic Theology: Volume I. Michael C. Rea, Oxford University Press (2020). © Michael C. Rea. DOI: 10.1093/oso/9780198866800.003.0006

    

99

comfortable participation in their own religious communities. This fact alone constitutes reason to adopt more inclusive ways of characterizing God, unless there is some reason for thinking that the traditional pattern is theologically mandatory. But what reason could there be? Why endorse traditionalism, the view that the pattern of characterizing God in predominantly masculine terms is theologically mandatory, rather than inclusivism, the view that feminine characterizations of God are just as legitimate as masculine ones? There is no explicit biblical affirmation of traditionalism. Nor can traditionalism be defended by appeal to facts about God’s body; for, incarnational considerations aside, the Christian tradition speaks virtually unanimously against the idea that God has a male or female body. Consequently, defences of traditionalism tend to argue either for the singular theological fittingness of masculine characterizations of God, or for the conclusion that the biblical pattern of characterization indicates that God wants to be characterized in predominantly masculine terms. Both kinds of argument would be undercut, however, if it could be shown that masculine characterizations of God are no more metaphysically accurate than feminine ones. Given the perceived harms of characterizing God in predominantly masculine terms, why would a perfectly loving God insist on being characterized in this way if feminine characterizations are equally accurate? Why would it be theologically fitting to continue in a metaphysically misleading pattern of characterization that raises obstacles to faith and the religious life? It is hard to see good answers to these questions. Consequently, in defending the claim that masculine characterizations of God are no more (or less) accurate than feminine ones, I take myself to be arguing against the best available underpinning for traditionalism. I am hardly the first to oppose the claim that masculine characterizations of God are more accurate than the alternatives. One of the main projects of so-called second-wave feminist theology in the 1980s and early 1990s was the idoloclastic attempt to show that the traditional Judeo-Christian concept of God as masculine is a pernicious misrepresentation of God.⁵ But, broadly speaking, the parties to the feminist side of the debate offered just two main strategies for defending the inclusivist position.⁶ Some dismissed the traditional concept of God as an ⁵ For detailed discussion of this point, see Raphael 2014. ⁶ I do not mean to suggest that feminism as such has been monolithically supportive of an unqualified or unmoderated inclusivism. Nor do I mean to suggest that merely embracing inclusivism suffices to address feminist concerns about traditional religious language. Soskice (1992) and Coakley (2013) both defend moderated inclusivist positions—inclusivist because they do not think that it is theologically mandatory to portray God in predominantly masculine terms, but moderated because they both think (for different reasons) that the language of divine fatherhood has an important place in Christian liturgy and devotion. Grace Jantzen (1998, ch. 8) has argued that the masculinism inherent in traditional religious language is embedded at the level of methodology: appeals to scripture and theological authorities, assumptions about God’s intelligibility, the traditional method of perfect being theology, and, indeed, the entire framework of symbols within which God has traditionally been imagined are all, on her view, rife with masculinist assumptions and imagery. So far as I can tell,

100

   :  

idolatrous invention, a mere projection onto the divine of male-centric values and fantasies.⁷ Others appealed to divine transcendence. A suitably strong theology of transcendence will imply that relatively few (if any) characterizations of God are strictly metaphysically accurate, thus making it extremely difficult to support the view that masculine characterizations are metaphysically privileged.⁸ On both ways of thinking, the traditional pattern of characterization is not due to its greater metaphysical accuracy but rather to the unjust suppression of feminine characterizations by sinful men in power, speaking and writing under the influence of entrenched biases against women. Feminists have largely moved on from these particular debates, in part because some of the underlying assumptions that helped determine the contours of the debates (e.g. that sex and gender are binary categories, or that gender supervenes on biological and psychological traits) are now more contested than they previously were. Even so, most inclusivists seem content with some variation on one of the two strategies I have just described. Most, that is, except for those working in the analytic tradition of philosophy. As is well known, analytic theologians tend to have very little sympathy either for projectionism or for strong theologies of divine transcendence. What, then, have they said about the feminist critique of the traditional concept of God? Remarkably little. Although frequently and in various ways targeted by inclusivists for uncritically embracing a masculine conception of God, those working in the analytic tradition have mostly opted simply to ignore the critique rather than to explore their own distinctive resources for addressing it. As a result, there is now a significant gap in the literature. What might one say if one is sympathetic towards feminist concerns about traditionalism but unsympathetic towards projectionism and strong theologies of transcendence? This is the question that animates the present chapter. My point of departure is the following insight from Elizabeth Johnson’s landmark book, She Who Is: If women are created in the image of God, then God can be spoken of in female metaphors in as full and limited a way as God is imaged in male ones. (Johnson 1993: 54)

nothing I say in the present chapter conflicts with the views of Soskice and Coakley. However, as shall emerge, the arguments of this chapter do rely on much of the methodology and (I think) the same general symbolic framework that Jantzen is concerned to criticize. Even if she is correct, however, I think that it is both important and novel to show that inclusivism can be supported from within the methodological and symbolic framework that I take for granted. ⁷ Cf. Ruether 1983, ch. 2; Hampson 1996, ch. 4. ⁸ Cf. Carr 1988, Johnson 1993, and McFague 1987.

    

101

Out of context, the remark seems to suggest that, on Johnson’s view, the primary rationale for inclusivism is the imago Dei doctrine. In fact, however, this is not the primary rationale that we find in the rest of She Who Is. Although the imago Dei doctrine certainly plays a role in her overall case, the metaphysical underpinning for her brand of inclusivism is, as I have already noted, the doctrine of divine transcendence.⁹ My own goal in the present chapter is to provide metaphysical underpinning for inclusivism that dispenses with the latter doctrine and rests more theoretical weight on the imago Dei doctrine.

1. The Main Argument Summarized In the following sections, I argue from a particular understanding of the imago Dei doctrine (elaborated below) via the method of perfect being theology for the conclusion that masculine characterizations of God are no more or less accurate than feminine ones. To say that one way of characterizing God is more accurate than another is just to say that it comes closer to telling the straightforward, literal truth about what God is really like (intrinsically or extrinsically, essentially or contingently). Accuracy can, of course, come apart from legitimacy, permissibility, or propriety. It doesn’t automatically follow from the fact that masculine and feminine characterizations come equally close to telling the truth about God that it is morally permissible or liturgically appropriate to characterize God in feminine terms. Accuracy is a matter of metaphysics; propriety is a matter of what is overall most fitting in light not only of what we know about the relevant metaphysics but also of what we know about God’s preferences, likely harms and benefits to human beings and their relationships with God, and so on. Further premises, therefore, would be required to establish inclusivism, which asserts not the equal accuracy but rather the equal propriety of characterizing God in feminine terms. From the remainder of this section up until the beginning of the paper’s closing section, I shall be concerned with matters of metaphysical accuracy. In addressing these issues, I take myself indirectly to be addressing the question whether we are rationally required to characterize God in predominantly masculine terms. Insofar as we can see that it is not most accurate to do so, we are not rationally required to do so. I then turn to consider questions about the moral and liturgical propriety of feminine characterizations of God in the final section of the paper. My main argument is simple. (i) God is most accurately characterized as masculine (say) only if God is masculine and God is not equally feminine. However, so I shall argue, (ii) God is masculine or feminine only if God is equally masculine and feminine. Therefore, (iii) God is not most accurately characterized

⁹ See also Johnson 1984.

102

   :  

as masculine. By the same reasoning, mutatis mutandis, God is not most accurately characterized as feminine (or as belonging to any other gender). I take the first premise of this argument to be obviously true. Thus, my remarks in what follows focus on defending the second premise. Another way of stating that second premise is to say that gender is a divine attribute only if God belongs equally to both genders. I use the term genderism to refer to the negation of that premise—i.e. to the following thesis: gender is a divine attribute (i.e. God has gender) and God is either more masculine than feminine, or vice versa. My defence of premise (ii) consists in an argument for the conclusion that genderism is false. As I have indicated, my argument rests on what I take to be obvious implications of the thesis that women, like men, are created in the image of God.¹⁰ Does anybody in fact endorse genderism? Initially, one might have doubts. It is commonly said that the general consensus in the Christian tradition is and always has been that God is beyond gender; it is also, admittedly, rather difficult to find theologians who clearly endorse genderism. However, careful attention to the texts commonly adduced as evidence of consensus that God is beyond gender reveals that those texts in fact only support the thesis that there has been consensus that God is beyond sex (cf. Cooper 1998: 168–9). Moreover, even in the absence of explicit theological affirmations of genderism, the history of iconography—in which God and each person of the trinity are overwhelmingly regularly depicted as male—bears strong testimony to the fact that masculine genderism has, on some level, been the dominant view throughout Church history.¹¹ Finally, although it is indeed hard to find explicit affirmations of genderism as I have formulated it, it is easy to find affirmations of claims close enough in the neighbourhood. Robert Jenson (1992), for example, insists that there are neither linguistic nor metaphysical reasons for thinking that Jesus’ address of God as ‘Father’ is non-literal, and that the Church decided during the fourth-century Arian controversies that this mode of address is both non-literal and absolutely to be preferred over feminine or neutral forms of address (1992: 105).¹² Again, this is not an explicit affirmation of genderism; but it is a nearly explicit affirmation of

¹⁰ One might wonder whether 1 Cor. 11:7—which says that ‘[man] is the image and reflection of God; but woman is the reflection of man’ (NRSV translation)—raises trouble (or should raise trouble for people who take a high view of the authority and veracity of scripture) for the thesis that men and women are created equally in the image of God. It does not. Genesis 1 is explicit that human beings— male and female—were created in the image of God. Commentators divide on the question whether 1 Cor. 11:7 contradicts Genesis on this score. (Cf., e.g. Fee 1987: 515 and Hays 1997: 186.) I cast my lot with those who interpret the passage in light of Genesis 1 rather than as contradicting Genesis 1. But the important thing to note is that, for those who take a high view of the authority and veracity of scripture, it is simply not an option to say that Paul is outright contradicting the author(s) of Genesis. Nor, I should think, is it at all plausible to interpret Genesis 1 in light of a superficial reading of 1 Cor. 11:7. ¹¹ Thanks to Sarah Coakley for this point about the iconography. Cf. also Soskice 1992: 83. ¹² Jenson’s read on the matter is not uncontroversial, however. Soskice, for example, says just the opposite, maintaining that the ‘consensus’ since the conclusion of the Arian controversy has been that ‘calling God “father” is metaphor . . . ’ (1992: 83).

    

103

the claim that God is literally our Father rather than our Mother or our beyondgender Parent. It seems hardly a stretch, then, to say that Jenson affirms genderism.¹³ The rest of the chapter unfolds as follows. In Section 2, I lay out some of my assumptions about gender and the imago Dei doctrine that are needed to understand my argument against genderism. In Section 3, I present that argument. Then, in Section 4, I consider whether my argument against genderism has negative implications for the moral and liturgical propriety of characterizing God exclusively or predominantly in masculine terms.

2. Gender and the Imago Dei In accord with conventions that are presently fairly standard, I use the terms ‘male’ and ‘female’ when talking about sex, and ‘masculine’, ‘feminine’, ‘man’, and ‘woman’ when talking about gender. Furthermore, I do not presume that gender necessarily correlates with sex. Accordingly, my terminology allows that there might be male women and female men. It also allows that someone might be a man or woman, masculine or feminine, without being either male or female. I shall generally try to remain neutral on the metaphysics of sex and gender; but for convenience, I will talk as if they are binary categories and as if sex is not socially constructed whereas gender might be. Nothing in my argument depends on these choices, both of which are now contentious.¹⁴ There seem to be two different (but not mutually exclusive) things that one might mean by saying that a characterization of someone is gendered—i.e. masculine or feminine. First, one might mean that it involves the use of a pronoun or predicate that can be literally applied only to members of a particular gender. For example, it is a conceptual truth that the pronoun ‘he’ and predicates like ‘is a father’ or ‘is a king’ apply literally only to men. (Some will say that they apply literally only to males; but I imagine one would say this only if one thought that only males could be men.) I’ll refer to characterizations that use pronouns and predicates like this as strongly gendered. Second, one might mean that the characterization involves the application of a predicate that stereotypically applies to members of just one gender, even though ¹³ Not only does Jenson seem to endorse genderism, but he seems also to think that it has been canonized by the Church. However, he does not explain—and I cannot myself see—how one might get from the premise that the Church has decided that God is literally the Father of Jesus to the conclusion that the Church has prohibited addressing God as ‘Mother’ or ‘Parent’. Consequently, I shall not engage this idea further in the sequel. It is perhaps worth noting, however, that it is not at all clear how to square Jenson’s views on inclusivism with his view, expressed elsewhere, that ‘God is a decision’ and ‘God is a conversation’ (1997: 222, 223). Why it should be theologically mandatory to think of any decision or conversation (divine or not) as masculine is wholly opaque to me. ¹⁴ Cf. Butler 1990, esp. pp. 18–34.

104

   :  

it can be literally applied to members of both genders. Predicates like this I’ll refer to as weakly gendered. In the literature, for example, we find claims to the effect that characterizations of God as powerful and a se are masculine, whereas characterizations of God as nurturing or as brooding over God’s children are feminine. Those who make such claims do not generally mean to suggest (e.g.) that being powerful is a sufficient condition for being a man, or that being nurturing is a sufficient condition for being a woman. Rather, their view seems to be that these predicates are gendered by virtue of the stereotypes associated with each gender.¹⁵ Both strongly and weakly gendered characterizations have been seen to be problematic. Strongly masculine characterizations of God obscure (or perhaps even tacitly deny) that traits definitive or stereotypical of womanhood reflect the image of God. The same is true of weakly masculine characterizations when their use greatly overshadows the use of weakly feminine ones. Mary Daly famously said that ‘If God is male, then the male is God’ (1973: 19). By extension, we might also say that if God is (exclusively or predominantly) masculine, then the traits that are definitive or stereotypical of masculinity are more divine than those definitive or stereotypical of femininity. But, according to the second-wave feminist critiques, it is precisely this thought, together with the fact that gender typically (even if not uniformly) tracks sex, that has led to the subordination of females within the Christian tradition. So much for preliminary terminological points. Let me now highlight two assumptions. First, I assume that sex is not a divine attribute: God is neither male nor female. (The incarnate Christ was male; but presumably his maleness is no more a divine attribute than his tallness, his hairiness, his musculature, etc.) Accordingly, I also assume that in offering a gendered characterization of God, one does not thereby commit oneself to the claim that God has, or is even imagined to have, male or female bodily features. The first assumption I take to be obviously true; the second is a natural correlate of the first. Finally, I want to make some brief remarks on how I understand the version of the imago Dei thesis on which my arguments depend. There are various ways of interpreting the claim that women and men equally are created in the image of God, some of which put it in conflict with theories about gender that ought to be compatible with it. For example, one might read it as saying that some people have been created as women, others as men, and the intrinsic properties that make someone a woman are no less a part of the divine image than those that make someone a man. On that reading, one cannot affirm the equality thesis while also affirming (e.g.) that gender is a social construction, or that gender membership reflects something about where one stands in the structure of power ¹⁵ Cf., for example, Hampson 1996: 124–5 and Johnson 1993: 68. Note that the relevant stereotypes don’t necessarily arise out of or reflect common cultural beliefs about men and women. Often traits are characterized by an author as masculine or feminine because their association with men or women, or with their development and socialization, is posited by a developmental theory that is accepted by the author or her interlocutors. (Cf. Chodorow 1978, cited in Johnson 1993.)

    

105

and oppression within one’s society.¹⁶ But surely that is a bad result. The equality thesis ought to conflict with theories according to which some people by virtue of their gender bear the divine image to a greater degree than people of a different gender. It should not, however, necessarily conflict with theories according to which gender is socially constructed. Nor should it necessarily conflict with theories according to which gender is determined more by social position than by personality and behavioural dispositions—especially since the latter traits are clearly more relevant to the divine image, and were the sorts of traits primarily in view in the debates of the 1980s and early 1990s. In light of all of this, I shall understand the equality thesis to be equivalent to the following claim: All human beings are created equally in the image of God; and if mental and behavioural characteristics contribute at all to gender membership, then the ones that contribute to making someone a woman are no more or less relevant to her bearing the image of God than those that contribute to making someone a man. I hope, too, that in showing how one might recast the equality thesis in such a way as to eradicate certain contested assumptions about gender, I will have gestured towards ways of reformulating any other claims in this chapter that might unintentionally presuppose contested theses about gender. In the next section, I provide an argument against genderism. If genderism is false, the following disjunction is true: gender is not a divine attribute or God belongs equally to both genders. The first disjunct implies that strongly gendered terms do not literally apply to God—God transcends gender categories. Importantly, however, it does not follow from this that weakly gendered terms and predicates have no literal application to God. Thus, even if predicates like ‘is a se’ or ‘is nurturing’ are weakly gendered, one can still coherently affirm that gender is not a divine attribute. Likewise, one can affirm that gender is not a divine attribute while still maintaining that strongly gendered pronouns or predicates have an important place in our theology. For, after all, such terms might be applicable to God, and the relevant figures of speech might be theologically important. I affirm with scripture and the oldest creeds of Christendom that God is my heavenly Father. Likewise, I can agree with Julian of Norwich that ‘as truly as God is our Father, so truly is God our Mother’.¹⁷ Both claims are theologically important; both claims express truths. And this is so even if both are metaphors. ¹⁶ For example: Sally Haslanger (2000) maintains that a necessary condition for being a woman is having observed or imagined features of one’s body play a role in one’s systematic subordination and in one’s occupying an oppressed social position; and a necessary condition for being a man is having observed or imagined features of one’s body play a role in one’s being systematically privileged. If that theory is correct, then (a) one’s intrinsic characteristics do not by themselves make someone a man or a woman, and (b) it is doubtful at best whether those intrinsic characteristics, if any, that contribute to making someone a man or a woman equally image the divine (since they would be characteristics that contribute to some people being oppressed and others being oppressors). ¹⁷ Julian 1978: 296. The concept of the motherhood of God is, of course, not an invention of Julian’s, though she was an important figure in the development of that idea. For more on this, see Bynum 1982 and Jantzen 1988: 115–24.

106

   :  

3. The Argument against Genderism Every divine attribute is such that God has it either essentially (i.e. of necessity) or contingently. Every divine attribute is furthermore either intrinsic to God or extrinsic. Thus, every divine attribute falls into one of the following mutually exclusive and jointly exhaustive categories: (i) essential intrinsic, (ii) essential extrinsic, (iii) contingent intrinsic, or (iv) contingent extrinsic.¹⁸ Those who endorse the doctrine of divine simplicity, according to which God is identical to God’s attributes, will insist that God has no attributes outside the first category. Those who reject divine simplicity and endorse an abundant theory of properties have room in their ontology for divine attributes that fall into some of the other categories.¹⁹ In the argument that follows, I make three further assumptions. First, in order to give my argument widest scope, I assume that the doctrine of divine simplicity is false and that every non-paradoxical predicate that has a literal meaning corresponds to an attribute.²⁰ I take it that we can generally tell which predicates are meaningful; and a predicate has a literal meaning if, and only if, it is not irreducibly figurative. A predicate F is paradoxical iff, on the assumption that F is a meaningful predicate, there is something x such that Fx  ¬Fx. (For example, on the assumption that is not a predicate that applies to itself is a meaningful predicate, it is paradoxical.) Note that I am not assuming that paradoxical predicates do not correspond to properties; I simply do not wish to presuppose that they do. Second, I assume that the term ‘divine attribute’ refers only to attributes had by at least one of the three persons of the trinity independently of the incarnation. There is no question that Jesus of Nazareth was male; there can be little question that he was also masculine. But just as we would not want to say, on the grounds that Jesus was bearded (if indeed he was), that being bearded or being embodied is a divine attribute, so too we should not want to say simply on the grounds that Jesus was a man that being masculine is a divine attribute.²¹ ¹⁸ Philosophers and theologians sometimes mean different things by the term ‘essential property’. For purposes here, p is an essential property of x iff, necessarily, x exists only if x has p; and p is a contingent property of x iff x has p but it is possible that x exist and fail to have p. ¹⁹ An abundant theory of properties is one according to which every non-paradoxical predicate that has a literal meaning corresponds to an attribute. (I discuss this idea further in the course of explaining the first assumption of my argument in this section.) ²⁰ As I indicated at the end of the previous paragraph, rejecting this assumption simply restricts the range of attributes that God might have. Accordingly, rejecting it will either make no difference to my argument or will make it easier to defend my conclusion (depending on which part of the assumption one rejects, and for what reasons). ²¹ It is fairly standard to maintain that Jesus still has a body. Jesus’ post-resurrection body was clearly transformed; but perhaps it is, nonetheless, a male body. Perhaps, too, Jesus is still masculine. Even so, just as we would not say on the basis of the belief that Jesus still has a male body that being embodied and being male are divine attributes, so too we should not say on the basis of the belief that Jesus is still masculine that being masculine is a divine attribute.

    

107

Third, I assume that strongly gendered terms (either pronouns or predicates) literally apply to God only if God has an attribute corresponding to a strongly gendered predicate. With these assumptions in hand, I shall argue for each of the following three premises: P1. Strongly gendered attributes are among God’s essential attributes only if God belongs equally to both genders. P2.

The category of contingent intrinsic divine attributes is empty.

P3. Strongly gendered attributes are among God’s contingent extrinsic attributes only if God belongs equally to both genders. The three premises exhaust the four categories of divine attribute; thus, if all three are true, it follows that God has strongly gendered attributes only if God belongs equally to both genders. I take this conclusion to be equivalent to the thesis that genderism is false. Let us begin with P1. The first step is to argue that no strongly gendered attribute is essential to all three persons of the trinity unless each person is equally masculine and feminine. This conclusion is delivered in two ways by the method of perfect being theology deployed under the assumption that it is neither better to be masculine than feminine nor better to be feminine than masculine. Here again the conviction that women and men equally are created in the image of God is crucial; for it is precisely that conviction which provides theological underpinning for the view that it is neither better to be masculine than feminine nor better to be feminine than masculine. Here is the first way: It seems plausible that, for every attribute A that is essential to all three persons of the trinity, it is intrinsically better to be A than not-A.²² The principle seems clearly to be true for attributes other than strongly gendered ones; thus it seems most sensible to say that it holds of strongly gendered attributes as well. But it is not intrinsically better to be masculine than to be feminine, or vice versa; thus, no strongly gendered attribute is essential to all three persons of the trinity unless each person is equally masculine and feminine. One might object: Every member of the trinity has essentially the attribute not being identical to the number 2. But it seems neither better to have that attribute nor better to lack it. (The number 3 has that attribute whereas the number 2 lacks ²² Note that I am not here presupposing that God has no neutral properties whatsoever, but only that God has no neutral properties essentially. Suppose God has some attribute A such that it is neither better to be A than not-A nor vice versa. There is, then, no explanation in terms of divine perfection for why God should have A rather than not-A (or vice versa); thus, one would expect that if God in fact has A, God could have had not-A (and vice versa). Thus, being A is contingent. Later I shall argue that God has no contingent intrinsic attributes; but this still leaves open the possibility that God has contingent extrinsic attributes, and that some of these are neutral.

108

   :  

it, but neither number is thereby better than the other.) Thus, there seems to be reason to reject my principle that for every attribute that is essential to all three persons, it is better to have the attribute than to lack it. I reply as follows: It is intrinsically better to be distinct from the number 2 than to be identical to the number 2; but the reason we don’t think that the number 3 is thereby better than the number 2 is simply that it is also better to lack the property being identical to the number 3 than to have it. None of this implies that it is bad to be identical to the number 2 or to the number 3, just that something is intrinsically better if it is distinct from them (and from all other numbers, and, indeed, from anything else besides God). Here is the second way to arrive at the conclusion that no strongly gendered attribute is essential to all three persons of the trinity unless each person is equally masculine and feminine: Under the method of perfect being theology, the recipe for determining the essential divine attributes is (roughly) to sort the attributes we know of into two categories—great-making properties, and other properties. The divine attributes, then, comprise the best possible array of mutually compatible great-making properties. But, under the assumption that it is not better to be masculine than to be feminine, or vice versa, (together with the assumption that there is no hybrid option) we can readily see that strongly masculine attributes belong in the category of great-making properties only if strongly feminine attributes do. A fortiori, strongly masculine attributes belong in the best possible array of such properties only if strongly feminine ones do. It follows, then, that at least one person of the trinity is essentially masculine only if at least one person of the trinity is essentially feminine; and at least one is essentially feminine only if at least one is essentially masculine. For suppose every person of the trinity were masculine. If being masculine and being feminine are equally great-making properties, then God would have been no better or worse had every person of the trinity been feminine. But then God does not have the greatest possible array of compatible great-making properties, contrary to the hypothesis of perfect being theology. To avoid this consequence, one of the following must be true: (a) at least one person of the trinity is masculine whereas at least one other is feminine; (b) all persons of the trinity are both masculine and feminine; or (c) no person of the trinity is either masculine or feminine. Suppose it turns out that two persons of the trinity (Father and Son, presumably) are masculine whereas only one person of the trinity is feminine,²³ and suppose that each divine person’s gender is essential to him or her. In that case, does it follow that God is more masculine than feminine? In a word, no. This for two reasons.

²³ Cf. Congar 1983: 155–64.

    

109

First, and in general, a claim of the form ‘God is more F than G’ is true only if (a) every person of the Trinity is more F than G; (b) the divine nature is more F than G; or (c) the trinity as a whole is more F than G. Where masculine and feminine are substituted for F and G, (a) and (b) are ruled out by the imago Dei doctrine. If every person of the trinity were more masculine than feminine, or if the divine nature were more masculine than feminine, men and women would clearly not be equally created as such in the image of God—men as such would be greater image-bearers than women. And it seems that (c) is satisfied only when F and G are properties had by all three persons. For, after all, we don’t say (e.g.) that God is more discarnate than incarnate, more non-indwelling believers than indwelling believers, less distinct from Father and Son than identical to Father or Son, etc. But we have already established that if all three persons of the trinity are masculine, all are equally feminine (and vice versa). Thus, on the supposition that two members of the trinity are essentially masculine and one is essentially feminine, it is still not the case that God is more masculine than feminine. Second, and perhaps more importantly, to say that God is more masculine than feminine when two divine persons are masculine and one is feminine is equivalent to saying that the term ‘masculine’ more aptly describes the Godhead—more adequately gets at what God really is—than the term ‘feminine’. But to say this, in turn, is to subordinate the feminine person of the trinity—to say that she is somehow less real, or less really God, than the other persons. But, of course, this is just the heresy of subordinationism; it is not an option for Christian theology. If every person of the trinity were more masculine than feminine, men and women as such would not be equally created in the image of God—men would be greater image-bearers than women. Thus, I conclude that P1 is true. P2 says that the category of contingent intrinsic divine attributes is empty. I reach this conclusion as follows. Suppose (for reductio) that God is intrinsically F but might have been intrinsically G, where being intrinsically G implies not being intrinsically F, and vice versa. Now, if God were G, then God would be better, worse, or neither better nor worse than God in fact is. If God would be better by virtue of being G, then God is not perfect; if God would be worse by virtue of being G, then God might not have been perfect. But God is necessarily perfect. Thus, being G is neither better nor worse than being F. But if that is true, then it seems to be a matter of sheer happenstance whether God is F or G. If intrinsic features of the divine nature depend on the divine will, then (since it is neither better nor worse to be G rather than F) there is no explanation in the divine will for why God is F rather than G; and divine aseity implies that there is no explanation outside God for God’s intrinsic attributes. Thus, it is sheer happenstance that God is F rather than G; and, if so, then God’s being F rather than G is wholly independent of God’s will. But none of God’s intrinsic features is wholly independent of God’s will—this, too, is a consequence of divine aseity. Thus, it is not sheer happenstance that God is F rather than G. (Contradiction.)

110

   :  

If this argument is sound, it follows that God has no contingent intrinsic values or desires. If that is right, then one might think that either God’s having the values and desires God has is a contingent extrinsic attribute of God or else God’s creative options are severely constrained. The reason is that what God creates plausibly depends upon what God wants and what God values. Thus, if God might have created different things, it seems to follow that God might have wanted or valued different things, from which it seems to follow that God’s having the values and desires God has is a contingent matter. So, if God has no contingent intrinsic values and desires, and if God might have created different things, then it must be a contingent extrinsic matter what God desires and values. The latter conclusion seems incompatible with divine aseity; for it seems to imply that divine desires and values—and hence perhaps also divine goodness—depends somehow on factors external to God. But denying that God might have created different things seems to raise problems for a robust understanding of divine freedom. So we might seem to face a terrible dilemma. But I reject the initial disjunction. Suppose God’s values and desires are intrinsic. Nevertheless, they might (for all we know) be conditional on how certain matters of chance turn out; and God might (for all we know) necessarily value leaving certain matters—the truth values of counterfactuals of creaturely freedom for example, or the movements of fundamental particles—to chance. If that is right, then there is room to say that God might have created different things, and this despite the fact that divine values and desires are intrinsic. Finally, I turn to P3: God has strongly gendered contingent extrinsic attributes only if God belongs equally to both genders. (Other examples of contingent extrinsic attributes might include being worshipped by Paul or being the creator of the Milky Way galaxy.) Many theories of gender maintain that gender is extrinsic by virtue of being at least partly socially constructed. Moreover, any credible theory of gender according to which gender is both extrinsic and potentially a divine attribute will maintain that a person’s gender depends in part upon some combination of outwardly directed beliefs and desires, behavioural dispositions, or observed behaviour towards others. If gender depended entirely upon internal and internally directed mental states and behaviours, it is hard to see why it would not be intrinsic; and if it did not depend at all upon beliefs, desires, behavioural dispositions, or actual behaviours, then it is hard to see how it could possibly be an attribute of a disembodied personal God. Given all this, it is easy to see why P3 is true. We have seen that God has no strongly gendered intrinsic attributes; likewise, God has no strongly gendered essential attributes. The question, then, is whether we have reason to think that there is anything about God’s beliefs and desires or about the social significance of God’s observable behaviour that would ground the literal application of strongly masculine terms to God but not strongly feminine ones, or vice versa. Scripture seems to characterize God as behaving in ways stereotypical of both genders, and

    

111

scripture also seems to attribute mental states to God—maternal love and concern, for example—that are appropriately characterized by way of feminine imagery.²⁴ When theologians like Julian of Norwich apply strongly feminine terms to God, the characterizations usually resonate with rather than contradict what the Christian tradition seems to affirm about God’s beliefs, desires, and behaviour toward creation; and likewise when strongly masculine terms are applied to God. Given all of this, there is good reason to think that either God has no strongly gendered extrinsic attributes at all or that God belongs to both genders. To say that God belongs to both genders, however, is not yet to say that God belongs equally to both genders. Perhaps God belongs to both genders but nevertheless has a preponderance of masculine attributes and therefore counts as more masculine than feminine. As noted earlier, however, to say this is to fall afoul of the equality thesis. Again: If every person of the trinity were more masculine than feminine, or if the divine nature were more masculine than feminine, or if the trinity as a whole were more masculine than feminine, then men as such would be greater image-bearers than women as such, contrary to the equality thesis. Thus, if gender is extrinsic, either God has no gender, or God belongs equally to both genders.

4. Implications for Traditionalism In the previous section, I defended the conclusion that God is masculine or feminine only if God is equally masculine and feminine. The only other premise in my main argument is the obvious truth that God is most accurately characterized as masculine only if God is masculine but not equally feminine. Thus, I conclude that it is not the case that God is most accurately characterized as masculine. By the same reasoning, mutatis mutandis, it follows that it is not the case that God is most accurately characterized as feminine. In this section I want to briefly consider the implications of this conclusion for the moral and liturgical propriety of characterizing God in exclusively or predominantly masculine terms. The conclusion that God is not most accurately characterized as masculine poses a serious challenge to traditionalists who maintain that God ought to be characterized in predominantly masculine terms. The challenge is to explain why this should be so. The most natural defence would appeal to the idea that privileging masculine characterizations somehow gives us special insight into the divine nature that would be lost if feminine characterizations came to enjoy an equal place in our theology and liturgical practice. But it is hard to maintain ²⁴ See Johnson 1993, ch. 5 and Cooper 1998, chs 3–5 for references and discussion (from opposing perspectives).

112

   :  

this view while at the same time admitting that characterizing God in predominantly masculine terms is metaphysically misleading. The question for the traditionalist, then, comes to this: Why think that masculine characterizations are to be preferred, given that preferring masculine characterizations is apt to mislead people about God’s attributes? Broadly speaking, I see only two responses to this question that have any hope of being viable. First, one might think that the risks (in terms of harm to human beings in general, or to the furtherance of what we reasonably take to be God’s goals for creation) of de-privileging masculine characterizations of God outweigh the risks of continuing to privilege them. At best, I think, we have no reason to think that this claim is true; and feminists have, in effect, been arguing for quite some time now that the claim is resoundingly false. Second, one might think that, wholly independently of considerations about harm and (again) despite the fact that it would be metaphysically misleading, God simply wants to be characterized by human beings more in masculine terms than feminine ones. This seems to be the position of John Cooper (1998), and perhaps there is some historical precedent for it as well.²⁵ In the next several paragraphs I will explain why I think his defence of that view is a failure. Three chapters of Cooper’s Our Father in Heaven are devoted to examining the ‘biblical pattern’ of representation of God and establishing the claim that the ‘only reasonable conclusion’ one might draw from that pattern is that ‘Scripture intends to portray God as a masculine person’ (1998: 114). His argument for this conclusion rests mainly on two premises: (i) all proper names, basic titles, and common nouns that name God are masculine; and (ii) these forms of speech are to be privileged over mere ‘figures of speech’ in determining what the authors (human and divine) intend to convey about what sorts of gendered terms are appropriately applied to God. I won’t contest the first premise; and I need not contest the part of the second premise that says that proper names, basic titles, and common nouns are to be privileged over mere figures of speech in determining what the human authors of scripture intend to convey about divine gender. The salient question is whether the absolute prevalence of masculine names, titles, and common nouns indicates anything about the divine will. Cooper is explicit that it does indicate the divine will. After arguing (against some feminist authors) that the masculine portrayals of God in scripture were not necessary in order to accommodate the beliefs and values entrenched in ancient cultures, Cooper writes: The fact that God could easily have chosen a gender-inclusive self-presentation . . . and did not do so is a strong reason for supposing that the exclusively masculine

²⁵ Cf. Widdicombe 1994: 255–60.

    

113

language for God in Scripture is not merely a temporary accommodation. In fact it is a good indication that God is not interested in promoting gender-inclusive language for himself. (1998: 153; emphasis mine)

At the same time, however, he affirms—emphatically, and multiple times—that ‘ontologically, God is genderless’ (169, n. 3). But he thinks that this fact is ultimately not relevant to the debate about how God is (morally and liturgically) appropriately characterized. Thus, after noting that the claim that God transcends gender is often taken to be a ‘key theological ground for the doctrine that God has no gender and thus for dealing with the Bible’s masculine language for God’, he writes: This move [is not consistent with Scripture and] in effect makes a theologicalphilosophical conclusion about God’s nature more determinative of Christian language than the text and teaching of the Bible.²⁶ (1998: 168)

He goes on later to say that the doctrine that God is beyond gender ‘is not stated in Scripture’ and that, although it ‘has a substantial biblical basis . . . its derivation involves some debatable cultural and philosophical assumptions about sex, gender, body, and spirit’ (183). Thus, he finally concludes: I strongly affirm the Christian doctrine that God himself is beyond gender. But I also conclude that because this doctrine is derived from Scripture and relies on extrabiblical assumptions, it ought not to be placed prior to or above Scripture in shaping the language of the Christian faith. (1998: 184)

The upshot, then, is that scripture—the ‘biblical pattern’—testifies against inclusivism; and the testimony of scripture trumps metaphysics as evidence about how God prefers to be characterized.²⁷ Many theologians will reject the claim that the testimony of scripture trumps metaphysics as evidence about how God prefers to be characterized. But that is not the route I wish to take. As a general rule, if my considered judgment were ever to be that scripture affirms p whereas my best theory in metaphysics affirms not-p, I would side with scripture and cast aside the metaphysical theory. On this much, I think that Cooper and I agree. But, as one might expect, I have yet to reach the ²⁶ It is common to suppose that the claim that God transcends gender implies that God has no gender. I do not myself endorse this supposition; I would want to leave room for the possibility that God is polygendered. But this is an issue that I cannot pursue in detail here. ²⁷ Note that Cooper’s view is apparently not that our evidence from scripture corrects or undermines our metaphysical theory that God is beyond gender. For, as he is at pains to emphasize, he accepts the metaphysical claim that God is beyond gender. Thus, his position is clearly that, despite the fact that it is metaphysically misleading, God for some reason simply prefers to be characterized as masculine and not (equally) feminine.

114

   :  

verdict that scripture affirms a proposition that the overall best metaphysical theory denies. Nor have I ever reached the unfortunate conclusion—one that Cooper himself seems to have reached—that there is some proposition that our overall best metaphysical theory affirms but God would prefer for us to ignore in our characterizations of God. The reason, of course, is that my metaphysical theorizing and my interpretation of scripture proceed in dialogue with one another. Theory-building is a holistic process. The problem with Cooper’s argument is that he fails to appreciate the way in which philosophical reasoning is already involved at the level of figuring out what it means to take sides with scripture and to let it serve as evidence for a conclusion. Let us grant that ‘the biblical pattern’ of reference to God is one which portrays God overwhelmingly in masculine terms. Let us also grant (as seems plausible) that this pattern is prima facie evidence that God wants to be portrayed in overwhelmingly masculine terms. Still, just like the doctrine that God is beyond gender, the thesis that God wants to be portrayed in overwhelmingly masculine terms is not at all stated in scripture, but is instead merely derived from scripture with the help of additional, extrabiblical assumptions. The question, then, is whether our total evidence—scripture, metaphysics, and whatever other extrabiblical assumptions we might appropriately appeal to in defending our views about how to characterize God—supports the thesis that God prefers to be characterized in predominantly masculine terms despite the fact that it is metaphysically misleading. Against Cooper’s view, we might first note that the biblical pattern of reference to God is one that applies exclusively Aramaic, Hebrew, and Greek terms to God; but, of course, nobody would claim that this pattern is evidence that God wants to be talked about exclusively in Aramaic, Hebrew, and Greek. The reason, of course, is that we know independently—i.e. at the level of philosophical presupposition—that there is no reason why God should want this. There is nothing about God, or about the words of Aramaic, Hebrew, and Greek that make them more apt than (say) English words for describing God; and there is no other reason that we can see why God should want to be described only by way of such words. So likewise, we might think, with gendered words. Moreover, we know from scripture that God cares about the truth; God is no deceiver. This, it seems, speaks strongly in favour of a view about how God prefers to be characterized that attributes to God an interest in characterizations that are not metaphysically misleading. (This is why I said earlier that the best route for a traditionalist to take is to oppose my arguments for the conclusion that masculine characterizations of God are not more accurate than feminine ones.) Cooper maintains, in effect, that the moral and liturgical aptness of feminine characterizations of God depends on the proportion of feminine names, basic titles, and common nouns that are applied to God in scripture. He thinks that the fact that all of those are masculine is decisive evidence that God has no interest in gender-inclusivism (1998: 114, 153). But, again, this claim is no less an

    

115

extrabiblical philosophical thesis than its negation, and it is no less an extrabiblical philosophical thesis than those often invoked to support inclusivism (such as the claim that if feminine characterizations are no less accurate than masculine ones, then they are likewise no less morally or liturgically appropriate than masculine characterizations). Cooper’s ‘biblical pattern’ is just that—a pattern, and one that is open to interpretation in light of our best overall evidence, philosophical presuppositions included. I acknowledge that nothing I have said so far demonstrates that Cooper’s conclusion is false. I have only shown that his argument for that conclusion fails. I also acknowledge that there is something quite natural in taking not just the overall biblical pattern of divine characterization but Jesus’ own practice (as reported in scripture) as indicative of a preference on God’s part to be characterized in predominantly masculine ways. It is tempting, indeed, to think that sticking with the traditional masculine modes of characterizing God is the safe option—safe because it is non-revisionary and is backed by the weight of centuries of practice. I grant that all of this constitutes some reason (partly pragmatic, partly epistemic) for affirming the thesis that God, for unknown reasons, simply prefers to be characterized in masculine terms. Still, we must view these considerations in light of the very persuasive arguments for the conclusion that much harm has come from treating masculine characterizations of God as mandatory, and also in light of the very real possibility that implicit bias against women plays a non-trivial role in shaping our judgments about how the biblical pattern of characterization, including what we learn in scripture about Jesus’ own practice, is naturally interpreted.²⁸ We must also recognize that there is nothing at all safe about insisting on a pattern of characterizing God that is ultimately metaphysically misleading (as I think we have very good reason to believe that the traditional pattern is) unless we also have very good reason for thinking that God simply prefers that pattern of characterization. To do that is to risk idolatry. Indeed, it is to risk other sins as well, given that the pattern in question has contributed to the oppression of half the human race, and given that, if we know anything at all from scripture about divine preferences, we know that God cares about justice. Much rides, then, on the claim that God prefers to be portrayed as masculine despite its being metaphysically misleading. As

²⁸ We should also take seriously the possibility that bias against women influenced the pattern itself—and this even if we maintain a very high view of the authority and veracity of scripture. One who thinks that the Bible teaches nothing false might nevertheless think that erroneous views about cosmology, or about Christ’s imminent return, or about the morality of slavery exerted some influence on the way in which certain truths were ultimately expressed in scripture; and one might also think that this influence contributed to making the Bible difficult to interpret in just those places. Thus, one might hold similar views about the influence of bias against women. Those who endorse a high Christology will insist that Jesus himself was free of such bias; but it does not follow that his own pattern of reference to God is uninfluenced by human biases. Perhaps, for example, it was aimed at accommodating (for reasons presumably unknown) long-entrenched beliefs about the God of Israel.

116

   :  

I have argued in this section of the chapter, Cooper’s case for that claim is a failure; and, for my part, I can see no other convincing reason to endorse it. Typical defences of inclusivism assume either that the concept of God is a human invention or that divine transcendence precludes the literal application to God of just about any human term, thus making both masculine and feminine characterizations equally appropriate. My goal in this chapter has been to offer metaphysical underpinning for inclusivism that rests on neither of these assumptions, nor on controversial assumptions about the nature of gender. I have argued that if God is masculine or feminine at all, then God is equally masculine and feminine. I have also provided some reason to think that if this conclusion is true, then it is morally and liturgically appropriate to characterize God in feminine terms—the most salient reason just being that, judging from how God is portrayed in scripture, God seems to prefer not to be characterized in ways that are known to be misleading.²⁹

5. Postscript (2019) Soon after this chapter was originally published in 2016, I was invited to write a paper developing the same ideas for inclusion in an undergraduate textbook (cf. Rea 2019). In the course of revising this material to make it more accessible, I realized that I could formulate the argument of Section 3 (against ‘genderism’) a bit more cleanly and simply as follows. Let the equal gender thesis be the claim that God is masculine or feminine only if God is equally masculine and feminine. Genderism is the negation of the equal gender thesis; so any argument for the latter thesis will be an argument against the former. And let us say that strongly gendered attributes are attributes picked out by strongly gendered terms. Here, then, is the simplified argument for the equal gender thesis: P1.

If God is masculine or feminine, then God has strongly gendered attributes.

P2. If God has strongly gendered attributes, then God is equally masculine and feminine. P3. Therefore, if God is masculine or feminine, God is equally masculine and feminine. ²⁹ An earlier version of this chapter was discussed in the weekly reading group of the Center for Philosophy of Religion at the University of Notre Dame. I am grateful for the helpful suggestions that came out of that discussion, and would particularly like to thank Evan Fales, Kate Finley, Matt Getz, Jack Himelright, Amy Seymour, and James Sterba. I am grateful also to Alex Arnold, Michael Bergmann, C. L. Brinks, Jeff Brower, and two anonymous referees for the journal Religious Studies for their helpful comments on earlier drafts. Finally, special thanks are due to Pamela Sue Anderson, Sarah Coakley, Robin Dembroff, and Kathryn Pogin. I have benefitted greatly from their extensive comments on earlier drafts, and from their advice both about the literature on feminism and gender and about various issues discussed in the chapter.

    

117

P1 follows straightforwardly from the definition of a strongly gendered attribute and a reasonably liberal metaphysics of properties. (If God is masculine, for example, and if there are properties corresponding to most commonsense predicates, then God will at least have the strongly gendered attribute of being masculine.) And P2 can be defended by way of some of the same considerations that entered into the defence I previously offered for P1 and P3 in the argument against genderism. Suppose God has strongly gendered attributes. Then one of the following options is true: (a) Every person of the trinity has the same strongly gendered attributes, and so all three belong to the same gender. (b) The persons of the trinity have different strongly gendered attributes, where two of them share a gender and the third is different. (c) The persons of the trinity have different strongly gendered attributes and none of them shares a gender. In each case, reasons I gave in the original argument against genderism will support the claim that God is equally masculine and feminine. Thus, if God has strongly gendered attributes, God is equally masculine and feminine. Although the same considerations that entered into my defence of the first premise of my argument against genderism could equally be used to support P2 of the revised argument here, I did not in fact appeal to exactly those considerations in developing the revised argument. In particular, whereas in my original argument I appealed to the method of perfect being theology and the claim that it is not intrinsically better to be masculine than feminine, in the revised argument I relied more on the claim that everyone is created equally in the image of God. I noted that what I mean by this claim is that differences between persons as regards gender-relevant characteristics make no difference in the extent to which those persons bear the image of God; and then I reasoned roughly as follows.³⁰ Suppose all three persons of the trinity belong to the same gender—this is option (a) above. Suppose further that there are n genders—g₁, g₂, . . . gn, of which at least two are the familiar commonsense genders, masculine and feminine. If all three persons of the trinity are masculine (say), then God is clearly masculine, and so being masculine and doing things in a distinctively masculine way will be a superior way of resembling God, of representing God to others, of performing Godlike functions on earth, and of doing or being any of the various other things that theologians have taken to be central to being a creature in God’s image. In that case, people of non-masculine genders are not created equally in the image of God. Similar will be true if all three persons of the trinity are feminine, or if all three are some other gender. So if, as we are assuming here, people of all genders are equally created in the image of God, then, if all three persons belong to the same gender, they belong equally to all genders. ³⁰ I say ‘roughly’ because in that paper I still made a simplifying assumption about how many genders there are—I assumed that the only gender options were masculine, feminine, and both. Here I drop that assumption.

118

   :  

Once we see this, it is easy also to see what to say about options (b) and (c). Either those two options imply that God belongs more to one gender than to others, or not. If they do, then two things follow: First, it follows that the person(s) of the trinity who belong(s) to the non-dominant genders is/are less what God fundamentally is than the person(s) who belong to the dominant gender. Second, it follows that people who belong to the genders that are not dominant in the godhead image God less than people who belong to the dominant gender. Neither conclusion is acceptable; thus, we should conclude either that options (b) and (c) are untenable or they do not imply that God belongs more to one gender than to others. Either way, we reach the conclusion that, if God has any strongly gendered attributes, God belongs equally to all genders.

References Butler, Judith. 1990. Gender Trouble : Feminism and the Subversion of Identity. New York: Routledge. Bynum, Caroline Walker. 1982. ‘Jesus as Mother and Abbot as Mother: Some Themes in Twelfth-Century Cistercian Writing’. In Jesus as Mother: Studies in the Spirituality of the High Middle Ages, 110–69. Berkeley: University of California Press. Carr, Anne. 1988. Transforming Grace: Christian Tradition and Women’s Experience. San Francisco, CA: Harper & Row. Chodorow, Nancy. 1978. The Reproduction of Mothering Psychoanalysis and the Sociology of Gender. Berkeley: University of California Press. Coakley, Sarah. 2013. God, Sexuality, and the Self : An Essay ‘On the Trinity’. Cambridge, UK; New York: Cambridge University Press. Congar, Yves. 1983. I Believe in the Holy Spirit, Vol. III: River of Life Flows in the East and in the West. Translated by David Smith. New York: Seabury Press. Cooper, John W. 1998. Our Father in Heaven: Christian Faith and Inclusive Language for God. Grand Rapids, MI: Baker Books. Daly, Mary. 1973. Beyond God the Father: Toward a Philosophy of Women’s Liberation. Boston, MA, Beacon Press. Daly, Mary. 1986. The Church and the Second Sex. Boston, MA: Beacon Press. Fee, Gordon D. 1987. The First Epistle to the Corinthians. Grand Rapids, MI: Eerdmans. Hampson, Daphne. 1996. After Christianity. 1st North American edn. Valley Forge, PA: Trinity Press International. Haslanger, Sally. 2000. ‘Gender and Race: (What) Are They? (What) Do We Want Them to Be?’ Noûs 34: 31–55.

    

119

Hays, Richard B. 1997. First Corinthians. Louisville, KY: John Knox Press. Jantzen, Grace. 1988. Julian of Norwich: Mystic and Theologian. New York: Paulist Press. Jantzen, Grace. 1998. Becoming Divine: Towards a Feminist Philosophy of Religion. Manchester: Manchester University Press. Jenson, Robert. 1992. ‘ “The Father, He . . . ” ’. In Speaking the Christian God: The Holy Trinity and the Challenge of Feminism, edited by Alvin F. Kimel Jr. Grand Rapids, MI: Eerdmans. Jenson, Robert W. 1997. Systematic Theology, Vol. 1: The Triune God. New York: Oxford University Press. Johnson, Elizabeth. 1984. ‘The Incomprehensibility of God and the Image of God as Male and Female’. Theological Studies 45: 441–65. Johnson, Elizabeth. 1993. She Who Is: The Mystery of God in Feminist Theological Discourse. New York: Crossroad. Julian, of Norwich. 1978. Showings. New York: Paulist Press. McFague, Sallie. 1987. Models of God: Theology for an Ecological, Nuclear Age. Philadelphia, PA: Fortress Press. Raphael, Melissa. 2014. ‘A Patrimony of Idols: Second-Wave Jewish and Christian Feminist Theology and the Criticism of Religion’. Sophia 53: 241–59. Rea, Michael. 2019. ‘Is God a Man?’ In Contemporary Debates in Philosophy of Religion, edited by Raymond Vanarragon, 293–301. 2nd edn. London: Wiley. Ruether, Rosemary Radford. 1983. Sexism and God-Talk: Toward a Feminist Theology. Boston, MA: Beacon Press. Soskice, Janet Martin. 1992. ‘Can a Feminist Call God “Father”?’ In Speaking the Christian God: The Holy Trinity and the Challenge of Feminism, edited by Alvin F. Kimel Jr., 81–94. Grand Rapids, MI: Eerdmans. Widdicombe, Peter. 1994. The Fatherhood of God from Origen to Athanasius. New York: Clarendon Press.

6 God beyond Being Towards a Credible Account of Divine Transcendence

Theologians divide on the question of how best to understand God’s transcendence. Minimally, it contrasts with immanence and dependence. God is not part of the material universe, nor does God depend on it. It is also typically understood to imply otherness—God is holy, not just in the modern moral sense, but in the sense of being radically different from creation. But how different? For many theologians, doing justice to the grandeur of God means ratcheting up the notion of divine otherness as far as possible. From at least the Patristic period onwards, transcendence has come increasingly to be seen as implying incomprehensibility, ineffability, and unknowability. Famously, towards the end of his Mystical Theology, Pseudo-Dionysius the Areopagite goes so far as to say that God is ‘beyond every denial, beyond every assertion’ and that God ‘lies beyond thought and beyond being’ (1987: 133, 136). Many in his wake have followed him in these affirmations. But what do they mean? A standard gloss on the Dionysian ‘beyondness’ affirmations says that God can only be spoken of in images, or metaphors. Just as the scriptural portrayal of God as the chariot-riding Ancient of Days (Dan. 7: 9–10) is nothing more than an image, so too, the implication seems to be, even the portrayal of God as the omnipotent, omniscient, omnibenevolent creator is a mere image. William Alston applied the term panmetaphoricism to this view; others (including Alston) have called it pan-symbolism.¹ I prefer the latter label.² Pansymbolism is a widely endorsed gloss on divine transcendence, probably the most common unpacking of the idea that God is beyond characterization, with the apparent consequence that God is beyond being in the sense that not even ‘God exists’ is strictly and literally true. It is not only widely affirmed among

¹ Alston 1989: 35, and Rowe 1962: 278, citing Urban 1940: 35–6. ² The view that I am calling ‘pan-symbolism’ is commonly understood to imply that no true theological claim is literally true. I shall follow suit; but I deny that this is a consequence of the thesis that all true theological discourse is metaphorical. This is why I prefer not to call the view ‘panmetaphoricism’: that label in the context of this discussion presupposes a sharp contrast been the literally true and the metaphorical. I take pan-symbolism also to be distinct from the thesis that all theological discourse is analogical. For purposes here, I will make no commitment one way or the other on the relation between metaphor and analogy.

Essays in Analytic Theology: Volume I. Michael C. Rea, Oxford University Press (2020). © Michael C. Rea. DOI: 10.1093/oso/9780198866800.003.0007

  

121

contemporary theologians, but it is also routinely described as a traditional view.³ According to Alston, it is an ‘article of faith’ in contemporary theology (1989: 17). But is it credible? And if it isn’t, is there some other way of doing justice to the language of beyondness? These are the central questions of the present chapter. Contemporary analytic philosophers have mostly been reluctant to embrace the Dionysian language, and for good reason. The claim that God is beyond every assertion seems itself to be an assertion about God, and presumably one which God is not beyond. If God is truly beyond being, then it seems that God must not have being, and so must not exist—apophaticism implies atheism. It would be tempting to dismiss the remarks at the end of The Mystical Theology as overly exuberant and unintelligible hymnody were it not for the fact that similarly puzzling and apparently paradoxical remarks are to be found elsewhere in the Dionysian corpus, as well as in the writings of a great many other theologians in the apophatic tradition. The language of beyondness is deeply entrenched in the tradition, so much so that one is apparently forced to a dilemma: dismiss many of the greatest and most influential Christian theologians as purveyors of nonsense, or try somehow to do justice to that paradoxical language. I’m going to try to do it justice. My goal here is to try to develop an intelligible account of divine transcendence that accommodates the language of beyondness while at the same time remaining consistent with the truth of central Christian doctrines—for example, that God exists, is personal, became incarnate, and is appropriately characterized through the various images mentioned at the beginning of this talk. I do not endorse the view that I will develop here; but, in light of the prevalence of the beyondness affirmations in the tradition, the project of trying to make theoretical sense of them seems worthwhile all the same. I will begin with some brief remarks about Pseudo-Dionysius. I am not a historian, and I have no comprehensive reading of his work to offer. But I do think that there is some clear evidence in his corpus that his own ‘beyondness’ affirmations are not the sort of paradox-mongering nonsense that some contemporary thinkers see in his work. Instead, I think that they are hyperbolic or poetical expressions of intelligible truths about God. If I am right, then I think there is good reason to think that doing justice to the language of beyondness—at least to his language, and that of many of those who were influenced by him—does not require taking it with full, literal seriousness. I will then explain why I think that the task of making sense of the two central beyondness affirmations—that God is beyond being, and that God is beyond characterization—appears so intractable. Finally, I will make my own attempt at that task, which will, I think, amount to reconditioning a familiar old strategy in light of some recent ideas in metaphysics and philosophy of language.

³ See, e.g. Hampson 1996; Hick 2000; Johnson 1984; and McFague 1987.

122

   :  

1. I said in the previous section that I think many of the most paradoxical remarks that one finds in the work of Pseudo-Dionysius are not meant to be taken with full, literal seriousness. I think that the same can be said of many other theologians in the apophatic tradition; but for now I will focus only on Pseudo-Dionysius. Consider, for starters, The Celestial Hierarchy. There, Dionysius states outright that God is known by human beings: ‘God’, he says, is ‘known in all things and as distinct from all things. He is known through knowledge and through unknowing.’⁴ It is a real challenge to unpack the difference between ‘knowledge and . . . unknowing’. But that is neither here nor there for the present point, which is just that Dionysius fully acknowledges that God is both knowable and known by creatures. The point is even clearer in The Divine Names. In the opening chapter, he declares that we ‘learn’ of the mysteries of God from scripture, and from what the scripture writers have to say about the divine names (the predicates that are applicable to God); the problem is just that our knowledge is halting and incomplete, and comes mostly by way of ‘symbols’ (sumbόlois) and ‘images’ (eivkόnas) as opposed (presumably) to coming by way of clear and literal discourse.⁵ But, he says, . . . in time to come, when we are incorruptible and immortal . . . we shall be ever filled with the sight of God shining gloriously around us . . . and we shall have a conceptual gift of light from him and, somehow . . . we shall be united with him and, our understanding carried away, blessedly happy, we shall be struck by his blazing light . . . . But as for now, what happens is this. We use whatever appropriate symbols we can for the things of God. With these analogies we are raised upward toward the truth of the mind’s vision, a truth which is simple and one. (1987: 52–3)

Assuming human beings lose neither their humanity nor their status as creatures upon acquiring the beatific vision, and assuming God does not cease to be transcendent when human beings acquire the beatific vision, what this passage conveys is that creatures can under the right conditions attain clear knowledge and understanding of God despite God’s transcendence. The obstacles to such knowledge and understanding are, according to the picture suggested here, due only to divine transcendence in conjunction with our present, materially embodied circumstances. Keeping in mind that Pseudo-Dionysius was a neo-Platonist, it is entirely natural to read the quoted passage as indicating that the fundamental

⁴ Pseudo-Dionysius 1987: 108–9.

⁵ Cf. Pseudo-Dionysius 1990, 592D and 869D.

  

123

problem posed by transcendence is simply one of acquaintance: as embodied creatures in a material world, we do not have the right acquaintance with the immaterial deity and the divine attributes to have anything other than halting, incomplete, and merely symbolic or analogical understanding; but someday, in the beatific vision, we will have the right acquaintance. Of course, there is no denying that some of the claims in the Dionysian corpus seem, in isolation, to imply that God is unknowable, uncharacterizable, and even non-existent. But often enough those claims do not occur in isolation; they are instead tempered by nearby claims that suggest we ought to look for a nonparadoxical interpretation. So, for example, although he says ‘we must not dare to apply words or conceptions to this hidden transcendent God’, immediately afterwards he says ‘[w]e can use only what scripture has disclosed’ (50). Somehow, then, whatever he means by not applying words or conceptions to God, it is perfectly compatible with using in our theological discourse the words and images that scripture employs. Contrary to what is often said about him, Pseudo-Dionysius also seems clearly to think that God can be talked about. For example, he asks: ‘How then can we speak of the divine names . . . if the Transcendent surpasses all discourse and all knowledge, if it abides beyond the reach of mind and of being . . . ? How can we enter upon this undertaking if the Godhead is superior to being and is unspeakable and unnamable?’ (53). The question presupposes that it is possible to predicate things of God; and, indeed, The Divine Names is filled with discourse predicating things of God. What the question seems to be asking, then, is just this: ‘What exactly are we trying to say with apophatic language like “God is superior to being” or “God is unspeakable and unnamable”, given that we can and do speak freely about the divine names?’ I see no reason to doubt that PseudoDionysius thought that this question could receive an intelligible answer. So whatever Dionysian apophaticism comes to, it has to be consistent with the idea that it is true that God is in some sense that allows for true predicative discourse about God, that there are knowable and sayable truths about God, that human beings can convey these knowable truths at least in symbolic terms, and that it is in principle possible for human beings, even if only in the beatific vision, to have clear, non-symbolic knowledge of these truths. And, indeed, this seems to be a point of general agreement among theologians who emphasize divine transcendence on at least into the eighteenth century (after which, in the wake of Kant, things start to get a little weird).

2. Doing justice to the language of beyondness requires nothing more than a credible paraphrase or explanation that shows how claims like ‘God is beyond being’ and

124

   :  

‘God is beyond characterization’ might reasonably be thought to express truths. The trouble, however, is that even this modest task appears intractable. Consider the claim that God is beyond being. As regards God’s relation to being, we seem to have just four options. I will formulate these options in terms of ‘modes’ of being; but, if you prefer, you should feel free to imagine them formulated in terms of ‘categories’ or ‘kinds’ of being: (a) There is just one kind, or mode, of being and it is not the case that God either is being or has being. (b) There are multiple distinct modes of being and it is not the case that God is or has any of them. (c) There is just one kind, or mode, of being and God is or has it. (d) There are multiple distinct modes of being, and God is or has one of them. The problem, in short, is that options (a) and (b) appear to imply that God does not exist, whereas options (c) and (d) appear to imply that God is in no meaningful sense beyond being. Perhaps the most common gloss on the idea that God is beyond being is this one, which has well-known roots in medieval philosophical theology: God is being itself; as such, God is beyond creaturely being, which is distinct from but somehow grounded in God. Moreover, creaturely being is analogous to or participates in the being that God is. My objection to this view is that there is no way of unpacking the idea that God is being itself that seems at all promising. The exact problem one faces in saying that God is being itself depends on what one takes being to be (don’t say ‘God’), and on what one thinks it means to have being. Some think that being is a property; some do not. If being is not a property, then nothing has being—since being is clearly neither an object that can be had (like a purse, or a computer), nor a mass (like money, or hair).⁶ So we can rule out that option. Being is a property. If being is a property, then either it is a universal or it is not. Suppose that it is not. A survey of the various theories of properties on offer according to which properties are not universals reveals that if being is a property but not a universal, then either it is a set or it is some kind of concept or idea. But God is not any of those things. So being is a universal—the sort of thing that can have instances or be exemplified or (in Platonic language) the sort of thing in which something can ‘participate’. It seems quite odd to say that God has instances, or to say that something exemplifies God. Indeed, it seems more than odd—it seems false—to say (as we must, if God is the property of being) that everything is an instance of God or exemplifies God. Best, then, to retreat to ⁶ Perhaps it is a condition—like the flu, or hives? In that case, I think, being is a property. I take it that predicates like ‘has the flu’ or ‘has hives’ attribute properties rather than possessions (like purses, or money) to their subjects.

  

125

something a little more mysterious and say—as has traditionally been said by those who want to say that God is being itself—that God is akin to a Platonic form in which things participate, and that everything that has being somehow participates in God. If the foregoing line of reasoning is correct, then it looks as if the only promising way of unpacking the idea that God is being itself is the traditional Platonic way of unpacking it. One might try to depart from that traditional understanding; but doing so implies absurdities, like the claim that God is a set or that everything is an instance of God or that being is an object or a mass that can be had in the way that people have computers or money. Unfortunately, the traditional Platonic understanding of the claim that God is being itself is unworkable. The reason is twofold. First, if God is being itself, then God is the semantic value of the word ‘being’; when we predicate being of anything, we predicate God of it. That’s fine within a Platonic metaphysic; but it makes hay of the idea, central to the apophatic tradition, that God is, in a deep way, beyond our concepts. Being is one of our most basic concepts; and if God is the semantic value of our being-words—and if, as is traditionally affirmed, God’s being is God’s essence—then God and God’s essence are about as cognitively accessible as anything ever gets. Second, even if we deny that every Platonic form participates in itself, there is no in-principle reason for insisting that no Platonic form participates in itself. Thus, it is in principle open to us to say that being itself has being, in which case it is not beyond being. In other words, the thesis that God is being itself is not equivalent to and does not even straightforwardly imply that God is not ‘a thing among others’, something that transcends being rather than something that has being. To secure that claim, we have to assert it in addition to the claim that God is being itself. But then why bother endorsing that latter doctrine at all? In other words, the doctrine that God is being itself fails to imply the doctrine it is intended to gloss—namely, that God is beyond being— and, as a result, stands unmotivated. So much for the claim that God is beyond being. What about the claim that God is beyond characterization? Here we find varying glosses in the literature. I cannot survey all of them here; but I will say that, apart from pan-symbolism, I think that the basic problem with all of them is that they imply that God is no more ‘beyond characterization’ than any of a wide range of perfectly ordinary and eminently characterizable things—things like tables and chairs, people, and the like.⁷ ⁷ For example, I think that is the most important problem that besets Jonathan Jacobs’s 2015 characterization of ineffability. Jacobs defines ineffability as follows: : For any proposition p about God’s intrinsic properties, p is not fundamentally true and p is not fundamentally false. Jacobs doesn’t offer a formal definition of ‘fundamentally true’; but he explains it as follows: to say that a proposition p is fundamentally true is just to say that it ‘carves nature at its joints, or represents reality in a perfectly ontologically perspicuous way’ (2015: 162) But it would seem that a minimal

126

   :  

Pan-symbolism remains the most promising account on offer for doing justice to the idea that God is beyond characterization; but it faces some serious problems. I shall focus on two of them, one due to William Alston and one due to Daniel Howard-Snyder. Alston’s (1989) objection to pan-symbolism rests on the idea that it belongs to the very nature of a truth-apt metaphor to attribute a property to the subject of the metaphor, and to posit similarity between the subject and an ‘exemplar’ of the metaphorical predicate. So, for example, when Romeo says that Juliet is the sun, he obviously does not attribute being the sun to Juliet; but, if he is what Alston regards as a ‘typical’ metaphor user, he does employ the metaphor to attribute some property to her—a property which he has in mind and specifically intends to attribute to her. This attribution is part (even if not all) of the propositional content of the metaphor. But, says Alston, one can have a property in mind and intend to attribute it to something only if one has a concept of the property; and, having a concept of a property, however inchoate, suffices for that property’s in principle being available to be the semantic value—the literal content—of a term in one’s language. So, on the assumption that at least some of our speech about God is true, at least some of it will have propositional content part of which will be, in principle, literally expressible. To put it another way: pan-symbolism implies that all speech about God is either untrue or devoid of content, which is a bad result. Howard-Snyder (2017) objects that pan-symbolism, or what he seems to think of as the most promising versions thereof, implies that theism is not true. His defence of this claim is of a familiar sort—what one might call ‘argument by pathwinnowing’—wherein one floats an argument against a position, considers a variety of replies (some involving modifications to the position, others involving substantive auxiliary assumptions), and then arrives at what looks to be the best path forward for the view, albeit one that still arrives at problematic consequences. condition on a proposition representing reality in an ontologically perspicuous way is that it be expressible by a sentence that correctly represents the fundamental structure of reality—i.e. its component names refer to fundamental entities, its component predicates express fundamental properties, and any other terms in the sentence (e.g. terms like ‘and’ or ‘there are’) ‘carve nature at the joints’ as well. But here is where the problem begins to emerge. Everyone who posits a metaphysically serious distinction between fundamental and non-fundamental entities wants to say that the world includes non-fundamental things. If one did not believe that there are non-fundamental things, there would be no point in making the distinction. Moreover, pretty much everyone also believes that a great many perfectly ordinary things—things we understand quite well, things that we purchase at supermarkets and electronics or home improvement stores, and things we make for ourselves at home—are non-fundamental. If that is right, then no sentence about such things could be either fundamentally true or fundamentally false, since no sentence about such things would contain only terms that refer to fundamental entities or properties. As a result, all such things count as ineffable. But the whole point of a doctrine of transcendence is to set God apart from mundane things in some metaphysically important way. If it turns out that God is no more or less ineffable—no more or less beyond characterization, or beyond assertion and denial—than a table, a computer, or a sack of flour, this goal has not been accomplished. For further objections against Jacobs’s proposal, see Lebens 2014, 2016.

  

127

Providing a faithful summary of the entire argument would take me too far afield; so I will restrict my focus to just one of its key threads. The formulation below is mostly my own; so, although I believe Howard-Snyder is to be credited for whatever insights it contains, and although I know he endorses a similar line of reasoning, readers will have to judge for themselves whether he can be held responsible for its vulnerability to the particular objections I will raise.⁸ Consider the claim that God exists. Pan-symbolism implies that this claim is not literally true. So, pan-symbolism implies either that it is not true at all that God exists—atheism is true—or that the following claim is true: (1)

‘God exists’ is true, but not literally true.

But (1) looks to be incoherent: the first conjunct implies the negation of the second. To see why, note first that (1) seems to imply (2): (2)

‘exists’ applies non-literally to God.

But since, in general, a predicate applies to x only if there is such a thing as x, (2) seems to imply (3): (3)

There is such a thing as God.

(3),

in turn, seems to imply (4):

(4)

‘God exists’ is literally true,

which is the negation of the second conjunct of (1). So (1) is incoherent. I hope it is clear now why one might think that the task of providing a credible paraphrase or explanation of the beyondness doctrines is intractable. In short: the most credible glosses on the thesis that God is beyond being seem to imply atheism or absurdities like ‘God is a set’ or ‘God is exemplifiable’; and the most credible gloss on the thesis that God is beyond characterization seems to imply both atheism and the thesis that talk about God is devoid of truth or content. In fact, however, I think that the task is not intractable, and I hope to show this over the course of the next two sections. In the remainder of this chapter, I will develop a version of pan-symbolism that does not fall prey to the two objections I have just mentioned. In doing so, I will provide what seem to me to be the most promising glosses on the idea that God is ⁸ Howard-Snyder’s defence of the claim that pan-symbolism implies the negation of theism starts in earnest, I think, on p. 31 and finally wraps up (after various detours through unpromising revisions to pan-symbolism) on p. 47. The thread I am most interested in appears on pp. 33–4.

128

   :  

beyond characterization and beyond being. Section 3 focuses on Alston’s claim that pan-symbolism implies that discourse about God is devoid of truth or content, and ends by providing an account of what it might be for God to be beyond characterization. Section 4 focuses on Howard-Snyder’s argument for the conclusion that pan-symbolism implies atheism, and ends by providing an account of what it might mean to say that God is beyond being.

3. Let me begin by saying a bit more about metaphor. On Alston’s view (and Howard-Snyder seems to be following Alston), a truth-apt metaphor uses terms in a semantically deviant way; it presents an exemplar of the metaphorical predicate as a useful ‘model’ of the metaphor’s subject; and part of its propositional content, in the typical case, is a specific property-attribution which the metaphor-user has explicitly in mind. But all of these claims about metaphor strike me as incorrect. For one thing, it seems clear (and is now generally taken for granted) that metaphor occurs not just at the level of terms, but also at the level of sentences, paragraphs, and even larger chunks of text.⁹ It also seems clear that some metaphors use terms fully in accord with their standard or stipulated meanings, do not present any particular ‘model’ of anything, and do not necessarily attribute to anything a specific property ostensibly had in mind by the speaker or hearer of the metaphor. Consider, for example, the metaphor embedded in the following line from Metallica’s No Leaf Clover: And it comes to be that the soothing light at the end of your tunnel is a freight train coming your way.

This is a truth-apt metaphor—one might readily agree with it or disagree with it. But it is not at all clear that any of its constitutive words is being used in a semantically deviant way. Nor is it clear that any particular thing is being modelled on a freight train (or a train’s headlight), or that the metaphor-user has in mind some specific property that he means to be attributing to the metaphor’s subject. Or consider the sentence, ‘A work of art is not an egg’, which serves one of the more important load-bearing examples in David Cooper’s (1986) study on metaphor. The sentence is clearly a metaphor. But no term is used in a semantically deviant way; works of art are clearly not being modelled on non-eggs; and there seems to be no particular property of non-eggs

⁹ Cf. White 1996: 56–8.

  

129

that the metaphor-user means to attribute to works of art in general. Of course, one might be tempted to respond on Alston’s behalf by saying that these metaphors are atypical; but the fact is that they are not. So what is going on in metaphor? One common view, which I endorse but will not try to defend here, is that metaphorical expressions—most saliently sentences or paragraphs, but perhaps also terms—convey content that differs from their semantic content, and what they say may or may not also include that semantic content.¹⁰ So, for example, ‘A work of art is not an egg’ has a true semantic content; and that content is part of what the metaphor says. But what it says also includes more than that semantic content. The line from No Leaf Clover has a false semantic content for most of us. (Few of us, if any, have found ourselves in a tunnel being soothed by the headlight of a freight train.) The semantic content of that line, furthermore, is no part of what the metaphor says. However, it does say something. In fact, it conveys a content that will ring true for many listeners. So metaphors are expressions that say something other than what they literally mean. Sometimes they say what they literally mean and more; sometimes they say something different without saying what they literally mean. Another way of putting it, then: On the assumption that an expression is semantically deviant if and only if it is not being used to say what it semantically expresses, metaphors are pragmatically enriched forms of speech that are often but not always semantically deviant. Metaphors are not the only expressions that work this way; but that is neither here nor there for present purposes. Now let us return to pan-symbolism. As I see it, pan-symbolism is best characterized as the view that true intrinsic predications of God are both semantically deviant and pragmatically enriched. In other words, true statements that predicate intrinsic properties of God do not say what (if anything) they semantically express; rather, what they say is just the content that they pragmatically convey. Note that, on this view, pan-symbolism itself is not a thesis about God so much as a thesis about the semantics and pragmatics of theological discourse. Note, too, that, as I have characterized it, pan-symbolism is fully consistent with the idea that some or maybe even all discourse about God has intelligible propositional content. Alston, as we have seen, maintains that if a metaphor has propositional content, then the metaphor will be at least partially eliminable; for part of the content will include a property attribution, and the property in question will be one of which the metaphor-user has some determinate concept. This is the heart of his objection to pan-symbolism. But, on the view of metaphor that I have sketched, the objection is unmotivated. A metaphor’s having propositional content does not depend upon the speaker’s having concepts of specific properties that she intends ¹⁰ This view is developed and defended in various papers by Elisabeth Camp—most importantly Camp 2006a, 2006b, and Reimer and Camp 2006.

130

   :  

to attribute to anything; and the question whether a metaphor is eliminable (partially or wholly) just boils down to the question whether and to what extent the content conveyed by the metaphor can be the semantic value (rather than simply the pragmatic import) of a sentence in our language. Given that every language has finite expressive resources, there is no obvious reason to think that just any content that can be conveyed by a sentence can also be semantically expressed in the language. Thus, contrary to what Alston has argued, there is no obvious reason to think that pan-symbolism guarantees that speech about God is devoid of truth or content. One might object at this point that, as I have characterized it, pan-symbolism does not do justice to the idea that God is ineffable. On my characterization, pansymbolism is fully consistent with the thesis that we can use human concepts and categories to convey truths about God. Our only limitation is that we are, for some reason, unable to use sentences whose semantic content is a truth about God. But, one might object, this all by itself is not much of a limitation at all. It implies no very interesting limitation on our ability to understand what God is like, or to say what God is like. In response, however, I submit that few if any of the apophatic theologians to whose views I’m most interested in doing justice—theologians like Gregory of Nyssa, Pseudo-Dionysius, the author of The Cloud of Unknowing, and others— ever thought that we can’t say what God is like. After all, many of them spilt a lot of ink in trying to say what God is like. What they denied, I believe, is that human sentences have or can have as their semantic values propositions that describe what God is intrinsically like. We can gesture at such propositions; we can understand them to some degree when they are conveyed in metaphor; but the expressive power of human language is too limited to accommodate them.

4. What about the charge that pan-symbolism implies atheism? The issue turns on whether it makes sense to suppose that (1) is both true and consistent with theism: (1)

‘God exists’ is true, but not literally true.

Recall that the argument given earlier for the incoherence of (1) rested on the claim that (1) implies (2), (2) implies (3), and (3) implies (4): (2)

‘Exists’ applies non-literally to God.

(3)

There is such a thing as God.

(4)

‘God exists’ is literally true.

  

131

I won’t contest the claim that (2) implies (3); but I deny that (1) implies (2) and that (3) implies (4). As noted earlier, it does not follow from the fact that a sentence is non-literal that some particular term in the sentence is being applied non-literally to the subject of the sentence.¹¹ So (1) does not imply (2). Moreover, it seems that ‘God exists’ and ‘There is such a thing as God’ say the same thing, whatever that might be; so if one is non-literal then the other might be as well. (3), therefore, doesn’t imply (4)—or, at any rate, it doesn’t apart from the further premise that ‘There is such a thing as God’ is literally true if true at all. So the Howard-Snyder-inspired argument fails. However, I can’t leave the matter there; for, after all, (1) sure sounds a lot like a devious form of atheism. It sounds as if, in affirming (1), a person is gearing up to affirm some sort of Kaufmanian reduction of the thesis that God exists—e.g. a claim like ‘ “God is real” [means] that in fact there are forces in the world that contribute to human flourishing.’¹² Perhaps some who endorse pan-symbolism would be happy with this kind of view; but those who are drawn to pansymbolism for reasons of piety and respect for the Christian tradition will not wish to affirm it, or anything like it; so more must be said. My suggestion is that theistic pan-symbolists might develop a credible, nonreductive, non-literal reading of ‘God exists’ by appeal to doubts about the possibility of absolutely unrestricted quantification.¹³ There are well known and persuasive reasons for thinking that, for any domain associated with a quantifier, or for any extension associated with the predicate exists, there will be (in some sense of ‘be’) things (in some sense of ‘things’) that fall outside the relevant class. Some have to do with technical concerns about absolute generality: paradoxes arise out of the supposition that there is an absolutely unrestricted quantifier, and these can be avoided only by taking on certain controversial commitments. Others arise out of the familiar Aristotelian idea that there are different modes of being, that quantifiers and other ‘being words’ refer ambiguously to these different modes, and that there is no absolutely general quantifier or existence predicate that encompasses all of them. I don’t claim that these reasons are compelling; and I will offer no defence of them here. My claim is simply that if we are persuaded by the arguments against absolute generality, then there is good reason to think that the sentence ‘Something doesn’t exist’ can be used to convey a truth that cannot be expressed. Generality relativism implies that something falls outside the scope of every quantifier. This, in turn, implies that, whatever we might mean by ‘exist’, it is true that something doesn’t exist. But where no explicit or contextual restriction has been imposed on our quantifiers, either ‘something’ and ‘exist’ in the sentence ‘Something does not exist’ have the same meaning, in which case the sentence is ¹¹ This is contrary to what Howard-Snyder explicitly assumes. Cf. 2017: 26. ¹² Plantinga 2000: 51, discussing quotations from Kaufman 1981. ¹³ Cf. Williamson 2003 and the essays in Rayo and Uzquiano 2007, especially Hellman 2007.

132

   :  

contradictory, or they have no definite meaning, in which case the sentence expresses nothing. Nevertheless, we understand it, and we see that it is true. A truth is conveyed, despite being inexpressible. Once we have granted this, we should have no objection to saying that, in at least some contexts, (5) is true: (5)

Something is such that it is true, but not literally true, that it exists.

The reason is that (5) seems to be equivalent to (6): (6)

Something falls outside the semantic value of ‘exists’.

(6), of course, expresses a falsehood—again, if ‘something’ is meaningful in that sentence, it is a quantifier that ranges over the extension that is the semantic value of ‘exist’. Nevertheless, (6) seems intelligible, and seems to convey a truth that is implied by generality relativism. Once we have granted (5) and (6), however, there seems to be no obstacle to a robustly theistic affirmation of the thesis that there is—there really is—such a thing as God even if ‘God exists’ is not literally true. Here, then, is my overall account of what it might be for God to be beyond being. Suppose, following Aristotle, we recognize different senses for each of the words that we take to express being—senses that correspond to different modes of being, or different ways in which a thing might ‘exist’. Aristotle spoke of the homonymy of being, but we might do better to follow the medieval philosophertheologians who spoke instead of analogy. In particular, what I want to say is that being-words have multiple analogous senses, each of which is intelligibly related to the others (in the way in which homonymous words, like ‘[river] bank’ and ‘[financial] bank’ often are not necessarily), and I want to say that corresponding to each of the different senses for being-words is a different mode of being. So, for example, exists might mean one thing when predicated of a material substance (e.g. a horse) and something different when predicated of an abstract object (e.g. a number). So far, so familiar. Earlier in this chapter I said that one common gloss on the idea that God is beyond being maintains that God has a mode of being that is different from the familiar creaturely modes (whatever they are) but that is intelligibly related to them as a mode of being. The problem with this gloss was that it seemed straightforwardly to imply that God has being, and so is not beyond being. What I propose to add to the gloss, however, is that God’s mode of being is not expressible by human ‘being’ words. It cannot be the semantic value of any human term. So the thesis that God has a mode of being that differs from creaturely modes is true, on this view, but semantically deviant. It is semantically deviant because the word ‘being’ can only have as its semantic value creaturely modes of being (or disjunctions thereof); so, what it expresses (viz. that there is a mode of

  

133

creaturely being had by God that differs from all modes of creaturely being) is false—indeed paradoxical, as critics readily point out—even though the sentence is being used to convey a truth.¹⁴ In saying this, I place a heavy burden on the distinction between expressing and conveying; and one might well ask how it is that we can convey propositions that are literally impossible to express. The answer, I think, is that expression, unlike conveyance, is solely a function of the meanings of words. What we can express is limited by the semantic values of our words; but what we can convey is limited only by the variety of thoughts that might be (reliably) stimulated by a particular string of words. The propositions we can express are thus a subclass of the propositions we can convey. Are we now perhaps in danger of affirming atheism? I have already acknowledged that the thesis that God is beyond being implies that it is not literally true that God exists; and one might reasonably think that, if it is not literally true that God exists, then it is literally true that nothing is God. And that truth, one might think, is identical to atheism. In fact, however, generality relativism implies that it is not. The thesis that nothing is God expresses atheism only if the quantifier is absolutely unrestricted. Generality relativism implies that it isn’t; thus, ‘nothing is God’ falls short of expressing atheism. Atheism is, plausibly, the proposition that is conveyed by the sentence ‘nothing is God’; but there is no reason to think that that proposition must be affirmed by someone who accepts that God is beyond being.

5. In closing, let me note that my account of the language of beyondness places God beyond being in pretty much exactly the same sense in which God is beyond characterization. God is beyond being in that God’s mode of being is one that is not expressible by human being-words or by sentences containing those words. Likewise, God is beyond characterization in that none of our sentences have, or can have, intrinsic predications of God as their semantic values. Even so, however, many of our words—‘wise’, ‘loving’, ‘exists’, and so on have semantic values that (by hypothesis) are at least intelligibly analogically related to properties intrinsic to God. ¹⁴ This view comports well with, but is not obviously equivalent to, the ‘illuminating falsehood’ view of apophatic speech defended by Samuel Lebens (2014, 2016). Quantifier pluralism and Aristotle’s views on being are not drawn into the development of Lebens’s view in the way that they are in mine; and Wittgensteinian considerations play no role in the development of my views as they do in the development of Lebens’s. Moreover, on my account, but not obviously on Lebens’s, apophatic claims convey true content; so it is misleading (I would say) to describe them as ‘illuminating falsehoods’. That said, however, there is much that is similar in the two views; and Lebens tells me in correspondence that the view I am developing here ‘is precisely in the family of options [he] was trying, somewhat tentatively, to sketch’.

134

   :  

Does it follow from this that God is not really wise, not really loving, and not really even existent? Again, no. It is apt to say that God exists and is wise and loving; it is true to say that God exists and is wise and loving. Indeed, although I have talked a lot in this chapter about metaphor, I am not even suggesting that the pan-symbolist ought to think such claims are merely metaphorical. What is true, according to the pan-symbolist that I envision, is that claims like ‘God exists’, ‘God is wise’, and so on resemble metaphors by virtue of being semantically deviant, pragmatically enriched forms of speech. In saying this, the pan-symbolist does not categorically deny that God exists, is wise, and so on. Rather, she simply affirms the familiar theological truth that God is wise, but that God’s wisdom is no creaturely wisdom. Hence the familiar refrain of apophatic ascent: ‘God is wise; God is not wise; God is beyond wise.’

6. Postscript (2019) I wrote this text with the intention of developing the view defended herein at greater length in my book, The Hiddenness of God. At the time, I had been reading—partly for research, and partly for personal reasons—a variety of patristic and medieval mystical texts, including the works of Pseudo-Dyonisius, The Cloud of Unknowing, and various works by Gregory of Nyssa, St Theresa of Ávila and St John of the Cross. Although still not inclined to make a fully apophatic turn myself, I was becoming quite sympathetic to the robust conception of divine transcendence that one finds in some of the aforementioned works. I wanted to see if good sense could be made of even the darkest apophatic sayings—especially the ones I describe in the chapter as the Dyonisian beyondness affirmations. Pan-symbolism seemed promising, the objections of Alston and Howard-Snyder notwithstanding, and this partly because their objections seemed predicated on misguided views about metaphor. Shortly after delivering this text as the William P. Alston Memorial Lecture at the 2016 Central Division Meeting of the APA, however, I abandoned the view that I was exploring herein. This for two reasons; one philosophical and one pastoral. Philosophically speaking, I became convinced that the objection that pansymbolism is inconsistent with theism sticks, albeit not quite for the reasons Howard-Snyder gives. Start again with a rough statement of the initial objection: Theism implies that something is God; but, since this is a theological thesis, pansymbolism implies that it is not literally true; so pan-symbolism implies that theism is not true. I still think that the pan-symbolist’s best reply is to deny that ‘not literally true’ implies ‘not true’, and so to say that ‘God exists’ is true but not literally true. But, as I explain in The Hiddenness of God, even if this is the pan-symbolist’s best reply, it is not a good reply. Howard-Snyder argues that this reply (depending on how it is deployed) results either in incoherence or

  

135

atheism (2018: 33–6). My own reason for rejecting it is different. The reason I reject the reply is, in short, that it is completely unmotivated: there is simply no reason to think that ‘God exists’ must be understood non-literally; and if there is no reason to think that it must be understood non-literally, there is likewise no reason to think that one who affirms it as a non-literal truth counts as a theist. Compare: Suppose someone maintains that the core theses of Darwinism are true, but not literally true. Do we count them among the Darwinists? I say no. There is no reason to think that the core theses of Darwinism are offered up as mere metaphors—despite the prevalence of metaphor in science. Thus, even if the claim that the core theses of Darwinism are true but not literally true counts as an interesting way of affirming those theses, it is not an affirmation of Darwinism and is, in fact, inconsistent with Darwinism. Why think that there is no reason to think that ‘God exists’ is mere metaphor? The answer is just that, unlike other predicates (like ‘is good’ or ‘is loving’) that are ripe for apophatic treatment, there is no theological reason for thinking that existence-words express merely ‘creaturely’ modes of being. Of course, as noted above, one can appeal to puzzles about unrestricted quantification to make a case for the claim that (roughly put) there are modes of being with which we are unacquainted, and that one of these is or might be a divine mode of being. (And then one can make every effort to try to state this precisely in a way that doesn’t inadvertently presuppose that some of our words semantically express the divine mode of being.) But none of this motivation for the idea that existence-words only express creaturely modes of being is theological; and even within metaphysics, it would all be quite controversial. Accordingly, though I deny that the claim that ‘God exists’ is true but not literally true strictly implies atheism, it also seems wrong to suggest that one who affirms it thereby counts as affirming theism. Having said that, I do admit that someone who maintains that ‘God exists’ is true but not literally true might well insist either that the boundaries of theism should be expanded to include their view, or that it doesn’t really matter whether they count as a ‘theist’ so long as they affirm—as they do—that it is true that God exists. But this brings me to my second, and more pastorally oriented, reason for abandoning the position described in this chapter. At the time that I wrote it, I was President of the Society of Christian Philosophers; and, as I further considered endorsing the position I developed here, I found that I simply could not be comfortable with the thought that the President of the Society of Christian Philosophers says that it is not literally true that God exists. Nor could I imagine being comfortable with a minister denying from the pulpit that it is literally true that God exists (however much he or she qualified the claim afterwards); nor could I imagine teaching the view to my children. Despite the fact that the view seemed philosophically defensible (at least to the limited extent that I have defended it here), and despite the fact that it seemed to have the weight of important parts of the Christian tradition behind it, it did not seem true and,

136

   :  

more importantly, it seemed in no way biblical. Furthermore, it seemed apt to mislead in important ways—indeed, it struck me as theologically corrosive. In the apophatic tradition it is affirmed with good motives: people seem to think that it elevates God in some way to say that God is beyond being. But I think it goes too far. Mightn’t one say the same thing about other apophatic claims? Is it any better to say that ‘God is good’ or ‘God loves us’ is mere metaphor, or even that it is mere analogy? At present, I am inclined to think that these sorts of claims are on very different footing from the claim that ‘God exists’ is metaphorical. The idea that ‘God is good’ and ‘God loves us’ are somehow metaphorical or analogical can be defended plausibly by appeal to scripture. ‘For my thoughts are not your thoughts, nor are your ways my ways, says the Lord’ (Isa. 55:8). But nowhere in scripture do we find God saying anything like ‘my mode of existence is not your mode of existence’. Accordingly, although I think that reflection on passages like the one just quoted (together with certain philosophical-theological considerations) ought to push us towards a reasonably strong conception of divine transcendence, I see no grounds in scripture or elsewhere for moving to a conception that would deny the literal truth of ‘God exists’.

References Alston, William P. 1989. ‘Irreducible Metaphors in Theology’. In Divine Nature and Human Language, 17–38. Ithaca, NY: Cornell University Press. Camp, Elisabeth. 2006a. ‘Contextualism, Metaphor, and What Is Said’. Mind and Language 21 (3): 280–309. Camp, Elisabeth. 2006b. ‘Metaphor and That Certain “Je Ne Sais Quoi” ’. Philosophical Studies 129 (1): 1–25. Cooper, David Edward. 1986. Metaphor. Aristotelian Society Series, vol. 5. New York: Basil Blackwell. Hampson, Daphne. 1996. After Christianity. 1st North American edn. Valley Forge, PA: Trinity Press International. Hellman, Geoffrey. 2007. ‘Against “Absolutely Everything”!’ In Absolute Generality, edited by Augustin Rayo and Gabriel Uzquiano, 75–97. Oxford: Clarendon Press. Hick, John. 2000. ‘Ineffability’. Religious Studies 36: 35–46. Howard-Snyder, Daniel. 2017. ‘Panmetaphoricism’. Religious Studies 53: 25–49. Jacobs, Jonathan D. 2015. ‘The Ineffable, Inconceivable, and Incomprehensible God: Fundamentality and Apophatic Theology’. Oxford Studies in Philosophy of Religion 6: 158–76.

  

137

Johnson, Elizabeth A. 1984. ‘The Incomprehensibility of God and the Image of God as Male and Female’. Theological Studies 45 (3): 441–65. Kaufman, Gordon. 1981. The Theological Imagination: Constructing the Concept of God. Philadelphia, PA: Westminster Press. Lebens, Samuel R. 2014. ‘Why So Negative about Negative Theology? The Search for a Plantinga-Proof Apophaticism’. International Journal for Philosophy of Religion 76: 259–75. Lebens, Samuel R. 2016. ‘Negative Theology as Illuminating and/or Therapeutic Falsehood’. In Negative Theology as Jewish Modernity, edited by Michael Fagenblat, 85–108. Bloomington: Indiana University Press. McFague, Sallie. 1987. Models of God: Theology for an Ecological, Nuclear Age. Philadelphia, PA: Fortress Press. Plantinga, Alvin. 2000. Warranted Christian Belief. New York: Oxford University Press. Pseudo-Dionysius. 1987. Pseudo-Dionysius: The Complete Works. Translated by Colm Luibheid. New York: Paulist Press. Pseudo-Dionysius, the Areopagite. 1990. Corpus Dionysiacum I: De Divinis Nominibus, edited by Beate Regina Suchla, Günter Heil, Adolf Martin Ritter, and Bishop of Scythopolis John. Patristische Texte und Studien; vol. 33. Berlin; New York: de Gruyter. Rayo, Augustin and Gabriel Uzquiano (eds). 2007. Absolute Generality. Oxford: Clarendon Press. Rea, Michael. 2018. The Hiddenness of God. Oxford: Oxford University Press. Reimer, Marga and Elisabeth Camp. 2006. ‘Metaphor’. In The Oxford Handbook of Philosophy of Language, edited by Ernest Lepore and Barry C. Smith, 845–63. Oxford: Oxford University Press. Rowe, William. 1962. ‘The Meaning of “God” in Tillich’s Theology’. Journal of Religion 42: 274–86. Urban, Wilbur. 1940. ‘A Critique of Professor Tillich’s Theory of the Religious Symbol’. Journal of Liberal Religion 2: 34–6. White, Roger W. 1996. The Structure of Metaphor: The Way the Language of Metaphor Works. Oxford: Basil Blackwell. Williamson, Timothy. 2003. ‘Everything’. Philosophical Perspectives 17: 415–65.

7 The Trinity One of the central mysteries of the Christian faith concerns the tri-unity of God. According to traditional Christian doctrine, God is three persons who are somehow consubstantial—one in substance. The persons are the Father, the Son, and the Holy Spirit. Each person possesses all of the traditional divine attributes— omnipotence, omniscience, perfect goodness, eternality, and so on. And yet (in the words of the Athanasian Creed), ‘they are not three eternals, but there is one eternal . . . there are not three almighties, but there is one almighty . . . there are not three Gods, but there is one God.’¹ But what does all of this really mean? And how could it possibly be true? In addressing these two very general questions, there are several more specific issues on which we might try to focus. One is interpretive: How were the central terms in the doctrine—terms like person, substance, and consubstantial—understood when the doctrine was first formulated, and how have they evolved throughout the history of the doctrine? Another is more straightforwardly philosophical: How could three distinct persons (in any reasonable sense of that term) be consubstantial in a way that would make them countable as one God? These questions are not wholly distinct from one another, and there are many others in the neighbourhood that are also worth pursuing. But the one on which I will focus is the second. More exactly, I’ll explore various attempts to show that the central statements in the doctrine, under some intelligible and orthodox interpretation, do not imply a contradiction. Along the way, I’ll touch a bit on the first question, but for the most part it will be set aside.² The question that concerns us here is commonly referred to by philosophers as ‘the logical problem of the Trinity’ and by theologians as ‘the threeness-oneness problem’. The goal of Section 1 is to explain in some detail just what the problem is supposed to be. I’ll begin by stating the central theses of the doctrine of the Trinity. Then I will formulate the logical problem of the Trinity and lay out the constraints that a solution must satisfy in order to preserve an orthodox understanding of the doctrine. In Section 2, I will briefly sketch a few of the most ¹ Quicumque vult (The Athanasian Creed); translation by Jeffrey Brower, quoted from Brower and Rea 2005a: 488. ² The most important and illuminating contemporary literature aimed at sorting out these difficulties has been written by theologians working in the field of Patristic studies. For a start into this literature, the following sources are especially useful: Ayres 2004; Barnes 1998; Coakley 1999, 2003; Stead 1977; Turcescu 2005; and Wolfson 1964.

Essays in Analytic Theology: Volume I. Michael C. Rea, Oxford University Press (2020). © Michael C. Rea. DOI: 10.1093/oso/9780198866800.003.0008

142

   :  

important solutions to the problem. Finally, in Section 3, I’ll present my own view of the Trinity and argue that it has both better historical pedigree and better prospects for solving the problem of the Trinity than the rival views presented in Section 2.

1. The Problem of the Trinity The two creeds to which contemporary Christians typically look for ‘official’ expressions of Trinitarian doctrine are the Nicene-Constantinopolitan Creed of  380/381 and the Athanasian Creed (c.  500). The fourth and fifth centuries witnessed a great deal of philosophical reflection and controversy over the doctrine of the Trinity, and the language of these two creeds is in large part a product of that discussion.³ The former creed is a revised and expanded version of the Creed of Nicaea, which was produced by the First Nicene Council in  325 (nowadays the Nicene-Constantinopolitan Creed is typically just called the ‘Nicene Creed’. I’ll follow this usage.) Like its predecessor, the Nicene Creed was written in Greek and subsequently translated into Latin. It includes the following words: We believe in one God, the Father, almighty, maker of heaven and earth, of all things visible and invisible; And in one Lord Jesus Christ, the only-begotten Son of God, begotten from the Father before all ages, light from light, true God from true God, begotten not made, of one substance (homoousion) with the Father . . . And in the Holy Spirit, the Lord and life-giver, Who proceeds from the Father, Who with the Father and the Son is together worshipped and together glorified, Who spoke through the prophets. (Schwartz 1914: 244–50, quoted in Kelly 1972: 297–8)

I have highlighted the Greek term ‘homoousion’ (o`moόusion) because that term—the term we translate as ‘consubstantial’ or ‘of the same essence’ or ‘of one substance’—was at the centre of some of the most important fourth-century debates about the doctrine of the Trinity. The Athanasian Creed—named after, but not written by, the famous fourthcentury defender of Nicene orthodoxy, Athanasius—was written in Latin, and includes the following statements: Now this is the catholic faith: that we worship one God in Trinity, and the Trinity in unity, nwithout either confusing the persons or dividing the substance. For the

³ For details on the relevant controversies, see especially Ayres 2004.

 

143

Father’s is one, the Son’s another, the Holy Spirit’s another; but the Godhead of the Father, the Son, and the Holy Spirit is one, their glory equal, their majesty coeternal. Such as the Father is, such is the Son, such also the Holy Spirit . . . . Thus the Father is God, the Son is God, and the Holy Spirit is God; and yet there are not three Gods, but there is one God. Thus the Father is Lord, the Son Lord, the Holy Spirit Lord; and yet there are not three lords but there is one Lord . . . . Because just as we are obliged by Christian truth to acknowledge each person separately both God and Lord, so we are forbidden by the catholic religion to speak of three Gods or Lords. (Kelly 1964: 17–20, repr. In Leith 1982: 705)

The Athanasian Creed is widely regarded as manifesting a bias towards ‘Latin’ theories of the Trinity (see Section 2 below). In the two passages just quoted, we have the three central tenets of the doctrine:⁴ (T1)

There is exactly one God, the Father Almighty.⁵

(T2)

Father, Son, and Holy Spirit are not identical.

(T3)

Father, Son, and Holy Spirit are consubstantial.

Each creed includes each of these tenets, if not explicitly, then by implication. The Nicene Creed affirms T1 in its first line (though the ‘exactly’ is only implied). It also explicitly affirms the Father-Son components of T2 and T3. The distinctness and consubstantiality of the Spirit are implicit in the remark that the Holy Spirit is ‘together worshipped and together glorified’ with the Father and the Son. Moreover, they are explicitly affirmed in the Synodical Letter written in  382 by the same council that produced the Nicene Creed.⁶ The Athanasian Creed enjoins us not to confuse the persons or divide the substance, thus committing to T2 and T3. It also affirms T1: we worship one God, and the Father is God. To be sure, it says a lot more than this (for example, it says that the Son is God too); but

⁴ It is more common nowadays to claim that the central tenets of the doctrine are these three: (i) there is exactly one God; (ii) the Father is not the Son, the Son is not the Spirit, and the Father is not the Spirit; and (iii) the Father is God, the Son is God, and the Spirit is God. This way of characterizing the doctrine more closely follows the Athanasian Creed. But the language of (iii) is less clear than the language of T3 above (due the fact that the predicate ‘is God’ can be, and has been, assigned a variety of different meanings). Just as importantly, this formulation obscures the centrality of the notion of consubstantiality in the doctrine—the very notion that lay at the centre of some of the most important controversies surrounding the First Nicene Council. My own formulation of the doctrine is more in accord with formulations found, for example, in the systematic theologies of Berkhof (1938/96) and Hodge (1873). ⁵ I take this straight from the first line of the Nicene Creed. The first line can be translated in ways different from what I have reproduced here; but every credible way of translating it strongly suggests that the Father is somehow the same as God. ⁶ Kelly 1972: 340–1; Schaff and Wace 1900/99, vol. 14: 189.

144

   :  

the more that it says is pretty clearly just a listing of some of the logical consequences of the conjunction of T1 and T3. The logical problem of the Trinity is just the fact that T1–T3 appear to be mutually inconsistent. There are various ways of trying to demonstrate the inconsistency, but the one I favour focuses on the (apparent) meaning of consubstantiality. To say that x and y are consubstantial, or of the same substance is, it seems, just to say that x and y share a common nature—i.e. they are members of one and the same kind. To say that two divine beings are consubstantial, then, would be to say that the two beings in question are identical with respect to their divinity: neither is subordinate to the other; they are not divine in different ways; and if one is a God, then the other one is too. (Note, by the way, that ‘God’ functions in T1 above as a kind-term, like man, and not as a name, like Fred. Thus, though it looks a bit odd, it makes perfect sense to speak of the Father as a God. If ‘God’ were functioning as a name, then T1 would be saying something very much like ‘There is exactly one YHWH’, which isn’t so much a monotheistic claim as a rather strange way of asserting the existence of YHWH.) Given all this, the logical problem of the Trinity can be expressed as follows: (LPT1)

There is exactly one God, the Father Almighty. (From T1)

(LPT2)

The Father is a God. (From LPT1)

(LPT3) The Son is consubstantial with but not identical to the Father. (From T2 and T3) (LPT4) If there are x and y such that x is a God, x is not identical to y, and y is consubstantial with x, then it is not the case that there is exactly one God. (Premise) (LPT5) Therefore: It is not the case that there is exactly one God. (From LPT2, LPT3, LPT4) ***Contradiction The only way out of the contradiction is either to give up one of the tenets of the doctrine of the Trinity or to give up LPT4. At this juncture, some will wonder why Christians don’t just give up on one of the tenets of the doctrine of the Trinity. After all, the doctrine isn’t explicitly taught in the Christian scriptures, and the precise language in which it is expressed wasn’t settled until the fourth century.⁷ So why not just abandon (say) T2 or T3? What reasons are there for accepting traditional Trinitarian doctrine?

⁷ The common story is that it was settled in 325 and then challenged by Arian rebels. Lewis Ayres (2004) and Michel Barnes (1998) tell a different story, however—one according to which the language wasn’t entirely settled in 325, but rather later in 380/1.

 

145

Some have argued that the doctrine of the Trinity (or something like it) can be established via a priori argument.⁸ But the main reason Christians take themselves to be committed to T1–T3 is that they seem to be implied both by Christian practice and by central claims in the Christian scriptures.⁹ For example, both the Old and New Testaments make it clear that there is only one being who deserves worship and who deserves titles like ‘God Almighty’ or ‘the one true God’; and Jesus refers to this being as ‘our heavenly Father’.¹⁰ Hence T1. Moreover, though Jesus says things like ‘I and the Father are one’, it is clear that, from the point of view of the New Testament, Jesus (the Son) and the Father are distinct.¹¹ Jesus prays to the Father; claims to submit to the Father’s will; is blessed by the Father; and so on. Likewise, the Holy Spirit is distinct from the Father and Son: the Spirit is sent by the Son and is said to intercede for us with the Father. Hence T2. And yet the New Testament advocates worshipping Jesus (the Son) and the Holy Spirit;¹² we find Jesus saying such things as, ‘Anyone who has seen me has seen the Father’; and we find the apostle Peter saying (of someone who has lied to the Holy Spirit), ‘You have not lied to men, but to God.’¹³ In short, there is pressure to say that the Son and Spirit are divine—and not in some derivative, or degenerate sense, but truly divine, like the Father. The only clear way to say this without contradicting T1, however, is to say that the Son and the Spirit are consubstantial with the Father: the divinity of the Father, which is the ‘substance’ of the Father (more on this later), is no different from the divinity of the Son. Hence T3.¹⁴ Thus we return to the challenge of unpacking the notion of consubstantiality in a way that enables us to reject LPT4 without incoherence or heterodoxy. The next two sections will be devoted to this task; but first let me say a bit about the boundaries of orthodoxy. Broadly speaking, there are three main ‘errors’ that the Church has condemned with respect to the doctrine of the Trinity: subordinationism, modalism, and polytheism. The task, then, is to explicate the doctrine in a way that avoids these errors. I’ll discuss each in turn. Subordinationism is the view that neither the Son nor the Spirit is truly and fully divine. Either they are not divine at all, or their divinity is somehow subordinate to that of the Father. They are gods of a sort, but lesser gods. Subordinationism is ruled out by language like ‘true God from true God’ or, more explicitly, ‘the Father is God, the Son is God, and the Holy Spirit is God’. ⁸ The best-known argument along these lines is one that is at least latent in the thought of a variety of medieval theologians, explicit in the work of Richard of St Victor, and developed in detail most recently by Richard Swinburne (1994, ch. 8). ⁹ As the fourth-century defenders of Nicaea were at pains to show. See, for starters, Athanasius’s ‘Defence of the Nicene Council’ Schaff and Wace 1900/99, vol. 4. ¹⁰ See, e.g. Exod. 20:3–5; Isa. 42:8; Matt. 5:9–13, 7:21; and John 2:16. ¹¹ Compare John 10:30, Matt. 24:36, Luke 22:42, and John 1:14, 18. ¹² e.g. Phil. 2:10–11. ¹³ John 14:9, Acts 5:3–4 (NIV translation). ¹⁴ Obviously the discussion here is highly compressed. For a fuller discussion of the biblical case for the doctrine, see (for example) O’Collins 1999, chs 1–4.

146

   :  

Modalism is the view that the Father, Son, and Holy Spirit are merely different aspects or manifestations of God—different modes of appearance by which God makes himself known. If modalism were true, then the terms ‘Father’, ‘Son’, and ‘Holy Spirit’ would be analogous to terms like ‘Superman’ and ‘Clark Kent’. The substance called ‘Superman’ is strictly identical to the substance called ‘Clark Kent’. But there is, nevertheless, a distinction to be drawn. The Superman disguise is different from the Clark Kent disguise; and so it makes perfect sense to say that Superman and Clark Kent are different manifestations of Kal El (the Kryptonian who is both Superman and Clark Kent), or different modes in which Kal El appears. If modalism is true, then precisely the same sort of thing can be said about the terms ‘God the Father’, ‘God the Son’, and ‘God the Holy Spirit’. Insofar as they are distinct, Father, Son, and Holy Spirit do not fall into the category of substance; rather, they fall into the category of ‘aspect’ or ‘property’. Polytheism is harder to characterize. According to standard dictionary definitions, ‘polytheism’ is the view that there are many gods, and a ‘god’ is any divine being. Given that Father, Son, and Holy Spirit are divine, these definitions plus T2 imply polytheism. Father, Son, and Holy Spirit are three divine beings; so they are three gods; so polytheism is true. If this is right, and if T1 and T3 are together meant to rule out polytheism, then the doctrine of the Trinity is incoherent. But, of course, it can’t be a simple matter of dictionary-definition that sinks the doctrine of the Trinity. The standard definitions require nuance. But how shall we modify them? I suggest the following: x is a god = df x is a divine substance polytheism = the view that there is more than one divine substance I won’t insist on these definitions here, though. For present purposes, I’ll leave the terms officially undefined and invite readers simply to consult their own intuitions about polytheism in making decisions about whether various models of the Trinity have managed to avoid it.

2. Solving the Problem In this section, I want to lay out some of the main strategies for solving the problem of the Trinity. Since the end of the nineteenth century, it has been common to divide the landscape of views into two camps: Latin (or Western) Trinitarianism (LT) and Greek (or Eastern) Trinitarianism (GT).¹⁵ Those ¹⁵ According to Michel Barnes, what I am calling the ‘standard’ classificatory scheme takes its cues from the work of the nineteenth-century theologian, Theodore de Régnon (1892/8). (Though Ayres notes that de Régnon’s own classificatory scheme wasn’t so much a division between Greek and Latin

 

147

who divide the territory this way tell roughly the following story about their classificatory scheme: The Latin tradition traces its historical roots through the western church. It is epitomized in the work of theologians like St Augustine and St Thomas Aquinas; it takes the unity of the Godhead as given and seeks to explain the plurality in God (rather than vice versa); and those in the tradition tend to gravitate towards psychological analogies. The Greek tradition, on the other hand, traces its roots through the eastern church; it is epitomized in the work of the Cappadocian Fathers; it takes the plurality of the divine persons as given and seeks to explain their unity; and those in the tradition tend to favour social analogies.¹⁶ (GT is commonly identified with ‘social trinitarianism’, discussed below.) In recent years, the standard way of dividing the territory has come under heavy attack, and I myself am inclined to reject it as well.¹⁷ Moreover, as we’ll see shortly, there is a lot more at stake in making a decision about the standard story than the viability of a mere heuristic device for classifying views about the Trinity. For purposes of this chapter, however, it seems prudent to start by presupposing the standard classificatory scheme, present some of the main views that fall under each heading, and only later subject the standard scheme to criticism. So that is how I shall proceed. In the two main parts of this section, I’ll present some of the more well-known LT and GT models of the Trinity. Though each of these models is intended to guide us towards a solution to the logical problem of the Trinity, we’ll see that all of them fall short. Moreover, I’ll argue that the most popular contemporary view, social trinitarianism, depends heavily for its plausibility upon one of the central claims involved in the LT–GT classificatory scheme: namely, the claim that the Greek tradition, starting with the Cappadocian Fathers, favoured ‘social analogies’ as ways of explicating the doctrine of the Trinity. This done, I’ll go on in Section 3 to present my own view of the Trinity—a version of the so-called ‘Relative Identity’ solution (which I’ll also save for discussion in Section 3) that transcends the alleged LT–GT divide. I’ll also present alternative readings of Augustine and the Cappadocian Fathers that identify both of their views as ancestors of mine. Ultimately, my conclusion will be that, of all the models considered herein, the view defended in Section 3 has the best claim to being orthodox and in accord with the views of the earliest defenders of the Creed of Nicaea.

models as between patristic and scholastic models. See Ayres 2004: 302ff.) Barnes subjects de Regnon’s scheme and its relatives to trenchant criticism in Barnes 1995a and 1995b. ¹⁶ Cf., e.g. Brown 1985; Cary 1995; Cross 2002a; LaCugna 1986, 1991; and Placher 1983: 79. ¹⁷ See, e.g. Ayres 2004; Barnes 1995a, 1995b, 1998; Cary 1995; Coakley 1999; and Cross 2002a. The idea isn’t that there is no difference whatsoever between eastern and western, or Latin and Greek, conceptions of the Trinity, but that the differences aren’t nearly as sharp as they are commonly construed, that they aren’t aptly characterized as differences over ‘starting points’ or as a fundamental difference in attitudes towards ‘social’ and ‘psychological’ analogies.

148

   :  

2.1 Latin Trinitarianism In On the Trinity, St Augustine provides several different analogies, or ‘images’ of the Trinity. According to one analogy, Father, Son, and Holy Spirit are like the mind, its understanding of itself, and its love for itself.¹⁸ Another—his preferred analogy—compares Father, Son, and Holy Spirit with the mind’s memory of itself, the mind’s understanding of itself, and the act(s) of will whereby the mind obtains self-understanding from its own memory of itself and delights in and makes use of what it remembers and understands.¹⁹ The usual glosses on these psychological models emphasize two points (not always discussed together, or connected very clearly): (i) The Son and the Spirit in the first analogy, and the Father as well in the second, are being compared to distinct faculties of the mind, or to distinct ways in which the mind operates. (ii) Both analogies indicate that the difference between at least two of the persons is fundamentally a relational difference: the persons are at least roughly analogous to different ways in which a subject might be related to itself, or different ways in which the mind relates to objects of thought in general.²⁰ As noted above, psychological models of the Trinity are common in the work of thinkers falling under the LT classification. So, too, is the idea that differences among persons are akin to differences among reflexive relations. Indeed, according to Richard Cross, the ‘vast consensus in the West’ is that the only distinguishing features among the persons are their relations—that, in the standard terminology, they are subsistent relations. (2002a: 287)

The view that the persons are subsistent relations is a development of the Augustinian models,²¹ and it is most notably associated with the name of St Thomas Aquinas. According to Aquinas, Distinction in God arises only through relations of origin . . . . But a relation in God is not like an accident inherent in a subject, but is the divine essence itself. So it is subsistent just as the divine essence is subsistent. Just as, therefore, the Godhead is God, so the divine paternity is God the Father, who is a divine person. Therefore, ‘divine person’ signifies a relation as subsistent. (Summa Theologiae 1.29.4 c, cited and translated in Cross 2002a: 286)

¹⁸ On the Trinity, bk 9, chs. 3ff.; Schaff and Wace 1887/1999, 127ff. ¹⁹ On the Trinity, bk 10, chs 10–12, and bk 15, ch. 3; Schaff and Wace 1887/1999, vol. 3: 140–3, 200–2. He characterizes the latter analogy as a ‘more subtle’ treatment of the matter than the former. ²⁰ Cf. Brown 1985, ch. 7; LaCugna 1991, ch. 3; O’Collins 1999, ch. 7; Richardson 1955. ²¹ Moreland and Craig say that Aquinas ‘pushes the Augustinian analogy to its apparent limit’ (2003: 585). But this makes it sound as if the ‘subsistent relations view’ isn’t present in Augustine. I myself don’t mean to suggest that, however. All I mean to suggest is that it is not explicit in Augustine.

 

149

What exactly it would mean for a relation to subsist, however, is open to interpretation. The usual objection to the views of Augustine and Aquinas (based on the characterizations just offered) is that they slip into modalism. Granted, Augustine and Aquinas both succeed in drawing distinctions among the persons of the Trinity. But in doing so, they seem to locate the persons in the wrong category. On some interpretations, the Augustinian view suggests that the persons are to be identified with cognitive faculties, or mental modes of operation, both of which are clearly mere aspects of a mind. But even on somewhat more careful interpretations, the Augustinian view at least suggests that the persons are to be identified with relations. And the Thomistic view is explicit on that score. But relations are commonly construed as properties of a certain kind (polyadic properties, to be specific) rather than as substances.²² But if neither Father, Son, nor Holy Spirit is a substance—if they are mere properties—then modalism is true. Aquinas, at least, tries to locate the persons in the right category by calling them ‘subsistent’ relations. But contemporary writers typically respond to this idea with bafflement. What, really, could it mean for a relation to subsist? Not having an answer, they then proceed to object that Aquinas has simply identified the divine persons with relations, which brings us back to modalism.²³ Similar problems beset contemporary psychological analogies, most of which can also be classified under the LT heading. For example, Thomas V. Morris suggests (without affirming) that the Trinity might be modelled on the different personalities of a patient suffering from multiple personality disorder.²⁴ Trenton Merricks, on the other hand, suggests that the persons might be thought of on analogy with the distinct ‘centres of consciousness’ that seem to be associated with the two hemispheres of a human brain. (In experimental situations, commissurotomy patients—people who have undergone a surgical procedure that severs the bundle of nerves that allows the two hemispheres of their brain to communicate with one another—show behaviour that seems to indicate that their consciousness is divided, as if there is a separate stream of thought associated with each hemisphere.) There is some pressure to classify these models under the GT heading instead; for, as we’ll shortly see, the standard lore about GT is that it is wedded to ‘social models’, which, allegedly in contrast to Latin views, regard the ²² A polyadic property is one that applies to several things together. A property like being loved is monadic: it applies to a single subject (even if there are many subjects to which it applies). Relations like ___ loves ___, or ___ thinks that ___ is ___, are polyadic. The former applies to a pair, the latter to a triad. Note, however, that, though it is nowadays common to treat relations as polyadic properties (as, e.g. in van Inwagen 2004), Jeff Brower argues that medieval philosophers did not treat them as such. See Brower 2001, 2005. ²³ I think that the objection can be met, and that the way to do it is to develop Aquinas’s view along the lines of the view described in Section 3. But I won’t attempt to do that here. ²⁴ Morris 1991, ch. 9.

150

   :  

persons of the Trinity as distinct psychological subjects. My own inclination, though, is to think that neither of these models compares the persons of the Trinity with distinct psychological subjects. Rather, they compare the Trinity as a whole with a fragmented psychological subject. The personalities of someone with multiple personalities are not substances; they are aspects of a substance. Plausibly, the same is true of the distinct ‘centres of consciousness’ that are elicited as a result of commissurotomy.²⁵ Thus, again, modalism looms. The final version of LT that I want to consider is Brian Leftow’s (2004).²⁶ Leftow regards his view as an instance of LT because it ‘begin[s] from the oneness of God’ (2004: 304). On his view, the persons of the Trinity might be thought of as analogous to a time traveller who appears thrice located at a single time. He offers us the example of Jane, a Rockette who is scheduled to dance in a chorus line but, at the last minute, discovers that two of her partners have failed to show up. Jane goes on stage and dances her part, then later enters a time machine (twice) so that she can (twice) go on stage with herself and dance the leftmost and rightmost parts as well. According to Leftow, there is a very clear sense in which this part of the chorus line contains three of something; and yet there is just one substance (Jane) in that part. Is the view adequate? It is hard to tell, because Leftow’s presentation is imprecise at a crucial juncture. Consider Jane’s part of the chorus line, and suppose we use the labels, ‘L’, ‘M’, and ‘R’ to refer to the occupants of the three positions in that part of the line (Figure 7.1).

L

M

R

Figure 7.1

²⁵ Admittedly, though, one might think of them as distinct (subsistent) minds or souls, embodied in or emergent upon the same material substance. If we think of them this way, then, as Jeff Brower and I note elsewhere (Brower and Rea 2005b) polytheism looms, rather than modalism. Whether the view actually falls into polytheism, however, depends entirely upon whether the relevant ‘centres of consciousness’ are in any sense the same substance. On this question, Merricks 2006 is silent. (This is not to say that Merricks simply ignores the ‘same substance’ issue. Indeed, he spends pages 319–22 discussing it; he says that he can ‘do justice’ to the claim that the Father is the same substance as the Son (320); and he says explicitly that ‘each divine person is a substance in a less-than-most-straightforward sense’ (322). But, for all this, he does not, so far as I can tell, say one way or another whether there is any sense of ‘same substance’ in which it is straightforwardly true (in that sense) that the Father is the same substance as the Son. It is in this respect that he is ‘silent’ on the question at issue here.) ²⁶ See also Leftow 2007.

 

151

Now, let us ask what the relations are between L, M, and R. Upon reflection, we face a problem: It is not at all clear what we mean by the occupants of the three positions. One might think that there is just one occupant, Jane, multiply located in three positions. In that case, ‘L’, ‘M’, and ‘R’ are all just names for Jane; and the relation between L, M, and R is identity: C 1:

L = M = R = Jane

Alternatively, we might think that what the names ‘L’, ‘M’, and ‘R’ refer to are three distinct events in the life of Jane. In that case, there are, after all, three Rockettes, and each Rockette is an event: C 2: L = the leftmost dancing event; M = the middle dancing event; R = the rightmost dancing event; and L ≠ M ≠ R Which case does Leftow have in mind? Unfortunately, it looks as if he has both in mind. He writes: If (as I believe) Jane has no temporal parts, then not just a temporal part of Jane, but Jane as a whole, appears at each point in the chorus line, and what the line contains many of are segments or episodes of Jane’s life-events. This may sound odd. After all, Rockettes dance. Events do not. But what you see are many dancings of one substance. What makes the line a line is the fact that these many events go on in it, in a particular set of relations. Each Rockette is Jane. But in these many events, Jane is there many times over. (2004: 308; emphasis mine)

The quoted passage says that ‘what there are many of ’ is events; and it speaks of Rockettes in the plural. Thus, C 2 looks like the correct interpretation. On the other hand, each Rockette is Jane. Thus, C 1 looks like the correct interpretation. But it is impossible for both interpretations to be correct. So it is hard to know what to make of what Leftow is saying here. Elsewhere, however, he says a bit more (Leftow 2007). ‘Perhaps’, he says, ‘the triune Persons are event-based persons founded on a generating substance, God’ (2007: 373f.). An event-based person is, roughly, a person whose existence is constituted by the occurrence of an event: what it is for the person to exist is for that event to occur in a particular substance (2007: 367f.). In the case of Jane, then, L, M, and R are presumably supposed to be analogous to event-based persons. Jane (the generating substance) exists in each of the three, as Leftow says at the end of the quoted passage above; but she is not strictly identical to any of the three. Likewise, God exists in each of the event-based persons that together constitute the Trinity.

152

   :  

If we take Leftow 2007 as definitive, then we have clarity on the relation between L, M, R, and Jane: all are distinct. But it is still hard to see how the view sheds light on the Trinity. Are L, M, and R consubstantial? It is hard to tell. They are events (or event-based things) involving a common substance; but that doesn’t guarantee consubstantiality. Suppose I paint a wall red. We then have two events: the wall’s becoming wet and the wall’s becoming red. But these two events are not consubstantial, for they are not substances that share a common nature. Indeed, they’re not substances at all; and even if they were, what they share isn’t a nature. So too, apparently, in the case of L, M, and R. Of course, Leftow might have more to say on the subject; but until more is said, it is, again, hard to know what to make of the view.²⁷ The reader will have noticed by now that, though I have offered several analogies under the heading of LT, I have not yet said how, exactly, any of the models solves the problem of the Trinity. The reason is simple: they don’t—at least not as they have been interpreted here. The reason they don’t is that, though they offer ‘senses’ in which God is both three and one, they do not explain how it is that numerically distinct consubstantial beings count as one God. Leftow’s model aside, the distinct items in so-called Latin models as they are interpreted here are pretty clearly not consubstantial. Indeed, they are not substantial at all. And, as we have seen, Leftow’s model is problematic for other reasons. This defect is not shared by GT models, which provide straightforward reasons for rejecting LPT4, grounded in a clear account of the consubstantiality of the persons (they are consubstantial by virtue of being members of a common kind) and varying analogies at explaining why we cannot conclude (as LPT4 effectively does) that distinct consubstantial divine beings would count as more than one God. But these models have other problems, as we shall now see.

2.2 Greek Trinitarianism In the contemporary literature on the doctrine of the Trinity, there is a family of views that fall under the label ‘social trinitarianism’ (ST).²⁸ Taking the LT–GT classification scheme for granted, ST is normally regarded as co-extensive with

²⁷ Thanks to Brian Leftow for helpful correspondence about his view. Naturally, if I have him wrong here, that is my fault and not his. My own view is that Leftow’s view can be made workable by developing it in the direction of the constitution view, described in Section 3. But I won’t attempt that here. ²⁸ Proponents of the ST strategy include Timothy Bartel (1993, 1994); David Brown (1985, 1989); Peter Forrest (1998); C. Stephen Layman (1988); J. P. Moreland and William Lane Craig (2003); Cornelius Plantinga, Jr (1986, 1988, 1989); Richard Swinburne (1994); Edward Wierenga (2004); and C. J. F. Williams (1994). Hasker (1970) claims to endorse ST; but some of what he says seems more in line with the constitution view (see esp. pp. 27–30). According to Richard Swinburne, the ‘most influential modern statement of social Trinitarianism’ is Moltmann 1981.

 

153

GT. What set the Cappadocian Fathers apart from the LT tradition, according to the common lore, was precisely their endorsement of ST; and contemporary versions of ST are basically just contemporary developments in the GT tradition. Owing in large part to misunderstandings that lead people to think of LT models as modalistic, ST models of the Trinity tend to be more popular in the contemporary literature. Critics, however, charge ST with polytheism (for reasons I shall explain shortly). This brings to light the fundamental import of the LT–GT classification scheme, and of recent criticisms of it. The Cappadocian Fathers are universally acknowledged to have played a vital role in the earliest defences and interpretations of the creed of Nicaea. They are among those who helped to define orthodoxy; so it would be surprising, to say the least, if whatever view they held turned out to be heretical. In short, saying ‘The Cappadocian Fathers endorsed model M’ is, if true, nearly decisive as a response to charges that model M is polytheistic or otherwise heretical. Hence the importance of the contemporary claim that there is such a thing as GT which (i) is found in the work of the Cappadocian Fathers and other key eastern theologians, and (ii) differs from LT by taking plurality rather than unity as its starting point, as evidenced by their endorsement of ST models. Historical challenges to the viability of the LT–GT classification scheme take issue with each of these points. They deny that the Cappadocians endorsed ST; they deny that there is reason to think that the Cappadocians took different ‘starting points’ from key Latin theologians like St Augustine; and they deny that there is any such thing as GT or LT as they are commonly conceived.²⁹ If the challenges are right, then contemporary criticisms of ST are suddenly much more forceful; for not only does ST lose the backing of the Cappadocians that is so vital to warding off those criticisms, but some other view—whatever view the Cappadocians actually endorsed—acquires their backing. In the next section, I will present a view that I take to be a contemporary development of what was held in common by (at least) Augustine and the Cappadocian Fathers. But for now, I want to focus on explaining what ST is and why it is inadequate as a way of understanding the Trinity. Contemporary social trinitarians have not been especially clear about what the central tenets of their view are supposed to be. Neither have their critics. First and foremost, ST theories are identified by their reliance on analogies that compare the persons of the Trinity to things that are numerically distinct but share a common nature—usually rational creatures of some sort, like human beings. These are called ‘social analogies’ because many of them (though not all) at least imply, if not explicitly state, that the unity among the divine persons is some sort of social unity: God is like a family, or a perfectly unified community of rulers, or whatever.

²⁹ See n. 17 for references.

154

   :  

This suggests that ST might fruitfully be characterized as committed to the following central tenets: (1) Father, Son, and Holy Spirit are not numerically the same substance. Rather, the persons of the Trinity are consubstantial only in the sense that they share a common nature; and the sharing is to be understood straightforwardly on analogy with the way in which three human beings share a common nature. (2) Monotheism does not imply that there is exactly one divine substance. Rather, it implies at most only that all divine substances—all gods, in the ordinary sense of the term ‘god’—stand in some particular relation R to one another, a relation other than being the same divine substance. (3) The persons of the Trinity stand to one another in the relation R that is required for monotheism to be true. Different versions of ST might then be distinguished in accord with differences over what relation R amounts to. There are many candidates in the literature for being monotheism-securing relations, but the most popular are the following: (a) (b) (c) (d) (e)

Being parts of a whole that is itself divine. Being the only members of the only divine kind. Being the only members of the community that rules the cosmos. Being the only members of a divine family. Being necessarily mutually interdependent, so that none can exist without the others. (f) Enjoying perfect love and harmony of will with one another, unlike the members of pagan pantheons.

Most social trinitarians in fact opt for a combination of these, and most (but not all) of the combinations include at least (a), (b), and (c). So, for example, Richard Swinburne (1994) focuses on the fact that YHWH is a composite individual or society whose parts or members stand in the relations identified in (e) and (f).³⁰ But, of course, he wouldn’t deny that they stand in (b) and (c) as well. J. P. Moreland andWilliam Lane Craig (2003) focus primarily on (a). On their view, YHWH is composed of the Persons in a sense analogous to the way in which the three-headed dog Cerberus, guardian of the underworld in Greek mythology, might be thought to be composed of three ‘centres of consciousness’ (2003: 593). On their view, the three conscious parts of Cerberus are not dogs; there is only one

³⁰ Cf. Swinburne 1994: 181.

 

155

dog—Cerberus. But the centres of consciousness are canine, just as any other part of Cerberus is (derivatively) canine. One dog, then; three derivatively canine individuals. Likewise in the Trinity: one God; three derivatively divine individuals. Monotheism is thus secured by the fact that the Persons are parts of a single fully divine being. Cornelius Plantinga Jr, on the other hand, argues that social trinitarians may ‘cling to respectability as monotheists’ simply by affirming that the persons are related in the ways described by (b), (c), and (d) (1989: 31). His idea seems to be that monotheism is true, no matter how many gods there are, so long as all gods derive their divinity from one source, or share a single divine nature (as we humans share a single human nature), or are joined together as a divine family, monarchy, or community. There are other suggestions in the literature; but they tend to run along very similar lines. A standard criticism of social trinitarianism is that these (re-)characterizations of monotheism are implausible. Those lodging the criticism typically do so in one of two ways. Some try to argue that statements in the Nicene and Athanasian Creeds that are commonly taken to rule out polytheism also speak against social trinitarianism.³¹ But from the point of view of those who accept the standard LT– GT distinction, there are two strikes against this objection. First, the Athanasian Creed was written in Latin and incorporates language from Augustine. So it can be dismissed as an LT treatise—one which we would antecedently expect to reflect an anti-ST bias. Second, the chief defenders of the creed of Nicaea were themselves social trinitarians (according to the common lore); so the claim that ST violates that creed is simply not credible. So, at this stage anyway, the creedal objection is indecisive. Others object to ST on the grounds that it is not plausible to think that (say) Greek polytheism would become monotheistic if only we added that Zeus and the other gods enjoyed perfect love, harmony, and mutual interdependence with one another; nor would it seem to help if we were to pare the pantheon down to a single divine family that rules the cosmos.³² I think that this objection is intuitively decisive against the suggestion that the relations described in (b), (c), (d), (e), or (f), singly or in combination, could possibly secure monotheism. But it has no implications for the more popular suggestion that Christianity is monotheistic because (a) the persons are parts of the one and only fully divine being. Moreover, it has the awkward implication that, if the Cappadocians really were social trinitarians, then they were polytheists after all. To close the case against ST, then, two tasks must be accomplished: (i) part–whole models must be shown to be problematic; (ii) one must either explain how the Cappadocians managed

³¹ See, especially, Brower 2004a, Clark 1996, Feser 1997, and Leftow 1999. ³² See, especially, Leftow 1999 and Merricks 2006.

156

   :  

inadvertently to fall into one of the very errors they were most concerned to avoid, or one must show that the Cappadocians really were not social trinitarians.³³ Various objections against part–whole trinitarianism are already present in the literature.³⁴ Rather than summarize those, however, I want here to present a new one.³⁵ I will take as my target the version of part–whole trinitarianism developed by J. P. Moreland and William Lane Craig (2003). I do this because theirs is the most developed version of part–whole trinitarianism currently available, but I think that substantially the same objection could be raised against any version. As I have already noted, Moreland and Craig compare God to the mythical dog Cerberus and the persons to Cerberus’s three heads (or, better, the souls that might be embodied in those heads). So, on their view, the divine persons—all of them—are parts of God. For obvious reasons, however, Moreland and Craig want to preserve the view that God is divine while denying that God is a fourth divine thing on a par with the persons. Thus, they distinguish two kinds of divinity: the full divine nature, which is possessed by God and implies tri-unity; and a derivative divine nature, possessed by each person. The distinction is analogous to the (allegedly) two ways in which something might be said to be feline: something can be feline by being a cat; or something can be feline in a derivative way by being part of a cat. Moreover, the making of this sort of distinction is implicit in every version of part–whole trinitarianism of which I am aware. Nobody wants to be pressed into affirming a quaternity in God; thus the composite is always treated, at least implicitly, as a different kind of thing from the parts—a non-person composed of persons, or a group-mind composed of single minds, and so on. So much for the view itself. But now we face two problems, both apparently devastating: First, Moreland and Craig cannot affirm the opening line of the Nicene creed: ‘We believe in one God, the Father, almighty.’ For, on their view, God is a fundamentally different thing from the Father.³⁶ Moreover, they cannot affirm the crucial homoousion clause in the same creed unless they reject the idea that there is exactly one divine nature—an idea which the Cappadocians and (so far as I am aware) every other major interpreter and defender of the creed of Nicaea in the fourth and fifth

³³ Or one might just try to sink it with the weight of other objections. And there are other objections in the literature—see, for example, Tuggy 2003 and 2004. My own view, though, is that even after these further objections have been piled on, the cumulative case is not dialectically strong enough apart from the two further tasks just mentioned. ³⁴ See, e.g. Howard-Snyder 2003, Leftow 1999, Rea 2006, and Tuggy 2003. ³⁵ My objection bears some similarity to objections based on the Athanasian Creed raised by Jeffrey Brower (2004a) against Wierenga’s (2004) social trinitarian view. ³⁶ Dale Tuggy (2003, and especially 2004) presses roughly this point against all versions of ST. Partwhole trintiarians could respond by saying that the Son and Spirit are parts of the Father (who is also a person). But in that event, their doctrine of two kinds of divinity (full and derivative) would commit them to denying that the Son and the Spirit (who would still be derivatively divine) have the same nature as the Father (who alone would have the full divine nature). Thus, they would run into direct conflict with the requirement that the Son be of the same substance with the Father.

 

157

centuries were in agreement about. Here is why: The only viable interpretations of the creedal claim that the Son is homoousion with the Father have it that the Son is either numerically the same substance as the Father or of the same nature as the Father. (Natures were also referred to as ‘substances’; hence, being consubstantial with something might just mean having the same nature.) The former, of course, they reject. The latter they accept; but in accepting it, they posit, effectively, two divine natures—one ‘genuine’, possessed only by God; the other derivative, but still divine, possessed by the two persons. Of course, they could deny that the derivative nature is a divine nature. But in so doing, they seem to strip the persons of their divinity, which would conflict with other parts of the Nicene and Constantinopolitan creeds. If all of this is right, then, part–whole trinitarianism is in serious trouble, at least if its proponents intend (as Moreland and Craig do) to be offering a view that respects Nicene orthodoxy. But now what of the Cappadocian Fathers? The initial justification for treating the Cappadocian Fathers as social trinitarians comes primarily from their own apparent employment of social analogies.³⁷ Under the usual interpretation, the Cappadocian view was that, ‘just as Peter, Paul, and Barnabas are each man, so the Father is God, the Son is God, and the Holy Spirit is God’ (Plantinga 1986: 329–30). In short: the persons of the Trinity are ‘of the same substance’ only in the sense that they share a common nature; and the sharing is to be understood straightforwardly on analogy with the way in which three men share a common nature. As I have already indicated, however, this way of interpreting the Cappadocian Fathers has, of late, come under heavy criticism. Space will not permit me to present the exegetical case against the standard interpretation; but in the next section, after laying out the view that I call Constitution Trinitarianism, I will present reasons for thinking that the Cappadocians were actually gesturing towards that view rather than towards ST. If this is right, then one of the most important reasons for accepting ST is undermined.

3. The Constitution Model I turn now to the model that I prefer.³⁸ Let me begin here by identifying a general strategy for solving the problem of the Trinity that I have so far omitted from our discussion: the Relative Identity strategy.³⁹ The problem, as I have characterized it,

³⁷ Plantinga 1986, 1988, and 1989 provide representative examples. ³⁸ See Brower and Rea 2005a, 2005b. See also Brower 2004b. ³⁹ Proponents of this strategy include G. E. M. Anscombe and P. T. Geach (1963, esp. pp. 118–20); James Cain (1989); A. P. Martinich (1978, 1979), and Peter van Inwagen (1988, 2003). I count myself as both a critic and a proponent, since I reject one way of deploying the strategy but endorse another. (See

158

   :  

is generated by the conjunction of the following premise (from Section 1 above) with the central tenets of trinitarian doctrine: (LPT4) If there are x and y such that x is a God, x is not identical to y, and y is consubstantial with x, then it is not the case that there is exactly one God. (Premise) But (LPT4) is true only if the following principle is also true: (P1) Necessarily, if x and y are not identical, then x and y are not numerically the same substance. If P1 is false, then LPT4 simply ignores the possibility that x and y are distinct but (perhaps by virtue of their consubstantiality) one and the same God. In other words, LPT4 presupposes that it is impossible for an object a and an object b to be numerically the same F without being absolutely identical. Give up that presupposition, and the argument that depends on LPT4 fails. But how can we reject P1? The way to do it is to endorse the doctrine of relative identity (RI). There is a weak version of this doctrine, and a strong version. The weak version says: (RI1) States of affairs of the following sort are possible: x is an F, y is an F, x is a G, y is a G, x is the same F as y, but x is not the same G as y. The strong version is just the weak version plus (RI2): (RI2) Either absolute (classical) identity does not exist, or statements of the form ‘x = y’ are to be analysed in terms of statements of the form ‘x is the same F as y’ rather than the other way around. RI2 is not needed for solving the problem of the Trinity; but some philosophers— notably, Peter Geach—endorse it for other reasons, and it serves as independent motivation for RI1.⁴⁰ Defenders of the relative identity solution have mostly occupied themselves with working out the logic of relative identity in an effort to show that the doctrine of relative identity itself is coherent, and to show that the doctrine of the Trinity can be stated in a way that is provably consistent given the assumption of relative

Rea 2003 and Brower and Rea 2005a, 2005b). For other criticism, see Bartel 1988; Cartwright 1987; Merricks 2006; and Tuggy 2003. ⁴⁰ See, e.g. Geach 1967; 1980, sections 30, 34, and 110.

 

159

identity.⁴¹ (What I have done here falls short of the latter goal. I have shown how accepting RI1 enables one to rebut a single argument for the inconsistency of the doctrine of the Trinity; but I have not shown that every argument against the consistency of the doctrine as I have stated it must fail.) Despite the efforts of its defenders, however, the relative identity solution has remained rather unpopular. RI2 in particular is widely rejected as implausible; and I have argued elsewhere that invoking it in a solution to the problem of the Trinity implies that the difference between the persons is theory-dependent, and so merely conceptual (Rea 2003). But without RI2, RI1 is (at first glance, anyway) unintelligible. The reason is simple: sameness statements are naturally interpreted as identity statements. So, the claim that ‘x and y are the same F’ seems logically equivalent to the claim that ‘x is an F, y is an F, and x = y’. RI1 is inconsistent with this analysis of sameness statements. But on its own, it doesn’t supply any replacement for that analysis. Thus, it renders sameness claims utterly mysterious. Appealing to RI1 without a supplemental story as a way of solving the problem of the Trinity, then, simply replaces one mystery with another. That is hardly progress. The constitution view supplies the relevant supplemental story. The story begins with an example: An artistic building contractor fashions a marble statue that is to be used as a pillar in the building he is constructing. So he has made a statue; he also has made a pillar. It would be strange to say that he has made two material objects that are simply located in exactly the same spot at the same time (though many philosophers do in fact say such a thing). What we are inclined to say is that the statue and the pillar are one and the same material object, not two. And yet they are distinct. Surface erosion will destroy the statue without destroying the pillar. Internal corruption that preserves the surface but undermines the statue’s capacity to support the weight of a building will destroy the pillar but (if the statue is removed from its position as a load-bearing structure) will not destroy the statue. Thus, what we want to say is that the statue and the pillar are the same material object, even though they are not identical. If we do say this, we commit ourselves to RI1. But we can make RI1 intelligible by adding that all it means to say that two things are the same material object is that those two things share all of the same physical matter. Let us flesh out the story just a bit further. Aristotle maintained that every material object is a compound of matter and form. The form might be thought of as a complex organizational property—not a mere shape, as the term suggests in English, but something much richer. For Aristotle, the form of a thing was its nature; and forms, like concrete things (though not in exactly the same sense), count as substances. Thus, on his view, St Peter would be a compound whose constituents were some matter and the form, humanity, or human nature. St Paul would be a compound whose constituents were that same form, but different

⁴¹ See, especially, van Inwagen 1988, 2003.

160

   :  

matter.⁴² Peter and Paul would thus be ‘of one substance’ on the Aristotelian view; though (unlike Gregory of Nyssa) Aristotle would not have spoken of Peter and Paul as being numerically the same man, nor would he have regarded them as numerically the same substance. With some minor modifications (plus the non-Aristotelian assumption that statues and pillars are substances, just like men are) Aristotle’s view would permit us to say that the statue and the pillar are numerically the same substance, even though they would not be ‘of the same substance’ since they would not share the same form, or nature. They would be numerically the same substance, one material object, but distinct matter-form compounds. They would be the same without being identical. In the case of our statue, then, we have two complex properties—being a statue and being a pillar—both had by the same underlying subject, some undifferentiated matter.⁴³ This gives us two compounds, a statue and a pillar. Each is a substance. Thus, the statue and the pillar are emphatically not mere aspects of a common substance. They are not properties or relations or anything of the sort. Furthermore, each is distinct from the other. But they are, nevertheless, the same substance. Note, too, that the underlying matter is not a substance, since it is not a matter-form compound. But even if it were, it would not be a fourth substance—it would be the same substance as the statue and the pillar (by virtue of sharing all of its matter in common with them). Nor (obviously) would it be a third compound. Hence, there is just one substance and two compounds. You might think that if there is just one substance, then we ought to be able to ask whether it—the one substance—is essentially a mere statue or essentially a mere pillar (or perhaps essentially a statue-pillar). But this thought is incorrect. On the present view, terms like ‘it’ and ‘the one substance’ are ambiguous: they might refer to the statue, or they might refer to the pillar. For, again, statue and pillar are distinct, though not distinct substances. You might think that if there are three matter-form compounds, then there are three ‘primary substances’ but only one ‘secondary substance’ (where a primary substance is a concrete particular, like a human being, and a secondary substance is a nature, or form). But this is also incorrect. On the view I am defending, if x and y share the same matter in common, and x is a primary substance and y is a primary substance, then x and y are the same primary substance, despite being different matter-form compounds.⁴⁴ By now the relevance of all of this to the Trinity should be clear: almost everything that I just said about the statue and pillar could likewise be said about the divine persons. In the case of the divine persons, we have three ⁴² There’s controversy in the literature on Aristotle over whether distinct human beings really share numerically the same form or not. Here I assume that they do; but since I feel no special burden to follow Aristotle on that point, nothing really depends on the assumption. ⁴³ Or, if you like, a lump of matter. See Brower and Rea 2005a: 504, n. 10 for discussion. ⁴⁴ Cf. Cary 1992: 381.

 

161

properties—being the Father, being the Son, and being the Spirit; or, perhaps, being Unbegotten, being Begotten, and Proceeding—all had by something that plays the role of matter. (It can’t really be matter, since God is immaterial. Suppose, then, that it is the divine substance, whatever that is, that plays the role of matter.)⁴⁵ Each divine person is a substance; thus, they are not mere aspects of a common substance. Furthermore, each is distinct from the other. (So modalism is avoided and T2 is preserved.) But they are nevertheless the same substance. (Hence T3.) Thus, there is just one divine substance, and so the view allows us to say, along with the creed, ‘We believe in one God, the Father almighty . . . and in one Lord Jesus Christ . . . begotten, not made, being of one substance with the Father’. Since Father, Son, and Holy Spirit count, on this view, as numerically the same substance despite their distinctness, LPT4 is false, and the problem of the Trinity is solved. So far so good; but there is one loose end that remains to be tied. What is it that plays the role of matter in the Trinity? And is it a substance itself? Here I want to offer only a partial view that might be developed in a variety of different ways. What plays the role of matter in the Trinity is the divine nature; and the divine nature is a substance. It is not a fourth substance, for reasons already discussed; nor is it a fourth person (since it is not a compound of ‘matter’ plus a persondefining-property). But it is a substance, since (again, taking cues from Aristotle) natures are substances. What I don’t want to take a position on here is the question of what, exactly, a nature is. Is it concrete or abstract? Is it particular or universal? Is it a property or something else? These questions I will not answer. I think that they must be answered in a way that allows the divine persons to be concrete particular non-properties; but I think that there are various ways of answering these questions that are compatible with that view. This completes my presentation of the constitution view. We are now in a position to see how the view connects with the views of fourth- and fifth-century defenders of Nicaea. At the heart of the constitution view is the idea that the divine persons are compounds whose constituents are a shared divine nature, which plays the role of Aristotelian matter, and a person-defining property (like being the Son, or being Begotten) that plays the role of form. But this is almost exactly the view that Richard Cross identifies as the fundamental point of agreement between eastern and western views of the Trinity. According to Cross, East and West agreed that (a) the divine nature is a property, and (b) that one and the same divine nature is a constituent of each of the divine persons—i.e. it is the point at which they overlap.⁴⁶

⁴⁵ One might want to say that it is the Father who plays the role of underlier, that the Son is a compound of the Father and the property being Begotten, and that the Spirit is a compound of the Father (or the Son) and the property of Proceeding. This view is suggested by passages in Augustine, and was first brought to my attention by Anne Peterson. I take it as a variation on the present theme, rather than a genuine rival to the view I am defending. ⁴⁶ Cross 2002a: 284. See also Cary 1992.

162

   :  

So far so good; but mere overlap is not sufficient to suggest the constitution model. For if nature-sharing is a kind of overlap (as it seems to be for the Cappadocians, at least),⁴⁷ then two men overlap on a constituent as well. But, of course, they don’t constitute one another. In order to attribute something like a constitution model to Augustine and the Cappadocians, we would have to show that their view posited not only overlap among the persons but a kind of overlap that guaranteed the sort of absence of separation that we find in objects (like the statue and the pillar) that constitute one another. Encouragingly, Michel Barnes insists that this idea was as crucial to both East and West as Cross thinks the idea of overlap was. Thus, he writes (speaking of both Augustine and the Cappadocians): the most fundamental conception and articulation in ‘Nicene’ Trinitarian theology of the 380s of the unity among the Three is the understanding that any action of any member of the Trinity is an action of the three inseparably. (Barnes in Davis et al. 1999: 156; emphasis in original)⁴⁸

Moreover, as we’ll see, both Augustine and the Cappadocians endorsed models that reinforce this reading and the Cappadocians, at least, rejected models that conflict with it. Let us return again (tfor starters) to Augustine’s favoured analogy: selfmemory, self-understanding, and self-directed willing. At the end of Book 10 of On the Trinity, after introducing the favoured analogy, Augustine raises some concerns and then says that the ‘discussion demands a new beginning’. In Book 11, he opens with two new analogies, the first of which compares the divine persons with an external body, vision of the body, and the attention of mind that joins these two. If Augustine’s goal really were to model the Trinity on different faculties or modes of operation for the human mind, returning to an analogy like this would be bizarre. For, after all, an external body has nothing essentially to do with the mind of one who beholds it. But, of course, this isn’t really Augustine’s goal at all. The key to understanding this is to attend carefully to the theory of vision that gets presented along with the analogy. For Augustine, vision involves a kind of reproduction in the viewer of the nature or form of the thing viewed.⁴⁹ So when an external body is seen, the form of that body is present both in the body and in the visual faculty of the viewer. Moreover, he seems to think that, in directing vision towards the external body and retaining the image of it in the visual faculty, the will takes on the same nature as well—just as ‘the little body of a ⁴⁷ On Gregory of Nyssa’s view of universals, see Cross 2002b. ⁴⁸ I want to avoid views that rule out shared agency; but I don’t myself want to commit to it. For, of course, it raises serious questions. The Son prays to the Father. Does the Father do so as well? Proponents of shared agency will say ‘yes, but not in the same way’ (see, e.g. Hasker 1970: 29). Trying to make good sense of this response, however, would require another paper. ⁴⁹ On the Trinity, bk 11, ch. 2; in Schaff and Wace 1887/1999, vol. 3: 145–6.

 

163

chameleon [varies] with ready change, according to the colors which it sees’.⁵⁰ Once we see this, however, the point of the new analogy becomes clear. The case at hand is one in which three very different items are unified by virtue of a kind of overlap: they have (in a way) the very same nature present in them. Of course, the analogy is imperfect; for, among other things, the external body’s nature is present in it in a way different from the way in which it is present in the viewer’s visual faculty or will. Augustine therefore offers another analogy to correct this defect; but, in the end, the best analogy is the one that I have called ‘the favoured analogy’. It is the best because, in that analogy, one and the same substance is present in three distinct things which are distinguished from one another only by their relations to that one substance. The things in question are not faculties or modes of operation. They are, rather, concrete events, each of which is, in effect, a complex whose constituents are a substance and a reflexive relation: the mind remembering itself, the mind understanding itself, and the mind willing itself. These events are not identical; and, as events, they are at least closer to the category of substance than to the category of property or aspect. Moreover, since it is the mind itself that is the common constituent in each complex, and since it is the mind that is doing the self-remembering, self-understanding, the view strongly suggests the sort of absence of separation that one finds in the constitution view; and it even suggests shared agency among the persons. What of the Cappadocians? The Cappadocians clearly thought of consubstantiality as the sharing of a nature—on this I think there is no substantive disagreement. Moreover, it seems clear that they thought of nature-sharing as the sharing of a common constituent. Gregory of Nyssa claims (in On Not Three Gods: To Ablabius) that ‘the man in [a group of men] is one’ (Schaff and Wace 1900/99, vol. 5: 332). Likewise, in his Fifth Theological Oration, Gregory Nazianzen says that ‘the Godhead [or divine nature] is One’, and he speaks of the Persons as those ‘in Whom the Godhead dwells’, which suggests that the Godhead—i.e. the divine nature—is a common constituent of the three (Schaff and Wace 1900/99, vol. 7: 322). Moreover, in the following passage he seems to indicate outright that the relation between the persons involves overlap on a common nature and differentiation by three properties: the very fact of being Unbegotten or Begotten, or Proceeding has given the name of Father to the First, of the Son to the Second, and [to] the Third . . . of the Holy Ghost, that the distinction of the Three Persons may be preserved in the one nature and dignity of the Godhead . . . . The Three are One in Godhead, and the One three in properties. (Fifth Theological Oration: On the Holy Spirit ix, Schaff and Wace 1900/99, vol. 7: 320)

⁵⁰ On the Trinity, bk 11, ch. 5; in Schaff and Wace 1887/1999, vol. 3: 146–7.

164

   :  

But what of inseparability? I said earlier that not only do the Cappadocians affirm models that posit the inseparability of the persons, but they reject models that conflict with it. Thus, though Gregory Nazianzen acknowledges that Adam, Eve, and Seth are consubstantial, he goes on to make it very explicit that his point in saying that they are is only to show that distinct things can be consubstantial, not to embrace what would now be called a social analogy: were not [Adam, Eve, and Seth] consubstantial? Of course they were. Well then, here it is an acknowledged fact that different persons may have the same substance. I say this, not that I would attribute creation or fraction or any property of body to the Godhead . . . but that I may contemplate in these, as on a stage, things which are objects of thought alone. For it is not possible to trace out any image exactly to the whole extent of the truth. (Fifth Theological Oration xi, Schaff and Wace 1900/99, vol. 7: 321)

Elsewhere, he is even clearer about the importance of inseparability, and the inadequacy of the analogy with three human beings: To us there is One God, for the Godhead is One . . . For one [Person] is not more and another less God; nor is One before and another after; nor are They divided in will or parted in power; nor can you find here any of the qualities of divisible things; but the Godhead is, to speak concisely, undivided in separate Persons . . . . [D]o not the Greeks also believe in one Godhead, as their more advanced philosophers declare? And with us, Humanity is one, namely the entire race; but yet they have many gods, not One, just as there are many men. But in this case the common nature has a unity which is only conceivable in thought; and the individuals are parted from one another very far indeed, both by time and by dispositions and by power. For we are not only compound beings, but contrasted beings, both with one another and with ourselves; nor do we remain entirely the same [over time]. (Fifth Theological Oration xiv, xv, Schaff and Wace 1900/99, vol. 7: 322)

As I see it, the upshot of these passages is that the analogy with three humans breaks down at precisely the point where the social model and the constitution model differ. Three humans are separated in matter, time, space, and agency; the three Persons are not. There are other indications in both Gregory of Nyssa and Gregory Nazianzus that their view of the Trinity was along the lines of the constitution model.⁵¹ ⁵¹ See, e.g. Gregory Nazianzen’s analogy of three suns in his Fifth Theological Oration, ch. 14, and the rainbow analogy in Basil of Caesarea’s Epistle 38 (generally regarded as having been written by Gregory of Nyssa rather than by Basil).

 

165

I want to close, however, by dismissing an argument for the claim that they rejected the constitution model. In Epistle 52, Basil of Caesarea writes: For they maintained that the homoousion set forth the idea both of essence and of what is derived from it, so that the essence, when divided, confers the title of co-essential on the parts into which it is divided. This explanation has some reason in the case of bronze and coins made therefrom, but in the case of God the Father and God the Son there is no question of substance anterior or even underlying both; the mere thought and utterance of such a thing is the last extravagance of impiety. (Schaff and Wace 1900/99, vol. 8: 155)

Commenting on this passage, William Lane Craig (2005: 80) takes Basil to be ‘vehemently’ rejecting constitution models outright. But this seems to me to be exactly the wrong interpretation. Though constitution theorists would certainly want to say of each coin that it is constituted by a piece of bronze, Basil is not here considering a case in which multiple bronze items constitute one another. Indeed, what Craig does not recognize is that the coins in this example are consubstantial (if at all) in the way that Father, Son, and Holy Spirit would be on the social trinitarian view. So, if any contemporary model is being rejected, it is that one (the one Craig himself favours) and not the constitution view. Moreover, what is at issue for Basil here is just the question whether there is an anterior substance underlying the Persons. But the constitution model is not committed to an anterior substance. An anterior substance would be one that is prior to and not the same substance as the persons. But, on the constitution view, the divine nature is neither. If the foregoing is right, then, the Cappadocians did not in fact favour anything like social trinitarianism; rather, their view was much more in line with constitution trinitarianism. Likewise, Augustine did not favour the modalistic-sounding view that was attributed to him in Section 2.1; rather, he too was defending a view in line with the constitution view. This is substantial support; and, to my mind, the constitution view wins out on purely intuitive grounds as well.⁵²

References Anscombe, G. E. M. and P. T. Geach. 1963. Three Philosophers. Ithaca, NY: Cornell University Press. Ayres, Lewis. 2004. Nicaea and its Legacy: An Approach to Fourth-Century Trinitarian Theology. New York: Oxford University Press. ⁵² I am grateful to Jeff Brower and Tom Flint for very helpful comments on an earlier draft of this chapter.

166

   :  

Barnes, Michel René. 1995a. ‘Augustine in Contemporary Trinitarian Theology’. Theological Studies 56: 237–50. Barnes, Michel René. 1995b. ‘De Regnon Reconsidered’. Augustinian Studies 26: 51–80. Barnes, Michel René. 1998. ‘The Fourth Century as Trinitarian Canon’. In Christian Origins: Theology, Rhetoric, and Community, edited by Lewis Ayres and Gareth Jones, 47–67. London: Routledge. Barnes, Michel René. 1999. ‘Rereading Augustine’s Theology of the Trinity’. In The Trinity, edited by Stephen T. Davis, Daniel Kendall, and Gerald O’Collins, 145–76. New York: Oxford University Press. Bartel, Timothy. 1988. ‘The Plight of the Relative Trinitarian.’ Religious Studies 24: 129–55. Bartel, Timothy. 1993. ‘Could There Be More Than One Almighty?’ Religious Studies 29: 465–95. Bartel, Timothy. 1994. ‘Could There Be More Than One Lord?’ Faith and Philosophy 11: 357–78. Berkhof, Louis. 1938/96. Systematic Theology: New Combined Edition. Grand Rapids, MI: Eerdmans. Brower, Jeffrey. 2001. ‘Relations without Polyadic Properties: Albert the Great on the Nature and Ontological Status of Relations’. Archiv fur Geschichte der Philosophie 83: 225–57. Brower, Jeffrey. 2004a. ‘The Problem with Social Trinitarianism: A Reply to Wierenga’. Faith and Philosophy 21: 295–303. Brower, Jeffrey. 2004b. ‘Abelard on the Trinity’. In The Cambridge Companion to Abelard, 223–57. Cambridge: Cambridge University Press. Brower, Jeffrey. 2005. ‘Medieval Theories of Relations’. Stanford Encyclopedia of Philosophy. http://plato.stanford.edu/entries/relations-medieval/#3.3. Brower, Jeffrey and Michael Rea. 2005a. ‘Material Constitution and the Trinity’. Faith and Philosophy 22: 487–505. Brower, Jeffrey and Michael Rea. 2005b. ‘Understanding the Trinity’. Logos 8: 145–57. Brown, David. 1985. The Divine Trinity. La Salle, IL: Open Court Publishing Co. Brown, David. 1989. ‘Trinitarian Personhood and Individuality’. In Trinity, Incarnation, and Atonement, edited by R. Feenstra and C. Plantinga, 48–78. Notre Dame, IN: University of Notre Dame Press. Cain, James. 1989. ‘The Doctrine of the Trinity and the Logic of Relative Identity’. Religious Studies 25: 141–52. Cartwright, Richard. 1987. ‘On the Logical Problem of the Trinity’. In Philosophical Essays, 187–200. Cambridge, MA: MIT Press. Cary, Philip. 1992. ‘On Behalf of Classical Trinitarianism: A Critique of Rahner on the Trinity’. The Thomist 56: 365–404.

 

167

Cary, Philip. 1995. ‘Historical Perspectives on Trinitarian Doctrine’. Religious and Theological Studies Fellowship Bulletin, Nov./Dec. http://www.scribd.com/doc/ 2385279/Historical-Perspectives-on-Trinitarian-Doctrine-by-Phillip-Cary. Clark, Kelly James. 1996. ‘Trinity or Tritheism?’. Religious Studies 32: 463–76. Coakley, Sarah. 1999. ‘Persons’ and the “Social” Doctrine of the Trinity: A Critique of Current Analytic Discussion’. In The Trinity, edited by Stephen T. Davis, Daniel Kendall, and Gerald O’Collins, 123–44. New York: Oxford University Press. Coakley, Sarah (ed.). 2003. Re-Thinking Gregory of Nyssa. Oxford: Blackwell. Craig, William Lane. 2005. ‘Does the Problem of Material Constitution Illuminate the Doctrine of the Trinity?’ Faith and Philosophy 22: 77–86. Cross, Richard. 2002a. ‘Two Models of the Trinity?’ Heythrop Journal 43: 275–94. Cross, Richard. 2002b. ‘Gregory of Nyssa on Universals’. Vigiliae Christianae 56: 372–410. Davis, Stephen T., Daniel Kendall, and Gerald O’Collins (eds). 1999. The Trinity. New York: Oxford University Press. De Régnon, Theodore. 1892/8. Études de théologie positive sur la Sainte Trinité, 4 vols. Paris: Victor Retaux. Feser, Edward. 1997. ‘Swinburne’s Tritheism’. International Journal for Philosophy of Religion 42: 175–84. Forrest, Peter. 1998. ‘Divine Fission: A New Way of Moderating Social Trinitiarianism’. Religious Studies 34: 281–97. Geach, Peter. 1967. ‘Identity’. Review of Metaphysics 21: 3–12. Reprinted in Logic Matters. Berkeley: University of California Press, 1972. Geach, Peter. 1980. Reference and Generality, 3rd edn. Ithaca, NY: Cornell University Press. Hasker, William. 1970. ‘Tri-Unity’. Journal of Religion 50: 1–32. Hodge, Charles. 1873. Systematic Theology, vol. 1. New York: Scribner, Armstrong, and Co. Howard-Snyder, Daniel. 2003. ‘Trinity Monotheism’. Philosophia Christi 5: 375–403. Kelly, J. N. D. (trans.) 1964. The Athanasian Creed. London: Adam Charles Black. Kelly, J. N. D. 1972. Early Christian Creeds, 3rd edn. New York: Longman. LaCugna, Catherine Mowry. 1986. ‘Philosophers and Theologians on the Trinity’. Modern Theology 2: 169–81. LaCugna, Catherine Mowry. 1991. God For Us: The Trinity and Christian Life. San Francisco, CA: Harper San Francisco. Layman, C. Stephen. 1988. ‘Tritheism and the Trinity’. Faith and Philosophy 5: 291–8. Leftow, Brian. 1999. ‘Anti Social Trinitarianism’. In The Trinity, edited by Stephen T. Davis, Daniel Kendall, and Gerald O’Collins, 203–49. New York: Oxford University Press.

168

   :  

Leftow, Brian. 2004. ‘A Latin Trinity’. Faith and Philosophy 21: 304–33. Leftow, Brian. 2007. ‘Modes without Modalism’. In Persons: Human and Divine, 357–75. Oxford: Oxford University Press. Leith, John H. (ed.) 1982. Creeds of the Churches, 3rd edn. Louisville, KY: John Knox. Martinich, A. P. 1978. ‘Identity and Trinity’. Journal of Religion 58: 169–81. Martinich, A. P. 1979. ‘God, Emperor, and Relative Identity’. Franciscan Studies 39: 180–91. Merricks, Trenton. 2006. ‘Split Brains and the Godhead’. In Knowledge and Reality: Essays in Honor of Alvin Plantinga, edited by Thomas Crisp et al., 299–326. Dordrecht: Kluwer. Moltmann, Jürgen. 1981. The Trinity and the Kingdom: The Doctrine of God. Translated by Margaret Kohl. San Francisco, CA: Harper San Francisco. Moreland, J. P. and William Lane Craig. 2003. Philosophical Foundations for a Christian Worldview. Downers Grove, IL: InterVarsity Press. Morris, Thomas V. 1991. Our Idea of God. Downers Grove, IL: InterVarsity Press. O’Collins, Gerald. 1999. The Tripersonal God. Mahwah, NJ: Paulist Press. Placher, William. 1983. A History of Christian Theology. Philadelphia, PA: Westminster Press. Plantinga, Cornelius Jr. 1986. ‘Gregory of Nyssa and the Social Analogy of the Trinity’. Thomist 50: 325–52. Plantinga, Cornelius Jr. 1988. ‘The Threeness/Oneness Problem of the Trinity’. Calvin Theological Journal 23: 37–53. Plantinga, Cornelius Jr. 1989. ‘Social Trinity and Tritheism’. In Trinity, Incarnation, and Atonement, edited by R. Feenstra and C. Plantinga, 21–47. Notre Dame, IN: University of Notre Dame Press. Rea, Michael. 2003. ‘Relative Identity and the Doctrine of the Trinity’. Philosophia Christi 5: 431–46. Rea, Michael. 2006. ‘Polytheism and Christian Belief ’. Journal of Theological Studies 57: 133–48. Richardson, Cyril. 1955. ‘The Enigma of the Trinity’. In Companion to the Study of St. Augustine, edited by Roy Battenhouse, 235–56. New York: Oxford University Press. Schaff, Philip and Henry Wace. 1887/1999. Nicene and Post-Nicene Fathers, First Series, 14 vols. Peabody, MA: Hendrickson Publishers. Schaff, Philip and Henry Wace. 1900/99. Nicene and Post-Nicene Fathers, Second Series, 14 volumes. Peabody, MA: Hendrickson Publishers. Schwartz, E. (ed.) 1914. Acta Conciliorum Oecumenicorum. Berlin: W. de Gruyter. Stead, Christopher. 1977. Divine Substance. New York: Oxford University Press. Swinburne, Richard. 1994. The Christian God. Oxford: Clarendon Press.

 

169

Tuggy, Dale. 2003. ‘The Unfinished Business of Trinitarian Theorizing’. Religious Studies 39: 165–83. Tuggy, Dale. 2004. ‘Divine Deception, Identity, and Social Trinitarianism’. Religious Studies 40: 269–87. Turcescu, Lucian. 2005. Gregory of Nyssa and the Concept of Divine Persons. New York: Oxford University Press. van Inwagen, Peter. 1988. ‘And Yet They Are Not Three Gods But One God’. In Philosophy and the Christian Faith, edited by Thomas Morris, 241–78. Notre Dame, IN: University of Notre Dame Press. van Inwagen, Peter. 2003. ‘Three Persons in One Being: On Attempts to Show that the Doctrine of the Trinity is Self-Contradictory’. In The Holy Trinity, edited by Melville Stewart, 83–97. Dordrecht: Kluwer. van Inwagen, Peter. 2004. ‘A Theory of Properties’. Oxford Studies in Metaphysics 1: 107–38. Wierenga, Edward. 2004. ‘Trinity and Polytheism’. Faith and Philosophy 21: 281–94. Williams, C. J. F. 1994. ‘Neither Confounding the Persons nor Dividing the Substance’. In Reason and the Christian Religion: Essays in Honor of Richard Swinburne, edited by Alan Padgett, 227–43. Oxford: Clarendon Press. Wolfson, H. A. 1964. The Philosophy of the Church Fathers, Vol. 1: Faith, Trinity, Incarnation. Cambridge, MA: Harvard University Press.

8 Polytheism and Christian Belief Christians are monotheists; but they believe in three fully divine beings—the three Persons of the Godhead: Father, Son, and Holy Spirit. The tension here is obvious and well known. ‘Polytheism’ is most commonly defined simply and without qualification as ‘belief in more than one god’, and a god is most commonly understood to be any being that is fully divine. Thus, on the most common way of understanding polytheism, orthodox Christian belief is not monotheistic, but quite clearly polytheistic. Christian philosophers and theologians have long been concerned with the question of how to reconcile their belief in three divine Persons with their commitment to monotheism. One strategy is to insist that, despite being three in some sense, the divine Persons are somehow also the same god.¹ By far the most popular strategy, however, involves denying that Father, Son, and Holy Spirit are the same god and then arguing that monotheism is secured by the special relationship that obtains among the divine Persons. Proponents of this latter sort of view are typically labelled ‘social trinitarians’ (because they conceive of the divine Trinity on analogy with a society of human persons).² One of the basic ideas that social trinitarians share in common, then, is that the Christian commitment to monotheism is not a commitment to the claim that there is only one divine being, but rather a commitment to the claim that all divine beings—all gods, in the ordinary sense of the term—stand in a particular relation to one or more of the members of the holy Trinity. What sort of relation do social trinitarians have in mind? Various suggestions have been offered in the literature. Cornelius Plantinga argues that social trinitarians may ‘cling to respectability as monotheists’ by noting what he takes to be three perfectly good senses in which there is only one God: (a) there is only one font of divinity (the Father), (b) there is ‘only one generic divinity . . . one Godhood or Godhead or Godness’, and (c) there is ‘only one divine family or monarchy or community, namely, the Holy Trinity itself ’ (1989: 31). The idea, then, seems to be that monotheism is true, no matter how many gods there are, so long as all gods derive their divinity from one source, or share a single divine

¹ For further discussion of this strategy, and references, see Rea 2003, and also Brower and Rea 2005. ² Defenders of social trinitarianism include Davis 1999, 2003; Forrest 1998; Layman 1988; Plantinga 1986, 1988, 1989; Swinburne 1994; and Williams 1994.

Essays in Analytic Theology: Volume I. Michael C. Rea, Oxford University Press (2020). © Michael C. Rea. DOI: 10.1093/oso/9780198866800.003.0009

   

171

nature (as we humans share a single human nature), or are joined together as a divine family, monarchy, or community. Richard Swinburne (1994), on the other hand, focuses on the fact that Yahweh is (as most social trinitarians have it) a composite individual or society whose parts or members are bound together by perfect love, harmony of will, and necessary mutual interdependence—i.e. none can exist without the others or come into conflict with the others. He then asserts that the fact that the Persons of the Trinity stand in these relations is sufficient to secure monotheism. Thus, the implication is that monotheism is nothing more than the view that all of the gods that exist are parts or members of a divine individual or community that does not depend for its existence on anything outside the community, and all are bound together by perfect love, harmony of will and action, necessary mutual interdependence.³ J. P. Moreland and William Lane Craig (2003) take a similar approach. On their view, Yahweh is composed of the Persons in a sense analogous to the way in which Cerberus, guardian of the underworld in Greek mythology, might be thought to be composed of three individual dogs. Here again the idea seems to be that monotheism is secured by the mere fact that the Persons are parts of a single fully divine being. There are other suggestions in the literature; but they tend to run along very similar lines. A standard criticism of social trinitarianism is that these (re-)characterizations of monotheism are implausible. Those lodging the criticism typically do so in one of two ways. Some try to argue that statements in the Nicene and Athanasian Creeds that are commonly taken to rule out polytheism also speak against social trinitarianism.⁴ But creedal exegesis is controversial, especially insofar as there is good reason to think that it is sometimes driven in part by precisely those considerations that fuel the controversy between social trinitarianism and its opponents. As a means of settling this latter controversy, therefore, it is hard to see matters of creedal exegesis as decisive. Others object to social trinitarianism on the grounds that it is not plausible to think that (say) Greek polytheism would become monotheistic if only we added that Zeus and the other gods enjoyed perfect love, harmony, and mutual interdependence with one another.⁵ This objection probably is decisive against the suggestion that the perfect love, harmony, and mutual interdependence of the Persons is what makes them one rather than three in the sense relevant to monotheism. But it has no implications for the more popular suggestion that Christianity is monotheistic because the Persons are parts of the one and only fully divine being. Thus, on the whole, the case currently available in the literature for the conclusion that social trinitarianism is polytheistic is, at best, incomplete. It would be nice if a stronger, more compelling case could be made; and I think that one can.

³ Cf. Swinburne 1994: 181. ⁴ See especially Clark 1996, Feser 1997, and Leftow 1999. ⁵ See Leftow 1999 and Merricks 6.

172

   :  

In what follows, I will argue that if the social trinitarian understanding of the doctrine of the Trinity is correct, then Christianity is not interestingly different from the polytheistic Amun-Re theology of Egypt’s New Kingdom period.⁶ In short, my argument is this: (a) Christianity as understood by the social trinitarian (henceforth ST-Christianity) resembles Amun-Re theology in the relations it takes to obtain among its gods; (b) Amun-Re theology was polytheistic; therefore, (c) ST-Christianity is polytheistic as well. Since the relations mentioned in premise (a) are both familial and mereological, the conclusion of my argument will apply equally to both versions of social trinitarianism described above. As we shall see, there is some controversy about whether the religion of ancient Egypt (during any dynasty) was genuinely polytheistic. However, as we will also see, the considerations invoked by those who think that Egyptian religion was monotheistic cut as much against premise (a) as against premise (b). Moreover, those who urge a monotheistic interpretation of Egyptian religion seem actually to agree that if premise (a) of my argument were true, premise (b) would be as well. Thus, I will argue, the light shed on the nature of polytheism by the controversy over the polytheistic character of Egyptian religion in fact strengthens rather than weakens my case for the conclusion that social trinitarianism is not a form of monotheism.

1. New Kingdom Amun-Re Theology Egyptian religion was constantly evolving. New gods were periodically added to the pantheon; attributes of the gods and the relations among gods were continually changing; and several conflicting creation myths were developed with different stories about the origins of the gods and the creation of the world. The present discussion will focus primarily on the Amun-Re theology that dominated Egypt’s 18th–20th Dynasties (outside the Amarna period),⁷ and on the features of Egypt’s supreme god during those dynasties, Amun-Re. Before turning to Amun-Re, however, it will be useful first to say a few words about some of the Egyptian deities that will occupy centre stage in our discussion and about some of the conflicting creation myths that serve as background for it. Egyptians worshipped at least four different sun gods: Khepri, god of the morning sun; Re, god of the noonday sun; Atum, god of the setting sun; and

⁶ The New Kingdom period spans Egypt’s 18th–20th Dynasties, 1550–1070 . For detailed discussion of Egyptian religion during that period, see especially Assmann 1995. ⁷ The Amarna period was the period during Egypt’s 18th Dynasty when Amenhotep IV came to power, changed his name to ‘Akhenaten’, moved the nation’s capital to Amarna, and sought to promote the worship of Aten to the exclusion of other deities in the pantheon. After Akhenaten’s reign, Egyptians quickly sought to undo the religious changes he had tried to institute. For brief and useful discussion, see Baines 1991, 2000, and Assmann 1995.

   

173

Aten, the solar disk. All of these gods were recognized at some point or other as self-created creator deities. Ptah and Amun were also among the gods who were sometimes said to be self-created creators. Ptah was primarily associated with arts and crafts and was the chief god of Memphis. Amun was known as the ‘hidden god’. He began his career as a relatively minor deity—not even the chief deity of Thebes, the city in which he was primarily worshipped. As Thebes rose to power, however, so too did Amun’s stature. During Egypt’s 12th Dynasty, Amun replaced Montu (a minor sun god) as the chief deity of Thebes. In the earliest and best known creation myths, it is usually Atum who is listed as the self-made creator of the world and of the other gods and goddesses.⁸ For example, according to the so-called Heliopolitan cosmogony, Atum appeared out of the primeval waters of Nun and brought forth a son and a daughter, Shu and Tefnut. One version of the story has it that Shu and Tefnut are the products of Atum’s act of masturbation; another claims that they sprang forth from his eyes and mouth. Shu and Tefnut later gave birth to Geb and Nut, who in turn gave birth to Osiris, Isis, Seth, and Nephthys. These nine gods and goddesses together constitute the Heliopolitan Ennead; and (importantly, for our purposes), the members of the Ennead comprise a divine family with a single father who is the source of being and of divinity. But it is not always Atum who is taken to be the father of the gods. In some cases, composite deities like Atum-Re, or other solar deities like Re or Khepri are said to play Atum’s role in the basic Heliopolitan story. But we also find dramatically different stories. For example, the cosmogonic tradition associated with Hermopolis maintains that the creator of the world and of the other gods emerged from an egg brought forth by eight primordial divinities (the Hermopolitan Ogdoad): Amun and Amaunet (hiddenness), Huh and Hauhet (formlessness), Kuk and Kauket (darkness), and Nun and Naunet (the watery abyss). Often Atum is listed as the creator who emerged from the egg; but, again, sometimes other deities are placed in that role. Alternatively, the cosmogonic tradition associated with Memphis lists Ptah as the uncreated creator, bringing forth all of the members of the Heliopolitan Ennead by an act of speech.⁹ As I have already indicated, the god Amun grew in significance as Thebes rose to power at the end of Egypt’s 11th Dynasty. Later, he came to be worshipped in conjunction with Re as Amun-Re. Amun-Re was eventually recognized as king of the gods and worshipped as the chief national deity. By the time of Egypt’s 18th Dynasty, we find a new creation story according to which it is Amun-Re who plays

⁸ For the most part, I have followed rather closely Lesko 1991 in my descriptions of the various creation stories and in my claims about the functions of the gods and goddesses mentioned therein. ⁹ Cf. Lichtheim 1973.

174

   :  

the role of uncreated creator and source of being for the world and all of the gods. The following is a passage from the Khonsu Cosmogony, which expresses a Theban version of the creation story: Words spoken by Amun-Re, King of the Gods, august being, chief of all the gods, the Great God, lord of the sky, earth, the other world, water and mountains, the august soul of Kem-atef serpent, father of the semen, mother of the egg, who engendered everything living, the hidden soul who made the gods, who formed the land with his semen, father of the fathers of the Ogdoad in the tomb chamber in the necropolis in the place Djeme, who created this place in Nun, overflowing seed the first time . . . . Amun in that name of his called Ptah created the egg that came forth from Nun . . . as Ptah of the Heh gods and the Nenu goddesses who created heaven and earth . . . ¹⁰

Here, we see an assimilation of the three basic creation traditions described above (Heliopolitan, Hermopolitan, and Memphite), with Amun-Re being identified in the roles of Atum and Ptah. The assimilation with Atum, Ptah, and other creator gods, and the elevation of Amun-Re to a position of supreme authority is explicit elsewhere as well. Thus, for example, in one of the 18th Dynasty Cairo Hymns to Amun-Re we find the following: This splendid God, Lord of all gods, Amun-Re Lord of the thrones of the two Lands, Foremost in Ipet-Sut Splendid Soul who came to be in the Beginning, great God who dwells in Truth, Primordial God who engendered the first gods, through whom every god came to be, Most unique of the unique, who made all that is, who began the world back in the First Time; Whose features are hidden, yet frequent his appearances, and there is no knowing how he flowed forth; Gloriously powerful, beloved, majestic, Mighty in his theophanies, magnificent; Powerful Being through whose Being each Being came to be, Who began Becoming with none but himself.¹¹

¹⁰ Parker and Lesko 1988: 169–70, quoted in Lesko 1991: 105–6.

¹¹ Foster 1995: 65.

   

175

In the Leiden Hymn to Amun-Re, we find the following startling passage: All gods are three: Amun, Re, Ptah, they have no equal. His name is hidden as Amun, he is Re in the face, and his body is Ptah.¹²

But now various questions arise: What is the relation, according to the New Kingdom Amun-Re theology, between deities like Amun-Re, Amun, Re, Atum, and Ptah? Is Amun-Re simply absorbing the functions of the others? Is the religion of Amun-Re, in effect, treating Atum, Ptah, Amun, and Re as alternate manifestations of a single unified being called by the name ‘Amun-Re’? Or do at least some of these deities stand in part–whole relations with one another, as the Leiden Hymn seems to suggest? As it turns out, syncretism (the combination of two or more gods into an apparent composite like Amun-Re) was not an uncommon phenomenon in Egyptian religion. Thus, we find, in addition to Amun-Re, other combinations like Atum-Re, Amun-Re-Atum, Amun-Re-Montu, Min-Amun, and so on. But Egyptologists are divided on the question of what to make of this phenomenon. The view that seems generally to be taken for granted in semi-popular works on Egyptian mythology as well as in some scholarly works is that the composite names represent the identification of one deity with another, or the absorption of one deity by another, or the unification of powers previously ascribed to separate deities in a single divine reality.¹³ According to this view, in coming to worship Amun-Re as a fusion of Amun and Re, and in identifying him in roles played by Amun, Re, Atum, and Ptah, the Egyptians were, in effect, replacing these other deities in their pantheon with a single deity who possessed in himself the (salient) attributes of all of them. Despite its popularity, however, this view is typically rejected by experts on Egyptian syncretism. Instead, scholars like John Baines (1991, 1998, 1999, 2000), Hans Bonnet (1939/99), Erik Hornung (1982), and Siegfried Morenz (1960) argue that syncretistic combinations represented not the absorption of one god by another or the identification of one god with another, but rather the introduction of a new composite deity who had the others as parts.¹⁴ Thus, for example, Erik Hornung writes: Quite often not just two but three or four gods form a new unit that is the object of a cult. Besides tripartite forms such as Ptah-Sokar-Osiris there are

¹² Baines 2000: 56 ¹³ See, for example, Tobin 2002. ¹⁴ Though Bonnet speaks more of one god inhabiting another—a locution that, in the relevant contexts, suggests but does not clearly imply part/whole relations.

176

   :  

quadripartite ones like Amon-Re-Harakhte-Atum or Harmachis-Khepry-ReAtum . . . . In such cases one is reminded of chemical compounds; like them, syncretistic combinations can be dissolved at any time into their constituent elements, which can also form part of other combinations without sacrificing their individuality. Is the purpose of these combinations a clever priestly ‘equalization’ of conflicting religious claims . . . ? Must gods be ‘equated’ with one another until one finishes with a vague, solar-tinged pantheism? Such an interchange of attributes, which leads toward uniformity, is un-Egyptian; if anything it is Hellenistic. The Egyptians place the tensions and contradictions of the world beside one another and then live with them. Amon-Re is not the synthesis of Amon and Re but a new form that exists along with the two older gods.¹⁵

Hornung goes on to make the point that composite deities like Amun-Re are, like their individual parts, objects of worship and cultic devotion. Thus, there is good reason to think that Amun-Re is no less divine than either Amun or Re individually. In sum, then, it seems not far off the mark to say that, according to the AmunRe theology of the New Kingdom, one god (Amun-Re) is the creator, the king of heaven, and the font of divinity; there are multiple divine beings but a single divine family; and at least some of the members of the divine family stand in genuine part–whole relations to the creator and ultimate source of being.¹⁶ In these respects, the Amun-Re theology is very much like ST-Christianity (in both its familial and mereological varieties). And yet it seems quite clear that the Amun-Re theology falls squarely in the category of polytheism. Worshippers of Amun-Re cannot ‘cling to respectability as monotheists’ simply by affirming that there is but one divine family, one ‘Godhead’ or ‘Godness’, one font of divinity. Nor would the claim that several gods are parts of Amun-Re and that all others are manifestations either of Amun-Re or his parts (as suggested in the quoted passage from the Leiden Hymn) make the religion monotheistic. But, if that is so, then it seems that ST-Christianity should be viewed as polytheistic as well.

2. Monotheism or Polytheism? Of course, many superficial differences between the Amun-Re theology and STChristianity remain. Most notably, there are these three: (a) the Theban creation

¹⁵ Hornung 1982: 96–7. ¹⁶ As discussed below, and as indicated above, this interpretation of the Amun-Re theology is not uncontroversial. But what matters most for my analogical argument is simply the fact that this is one available interpretation, and if it is correct, then the Amun-Re theology was polytheistic.

   

177

myth seems to recognize primordial divinities (like Nun) apart from Amun-Re, (b) the divine beings of ST-Christianity, but not (obviously) of Amun-Re theology, stand in relations of necessary mutual interdependence, and (c) the pantheon apparently still includes various divinities (e.g. Osiris, Isis) who seem not to be among the proper parts of Amun-Re.¹⁷ But it is not clear that these differences are at all critical to the assessment of the Amun-Re theology as polytheistic. Regarding the first two differences, it would not be much of a departure to suppose that Amun stands with Re and Ptah in relations of necessary mutual interdependence, to suppose that Amun-Re created rather than emerged from Nun, and to posit an ‘initial’ state exactly parallel to that described in the opening lines of the book of Genesis: In the beginning, God created the heavens and the earth. Now the earth was formless and empty, darkness was over the surface of the deep, and the Spirit of God was hovering over the waters. (NIV)

As to the third difference, it is actually not at all clear how much of a difference it really amounts to. After all, Christians both early and modern have believed in a vast array of superhuman and supernatural beings in addition to the Persons of the Trinity, most notably angels and demons. By comparison with Egyptian religion, relatively fewer of these sorts of beings are given names in the books of the Old and New Testaments; but they are there nonetheless. But nobody thinks that the recognition of such beings is a threat to the monotheistic character of Christianity. The reason is that, though the term ‘god’ in its ordinary sense might reasonably apply to such beings, none of them are taken to be gods on a par with God. But, arguably, the same is true of the Amun-Re theology. Granted, there are such beings as the members of the Heliopolitan Ennead; but, according to the theology, they are not on a par with their creator, the king of the gods, Amun-Re. Nevertheless, it is important to acknowledge here that there is controversy about whether and to what extent ancient Egyptian religion was polytheistic. Most commonly, the focus of these discussions is on the Amarna period, during which time Amenhotep IV changed his name to Akhenaton and undertook a campaign to focus all worship on Aten alone. But there is also considerable discussion of the question whether Egyptian religion at times outside the Amarna period might properly be construed as monotheistic. In the opening chapter of Conceptions of

¹⁷ Here is a fourth difference: Amun-Re theology arose out of a polytheistic background, but STChristianity did not. I grant that this difference may partly explain some of our initial intuitions about the Amun-Re theology; but, at least in my own case, the intuitions about the polytheistic nature of Amun-Re theology remain strong even after attending to this difference. Moreover, the monotheistic background of ST-Christianity does not make me even the least bit inclined to see it as a version of monotheism. Thus, I am inclined to think that this difference is not a relevant one.

178

   :  

God in Ancient Egypt: The One and the Many, Erik Hornung devotes several pages to documenting the widespread nineteenth-century belief that Egyptian religion was, in the words of Champillion-Figeac, ‘a pure monotheism, which manifested itself externally by a symbolic polytheism’.¹⁸ And some texts do seem strongly to suggest that sort of view. For example, in one myth we find Re declaring ‘At dawn, I am called Khepri; at midday, Re; and in the evening, Atum.’¹⁹ But Hornung also argues that by the early twentieth century, most scholars were inclined to reject the view that Egyptian religion was monotheistic.²⁰ Still, the controversy remains unsettled.²¹ Obviously I cannot offer here a sustained defence of the claim that Egyptian religion in the period I am considering was polytheistic. But what will be useful for our purposes is a look at what issues the debate turns on. Consider, for example, Hans Bonnet’s remarks on the effects of syncretism: [Syncretism] . . . dissolves polytheism. Of course the gods remain; but the individual features that characterize them and distinguish one from another lose their significance . . . . The conception of god, which syncretism introduced into local cults . . . had everywhere the same basic tenor. Everywhere there prevailed the idea of the primeval deity and cosmic ruler who at the beginning created and who continues to influence and sustain everything. The pious encountered this One, in whom all divine efficacy ultimately resides and is based, in forms that varied from place to place. This experience left the way open for the insight that under all names and manifestations only the one, intrinsically identical reality is apprehended and worshipped.²²

For Bonnet, the ‘dissolution of polytheism’ consists precisely in the growing view that the various gods and goddesses of the pantheon are all just manifestations of a single divine reality. Erik Hornung disagrees with Bonnet’s claim that syncretism dissolves polytheism. But, notably, he does not take issue with the suggestion that polytheism would dissolve if the various gods and goddesses of the pantheon came to be seen as manifestations of a single divine reality. Rather, he takes issue with the suggestion that syncretistic combinations represent the apprehension of multiple gods as manifestations of a single divine reality. Thus, he writes:

¹⁸ Quoted in Tobin 1989: 160. ¹⁹ Watterson 1984: 61; cf. Spence 1915/90: 259. ²⁰ See, especially, Hornung 1982, ch. 3 on syncretism and the conclusion in ch. 8. ²¹ For illuminating discussion and arguments on both sides, see Allen 1999a, 1999b; Assmann 1995, 2001; Baines 1991, 1998, 2000; Budge 1934/88; Černý 1952; Morenz 1960; Tobin 1989; Wilkinson 2003. ²² Bonnet 1939/99: 194.

   

179

It is clear that syncretism does not contain any ‘montheistic tendency,’ but rather forms a strong counter-current to monotheism—so long as it is kept within bounds. Syncretism softens henotheism, the concentration of worship on a single god, and stops it from turning into monotheism, for ultimately syncretism means that a single god is not isolated from the others: in Amun one apprehends and worships also Re, or in Harmachis other forms of the sun god. In this way, the awareness is sharpened that the divine partner of humanity is not one but many.²³

Here too, then, the question whether syncretistic combinations reflect a monotheistic trend in Egyptian religion is taken to turn on the question whether Egyptians viewed the ‘divine partner of humanity’ as one or many. Hornung, recall, is one of the champions of the view that syncretistic combinations involve multiple gods coming to be seen as parts of some further composite deity. But nowhere is this fact taken by him to be a consideration in favour of a monotheistic interpretation of the religion. And, indeed, none of the parties to the controversy over whether Egyptian religion is properly construed as monotheistic seem to take this fact as relevant. Consistently what matters most is either the simple question of whether the religion recognizes one or many divine beings, or the somewhat more nuanced question addressed by Bonnet and others of whether the multiple gods of the religion were viewed simply as different manifestations of a single divine reality.²⁴ This latter question seems clearly to be the right one to focus on. We have already acknowledged that orthodox Christian belief is straightforwardly committed to the existence of multiple divine beings. Thus, simply to define polytheism as ‘belief in multiple divine beings’ is, as social trinitarians rightly point out, to fail from the outset to take the Christian claim to monotheism seriously. But, at the same time, we must also take seriously the normal sense of the term ‘polytheism’; and any answer to the question of what polytheism is must respect that normal sense. Social trinitarians claim that their reconstruals of polytheism do respect the normal sense of the term. But my point here is that experts on religion with no trinitarian axes to grind seem not to agree. No participant in the debate about whether Egyptian religion is polytheistic seems to think that monotheism would be secured simply by positing the sorts of relations among gods that social trinitarians think obtain between Father, Son, Holy Spirit, and the Christian God. Indeed, the Egyptologists who argue most strongly in favour of a polytheistic interpretation of Egyptian religion are precisely the ones who interpret the religion in the way that I have summarized at the end of Section 1—in the way that makes it most strongly analogous to ST-Christianity.

²³ Hornung 1982: 98.

²⁴ See, especially, Watterson 1984: 61; cf. Spence 1915/90: 259.

180

   :  

It might be objected that I have placed undue weight on the opinions of Egyptologists about the nature of polytheism and monotheism. Thus, the social trinitarian might simply dig in her heels and insist that Egyptologists are generally using the wrong criteria for determining whether Egyptian religion (in any dynasty) was polytheistic. But I think that this objection would be convincing only if there were a substantial tradition of reflection on the nature of monotheism and polytheism apart from the writings of social trinitarians themselves that supported the claim that Egyptologists were using the wrong criteria. Unfortunately, there seems not to be. Indeed, so far as I can tell, what reflection there is on the nature of monotheism and polytheism outside the literature on social trinitarianism tends generally to support the thesis that I have been urging—namely, that mereological relations, familial relations, and relations of necessary mutual interdependence among deities are neither alone nor in conjunction sufficient to secure monotheism.²⁵ This is not to say that there is anything like consensus on the nature of either monotheism or polytheism. Far from it. But amidst the disagreement there is a remarkable absence of support outside the writings of social trinitarians themselves for the conclusion that social trinitarianism is monotheistic. And it is largely because of this fact that I take the resemblance between ST-Christianity and Egypt’s Amun-Re theology to speak so compellingly in favour of the conclusion that ST-Christianity is not monotheistic.

3. Conclusion I have argued that, given the interpretation of Egyptian syncretism that is endorsed by scholars like Hornung, Baines, and others, Amun-Re theology is directly analogous to ST-Christianity. Since Amun-Re theology as I have interpreted it is clearly polytheistic, the close analogy with ST-Christianity strongly favours a view according to which ST-Christianity is polytheistic as well. We have also seen that there is controversy about whether Egyptian religion in the New Kingdom and during other periods was polytheistic; but, far from providing resources for a monotheistic understanding of social trinitarianism, that controversy seems only to seal the case in favour of the polytheistic nature of STChristianity. The reason, again, is that the controversy over the question whether Egyptian religion was polytheistic turns on the question whether the gods of Egypt

²⁵ To take just a few examples, see the essays in Athanassiadi and Frede 1999, Davies 1989, and Porter 2000. See also MacDonald 2003, Mavrodes 1995, Miller 1974, Owen 2002, and Smith 2001. I am grateful to Carl Mosser for pressing me to take explicit notice of the great diversity of opinions about the nature of monotheism and polytheism and of the great difficulty involved in providing satisfying characterizations of each. Despite the difficulty and diversity, however, I think that it is possible to discern broad lines of agreement; and, as indicated above, I think that once those lines have been discerned, it is easy to see that social trinitarians fall afoul of them.

   

181

are best regarded as manifestations of a single divine reality. On the interpretation of the Amun-Re theology that I have favoured, they are not. Likewise, on the social Trinitarian’s understanding of the doctrine of the Trinity, Father, Son, and Holy Spirit are not manifestations of a single divine reality either. Thus, again, both theologies seem best understood as polytheistic. If there were no orthodox understanding of the doctrine of the Trinity according to which Father, Son, and Holy Spirit might meaningfully be said to be manifestations of a single divine reality—if, in other words, there were no viable interpretation of that doctrine according to which they were distinct but still, somehow, the same God—then the conclusion of this chapter would probably be that Christians should learn to be content regarding themselves as in some sense polytheists. In that event, I would urge that attention be shifted away from defending the Christian claim to monotheism and towards an investigation of the question what sorts of polytheism Christianity means to oppose. If, however, there are interpretations of the doctrine of the Trinity which (while avoiding the heresy of modalism) are consistent with monotheism,²⁶ I say so much the better for those interpretations, and so much the worse for social trinitarianism.²⁷

References Allen, J. P. 1999a. ‘Monotheism: The Egyptian Roots’. Archaeology Odyssey 78: 44–54, 59. Allen, J. P. 1999b. Middle Egyptian: An Introduction to the Language and Culture of Hieroglyphs. New York: Cambridge University Press. Assmann, Jan. 1995. Egyptian Solar Religion in the New Kingdom: Re, Amun, and the Crisis of Polytheism, trans. Anthony Alcock. London: Kegan Paul. Assmann, Jan. 2001. The Search for God in Ancient Egypt, translated by David Lorton. Ithaca, NY: Cornell University Press. Athanassiadi, Polymnia and Michael Frede (eds). 1999. Pagan Monotheism in Late Antiquity. Oxford: Oxford University Press. Baines, John. 1991. ‘Society, Morality, and Religious Practice’. In Religion in Ancient Egypt: Gods, Myths, and Personal Practice, edited by Byron Shafer, 123–200. Ithaca, NY: Cornell University Press.

²⁶ And I think that there are: see Brower and Rea 2005. ²⁷ I would like to thank Michael Bergmann, Jeff Brower, Marian David, Daniel Howard-Snyder, Shieva Kleinschmidt, and James VanderKam for helpful comments on an earlier draft. An earlier version of this chapter was read at the 2003 Annual Meeting of the American Academy of Religion. I am grateful to my commentator, Stephen Davis, and also to Carl Mosser for valuable discussion during that session and for helpful correspondence on the issues discussed in this chapter.

182

   :  

Baines, John. 1998. ‘The Dawn of the Amarna Age’. In Amenhotep III: Perspectives on His Reign, edited by David O’Connor and Eric H. Cline, 271–312. Ann Arbor: University of Michigan Press. Baines, John. 1999. ‘Egyptian Syncretism: Hans Bonnet’s Contribution’. Orientalia 68: 199–214. Baines, John. 2000. ‘Egyptian Deities in Context: Multiplicity, Unity, and the Problem of Change’. In One God or Many? Conceptions of Divinity in the Ancient World, edited by Barbara Porter, 9–78. Chebeague, ME: Casco Bay Assyriological Institute. Bonnet, Hans. 1939/99. ‘On Understanding Syncretism’. Translated by John Baines. Orientalia 68: 181–98. Translated from ‘Zum Verständnis des Synkretismus’, Zeitschrift für Ägyptische Sprache und Altertumskunde 75 (1939): 40–52. Brower, Jeffrey and Michael Rea. 2005. ‘Material Constitution and the Trinity’. Faith and Philosophy 22: 1–20. Budge, E. A. W. 1934/88. From Fetish to God in Ancient Egypt. New York: Dover. Černý, Jaroslav. 1952. Ancient Egyptian Religion. Westport, CT: Greenwood Press. Clark, Kelly. 1996. ‘Trinity or Tritheism?’. Religious Studies 32: 463–76. Davies, Glenys (ed.). 1989. Polytheistic Systems. Edinburgh: Edinburgh University Press. Davis, Stephen. 1999. ‘A Somewhat Playful Proof of the Social Trinity in Five Easy Steps’. Philosophia Christi 1: 103–5. Davis, Stephen. 2003. ‘Perichoretic Monotheism’. in The Holy Trinity, edited by Alexander Kirlezjev and Melville Stewart. Dordrecht: Kluwer. Feser, Edward. 1997. ‘Swinburne’s Tritheism’. International Journal for Philosophy of Religion 42: 175–84. Forrest, Peter. 1998. ‘Divine Fission: A New Way of Moderating Social Trinitarianism’. Religious Studies 34: 281–97. Foster, John (trans.). 1995. Hymns, Prayers, and Songs: An Anthology of Ancient Egyptian Lyric Poetry, edited by Susan Tower Hollis. Atlanta, GA: Scholars Press. Hornung, Erik. 1982. Conceptions of God in Ancient Egypt: The One and the Many, translated by John Baines. Ithaca, NY: Cornell University Press. Layman, C. Stephen. 1988. ‘Tritheism and the Trinity’. Faith and Philosophy 5: 291–8. Leftow, Brian. 1999. ‘Anti Social Trinitarianism’. In The Trinity, edited by Stephen T. Davis, Daniel Kendall, and Gerald O’Collins, 203–49. New York: Oxford University Press. Lesko, Leonard. 1991. ‘Ancient Egyptian Cosmogonies and Cosmology’. In Religion in Ancient Egypt: Gods, Myths, and Personal Practice, edited by Byron Shafer, pp. 88–122. Ithaca, NY: Cornell University Press. Lichtheim, Miriam. 1973. Ancient Egyptian Religion: A Book of Readings, Vol. 1: The Old and Middle Kingdoms. Berkeley: University of California Press.

   

183

MacDonald, Nathan. 2003. Deuteronomy and the Meaning of ‘Monotheism’. Tübingen: Moher Siebeck. Mavrodes, George. 1995. ‘Polytheism’. in The Rationality of Belief and the Plurality of Faith, edited by Thomas Senor. Ithaca, NY: Cornell University Press. Merricks, Trenton. 2006. ‘Split Brains and the Godhead’. In Knowledge and Reality: Essays in Honor of Alvin Plantinga, edited by Thomas Crisp et al., 299–326. Dordrecht: Kluwer. Miller, David. 1974. The New Polytheism: Rebirth of the Gods and Goddesses. New York: Harper Collins. Moreland, J. P. and William Lane Craig. 2003. Philosophical Foundations for a Christian Worldview. Downers Grove, IL: InterVarsity Press. Morenz, Siegfried. 1960. Egyptian Religion, translated by Ann E. Keep. Ithaca, NY: Cornell University Press. Owen, Paul. 2002. ‘Monotheism, Mormonism, and the New Testament Witness’. In The New Mormon Challenge, edited by Francis J. Beckwith, Carl Mosser, and Paul Owen, 271–314. Grand Rapids, MI: Zondervan. Parker, Richard and Leonard Lesko. 1988. ‘The Khonsu Cosmogony’. In Pyramid Studies and Other Essays Presented to I. E. S. Edwards, edited by John Baines et al., 168–75. London: Egypt Exploration Society. Plantinga, Cornelius Jr. 1986. ‘Gregory of Nyssa and the Social Analogy of the Trinity’. Thomist 50: 325–52. Plantinga, Cornelius Jr. 1988. ‘The Threeness/Oneness Problem of the Trinity’. Calvin Theological Journal 23: 37–53. Plantinga, Cornelius Jr. 1989. ‘Social Trinity and Tritheism’. In Trinity, Incarnation, and Atonement: Philosophical and Theological Essays, edited by Ronald Feenstra and Cornelius Plantinga, Jr., 21–47. Notre Dame, IN: University of Notre Dame Press. Porter, Barbara (ed.). 2000. One God or Many? Conceptions of Divinity in the Ancient World. Chebeague, ME: Casco Bay Assyriological Institute. Rea, Michael. 2003. ‘Relative Identity and the Doctrine of the Trinity’. Philosophia Christi 5: 431–46. Smith, Mark. 2001. The Origins of Biblical Monotheism. Oxford: Oxford University Press. Spence, Lewis. 1915/90. Ancient Egyptian Myths and Legends. New York: Dover. Swinburne, Richard. 1994. The Christian God. Oxford: Clarendon Press. Tobin, Vincent. 1989. Theological Principles of Egyptian Religion. New York: Peter Lang. Tobin, Vincent. 2002. ‘Amun and Amun-Re’. In The Ancient Gods Speak: A Guide to Egyptian Religion, edited by Donald B. Redford, 18–20. Oxford: Oxford University Press.

184

   :  

Watterson, Barbara. 1984. The Gods of Ancient Egypt. New York: Facts on File Publications. Wilkinson, Richard. 2003. The Complete Gods and Goddesses of Ancient Egypt. London: Thames & Hudson. Williams, C. J. F. 1994. ‘Neither Confounding the Persons nor Dividing the Substance’. In Reason and the Christian Religion: Essays in Honor of Richard Swinburne, edited by Alan Padgett, 227–43. Oxford: Clarendon Press.

9 Relative Identity and the Doctrine of the Trinity The doctrine of the Trinity maintains that there are exactly three divine Persons (Father, Son, and Holy Spirit) but only one God. The philosophical problem raised by this doctrine is well known. On the one hand, the doctrine seems clearly to imply that Father, Son, and Holy Spirit are numerically distinct. How else could they be three Persons rather than one? On the other hand, it seems to imply that Father, Son, and Holy Spirit are identical. If each Person is divine, how else could there be exactly one God? But Father, Son, and Holy Spirit cannot be both distinct and identical. Thus, the doctrine appears to be incoherent. In the contemporary literature, there are two main strategies for solving the problem: the relative identity (RI) strategy, and the social trinitarian (ST) strategy.¹ Both of these strategies solve the problem by affirming the divinity of Father, Son, and Holy Spirit while denying their absolute identity either with God or with one another. According to the RI strategy (which will be explained more fully below), the divine Persons stand in various relativized relations of sameness and distinctness. They are, for example, the same God as one another, but they are not the same Person. They are, we might say, God-identical but Person-distinct. Peter Geach has argued for reasons independent of the problem of the Trinity that there is no such thing as absolute identity, that all well-formed identity statements are at least implicitly relativized, and that there is no in-principle obstacle to there being x, y, F, and G such that x is the same F as y but not the same G.² Not surprisingly, then, some philosophers who have embraced the RI strategy have endorsed Geach’s theory of relative identity along with it. But, as we shall see, Geach’s theory is just one among several, and it is even possible to pursue the RI strategy without endorsing a theory of relative identity at all. The ST strategy, on the other hand, maintains that the relation between God and the Persons is not any sort of identity or sameness relation at all. Rather, it is something like

¹ Proponents of the ST strategy include Timothy Bartel (1993, 1994); David Brown (1985, 1989); Stephen Davis (1999); Peter Forrest (1998); C. Stephen Layman (1988); Cornelius Plantinga, Jr. (1986, 1988, 1989); Richard Swinburne (1994); and C. J. F. Williams (2002). The position is also commonly attributed to the Cappadocian Fathers. See, especially, Brown 1985; Plantinga 1986; and Wolfson 1964. Proponents of the RI strategy include G. E. M. Anscombe and P. T. Geach (1961, at pp. 118–20); James Cain (1989); A. P. Martinich (1978, 1979); and Peter van Inwagen (1988). ² See Geach 1967, 1969, 1973, and 1980, sections 30, 34, and 110.

Essays in Analytic Theology: Volume I. Michael C. Rea, Oxford University Press (2020). © Michael C. Rea. DOI: 10.1093/oso/9780198866800.003.0010

186

   :  

parthood (God is a composite being whose parts are the Father, Son, and Holy Spirit) or membership (God is a community whose members are the divine Persons). The main challenge for both strategies is to find some way of respecting the Christian commitment to monotheism without incurring other problems in the process. The ST strategy has been roundly criticized in the literature, and many of the criticisms I wholeheartedly endorse.³ The RI strategy, on the other hand, has been left largely untouched. Opponents of the strategy typically either dismiss it outright as unintelligible or else criticize it for reasons that simply do not apply to all theories of relative identity. Of course, if the RI strategy is unintelligible, or if the theories of relative identity that escape the usual criticisms are untenable, then these are serious objections that must be reckoned with. But, so far as I can tell, neither of these latter two claims has been adequately supported in the literature. Thus, it is not surprising that RI trinitarians seem largely unmoved by the objections that have so far been raised against their position. In the present chapter, I hope to offer a more persuasive line of criticism. In particular, I will argue for the following two conclusions: (i) It is possible to pursue the RI strategy without endorsing a Geach-style theory of relative identity. But doing so without telling an appropriate supplemental story about the metaphysics underlying RI relations leaves one with an incomplete solution to the problem of the Trinity and also leaves one vulnerable to the charge of polytheism. (ii) Pursuing the RI strategy under the assumption that a Geach-style theory of relative identity is correct commits one to the view that the very existence of the divine Persons is a theory-dependent matter. The consequences mentioned in (i) and (ii) are not acceptable. Thus, the RI strategy is unsuccessful as a stand-alone solution to the problem of the Trinity.⁴ I will begin in Section 1 by describing the doctrine of relative identity in some detail. In Section 2, I will identify and describe two versions of the RI strategy— what I will call the pure RI strategy and the impure RI strategy. I will also discuss briefly what I take to be the standard criticisms of the RI strategy. My own objections to that strategy will then be presented in Section 3.

³ See, e.g. Bartel 1988; Brower 2004; Clark 1996; Feser 1997; Leftow 1999; and Merricks 2006. Some of what I would be inclined to add to these criticisms is expressed in Chapters 7 and 10 in this volume. ⁴ The solution that I favour is plausibly seen as a conjunction of what I will later call the ‘impure’ RI strategy with a supplemental story about the metaphysics of RI relations. Jeffrey Brower and I defend this solution in Chapter 10 in this volume.

       

187

1. Relative Identity Classically understood, identity is an absolute relation that obeys Leibniz’s Law.⁵ To say that identity is an absolute relation is to say (at least) that unqualified sentences of the form ‘x = y’ are well-formed and meaningful and that they are not to be analysed in terms of sentences of the form ‘x is the same F as y’. To say that it obeys Leibniz’s Law is to say that it is governed by the following principle: (LL) For all x and y, x is identical to y only if, for all predicates φ, x satisfies φ if and only if y satisfies φ; or, in symbols: 8x8y[x = y  (8φ)(φ(x)  φ(y))] As is well known, some philosophers reject the classical understanding of identity. Among those who do, some hold that identity is relative. Various different views have been advertised in the literature under the label ‘relative identity’. Peter Geach, the earliest and most well-known contemporary advocate of a theory of relative identity, endorses both of the following theses:⁶ (R1) Statements of the form ‘x = y’ are incomplete and therefore ill-formed. A proper identity statement has the form ‘x is the same F as y’. (R2) States of affairs of the following sort are possible: x is an F, y is an F, x is a G, y is a G, x is the same F as y, but x is not the same G as y. Together, R1 and R2 imply, among other things, the following: (C1)

Classical identity does not exist.

(C2)

LL is ill-formed.

(C3) In general, x’s being the same F as y does not guarantee that x is indiscernible from y. (C4)

‘x is the same F as y’ is not analysable as ‘x is an F, y is an F, and x = y’.

Let us refer to the conjunction of R1 and R2, together with the consequences C1–C4, as Geach’s theory of relative identity. Not everyone who claims to embrace relative identity endorses Geach’s theory, however. For example, Eddy Zemach accepts R1 but proposes a replacement for R2, and Leslie Stevenson rejects both R1 and R2, but declares himself a relative identity theorist on the grounds that, on his view, ‘x = y’ just means that, for some

⁵ Note that I am here using the label ‘Leibniz’s Law’ to refer simply to the principle of the indiscernibility of identicals. The converse principle—the principle of the identity of indiscernibles— is more controversial and plays no substantive role in the discussion that follows. ⁶ See Geach 1967, 1969, 1973, and 1980, sections 30, 34, and 110.

188

   :  

count noun F, x is the same F as y.⁷ And so not everyone who claims to be a relative identity theorist is committed to C1–C4. I will shortly offer reasons for doubting that some of the views just mentioned actually deserve to be called theories of relative identity. But for now the point is just that, in the literature on relative identity, there are alternatives to Geach’s theory that are advertised by their proponents as theories of relative identity. And some of these alternatives do deserve the label. Among the alternatives to Geach’s theory, probably the most important is the one defended by Nicholas Griffin.⁸ Griffin rejects R1, but accepts R2 along with two further theses. The first is a thesis about the relationship between relative and absolute identity statements: (R3) Sortal-relative identity statements are more fundamental than absolute identity statements. One implication of R3 is that absolute identity statements are to be analysed or defined in terms of more primitive sortal-relative identity statements, rather than the other way around. The second thesis is Griffin’s proposed substitutivity principle for relative identity predicates: (RLL)

x is the same F as y  (8φ 2 ΔF)(Fx & Fy & φx  φy)

Intuitively, RLL just says that, for each general noun F, being the same F implies indiscernibility with respect to the members of some class of predicates, ΔF. Note that RLL leaves open the possibility that there are predicates not in ΔF. Griffin introduces RLL because, as he notes, a theory of relative identity ought to licence inferences like (I) and (II) below without licencing inferences like (III): (I) (II) (III)

x is the same colour as y; x is red. Therefore, y is red. x is the same car as y; x is twelve feet long. Therefore, y is twelve feet long. x is the same colour as y; x is a car. Therefore, y is a car.

RLL is supposed to provide a way of getting what we want without contradicting R2; and, Griffin argues, adding it to a classical second-order logic generates a

⁷ Stevenson 1972; Zemach 1974. I take it that saying that ‘x = y’ means that x is the same F as y (for some count noun F), is not quite the same as saying that ‘x = y’ is incomplete or ill-formed. Likewise, I think that R1 is not implied by the claim that x = y if and only if there is some sortal F such that x is the same F as y. ⁸ Griffin 1977. See also Routley and Griffin 1979.

       

189

consistent and complete theory of relative identity. Moreover, the logic that results will be one within which classical identity is consistently definable.⁹ R3, on the other hand, is supposed to follow from Griffin’s view that individuation without sortals is impossible. He writes: It is hard to see how any sense can be made of the notion of an individual item without individuation, and it is hard to see how sense can be made of individuation without sortals which supply the principles which make individuation possible. In view of this, it seems to me that, while all types of identity statements are admissible, sortal-relative identity statements have the most fundamental role to play, for without them we cannot make sense of the notion of an individual item. Once we have individuated some items by means of a sortal and found, say, that the item named by ‘a’ and that named by ‘b’ are the same F, we can go on to ask if they share all their properties and are thus the same absolutely.¹⁰

It is, of course, tempting to ask here what our individuation practices could possibly have to do with the identity of the things that we individuate. Why should the fact that we individuate things by way of sortal concepts go any distance towards showing that nothing is distinct from anything else absolutely and independently of our sortal concepts? The answer should be obvious: It does not go any distance towards showing this unless we presuppose a decidedly antirealist metaphysic. And this is precisely the sort of metaphysic that Griffin thinks underlies relative identity theory. Michael Dummett brings this point out nicely. In discussing Geach’s theory of relative identity, Dummett raises the question of how quantifiers are to be interpreted. His worry is as follows: . . . there is a compelling feeling of incompatibility between the picture that we are accustomed to form of the classical interpretation of the quantifiers and the picture evoked by Geach’s doctrine on identity. . . . [On the classical treatment of quantifiers], the picture we have of what constitutes a domain of objects which can serve as the range of the individual variables is such that it is impossible to see how there could be any objection to supposing an absolute relation of identity to be defined on it: the elements of the domain are thought of as being, in Quine’s words, the same or different absolutely. . . . [But] it seems that Geach means us to picture that over which the variables range as an amorphous lump of reality, in itself not articulated into distinct objects. Such an articulation may be ⁹ One way is to define it as follows (where ‘x =F y’ abbreviates ‘x is the same F as y’): (=) x = y  (8F)[(Fx & Fy)  x = F y] Another way is simply to define it by way of LL. ¹⁰ Griffin 1977: 159.

190

   :  

accomplished in any one of many different ways: we slice up reality into distinct individual objects by selecting a particular criterion of identity.¹¹

The apparent tension between Geach’s theory of identity and the classical treatment of quantifiers arises because of Geach’s commitment to R1. But Griffin also stands against the classical treatment of quantifiers on the grounds that ‘it gets the semantics upside down’.¹² On the classical treatment, the individuation of objects in the domain is independent of our sorting procedures. But Griffin wants it the other way around. Thus, he endorses Dummett’s characterization of the metaphysic underlying Geach’s theory of relative identity as an accurate description of the metaphysic underlying his own. As I have already indicated, views other than Griffin’s and Geach’s sometimes go under the label ‘relative identity’. As I see it, however, no view deserves that label unless it is committed to either R1 or R3. Views according to which classical identity exists and is no less fundamental than other sameness relations are simply not views according to which identity is relative. Perhaps there are multiple sameness relations, and perhaps some of those relations are both sortal-relative and such that R2 is true of them. But so long as classical identity exists and is not to be analysed in terms of more primitive relative identity relations—there seems to be no reason whatsoever to think of other ‘sameness’ relations as identity relations.¹³ Thus, on views that reject both R1 and R3, there seems to be no reason for thinking that identity is non-absolute.

2. The RI Strategy The RI strategy for solving the problem of the Trinity comes in two varieties. The pure strategy endorses each of the following two claims: (RIA) Some doctrine of relative identity (that is, some doctrine that includes either R1 or R3) is true. (RIB) The words ‘is God’ and ‘is distinct from’ in trinitarian formulations express relativized identity and distinctness relations rather than absolute identity and distinctness. The impure strategy endorses RIB without endorsing RIA.

¹¹ Dummett 1973: 562–3. ¹² Griffin 1977: 158. ¹³ Aristotle’s ‘accidental sameness’, for example, appears to be a relation that fits this description. That is, it is plausibly taken as a ‘sameness’ relation; it is sortal-relative and obeys R2; but it is not an identity relation, and belief in accidental sameness does not preclude belief in absolute identity. For further discussion of accidental sameness, see Rea 1998 and Chapter 10 in this volume.

       

191

The ‘trinitarian formulations’ mentioned in RIB include statements like these: (T1)

Each Person of the Trinity is distinct from each of the others.

(T2)

Each Person of the Trinity is God.

According to both versions of the RI strategy, the relations expressed by ‘is God’ and ‘is distinct from’ in statements like T1 and T2 are relations like being the same God as and being a distinct Person from, respectively. As relativized identity relations, they are not to be analysed in terms of classical identity, and they do not guarantee the indiscernibility of their relata—and this regardless of whether classical identity happens to exist. In other words, predicates like ‘is the same God as’ and ‘is the same Person as’ are predicates of which R2 is true, and they are governed by RLL rather than by LL. According to RI trinitarians, T1 and T2 are to be understood as equivalent to T1a and T2a, respectively: (T1a) No Person of the Trinity is the same Person as any of the others. (That is, the Father is not the same Person as the Son, the Father is not the same Person as the Spirit, and the Son is not the same Person as the Spirit.) (T2a)

Each Person of the Trinity is the same God as each of the others.

Since ‘same God as’ and ‘same Person as’ do not pick out relations that obey LL, no contradiction can be derived from their conjunction with T3: (T3) There is exactly one God. Thus, the RI strategy apparently manages to respect the ‘oneness’ of God without giving up the ‘threeness’ of the Persons. As indicated earlier, there has been relatively little by way of explicit criticism of this strategy in the literature on the Trinity; and, of the few critical remarks that have been offered, none seems especially persuasive. Broadly speaking, the two standard objections are these: (i) that the view requires its proponents to reject the principle of the Indiscernibility of Identicals;¹⁴ and (ii) that, on the assumption that statements of the form ‘A is the same F as B’ are not just equivalent to statements of the form ‘A and B are both Fs, and A = B’, RI predicates are unintelligible.¹⁵ As the previous section makes clear, objection (i) is simply false. Objection (ii) seems to me to be on target, but, by itself and without further development, it is unlikely to move those who think that they can understand the RI solution (which, of course, will be every single proponent of that solution, and ¹⁴ See Bartel 1988.

¹⁵ See Cartwright 1987 and Merricks 2006.

192

   :  

perhaps some fence-sitters as well). More needs to be said if the RI strategy is to be effectively undermined. It is to that task that I now turn.

3. Objections I will start by explaining why I think that what I have called the impure RI strategy ought to be rejected. Recall that this strategy involves embracing RIB without embracing RIA—that is, without embracing a doctrine of relative identity. This is the sort of strategy that is pursued by the most well-known proponent of the RI solution to the problem of the Trinity, Peter van Inwagen.¹⁶ His view, therefore, will be my focus in the first part of the discussion that follows. Van Inwagen remains explicitly neutral on the question whether absolute identity exists, and he also refrains from committing himself to anything like R1 or R3. Instead, he simply starts with two undefined RI predicates—‘is the same person as’ and ‘is the same being as’—and assumes, in effect, that R2 is true of both.¹⁷ From there, he proceeds to construct translations of T1–T3 in a language devoid of singular referring terms and involving no RI predicates other than those defined in terms of the two primitives and various one-place predicates like ‘is divine’, ‘begets’, ‘is begotten’, and ‘proceeds’.¹⁸ In brief, the task of translation is carried out as follows: Let B stand for the same being as relation. Let Gx abbreviate ‘x is divine and 8y(y is divine  xBy)’. Gx, then, is the RI equivalent of ‘x is God’.

¹⁶ See van Inwagen 1988. ¹⁷ Note that van Inwagen’s ‘same person as’ relation is different from the relation that would be expressed by the words ‘same Person as’. I have been treating ‘Person’ as a sui generis sortal term that applies only to the Father, Son, and Holy Spirit. But van Inwagen’s terminology makes no distinction as regards personhood between the divine Persons and human persons. ¹⁸ The Father begets; the Son is begotten; the Spirit proceeds; and all three are divine. Singular referring expressions are banned for the sake of neutrality with respect to the existence of classical identity. According to van Inwagen: The philosopher who eschews classical, absolute identity must also eschew singular terms, for the idea of a singular term is—at least in currently orthodox semantical theory—inseparably bound to the classical semantical notion of reference or denotation; and this notion, in its turn, is inseparably bound to the idea of classical identity. It is a part of the orthodox semantical concept of reference that reference is a many-one relation. And it is a part of the idea of a many-one relation . . . that if x bears such a relation to y and bears it to z, then y and z are absolutely identical. (That’s what it says on the label.) (1988: 259) Note, however, that there is reason to doubt that singular referring expressions must be banned. If the quoted line of reasoning is correct, RI theorists should have trouble with counting generally. But, assuming they do not have general problems counting, why could not a relative identity theorist just say, e.g. that reference is a relation between many words and one man, or between many words and one horse, or between many words and one . . . etc.? If the worry is that relativists cannot make sense of the notion that there is one horse standing in the ‘is referred to by’ relation to various words, then how can we expect them to count at all? And, importantly, how can we expect them to offer an adequate translation of T3—the thesis that there is exactly one God? On the other hand, if identifying one horse as the referent of various names is not the problem, it is hard to see what would be the problem with singular reference. Notably, neither Geach nor Griffin seems especially concerned about the issue.

       

193

Let F, S, and H stand for predicates (‘begets’, ‘is begotten’, and ‘proceeds’) that have the following properties: (a) they are satisfied, if at all, by a divine Person, (b) if x and y satisfy one of them, then x is the same person as y, and (c) if x satisfies one and y satisfies another, then x is not the same person as y. We can then stipulate that Fx, Sx, and Hx are, respectively, the RI equivalents of ‘x is the Father’, ‘x is the Son’, and ‘x is the Holy Spirit’.¹⁹ Finally, let P stand for the same person as relation. Given all this, a sentence like ‘The Father is God’ may be translated simply as ‘∃x(Fx & Gx)’. Thus, theses T1 and T2 may be translated as follows: (RT1)

∃x∃y∃z(Fx & Sy & Hz & ¬xPy & ¬xPz & ¬yPz)

(RT2)

∃x∃y∃z(Fx & Gx & Sy & Gy & Hz & Gz)

On the assumption that counting Gods is a matter of counting divine beings rather than counting (say) divine persons, Gx will also be the RI equivalent of ‘x is the one and only God’. Thus, we have the following translation for T3: (RT3)

∃xGx

RT3 is entailed by RT2; and since the only ‘identity’ predicates involved in RT1– RT3 are predicates that do not obey Leibniz’s Law, there is no way to derive a contradiction from the conjunction of RT1 and RT2. Admittedly, given our stipulations, we can derive the following claim: ∃x∃y(xBy & xPx & ¬xPy).²⁰ Given those same stipulations, however, that claim is not contradictory. But what has been accomplished? Let us grant that the above translations are plausible. Has van Inwagen shown that the doctrine of the Trinity is coherent? Surprisingly, the answer is no. To be sure, he has shown that, on one way of understanding them, no contradiction can be derived from T1–T3 alone (or from ¹⁹ Cf. van Inwagen 1988: 265–7. ²⁰ Here is the derivation: From RT2 and the definition of Gx we get RT4: (RT4) ∃x∃y∃z(Fx & x is divine & 8m(m is divine  xBm) & Sy & y is divine & 8m(m is divine  yBm) & Hz & z is divine & 8m(m is divine  zBm)). RT4 implies RT5: (RT5) ∃x∃y∃z(Fx & Sy & Hz & xBy & xBz & yBz). Given our stipulations about F, S, and H, RT5 implies RT6: (RT6) ∃x∃y∃z(Fx & Sy & Hz & xBy & xBz & yBz & ¬xPy & ¬xPz & ¬yPz). And, given those same stipulations, RT6 implies RT7: (RT7) ∃x∃y∃z(xPx & yPy & zPz & xBy & xBz & yBz & ¬xPy & ¬xPz & ¬yPz). Simplification of RT7, in turn, yields the desired conclusion, RT8: (RT8) ∃x∃y(xBy & xPx & ¬xPy).

194

   :  

suitably similar conjunctions of trinitarian claims). But by remaining neutral on the question whether absolute identity exists, he leaves open the possibility that Father, Son, and Holy Spirit are absolutely distinct, and if they are absolutely distinct, it is hard to see what it could possibly mean to say that they are the same being, as RT2 implies. Thus, in order for his argument to be convincing, it appears that van Inwagen must rule out the possibility that Father, Son, and Holy Spirit are absolutely distinct. But he cannot do this while remaining neutral on the question whether absolute identity exists. Let me put this point another way. If absolute identity exists, the following is a rather compelling principle: (8x8y)(x ≠ y  ¬xBy) ‘Being’ is plausibly the most general sortal, on a par with sortals like ‘entity’, ‘thing’, and ‘object’. Thus, ‘x is (absolutely) distinct from y’ seems to be synonymous with ‘x is not the same being (thing, entity, object) as y’. If this is right, then P is analytic. But the conjunction of P with RT1 and RT2 is incoherent.²¹ Thus, van Inwagen’s arguments show that the doctrine of the Trinity is coherent only on the assumption that P is not true.²² A trinitarian who accepts P will not escape the charge of incoherence by accepting RT1 and RT2 as translations of T1 and T2. Thus, given that P is highly intuitive, in order to show that the doctrine of the Trinity is coherent van Inwagen must give us some reason for thinking that P is not true. One who accepts R1 can easily provide such a reason: P is not true, she will say, because it is ill-formed. It is ill-formed because it includes the formula ‘x≠ y.’ But van Inwagen does not want to commit to R1, and he has offered us no other reason for believing that P is false. There are at least two morals to draw from this conclusion. One is that pursuing the impure R1 strategy offers at best an incomplete solution to the problem of the Trinity. At the very least, a story will have to be told that explains how R2 could be true of the sameness relations invoked in trinitarian formulations. Moreover, the story will have to be non-heretical. (Thus, for example, telling a story according to which Father, Son, and Holy Spirit can be the same God but different Persons in

²¹ Here is the proof. As stated above, from RT1 & RT2 we can derive RT8: (RT8) ∃x∃y(xBy & xPx & ¬xPy). But, allowing that there is such a thing as absolute distinctness, the conjunction of xPx & ¬xPy implies x≠ y. Thus, RT8 implies RT9: (RT9) ∃x∃y(xBy & x≠ y). But RT9, together with P, implies the contradictory RT10: (RT10) ∃x∃y(xBy & ¬xBy). ²² Van Inwagen acknowledges this point (1988: 262, 265), but he does not seem to see it as a source of any serious objection to his project.

       

195

just the same way that two cars can be the same colour but different cars clearly will not do the job.) The other moral is that even embracing a Griffin-like theory of relative identity—that is, a theory that does not rule out classical identity—will leave one vulnerable to polytheism, or worse. Monotheism requires that there be exactly one being who is God. But, as we have just seen, if classical identity makes its way into one’s logic, T1 and T2 together seem to imply that there are three distinct beings who ‘are’ God. And now the familiar problem is back: incoherence looms, unless we can tell a story (other than the ‘absolute identity’ story) about what it is to ‘be’ God that is both orthodox and plausible. Elsewhere, Jeffrey Brower and I offer such a story, and we do so without embracing any doctrine of relative identity.²³ But my point here is just that some such story is needed and has not so far been offered by anyone who wishes to pursue the impure RI strategy. I conclude, then, that absent some supplementary story explaining the metaphysics of RI relations, the impure RI strategy is unacceptable. But what about the pure RI strategy? As I see it, the consequences of that strategy are catastrophic. The reason is that, as noted above, the doctrine of relative identity seems to presuppose an anti-realist metaphysic. I have already quoted Dummett’s reasons for thinking that relative identity goes hand in hand with anti-realism. But we can bring the presupposition to light in another way by taking a brief look at Geach’s main argument for the relativity of identity.²⁴ Geach thinks that semantic paradoxes (for example, Richard’s and Grelling’s) prevent us from reading LL as saying that x = y if and only if whatever is true of x is true of y.²⁵ Thus, he says, we must read it instead as saying (roughly) that x = y if and only if x and y are indiscernible with respect to all of the predicates that form the descriptive resources of our theory. But if this is right, then identity is best construed as theory-relative. The reason is that whereas one theory might have the descriptive resources to distinguish x from y, another theory might not. By the lights of the first theory, then, x = y is false, whereas, by the lights of the second theory, x = y is true. But that is incoherent if x = y is understood as expressing absolute identity. Better, then, to treat it as expressing sortal-relative identity. Treating it that way, we can say that x and y are the same F but not the same ²³ See Chapter 10, this volume. ²⁴ What follows is a summary of the argument on behalf of R1 that is presented in Geach 1967. But see also Geach 1973. ²⁵ To see why one might think this, consider Grelling’s Paradox: Let the word ‘heterological’ mean ‘is not true of itself ’. Thus, ‘long’ is heterological, since it is not a long word; ‘unspeakable’ is heterological, since it can be spoken; etc. Now, by definition, ‘heterological’ is heterological only if it is not true of itself; but if it is not true of itself, then it is not heterological. So ‘heterological’ is heterological if and only if it is not heterological, which is contradictory. One way to solve this paradox (and others) is to hold that truth is relativized to a language, so that, e.g. we can speak in a language other than L of what is true-in-L of a thing x, but we cannot speak in L of what is true simpliciter of x. (Cf. Tarski 1983a, 1993b.) But if we adopt this solution, or one relevantly similar to it, then we must reject constructions like ‘whatever is true of x . . . ’.

196

   :  

G (where, presumably, G is a predicate in our first theory but not our second), and the incoherence dissolves. To see the point more clearly, consider the following example. Imagine two theories: one, T₁, which includes the sortal ‘lump of clay’ but no artefact sortals (like ‘statue’ or ‘bowl’), and another, T₂, which includes the sortals ‘lump of clay’, ‘statue’, and ‘bowl’, and, furthermore, treats statues and bowls that have been made from the same lump of clay as distinct items. Now suppose a T₁-theorist and a T₂-theorist watch a sculptor take a lump of clay and make first (what the T₂-theorist would call) a statue, then a bowl. By the lights of the T₁-theorist, the sculptor does not manage to generate or destroy anything. What the T₂-theorist would call ‘the statue’ and ‘the bowl’ are identical. By the lights of the T₂-theorist, however, statue and bowl are distinct. But, obviously, both cannot be right. Thus, we have a problem. One way out is to say that identity is theory-relative: the bowl and statue are the same lump of clay; they are not the same bowl or statue; and there is simply no fact about whether they are absolutely identical or distinct. But if we do say this (taking very seriously the claim that there is no theoryindependent fact about what there is or about how many things there are in the various regions occupied by what the T₁-theorist calls ‘the lump of clay’), then we commit ourselves to the view that the very existence of things like statues, bowls, and lumps of clay depends upon the theories that recognize them. This is antirealism. Many philosophers are attracted to anti-realism, but accepting it as part of a solution to the problem of the Trinity is disastrous. For clearly orthodoxy will not permit us to say that the very existence of Father, Son, and Holy Spirit is a theorydependent matter. Nor will it permit us to say that the distinctness of the divine Persons is somehow relative to our ways of thinking or theorizing. The latter appears to be a form of modalism.²⁶ And yet it is hard to see how it could be otherwise if Geach’s theory of relative identity is true. For what else could it possibly mean to say that there is simply no fact about whether Father, Son, and Holy Spirit are the same thing as one another, the same thing as God, or, indeed, the same thing as Baal? Perhaps there is a way for proponents of R1 to dodge commitment to antirealism. For perhaps there is some alternative metaphysic that might sensibly be thought to underlie a Geach-style theory of relative identity. (Though the quotations from Dummett and Griffin, and the argument on behalf of R1 supplied by Geach, suggest that R1 commits its adherents to anti-realism, they do not, after all, prove that conclusion.) But it is hard to see what that metaphysic might be.

²⁶ Modalism is the view that Father, Son, and Holy Spirit are not really distinct from one another. According to modalism, each Person is just God in a different guise, or playing a different role—much like Superman and Clark Kent are just the Kryptonian Kal-El in different guises, or playing different roles.

       

197

Thus, for the relative identity theorist who wants to avoid commitment to antirealism, the charge of incoherence becomes an objection seriously to be reckoned with. What motivates commitment to R1, if not reasoning of the sort described above (reasoning, again, that leads straight to anti-realism)? What can the would-be realist RI theorist say in response to Dummett’s worries about quantification in the context of an RI logic? I said earlier that the charge of unintelligibility is unpersuasive mainly because proponents of relative identity have taken pains to try to make their view intelligible, and the accounts they have offered along these lines have not themselves been shown to be unintelligible. But if one rejects what proponents of relative identity have to say about the motivation and metaphysics underlying their view, the charge of unintelligibility returns with a vengeance, and I cannot see how that charge can be rebutted without embracing anti-realism. Moreover, even if we concede that R1 can be squared with a realist metaphysic, the fact remains that it is extremely implausible, and what few arguments have been marshalled on its behalf have been strongly (and rightly, in my opinion) criticized in the literature.²⁷ Thus, even apart from its anti-realist consequences, I think that there are ample grounds for rejecting R1.

4. Conclusion I have argued, in effect, that RI theorists face a dilemma. If they pursue the impure RI strategy, or even if they pursue a pure strategy that falls short of endorsing R1, they leave open the possibility that absolute identity exists. In leaving open that possibility, and without telling an appropriate supplemental story about the metaphysics underling RI relations, they find themselves with an incomplete solution to the problem of the Trinity and they leave themselves vulnerable to polytheism. If they pursue the pure RI strategy, then they are committed to thinking that the existence and distinctness of the divine Persons is somehow a theory-dependent matter—a view that implies modalism, or worse. I have acknowledged that this latter commitment can be avoided by giving up standard views about the motivation and metaphysics underlying relative identity theory. But in giving up those standard views, the relative identity theorist incurs the burden of supplying alternative motivation and an alternative explanation of the metaphysic underlying her theory of identity. So far, this has not been done; and it seems to me that it cannot be done apart from a commitment to anti-realism. In

²⁷ See especially Dummett 1973, ch. 16; Dummett 1981, ch. 11; Dummett 1991; Feldman 1969a, 1969b; Griffin 1977, ch. 8; Nelson 1970; Odegard 1972; Perry 1972; Quine 1964. For additional, general criticisms of the doctrine of relative identity, see also Hawthorne 2003.

198

   :  

light of these considerations, I conclude that the RI solution to the problem of the Trinity, taken as a stand-alone solution to that problem, is unsuccessful.²⁸

References Anscombe, G. E. M. and P. T. Geach. 1961. Three Philosophers. Ithaca, NY: Cornell University Press. Bartel, Timothy. 1988. ‘The Plight of the Relative Trinitarian’. Religious Studies 24: 129–55. Bartel, Timothy. 1993. ‘Could There Be More than One Almighty?’ Religious Studies 29: 465–95. Bartel, Timothy. 1994. ‘Could There Be More than One Lord?’ Faith and Philosophy 11: 357–78. Brower, Jeffrey. 2004. ‘The Probem with Social Trinitarianism: A Repy to Wierenga’. Faith and Philosophy 21: 295–303. Brown, David. 1985. The Divine Trinity. La Salle, IL: Open Court Publishing Co. Brown, David. 1989. ‘Trinitarian Personhood and Individuality’. In Trinity, Incarnation, and Atonement, edited by Ronald J. Feenstra and Cornelius Plantinga Jr., 48–78. Notre Dame, IN: University of Notre Dame Press. Cain, James. 1989. ‘The Doctrine of the Trinity and the Logic of Relative Identity’. Religious Studies 25: 141–52. Cartwright, Richard. 1987. ‘On the Logical Problem of the Trinity’. In Philosophical Essays, edited by Richard Cartwright, 187–200. Cambridge, MA: MIT Press. Clark, Kelly James. 1996. ‘Trinity or Tritheism?’ Religious Studies 32: 463–76. Davis, Stephen. 1999. ‘A Somewhat Playful Proof of the Social Trinity in Five Easy Steps’. Philosophia Christi 1: 103–5. Dummett, Michael. 1973. Frege: Philosophy of Language. London: Harper & Rowe. Dummett, Michael. 1981. The Interpretation of Frege’s Philosophy. Cambridge, MA: Harvard University Press. Dummett, Michael. 1991. ‘Does Quantification Involve Identity?’ In Peter Geach: Philosophical Encounters, edited by Harry Lewis, 161–84. Dordrecht: Kluwer. Feldman, Fred. 1969a. ‘A Rejoinder’. Review of Metaphysics 22: 560–1. Feldman, Fred. 1969b. ‘Geach and Relative Identity’. Review of Metaphysics 22: 547–55. Feser, Edward. 1997. ‘Swinburne’s Tritheism’. International Journal for Philosophy of Religion 42: 175–84.

²⁸ This chapter has benefitted greatly from the advice of J. C. Beall, Michael Bergmann, Tom Crisp, John Hawthorne, Brian Leftow, Trenton Merricks, and especially Jeff Brower.

       

199

Forrest, Peter. 1998. ‘Divine Fission: A New Way of Moderating Social Trinitarianism’. Religious Studies 34: 281–97. Geach, Peter. 1967. ‘Identity’. Review of Metaphysics 21: 3–12. Geach, Peter. 1969. ‘Identity—A Reply’. Review of Metaphysics 22: 556–9. Geach, Peter. 1973. ‘Ontological Relativity and Relative Identity’. In Logic and Ontology, edited by Milton K. Munitz, 287–302. New York: New York University Press. Geach, Peter. 1980. Reference and Generality. 3rd edn. Ithaca, NY: Cornell University Press. Griffin, Nicholas. 1977. Relative Identity. Oxford: Clarendon Press. Hawthorne, John. 2003. ‘Identity’. In Oxford Handbook of Metaphysics, edited by Michael Loux and Dean Zimmerman, 99–130. Oxford: Oxford University Press. Layman, C. Stephen. 1988. ‘Tritheism and the Trinity’. Faith and Philosophy 5: 291–8. Leftow, Brian. 1999. ‘Anti Social Trinitarianism’. In The Trinity: An Interdisciplinary Symposium, edited by Stephen Davis, Daniel Kendall SJ, and Gerald O’Collins SJ, 203–49. Oxford: Oxford University Press. Martinich, A. P. 1978. ‘Identity and Trinity’. Journal of Religion 58: 169–81. Martinich, A. P. 1979. ‘God, Emperor, and Relative Identity’. Franciscan Studies 39: 180–91. Merricks, Trenton. 2006. ‘Split Brains and the Godhead’. In Knowledge and Reality: Essays in Honor of Alvin Plantinga, edited by Thomas Crisp, Matthew Davidson, and David Vander Laan, 299–326. Dordrecht: Springer. Nelson, Jack. 1970. ‘Relative Identity’. Noûs 4: 241–60. Odegard, Douglas. 1972. ‘Identity Through Time’. American Philosophical Quarterly 9: 29–38. Perry, John. 1972. ‘The Same F’. Philosophical Review 79: 181–200. Plantinga, Cornelius, Jr. 1986. ‘Gregory of Nyssa and the Social Analogy of the Trinity’. Thomist 50: 325–52. Plantinga, Cornelius, Jr. 1988. ‘The Threeness/Oneness Problem of the Trinity’. Calvin Theological Journal 23: 37–53. Plantinga, Cornelius, Jr. 1989. ‘Social Trinity and Tritheism’. In Trinity, Incarnation, and Atonement, edited by Ronald J. Feenstra and Cornelius Plantinga Jr, 21–47. Notre Dame, IN: University of Notre Dame Press. Quine, W. V. O. 1964. ‘Review of P. T. Geach, Reference and Generality’. Philosophical Review 73: 100–4. Rea, Michael. 1998. ‘Sameness without Identity: An Aristotelian Solution to the Problem of Material Constitution’. Ratio 11 (3): 316–28. Routley, Richard and Nicholas Griffin. 1979. ‘Towards a Logic of Relative Identity’. Logique et Analyse 22: 65–83.

200

   :  

Stevenson, Leslie. 1972. ‘Relative Identity and Leibniz’s Law’. Philosophical Quarterly 22: 155–8. Swinburne, Richard. 1994. The Christian God. Oxford: Clarendon Press. Tarski, Alfred. 1983a. ‘The Concept of Truth in Formalized Languages’. In Logic, Semantics, Metamathematics: Papers from 1923–1938, edited by Alfred Tarski, translated by J. H. Woodger, 152–278. Indianapolis, IN: Hackett. Tarski, Alfred. 1983b. ‘The Establishment of Scientific Semantics’. In Logic, Semantics, Metamathematics: Papers from 1923–1938, edited by Alfred Tarski, translated by J. H. Woodger, 401–8. Indianapolis, IN: Hackett. van Inwagen, Peter. 1988. ‘And Yet They Are Not Three Gods but One God’. In Philosophy and the Christian Faith, edited by Thomas V. Morris, 241–78. Notre Dame, IN: University of Notre Dame Press. Williams, C. J. F. 2002. ‘Neither Confounding the Persons nor Dividing the Substance’. In Reason and the Christian Religion: Essays in Honour of Richard Swinburne, edited by Alan G. Padgett, 227–43. Oxford: Clarendon Press. Wolfson, H. A. 1964. The Philosophy of the Church Fathers, Vol. 1: Faith, Trinity, Incarnation. Cambridge, MA: Harvard University Press. Zemach, Eddy. 1974. ‘In Defense of Relative Identity’. Philosophical Studies 26: 207–18.

Material Constitution and the Trinity [with Jeffrey Brower] Michael C. Rea

DOI:10.1093/oso/9780198866800.003.0011

Abstract and Keywords The Christian doctrine of the Trinity poses a serious philosophical problem. On the one hand, it seems to imply that there is exactly one divine being. On the other hand, it seems to imply that there are three divine beings. There is another well-known philosophical problem that presents us with a similar sort of tension: the problem of material constitution. This chapter argues that a relatively neglected solution to the problem of material constitution—an appeal to the Aristotelian doctrine of numerical sameness without identity—can be developed into a novel solution to the problem of the Trinity. Keywords: Trinity, material constitution, constitution trinitarianism, relative identity, social trinitarianism, Nicene Creed

As is well known, the Christian doctrine of the Trinity poses a serious philosophical problem. On the one hand, it affirms that there are three distinct Persons—Father, Son, and Holy Spirit—each of whom is God. On the other hand, it says that there is one and only one God. The doctrine therefore pulls us in two directions at once—in the direction of saying that there is exactly one divine being and in the direction of saying that there is more than one. There is another well-known philosophical problem that presents us with the same sort of tension: the problem of material constitution. This problem arises whenever it appears that an object a and an object b share all of the same parts and yet have different modal properties.1 To take just one of the many well-worn examples in the literature: Consider a bronze statue of the Greek goddess, Athena, and the lump of bronze that constitutes it. On the one hand, it would appear that we must recognize at least two material objects in the region occupied by the statue. For presumably the statue cannot survive the process of being melted down and recast whereas the lump of bronze can. On the other hand, our ordinary counting practices lead us to recognize only one material object in the region. As Harold Noonan aptly puts it, counting two material objects in such a region seems to ‘manifest a bad case of double vision’.2 Here, then, as with the doctrine of the Trinity, we are pulled in two directions at once. Admittedly, the analogy between the two problems is far from perfect. But we mention it because, as we shall argue below, it turns out that a relatively neglected response to the problem of material constitution can be developed into a novel solution to the problem of the Trinity. In our view, this new solution is more promising than the other solutions available in the contemporary literature. It is independently plausible, it is motivated by considerations independent of the problem of the Trinity, and it is immune to objections that afflict the other solutions. The guiding intuition is the Aristotelian idea that it is possible (p.202) for an object a and an object b to be ‘one in number’—that is, numerically the same—without being strictly identical. We will begin in Section 1 by offering a precise statement of the problem of the Trinity. In Section 2, we will flesh out the Aristotelian notion of ‘numerical sameness without identity’, explain how it solves the problem of material constitution, and defend it against what we take to be the most obvious and important objections to it. Also in that section we will distinguish numerical sameness without identity from two superficially similar relations. Finally, in Sections 3 and 4, we will show how the Aristotelian solution to the problem of material constitution can be developed into a solution to the problem of the Trinity, and we will highlight some of the more interesting consequences of the solution we describe. 3

1. The Problem of the Trinity The central claim of the doctrine of the Trinity is that God exists in three Persons—Father, Son, and Holy Spirit. This claim is not problematic because of any superficial incoherence or inconsistency with wellentrenched intuitions. Rather, it is problematic because of a tension that results from constraints imposed on its interpretation by other aspects of orthodox Christian theology. These constraints are neatly summarized in the following passage from the so-called Athanasian Creed: We worship one God in Trinity and Trinity in unity, neither confusing the Persons, nor dividing the substance. For there is one person for the Father, another for the Son, and yet another for the Holy Spirit. But the divinity of the Father, Son, and Holy Spirit is one…The Father is eternal, the Son is eternal, and the Holy Spirit is eternal; and yet they are not three eternals, but there is one eternal. Likewise, the Father is almighty, the Son is almighty, and the Holy Spirit (p.203) is almighty; and yet there are not three almighties, but there is one almighty. Thus, the Father is God, the Son is God, and the Holy Spirit is God; and yet there are not three Gods, but there is one God.4 The passage quoted here is widely—and rightly—taken to offer a paradigm statement of the orthodox understanding of the doctrine of the Trinity. Moreover, it tells us that the doctrine of the Trinity must be understood in such a way as to be compatible with each of the following theses: (T1) Each Person of the Trinity is distinct from each of the others. (T2) Each Person of the Trinity is God. (T3) There is exactly one God. Each of these theses is affirmed by the Creed in order to rule out a specific heresy. T1 is intended to rule out modalism, the view that Father, Son, and Holy Spirit are not really distinct from one another. According to modalism, each Person is just God in a different guise, or playing a different role—much like Superman and Clark Kent are just the Kryptonian Kal-El in different guises, or playing different roles. T2 is intended to rule out subordinationism, the view that not all of the Persons are divine, or that the divinity of one or more of the Persons is somehow unequal with, or subordinate to, that of the others. T3 is intended to rule out polytheism, the view that there is more than one God. The problem, however, is that the conjunction of T1–T3 is apparently incoherent. For on their most natural interpretation, they imply that three distinct beings are each identical with one being (since each of the Persons is God, and yet there is only one God). In the contemporary literature, there are two main strategies for solving the problem: the relative identity strategy, and the social trinitarian strategy. Both of these strategies solve the problem at least in part by denying that the words ‘is God’ in trinitarian formulations mean ‘is absolutely identical with God’. Thus both are well-poised to avoid the heresy of modalism.5 Furthermore, both affirm T2 (or some suitable variant thereof); thus, subordinationism is not a worry either. The real question is whether either manages to avoid polytheism without incurring other problems in the process. In our view, the answer is no—at least not as these solutions have been developed in the literature so far. Social trinitarianism we reject outright. The relative identity solution we reject as a stand-alone solution to the problem of the Trinity. (That is, we think that it is successful only if it is supplemented by a story about the metaphysics of relative identity relations. More on this at the end of Section 2 below). Since we have already explained elsewhere (p.204) why we find these solutions unsatisfying, we will not repeat the details of our objections here.6 Instead, we’ll simply summarize by saying that we reject both the social trinitarian

solution and existing versions of the relative identity solution because they fail to provide an account of the Trinity that satisfies the following five desiderata: (D1) It is clearly consistent with the view that Father, Son, and Holy Spirit are divine individuals, and that there is exactly one divine individual. (D2) It does not conflict with a natural reading of either the Bible or the ecumenical creeds. (D3) It is consistent with the view that God is an individual rather than a society, and that the Persons are not parts of God.7 (D4) It is consistent with the view that classical identity exists and is not to be analysed in terms of more fundamental sortal-relativized sameness relations like being the same person as. (D5) It carries no anti-realist commitments in metaphysics. The social trinitarian solution violates D1–D3. Extant versions of the relative identity solution violate D1, D4, or D5. As will emerge shortly, our solution, which may fruitfully be thought of as an appropriately supplemented version of (p.205) the relative identity solution, succeeds precisely where these others fail—namely, in satisfying all five desiderata.

2. Sameness without Identity and the Problem of Material Constitution The point of departure for our solution is Aristotle’s notion of ‘accidental sameness’. Elsewhere, we have proposed (for the sake of argument, at any rate) that the phenomenon of material constitution be understood in terms of accidental sameness.8 What we propose here is that the unity of the divine Persons also be understood in terms of this relation (or more accurately, in terms of the genus of which it is a species—namely, numerical sameness without identity9). In this section, therefore, we review the way in which appeal to accidental sameness provides a solution to the problem of material constitution and address what we take to be the most natural objections to it. 2.1 Accidental Sameness Characterized According to Aristotle, familiar particulars (trees, cats, human beings, etc.) are hylomorphic compounds— things that exist because and just so long as some matter instantiates a certain kind of form. Forms, for Aristotle, are complex organizational properties, and properties are immanent universals (or, as some have it, tropes). The matter of a thing is not itself an individual thing; rather, it is that which combines with a form to make an individual thing.10 Thus, for example, a human being exists just in case some matter instantiates the complex organizational property humanity. Each human being depends for its continued existence on the continued instantiation of humanity by some matter; and each (p.206) human being is appropriately viewed as a composite whose parts (at one level of decomposition) are just its matter and (its) humanity.11 On Aristotle’s view, living organisms are the paradigmatic examples of material objects. But Aristotle also acknowledges the existence of other hylomorphic compounds. Thus, books, caskets, beds, thresholds, hands, hearts, and various other non-organisms populate his ontology, and (like an organism) each one exists because and only so long as some matter instantiates a particular complex organizational property.12 Indeed, Aristotle even countenances what Gareth Matthews calls ‘kooky’ objects—objects like ‘seated-Socrates’, a thing that comes into existence when Socrates sits down and which passes away when Socrates ceases to be seated.13 Seated-Socrates is an ‘accidental unity’—a unified thing that exists only by virtue of the instantiation of an accidental (non-essential) property (like seatedness) by a substance (like

Socrates). The substance plays the role of matter in this sort of hylomorphic compound (though, of course, unlike matter properly conceived, the substance is a pre-existing individual thing), and the accidental property plays the role of form. Accidental sameness, according to Aristotle, is just the relation that obtains between an accidental unity and its parent substance.14 One might baulk at this point on the grounds that Aristotle’s accidental unities are just a bit too kooky for serious ontology. We see that Socrates has seated himself; but why believe that in doing so he has brought into existence a new object—seated-Socrates? Indeed, one might think it’s clear that we shouldn’t believe this. For there is nothing special about seatedness, and so, if we acknowledge the existence of seatedSocrates, we must also acknowledge the existence of a myriad other kooky objects: pale-Socrates, baldSocrates, barefoot-Socrates, and so on. But surely there are not millions of objects completely overlapping Socrates. Fair enough; and nothing here depends on our believing in seated-Socrates or his cohorts. But note that, regardless of what we think of seated-Socrates, we (fans of common sense) believe in many things relevantly like seated-Socrates. That is, we believe in things that are very plausibly characterized as hylomorphic compounds whose matter is a familiar material object and whose form is an accidental property. For example, we believe in fists and hands, bronze statues and lumps of bronze, cats and heaps of cat tissue, and so on. Why we should believe all this but not that sitting down is a way of replacing one kind of object (a standing-man) with another (a seated-man) is an interesting and surprisingly difficult question. But never mind that for now. The important point here is that, regardless of whether we go along with Aristotle in believing in what he calls accidental unities, (p.207) the fact is that many of us will be inclined to believe in things relevantly like accidental unities along with other things that are relevantly like the parent substances of accidental unities. This last point is important because the things we have listed as being relevantly like accidental unities and their parent substances are precisely the sorts of things belief in which gives rise to the problem of material constitution. Hence the relevance of Aristotle’s doctrine of accidental sameness. Aristotle agrees with common sense in thinking that there is only one material object that fills the region occupied by Socrates when he is seated. Thus, he says that the relation between accidental unities and their parent substances is a variety of numerical sameness. Socrates and seated-Socrates are, as he would put it, one in number but not one in being.15 They are distinct, but they are to be counted as one material object.16 But once one is committed to believing in such a relation, one has a solution to the problem of material constitution ready to hand. Recall that the problem arises whenever it appears that an object a and an object b share all of the same parts and yet have different modal properties. In such cases we are pushed in the direction of denying that the relevant a and b are identical and yet we also want to avoid saying that they are two material objects occupying the same place at the same time. Belief in the relation of accidental sameness solves this problem because it allows us to deny that the relevant a and b are identical without thereby committing us to the claim that a and b are two material objects. Thus, one can continue to believe that (e.g.) there are bronze statues and lumps of bronze, that every region occupied by a bronze statue is occupied by a lump of bronze, that no bronze statue is identical to a lump of bronze (after all, statues and lumps have different persistence conditions), but also that there are never two material objects occupying precisely the same place at the same time. One can believe all this because one can say that bronze statues and their constitutive lumps stand in the relation of accidental sameness: they are one in number but not one in being.

2.2 Accidental Sameness Defended But should we believe in accidental sameness? The fact of the matter is that this sort of solution to the problem of material constitution is probably the single most neglected solution to that problem in the contemporary literature; and it is not hard to see why. Initially it is hard to swallow the idea that there is a variety of numerical sameness that falls short of identity. But, in our view, the most obvious and serious objections are failures, and the bare fact that the doctrine of accidental (p.208) sameness is counterintuitive is mitigated by the fact that every solution to the problem of material constitution is counterintuitive (a fact which largely explains the problem’s lasting philosophical interest). In the remainder of this section, we will address what we take to be the four most serious objections against the doctrine of accidental sameness. We will also explain how the relation of accidental sameness differs from two other relations to which it bears some superficial resemblance. In doing all this, we hope to shed further light on the metaphysics of material objects that attends belief in accidental sameness. First objection: Most contemporary philosophers think that, for any material objects a and b, a and b are to be counted as one if and only if a and b are identical. Indeed, it is fairly standard to define number in terms of identity, as follows: (1F) there is exactly one F =df ∃x(Fx & ∀y(Fy ≡ y = x)) (2F) there are exactly two Fs =df ∃x∃y(Fx & Fy & x ≠ y & ∀z(Fz ≡ y = z ν x = z)) etc. But if that is right, then it is hard to see how there could be a relation that does not obey Leibniz’s Law but is nevertheless such that objects standing in that relation are to be counted as one. Obviously enough, a believer in accidental sameness must reject standard definitions like 1F and 2F. But this does not seem to us to be an especially radical move. As is often pointed out, common sense does not always count by identity.17 If you sell a piano, you won’t charge for the piano and for the lump of wood, ivory, and metal that constitutes it. As a fan of common sense, you will probably believe that there are pianos and lumps, and that the persistence conditions of pianos differ from the persistence conditions of lumps. Still, for sales purposes, and so for common sense counting purposes, pianos and their constitutive lumps are counted as one material object. One might say that common sense is wrong to count this way. But why go along with that? Even if we grant that 1F and its relatives are strongly intuitive, we must still reckon with the fact that we have strong intuitions that support the following: (MC) In the region occupied by a bronze statue, there is a statue and there is a lump of bronze; the lump is not identical with the statue (the statue but not the lump would be destroyed if the lump were melted down and recast in the shape of a disc); but only one material object fills that region. (p.209) If we did not have intuitions that support MC, there would be no problem of material constitution. But if MC is true, then 1F and its relatives are false, and there seems to be no compelling reason to prefer the latter over the former. Of course, if rejecting 1F and its relatives were to leave us without any way of defining number, then our move would be radical, and there would be compelling reason to give up MC. But the fact is, rejecting 1F and its relatives does not leave us in any such situation. Indeed, belief in accidental sameness doesn’t even preclude us altogether from counting by identity. At worst, it simply requires us to acknowledge a distinction between sortals that permit counting by identity and sortals that do not. For example, according to the believer in accidental sameness, we do not count material objects by identity. Rather, we

count them by numerical sameness (the more general relation of which both accidental sameness and identity are species). Thus: (1M) there is exactly one material object =df∃x(x is a material object & ∀y(y is a material object ≡ y is numerically the same as x)) (2M) there are exactly two material objects =df∃x∃y(x is a material object & y is a material object and x is not numerically the same as y and ∀z(z is a material object ≡ z is numerically the same as x or z is numerically the same as y)) etc. Perhaps the same is true for other familiar sortals. For example: Suppose a lump of bronze that constitutes a bronze statue is nominally, but not essentially, a statue.18 Then the lump and the statue are distinct, and both are statues. But, intuitively, the region occupied by the lump/statue is occupied by only one statue. Thus, given the initial supposition, we should not count statues by identity either. Nevertheless, we can still grant that there are some sortals that do allow us to count by identity. Likely candidates are technical philosophical sortals like ‘hylomorphic compound’, or maximally general sortals, like ‘thing’ or ‘being’. For such sortals, number terms can be defined in the style of 1F and its relatives. Admittedly, the business of defining number is a bit more complicated for those who believe in accidental sameness (we must recognize at least two different styles of defining number corresponding to two different kinds of sortal terms). The important point, however, is that it is not impossible. In saying what we have about the categories of hylomorphic compound, thing, and being, we grant that proponents of our Aristotelian solution to the problem of material constitution are committed to a kind of co-locationism. Although cases of material constitution will never, on the view we are proposing, present us with (p.210) two material objects in the same place at the same time, they will present us with (at least) two hylomorphic compounds or things in the same place at the same time. But we deny that this commitment is problematic. By our lights, it is a conceptual truth that material objects cannot be colocated; but it is not a conceptual truth that hylomorphic compounds (e.g. a statue and a lump, a fist and a hand, etc.) or things (e.g. a material object and an event) cannot be co-located. We take it as an advantage of the Aristotelian solution that it respects these prima facie truths. Second objection: To say that hylomorphic compounds, or mere things, can be co-located but material objects cannot smacks of pretence. For while it preserves the letter, it does not preserve the spirit of the intuition that material objects cannot be co-located. If counting two material objects in the same place at the same time ‘reeks of double counting’,19 then the same reek must attend the counting of two hylomorphic compounds or two things in the same place at the same time. At best, therefore, the Aristotelian solution is only verbally distinct from the co-locationist solution. For co-locationists and fans of accidental sameness will still have the same metaphysical story to tell about statues and their constitutive lumps—namely, that they are distinct, despite occupying precisely the same region of spacetime—and that metaphysical story is all that matters. But this objection is sound only on the assumption that the properties being a material object, being a hylomorphic compound, and being a thing are on a par with one another. From x is a hylomorphic compound & y is a hylomorphic compound & x ≠ y, we rightly infer that x and y are two hylomorphic compounds. And if, somehow, we come to believe that x and y are co-located, we’d have no choice but to conclude that x and y are two distinct hylomorphic compounds sharing the same place at the same time.

The reason is that the following seems to be a necessary truth about the property of being a hylomorphic compound: (H1) x is a hylomorphic compound iff x is a matter-form composite; exactly one hylomorphic compound fills a region R iff some matter instantiates exactly one form; and x is (numerically) the same hylomorphic compound as y iff x is a hylomorphic compound and x = y. According to the second objection, a parallel principle expresses a necessary truth about the property of being a material object: (M1) x is a material object iff x is a hylomorphic compound; exactly one material object fills a region R iff exactly one hylomorphic compound fills R; and x is (numerically) the same material object as y iff x is a material object and x = y. (p.211) Note that M1 is not a mere linguistic principle; it is a substantive claim about the necessary and sufficient conditions for having a material object in a region, having exactly one material object in a region, and having (numerically) the same material object in a region. But M1 is a claim that will be denied by proponents of the Aristotelian solution we have been describing here. As should by now be clear, proponents of that solution will reject M1 in favour of something like M2: (M2) x is a material object iff x is a hylomorphic compound; exactly one material object fills a region R iff at least one hylomorphic compound fills R; and x is (numerically) the same material object as y iff x and y are hylomorphic compounds sharing the same matter in common. M2 is equivalent to M1 on the assumption that no two hylomorphic compounds can share the same matter in common; but, short of treating the technical philosophical category hylomorphic compound as co-extensive with the common-sense category material object, it is hard to see what would motivate that assumption. Thus, there is room for disagreement on the question whether M2 is true or whether M2 is equivalent to M1; and, importantly, accidental-sameness theorists and co-locationists will come down on different sides of those questions. Thus, there is a substantive (as opposed to a merely verbal) disagreement to be had here after all. Two further points should be made before we move on to the third objection. First, though M2 is specifically a thesis about the property being a material object, the doctrine of accidental sameness makes it plausible to think that similar theses about various other properties will be true. In particular, if one thinks that sortals like ‘cat’, ‘house’, ‘lump’, ‘statue’, and so on can apply nominally to things that constitute cats, houses, lumps, or statues, then something like M2 is true of most familiar composite object kinds. Second, though it may be tempting to think that the relation of accidental sameness (or of numerical sameness without identity) is nothing other than the relation of sharing exactly the same matter, as we see it, this isn’t quite correct. On our view (though probably not on Aristotle’s), the relation of numerical sameness without identity can hold between immaterial objects, so long as the relevant immaterial objects are plausibly thought of on analogy with hylomorphic compounds. Thus, it is inappropriate to say (as might so far seem natural to say) that the relation of numerical sameness without identity is nothing other than the relation of material constitution. Rather, what is appropriate to say is that material constitution is a species of numerical sameness without identity. Third objection: The principles for counting that we have just described (i.e. H1 and M2) are apparently inconsistent with the doctrine of accidental sameness. To see why, consider the following argument. Let Athena be a particular bronze statue; let Lump be the lump of bronze that constitutes it. Let R be the region filled by Athena and Lump. Then:

(p.212) 1.

(1) Athena is identical with the material object in R whose matter is arranged statuewise. 2. (2) Lump is identical with the material object in R whose matter is arranged lumpwise. 3. (3) The material object whose matter is arranged statuewise is identical with the material object whose matter is arranged lumpwise. 4. (4) Therefore, Athena is identical with Lump (contrary to the doctrine of accidental sameness). The crucial premise, of course, is premise 3; and premise 3 seems to follow directly from a proposition that is entailed by the facts of the example in conjunction with our remarks about counting—namely, that there is exactly one object in R whose matter is arranged both statuewise and lumpwise. On reflection, however, it is easy to see that this objection is a nonstarter. For premise 3 follows only if the doctrine of accidental sameness is false. Numerical sameness, according to Aristotle, does not entail identity. Thus, if his view is correct, it does not follow from the fact that there is exactly one material object in R whose matter is arranged both statuewise and lumpwise that the object whose matter is arranged lumpwise is identical with the object whose matter is arranged statuewise. Simply to assume otherwise, then, is to beg the question. One might insist that the assumption is nevertheless highly intuitive, and therefore legitimate. But, again, the right response here is that every solution to the problem of material constitution is such that its denial is highly intuitive. That is why we have a problem. Successfully rejecting a solution requires showing that the intuitive cost is higher with the objectionable solution than with some other solution; but, with respect to the doctrine of accidental sameness, this has not yet been done. Fourth objection: We say that there is one (and only one) material object that fills a region just in case the region is filled by matter unified in any object-constituting way. So consider a region R that is filled by matter arranged both lumpwise and statuewise. What is the object in R? What are its essential properties? If there is exactly one object in R, these two questions should have straightforward answers. But they do not (at least not so long as we continue to say that there is a statue and a lump in R). Thus, there is reason to doubt that there could really be exactly one object in R. This is probably the most serious objection of the lot. But there is a perfectly sensible reply: To the first question, the correct answer is that the object is both a statue and a lump; to the second question there is no correct answer.20 If the doctrine of accidental sameness is true, a statue and its constitutive lump are numerically the same object. This fact seems sufficient to entitle believers in (p.213) accidental sameness to say that the object in R ‘is’ both a statue and a lump, so long as they don’t take this to imply either that the statue is identical to the lump or that some statue or lump exemplifies contradictory essential properties. But if this view is right, how could there be any correct answer to the question ‘What are its essential properties?’ absent further information about whether the word ‘it’ is supposed to refer to the statue or the lump? The pronoun is ambiguous, as is the noun (‘the object in R’) to which it refers. 21 Thus, we would need to disambiguate before answering the question. Does this imply that there are two material objects in R? It might appear to because we are accustomed to finding ambiguity only in cases where a noun or pronoun refers to two objects rather than one. But if the doctrine of accidental sameness is true, we should also expect to find such ambiguity in cases of accidental sameness. Thus, to infer from the fact of pronoun ambiguity the conclusion that there must be two objects in R is simply to beg the question against the doctrine of accidental sameness.

So much for objections. Now, in closing this section, we would like to make it clear how accidental sameness differs from two apparently similar relations. Those who have followed the recent literature on material constitution will know that, like us, Lynne Baker has spoken of a relation that stands ‘between identity and separate existence’ (2000: 29) and that this relation is (on her view) to be identified with the relation of material constitution. On hearing this characterization, one might naturally think that what Baker has in mind is something very much like accidental sameness. In fact, however, the similarity between accidental sameness and Baker-style constitution ends with the characterization just quoted. Baker’s definition of constitution is somewhat complicated; but for present purposes we needn’t go into the details. Suffice it to say that, according to Baker, the relation of material constitution is neither symmetric nor transitive whereas accidental sameness is both symmetric and transitive. (At least, it is synchronically transitive.) Lacking the same formal properties, the two relations could not possibly be the same.22 One might also naturally wonder whether what we call ‘numerical sameness without identity’ isn’t just good old-fashioned relative identity under a different name. Different views have been advertised in the literature under the label ‘relative identity’. But one doctrine that virtually all of these views (and certainly all that deserve the label) share in common is the following: (p.214) (R1) States of affairs of the following sort are possible: x is an F, y is an F, x is a G, y is a G, x is the same F as y, but x is not the same G as y. This is a claim that we will endorse too; and, like those who endorse the relative identity solution to the problem of the Trinity, it is a truth we rely on in order to show that T1–T3 are consistent with one another. It is for this reason, and this reason alone, that we say that our solution may fruitfully be thought of as a version of the relative identity strategy. Despite our commitment to R1, it would be a mistake to suppose that we endorse a doctrine of relative identity. Our solution to the problem of the Trinity is therefore importantly different from the relative identity solution in its purest form.23 How is it possible to accept R1 while at the same time rejecting relative identity? The answer, as we see it, is that identity is truly relative only if one of the following claims is true: (R2) Statements of the form ‘x = y’ are incomplete and therefore ill-formed. A proper identity statement has the form ‘x is the same F as y’. (R3) Sortal-relative identity statements are more fundamental than absolute identity statements. 24 R2 is famously associated with P. T. Geach (1967, 1969, 1973), whereas R3 is defended by, among others, Nicholas Griffin (1977).25 Views according to which classical identity exists and is no less fundamental than other sameness relations are simply not views according to which identity is relative. Perhaps, on those views, there are multiple sameness relations; and perhaps some of those relations are both sortalrelative and such that R1 is true of them. But so long as classical identity exists and is in no way derivative upon or less fundamental than they are, there seems to be no reason whatsoever to think of other ‘sameness’ relations as identity relations. Thus, on views that reject both R2 and R3, there seems to be no reason for thinking that identity is relative.

The difference between accidental sameness and relative identity is important, especially in the present context, because it highlights the fact that there is more than one way to make sense of sameness without identity. It is for this reason that endorsing R1 apart from R2 or R3 won’t suffice all by itself to solve the problem of (p.215) the Trinity. As we have argued elsewhere (Rea 2003), absent an appropriate supplemental story about the metaphysics underlying relative-identity relations, endorsing R1 apart from R2 or R3 leaves one, at best, with an incomplete solution to the problem of the Trinity and, at worst, with an heretical solution.26 We think that the doctrine of accidental sameness provides the right sort of supplemental story, and that the solution it yields (in conjunction with R1) is both complete and orthodox. We suspect, moreover, that failure to distinguish different ways of making sense of sameness without identity is partly responsible for the attraction that the relative identity solution holds for many. As is well known, respected Christian philosophers and theologians—such as Augustine, Anselm, and Aquinas— habitually speak of the Trinity in ways that require the introduction of a form of sameness that fails Leibniz’s Law. But this way of speaking, it is often assumed, can only be explained in terms of relative identity.27 In light of what has just been said, however, we can see that this assumption is false. Sameness without identity does not imply relative identity, and hence any appeal to such sameness either to determine the views of actual historical figures or to provide authoritative support for a (pure) relative identity solution is wholly misguided. Relative identity does provide one way of explaining (numerical) sameness without identity, but it does not provide the only way of explaining it.

3. Sameness without Identity and the Problem of the Trinity If we accept the Aristotelian solution to the problem of material constitution, then, as we have seen, the familiar particulars of experience must be conceived of as hylomorphic compounds—that is, as matterform structures related to other things sharing their matter by the relation of accidental sameness. The relevance of this Aristotelian solution to the problem of the Trinity is perhaps already clear. For like the familiar particulars of experience, the Persons of the Trinity can also be conceived of in terms of hylomorphic compounds. Thus, we can think of the divine essence as playing the role of matter; and we can regard the properties being a Father, being a Son, and being a Spirit as distinct forms instantiated by the divine essence, each giving rise to a distinct Person. As in the case of matter, moreover, we can regard the divine essence not as an individual thing in its own right but rather as that which, together with the requisite ‘form’, constitutes a Person. Each (p.216) Person will then be a compound structure whose matter is the divine essence and whose form is one of the three distinctive trinitarian properties. On this way of thinking, the Persons of the Trinity are directly analogous to particulars that stand in the familiar relation of material constitution. Of course, there are also some obvious disanalogies. For example, in contrast to ordinary material objects, the role of matter in the case of the Trinity is played by immaterial stuff, and so the structures or compounds constituted from the divine essence (namely, the divine persons) will be ‘hylomorphic’ only in an extended sense. Also, in the case of material objects, the form of a particular hylomorphic compound will typically only be contingently instantiated by the matter. Not so, however, in the case of the Trinity. For Christian orthodoxy requires us to say that properties like being a Father and being a Son are essentially such as to be instantiated by the divine essence. As we have seen, moreover, the relation of accidental sameness on which our solution is modelled is, in Aristotle anyway, paradigmatically a relation between a substance (e.g. a man) and a hylomorphic structure built out of the substance and an accidental property. The Persons, however, are not like this. Thus, it is at best misleading to say that the relation between them is one of accidental sameness. Better instead to go with the more general label we

have used throughout this paper: the Persons stand in the relation of numerical sameness without identity. As far as we can tell, none of these disanalogies are of deep import. It seems not at all inappropriate to think of the divine Persons on analogy with hylomorphic compounds; and once we do think of them this way, the problem of the Trinity disappears. Return to the analogy with material objects: According to the Aristotelian solution to the problem of material constitution, a statue and its constitutive lump are two distinct hylomorphic compounds; yet they are numerically one material object. Likewise, then, the Persons of the Trinity are three distinct Persons but numerically one God. The key to understanding this is just to see that the right way to count Gods resembles the right way to count material objects. Thus: (G1) x is a God iff x is a hylomorphic compound whose ‘matter’ is some divine essence; x is the same God as y iff x and y are each hylomorphic compounds whose ‘matter’ is some divine essence and x’s ‘matter’ is the same ‘matter’ as y’s; and there is exactly one God iff there is an x such that x is a God and every God is the same God as x. And, in light of G1, the following principle also seems reasonable: (G2) x is God iff x is a God and there is exactly one God. If these principles are correct, and if (as Christians assume) there are three (and only three) Persons that share the same divine essence, then we arrive directly at (p.217) the central trinitarian claims T1–T3 without contradiction. For in that case, there will be three distinct Persons; each Person will be God (and will be the same God as each of the other Persons); and there will be exactly one God. Admittedly, if G1 is taken all by itself and without explanation, it might appear just as mysterious as the conjunction of T1–T3 initially appeared. But that is to be expected. What is important is that once the parallel with M2 is appreciated, and the doctrine of numerical sameness without identity is understood and embraced, much of the mystery goes away. We are now in a position to see how our Aristotelian account of the Trinity meets the desiderata we set out earlier for an adequate solution to the problem of the Trinity (namely, D1–D5). As should already be clear, our solution resolves the apparent inconsistency of T1–T3 in the same basic way that relative identity and social trinitarian solutions do: namely, by rejecting the idea that the words ‘is God’ in trinitarian statements like ‘Each of the Persons is God’ mean ‘is absolutely identical with God’. According to our solution, these words should be interpreted to mean ‘is numerically the same as the one and only God’. But once this interpretation of T2 is adopted—together with a proper understanding of the relata of the relation of numerical sameness without identity—the apparent inconsistency of T1–T3 is resolved, and in a way that satisfies D1 and D2. For inasmuch as the Persons of the Trinity are distinct hylomorphic compounds, they are distinct from one another (hence T1 is true); and inasmuch as they are each numerically the same as the one and only God, each of them is God and there is only one God (hence T2 and T3 are true). Moreover, since our solution implies that each of the Persons is a divine individual who is one in number with each of the other two Persons, it is consistent with the claim that there are three Persons but exactly one divine individual (thus satisfying D1),28 and it also seems to preserve the intention of traditional formulations of the doctrine of the Trinity (thereby satisfying D2). It should also be clear how our solution meets the other desiderata. Unlike (pure) relative identity solutions, ours is compatible with the claim that classical identity exists and is as fundamental as any other sameness relation (and hence satisfies D4). Moreover, it supplies an explanation for why ‘x = y’ does not follow from ‘x is the same God as y’. Unlike social trinitarian strategies, on the other hand, ours is

clearly compatible with the view that God is an individual rather than a society, and that the Persons are not parts of God (and hence satisfies D3). Furthermore, our story about the unity of the Persons exploits what we take to be a plausible story about the unity of distinct hylomorphic compounds, whereas no similarly plausible analogy seems to be available to the social trinitarian. Finally, though we deny that it makes sense to say, unequivocally, that each of the Persons (p.218) is absolutely identical with God, our view—unlike either of the other two strategies—allows us to say that the Father is identical with God, the Son is identical with God, the Holy Spirit is identical with God, and yet the Father, Son, and Holy Spirit are distinct from one another. And it can do all of this without introducing any anti-realist commitments in metaphysics (thus satisfying D5). Consider a parallel drawn from one of our earlier examples: Athena is identical to the material object in R; Lump is identical to the material object in R; but Athena is distinct from Lump. Since ‘the material object in R’ is ambiguous, there is no threat of contradiction; and the doctrine of numerical sameness without identity blocks an inference to the claim that Lump and Athena are co-located material objects. Likewise in the case of the Trinity. For all these reasons, therefore, our Aristotelian solution to the problem of the Trinity seems to us to be the most philosophically promising and theologically satisfying solution currently on offer.

4. Important Consequences This completes our defence of the Aristotelian account of the Christian doctrine of the Trinity. As we see it, however, this account is not only interesting in its own right, but also has several important consequences. We close by calling attention to two of these. First, our solution suggests a revision in our understanding of the nature of the copula. Philosophers traditionally distinguish what is called the ‘is’ of predication from the ‘is’ of identity. It is sometimes added, moreover, that any solution to the problem of material constitution that denies that constitution is identity must introduce a third sense of ‘is’. As Lynne Baker says: If the constitution view [i.e. the view that constitution is not identity] is correct, then there is a third sense of ‘is’, distinct from the other two. The third sense of ‘is’ is the ‘is’ of constitution (as in ‘is (constituted by) a piece of marble’).29 Baker seems to think that if constitution is not identity, there will have to be three main senses of the copula, each co-ordinate with the other two. But we can now see that this is a mistake. If our account of the Trinity is correct, constitution can be explained in terms of something other than identity (namely, accidental sameness). Even so, there will be only two main senses of the copula, namely, the traditional ‘is’ of predication and a heretofore unrecognized sense of the copula, the ‘is’ of numerical sameness. There will still be an ‘is’ of identity and (p.219) an ‘is’ of constitution, as Baker suggests, but these will both be subsumed under the second of the two main senses just mentioned. Indeed, if we take into account all of the changes suggested by our account of the Trinity, we will get a fairly complex set of relations holding between the various senses of the copula, as Figure 10.1 makes clear.

Figure 10.1

Second, our solution helps to make clear that both the problem of material constitution and the problem of the Trinity are generated in part by the fact that we have incompatible intuitions about how to count things. Thus, both problems might plausibly be seen as special instances of a broader counting problem—a problem that arises whenever we appear to have, on the one hand, a single object of one sort (e.g. God or material object) and, on the other hand, multiple coinciding objects of a different sort

(e.g. Person or hylomorphic compound). One significant advantage of the Aristotelian solution to the problem of material constitution is that it alone seems to provide a unified strategy for resolving the broader problem of which it is an instance.30 References Bibliography references: Anscombe, G. E. M. and P. T. Geach. 1961. Three Philosophers (Ithaca, NY: Cornell University Press).

Baker, Lynne Rudder. 1999. ‘Unity without Identity: A New Look at Material Constitution’. Midwest Studies in Philosophy 23: 144–65.

(p.220) Baker, Lynne Rudder. 2000. Persons and Bodies. Cambridge: Cambridge University Press. Bartel, Timothy. 1988. ‘The Plight of the Relative Trinitarian’. Religious Studies 24: 129–55.

Bartel, Timothy. 1993. ‘Could There Be More than One Almighty?’ Religious Studies 29: 465–95.

Bartel, Timothy. 1994. ‘Could There Be More than One Lord?’ Faith and Philosophy 11: 357–78. Brower, Jeffrey E. 2004a. ‘Trinity’. In The Cambridge Companion to Abelard, edited by Jeffrey E. Brower and Kevin Guilfoy, 223–57. Cambridge: Cambridge University Press. Brower, Jeffrey E. 2004b. ‘The Problem with Social Trinitarianism: A Reply to Edward Wierenga’. Faith and Philosophy 21: 295–303.

Brown, David. 1985. The Divine Trinity. La Salle, IL: Open Court. Brown, David. 1989. ‘Trinitarian Personhood and Individuality’. In Trinity, Incarnation, and Atonement, edited by Ronald J. Feenstra and Cornelius Plantinga Jr., 48–78. Notre Dame, IN: University of Notre Dame Press. Cain, James. 1989. ‘The Doctrine of the Trinity and the Logic of Relative Identity’. Religious Studies 25: 141–52. Cartwright, Richard. 1987. ‘On the Logical Problem of the Trinity’. in Philosophical Essays, 187–200. Cambridge, MA: MIT Press.

Clark, Kelly. 1996. ‘Trinity or Tritheism?’ Religious Studies 32: 463–76. Davis, Stephen. 1999. ‘A Somewhat Playful Proof of the Social Trinity in Five Easy Steps’. Philosophia Christi 1: 103–5. Feser, Edward. 1997. ‘Swinburne’s Tritheism’. International Journal for Philosophy of Religion 42: 175– 84. Forrest, Peter. 1998. ‘Divine Fission: A New Way of Moderating Social Trinitarianism’. Religious Studies 34: 281–97. Geach, Peter. 1967. ‘Identity’. Review of Metaphysics 21: 3–12. Reprinted in Peter Geach, 1972. Logic Matters, 238–46. Berkeley: University of California Press. Geach, Peter. 1969. ‘Identity—A Reply.’ Review of Metaphysics 22: 556–9. Reprinted in Peter Geach, 1972. Logic Matters, 247–9. Berkeley: University of California Press. Geach, Peter. 1973. ‘Ontological Relativity and Relative Identity’. in Logic and Ontology, edited by Milton K. Munitz, 287–302. New York: New York University Press.

Griffin, Nicholas. 1977. Relative Identity. Oxford: Clarendon Press.

Layman, C. Stephen. 1988. ‘Tritheism and the Trinity’. Faith and Philosophy 5: 291–8. Leftow, Brian. 1999. ‘Anti Social Trinitarianism’. In The Trinity: An Interdisciplinary Symposium, edited by Stephen Davis, Daniel Kendall SJ, and Gerald O’Collins SJ, 203–49. Oxford: Oxford University Press.

(p.221) Lewis, David, 1986. On the Plurality of Worlds. Oxford: Basil Blackwell. Lewis, David 1993. ‘Many but Almost One’. In Ontology, Causality, and Mind: Essays on the Philosophy of D. M. Armstrong, edited by K. Campbell, J. Bacon, and L. Reinhardt. (Cambridge: Cambridge University Press). Reprinted in David Lewis, Papers in Metaphysics and Epistemology, 163–82 (Cambridge: Cambridge University Press, 1999).

Martinich, A. P. 1978. ‘Identity and Trinity’. Journal of Religion 58: 169–81.

Martinich, A. P. 1979. ‘God, Emperor and Relative Identity’. Franciscan Studies 39: 180–91. Matthews, Gareth. 1982. ‘Accidental Unities’. In Language and Logos, edited by M. Schofield and M. Nussbaum, 223–40. Cambridge: Cambridge University Press. Matthews, Gareth. 1992. ‘On Knowing How to Take Aristotle’s Kooky Objects Seriously’. Paper presented at the Pacific Division Meeting of the APA, Portland. Merricks, Trenton. 2005. ‘Split Brains and the Godhead’. in Knowledge and Reality: Essays in Honor of Alvin Plantinga, edited by Thomas Crisp, Matthew Davidson, and David Vander Laan, 299–326. Dordrecht: Springer.

Noonan, Harold. 1988. ‘Reply to Lowe on Ships and Structures’. Analysis 48: 221–3. Pereboom, Derk. 2002. ‘On Baker’s Persons and Bodies’. Philosophy and Phenomenological Research 64: 615–22. Plantinga, Cornelius Jr. 1986. ‘Gregory of Nyssa and the Social Analogy of the Trinity’. Thomist 50: 325– 52. Plantinga, Cornelius Jr. 1988. ‘The Threeness/Oneness Problem of the Trinity’. Calvin Theological Journal 23: 37–53. Plantinga, Cornelius Jr. 1989. ‘Social Trinity and Tritheism’. In Trinity, Incarnation, and Atonement, edited by Ronald J. Feenstra and Cornelius Plantinga Jr, 21–47. Notre Dame, IN: University of Notre Dame Press. Rea, Michael. 1998. ‘Sameness without Identity: An Aristotelian Solution to the Problem of Material Constitution’. Ratio 11: 316–28. Reprinted in Form & Matter: Contemporary Themes in Metaphysics, edited by D. S. Oderberg, 103–16. Oxford: Basil Blackwell, 1999.

Rea, Michael. 2000. ‘Constitution and Kind-Membership’. Philosophical Studies 97: 169–93. Rea, Michael. 2002. ‘Lynne Baker on Material Constitution’. Philosophy and Phenomenological Research 64: 607–14.

Rea, Michael. 2003. ‘Relative Identity and the Doctrine of the Trinity’. Philosophia Christi 5: 431–46.

Robinson, Denis. 1985. ‘Can Amoebae Divide Without Multiplying?’ Australasian Journal of Philosophy 63: 299–319. Routley, Richard and Nicholas Griffin. 1979. ‘Towards a Logic of Relative Identity’. Logique et Analyse 22: 65–83. (p.222) Sider, Theodore. 2002. ‘Review of Lynne Rudder Baker’s Persons and Bodies’. Journal of Philosophy 99: 45–8. Swinburne, Richard. 1994. The Christian God. Oxford: Clarendon Press. van Inwagen, Peter. 1988. ‘And Yet They Are Not Three Gods but One God’. In Philosophy and the Christian Faith, edited by Thomas Morris, 241–78. Notre Dame, IN: University of Notre Dame Press.

Williams, C. J. F. 1994. ‘Neither Confounding the Persons nor Dividing the Substance’. In Reason and the Christian Religion: Essays in Honor of Richard Swinburne, edited by Alan Padgett, 227–43. Oxford: Clarendon Press. Wolfson, H. A. 1964. The Philosophy of the Church Fathers, Vol. 1: Faith, Trinity, Incarnation. Cambridge, MA: Harvard University Press. Zimmerman, Dean. 1995. ‘Theories of Masses and Problems of Constitution’. Philosophical Review 104: 53–110. Zimmerman, Dean. 2002. ‘Persons and Bodies: Constitution without Mereology?’ Philosophy and Phenomenological Research 64: 599–606. Notes: (1) For purposes here, an object x and an object y stand in the relation of material constitution just in case x and y share all of the same material parts. Thus, on our view, material constitution is both symmetric and transitive. Contrary to some philosophers (e.g. Lynne Baker, discussed below) who treat material constitution as asymmetric, we think that there are good theoretical reasons for regarding it as a symmetric relation; but we will not attempt to defend that view here. (2) Noonan 1988: 222.

(3) Note, however, that we stop short of actually endorsing the solution that we describe. There are three reasons for this. First, our solution, like most others, attempts to provide a metaphysical account of the ultimate nature of God. But surely here, if anywhere, a great deal of circumspection is warranted. Second, the contemporary trinitarian debate, as we see it, is still in its infancy; hence a definitive stand on any particular solution, including our own, strikes us as a bit premature. Third, the solution we develop strongly supports a specific understanding of material constitution (as will become clear in Section 4)— one that is at odds with some of our previously considered views on the matter. (See, e.g. Rea 2000.) But, given the current state of the trinitarian debate, we are uncertain whether this fact should motivate us to change our views about material constitution or to continue exploring yet other alternatives to the currently available accounts of the Trinity. Thus, it is important to understand that we are not here aiming to resolve the contemporary trinitarian debate once and for all, but rather to advance it by introducing what seems to us to be the most promising solution to the problem of the Trinity developed so far. (4) Quicumque vult (our translation). (5) Denying that ‘is God’ means ‘is absolutely identical with God’ doesn’t guarantee that modalism is false; but making the denial removes any pressure toward modalism that might arise out of T1–T3. (6) See Rea 2003 and Brower 2004b. Proponents of the relative identity strategy include Anscombe and Geach 1961: 118–20; Cain 1989; Martinich 1978, 1979; and van Inwagen 1988. Proponents of the more typical versions of the social trinitarian strategy include Bartel 1993, 1994; Brown 1985, 1989; Davis 1999; Layman 1988; C. Plantinga 1986, 1988, 1989; and Swinburne 1994. The position is commonly attributed to the Cappadocian Fathers. (See, esp. Brown 1985, Plantinga 1986, and Wolfson 1964.) It is against these relatively typical versions of ST that our previously published objections most straightforwardly apply. Among the less typical versions of ST are, for example, Peter Forrest’s (1998), according to which the Persons are three ‘quasi-individuals’ that result from an event of divine fission, and C. J. F. Williams’s (1994), according to which ‘God is the love of three Persons for each other.’ We reject Forrest’s view because it implies (among other things) that there is no fact about whether there are one or many Gods, and there is no fact about whether there are three or many more than three Persons. On his view, ‘one’ is the lowest correct answer to the question ‘How many Gods are there?’ and ‘three’ is the lowest correct answer to the question ‘How many persons are there?’; but it is sheer convention that allows us to say that ‘one’ and ‘three’—rather than, say ‘twenty’ and ‘two hundred and forty one’—are the correct answers to those questions. As for Williams’s view, we take it that his, along with other less common versions of ST, will fall prey to objections similar to those we raise against the more typical versions. For further critical discussion of both the relative identity strategy and the social trinitarian strategy, see Bartel 1988, Cartwright 1987, Clark 1996, Feser 1997, Leftow 1999, and Merricks 2005. (7) Note that the point of D3 isn’t to deny that the Persons compose a society. Of course they do, if there are genuinely three Persons. Rather, the point of D3 is to deny both that the name ‘God’ refers to the society composed of these Persons and that the Persons are proper parts of God. But if the society of Persons is the Trinity, and the Trinity is God, doesn’t it follow that ‘God’ refers to the society of Persons after all? No. Each member of the Trinity is God, and God ‘is a Trinity’ (that is, He exists in three Persons). But nothing in orthodoxy seems to require that the Trinity is itself a whole composed of three Persons and referred to by the name ‘God’. Moreover, in light of objections to social trinitarianism raised here and elsewhere, it seems that orthodoxy actually precludes us from saying such a thing (which is part of why we reject social trinitarianism).

(8) See Rea 1998 and Brower 2004a. (9) For reasons that we shall explain below, the label ‘accidental sameness’ is not appropriate in the context of the Trinity. (10) This claim is negotiable; and, in fact, there are independent (non-Aristotelian) reasons for thinking that ‘masses of matter’ must be treated as individuals. (See, e.g. Zimmerman 1995.) But the view of matter articulated here seems to comport best with Aristotle’s metaphysics and with the solution to the problem of the Trinity that we will propose, and so we will go ahead and endorse it here. Those who think of masses of matter as individuals may be inclined (in Section 3 below) also to think of what we will call ‘the divine essence’ as an individual. Were we to endorse this view, we would deny that the divine essence is a fourth Person or a second God (just as we would deny that Socrates’s matter is a second man co-located with Socrates). Rather, we would say that the divine essence is one in number with God, a sui generis individual distinct from the Persons and, indeed, nothing other than a substrate for the Persons. We would also deny that there is any sense in which the divine essence is prior to or independent of God. (11) We place ‘its’ in parentheses to signal our neutrality on the question whether, say, the humanity of Plato is a special kind of trope or a multiply instantiated universal. (12) See, e.g. Metaphysics H2, 1042b15–25. (13) Matthews 1982, 1992. (14) Topics A7, 103a23–31; Physics A3, 190a17–21, 190b18–22; Metaphysics D6, 1015b16–22, 1016b32– 1017a6; Metaphysics D9, 1024b30–1. (15) Topics A7, 103a23–31; Metaphysics D6, 1015b16–22, 1016b32–1017a6. (16) And, we might add, the same would hold true for Socrates and his matter, if indeed the matter of a thing were to be understood as an individual distinct from that thing. (17) See, e.g. Lewis 1993: 175, and Robinson 1985. (18) An object belongs to a kind in the nominal way just in case it displays the superficial features distinctive of members of that kind. (19) Lewis 1986: 252. (20) We assume that ‘object’ in the context here means ‘material object’. (21) Here is why ‘the object in R’ is ambiguous. There aren’t two material objects in R; and the material object in R isn’t a third thing in addition to Athena and Lump. Thus, ‘Athena = the material object in R’ and ‘Lump = the material object in R’ must both express truths. But they can’t both express truths unless either Lump = Athena (which the doctrine of accidental sameness denies) or ‘the material object in R’ is ambiguous. (22) Baker’s definition appears in both Baker 1999 and Baker 2000. For critical discussion, see Pereboom 2002, Rea 2002, Sider 2002, and Zimmerman 2002.

(23) Elsewhere we distinguish between pure and impure versions of the relative identity strategy (see Rea 2003). Impure versions endorse R1 without endorsing a doctrine of relative identity; pure versions endorse R1 in conjunction with either R2 or R3 below. Our solution is thus an impure version of the relative identity solution. (24) To say that sortal-relative identity statements are more fundamental than absolute identity statements is, at least in part, to say that absolute identity statements are to be analysed or defined in terms of more primitive sortal-relative identity statements, rather than the other way around. See Rea 2003 for further discussion of views that endorse R3. (25) See also Routley and Griffin 1979. (26) This is, roughly, the problem that we think Peter van Inwagen’s solution to the problem of the Trinity faces. (Cf. Rea 2003.) (27) For example, Cartwright (1987: 193) claims to detect an appeal to relative identity in a letter of Anselm, as well as the Eleventh Council of Toledo, on just these grounds. The same sort of reasoning may also help to explain Anscombe and Geach’s (1961: 118) attribution of the relative identity solution to Aquinas. (28) Assuming, anyway, that counting divine individuals is more like counting Gods than counting Persons. But this assumption seems clearly legitimate in context of D1. (29) Baker 1999: 51. (30) This chapter has benefitted greatly from the advice and criticism of Michael Bergmann, Jan Cover, Tom Crisp, William Hasker, John Hawthorne, Michael Jacovides, Brian Leftow, Trenton Merricks, Laurie Paul, William Rowe, and two anonymous referees for Faith and Philosophy.               

Index For the benefit of digital users, indexed terms that span two pages (e.g., 52–53) may, on occasion, appear on only one of those pages. absolute generality 131–2 accidental sameness 205, 215–18 Achtemeier, Paul 53n.1 Allen, J. P. 178n.21 Alston, William 68–9, 120–1, 125–6, 129–30 Amenhotep IV 177–8 Amun-Re theology 172–6 analogy 120n.2, 124, 132 analytic metaphysics, see metaphysics analytic philosophy 3–4, 36, 84–5 analytic theology and analytic philosophy of religion 4–5 as a type of theology 85–6, 88, 90–7 concept of 1, 3–5 history of 1–2 misconceptions 5–10 Anderson, Pamela 96 Anglo-American philosophy, see analytic philosophy Anscombe, G. E. M. 157n.39, 185n.1, 204n.6, 215n.27 Anselm 215 anti-realism 51, 196–7, 204, 217–18; see also realism, in theology and metaphysics apophatic theology 36 apophaticism 12, 121–3, 125, 130, 134 Aquinas, St. Thomas 146–9, 215 Aristotle 13–14, 132, 133n.14, 159–61, 190n.13, 201–2, 205–15 Assmann, Jan 172n.7, 178n.21 Athanasian Creed 141–4, 155–6, 156n.35, 202–3 Athanassiadi, Polymnia 180n.25 atheism 121, 127, 130–3 Augustine 146–7, 152–3, 161n.45, 162–3, 215 authority of texts, see texts, authority of authority-defeat 56–7, 67–8 authority de facto and de jure 61–3 defeated, see authority-defeat foundational 61–3, 74–5

more authoritative than relation 59–60 practical 55, 61, 73 relationship to truth theoretical 55, 61, 73–4 Ayres, Lewis 141n.2, 142n.3, 144n.7, 146n.15, 147n.17 Baines, John 172n.7, 175, 178n.21 Baker-Hytch, Max 5n.5 Baker, Lynne 213, 218–19 Barnes, Michel 141n.2, 144n.7, 146n.15, 147n.17 Bartel, Timothy 152n.28, 185n.1, 186n.3, 204n.6 Barth, Karl 83, 91–3 Basil of Caesarea 164–5 Bavinck 8n.11 Being itself 37–8 being of beings 37–8 ultimate ground of 94 Belgic Confession 10 belief 19–20 reliable 21–3 Berkhof, Louis 143n.4 beyondness, see divine transcendence, as beyondness Bible 53–5, 63–4, 72–3 biblical authority 53–4, 63–4 biblical inerrancy 53–4, 76–7 Bitar, Ray Paul 5n.6 Blanshard, Brand 67–8 Bonnet, Hans 175, 178 Brower, Jeffrey 141n.1, 149n.22, 150n.25, 155n.31, 156n.35, 157n.39, 160n.43, 181n.26, 186n.3, 195, 204n.6 Brown, David 152n.28, 185n.1, 204n.6 Bryant, Robert 63–4 Budge, E. A. W. 178n.21 Burggraeve, Roger 40–1 Butler, Judith 103n.14 Bynum, Caroline Walker 105n.17 Byrne, Peter 19–28, 22n.3

224



Cain, James 157n.39, 185n.1, 204n.6 Camp, Elisabeth 131n.11 Canons and Decrees of the Council of Trent 22–3 Cappadocian Fathers 146–7, 152–7, 162–5, 185n.1, 204n.6 Caputo 39, 41, 49–50 Carr, Anne 98n.1 Cartwright, Richard 204n.6, 215n.27 Cary, Philip 147n.17 Cataphatic theology 36, 50 Catholic Church 22–3 Černý, Jaroslav 178n.21 Champillion Figeac 177–8 Chicago Statement on Inerrancy 76 Chodorow, Nancy 104n.15 Christ 43–5, 142, 145 Christian Reformed Church 10–11, 142–3 Clark, Kelly James 155n.31, 186n.3, 204n.6 co-locationism 209–10 Coakley, Sarah 99n.6, 102n.11, 141n.2, 147n.17 concepts, human 36, 38–52 conceptual scheme 38–40 constitution trinitarianism 157–65 consubstantiality 13, 141–6, 152, 154, 158, 163–5 Cooper, David 111n.24, 112–16, 128–9 Corinthians 11:7 102n.10 Couenhoven, Jesse 5n.6 counting problem 219 Craig, William Lane 148n.21, 152n.28, 154–7, 170–1 creation myths 172–4, 176–7 Crisp, Oliver 1 Cross, Richard 147n.17, 148, 161–2 Daly, Mary 83, 98n.1, 104 Davies, Glenys 180n.25 Davis, Stephen 185n.1, 204n.6 decisive reasons for action and for believing 55–6, 62 defeaters of authority, see authority, defeated Dionysius, see Pseudo-Dionysius the Areopagite divine, the experience of 39 divine attribute as a topic in analytic philosophy of religion 83–4 as a topic in analytic theology 92–3 gender as 102, 107–11; see also God, gender of sex as 104 nature of 106 perfection 92–3 transcendence, see divine transcendence divine image, see imago dei doctrine

divine transcendence nature of 120–3 as beyondness 123–8; see also God, as beyond being, and God, as beyond characterization Dole, Andrew 89–90 Dummett, Michael 189–90, 196–7 Eastern Church 146–7 Eastern Trinitarianism, see Greek Trinitarianism Egyptian religion 172–80 empirical stance 20–1, 32–4 empiricism, see empirical stance equal gender thesis 116–17 equality thesis 104–5 essence, of material objects 37 essentialist-correspondentist metaphysics 37 explanatory virtues 28–30, 32–3 feminist theology, see theology, feminist Feser, Edward 155n.31, 186n.3, 204n.6 Finch, Alicia 33–4 Forrest, Peter 152n.28, 185n.1, 204n.6 Frede, Michael 180n.25 Frei, Hans 85–92 Gale, Richard 95 Geach, P. T. 157n.39, 158, 185–90, 185n.1, 195–7, 204n.6, 214, 215n.27 gendered pronouns and predicates 103–5, 107–11, 116–18 genderism 102–3, 106–11 generality relativism 131–3 Genesis 177 Gnuse, Robert 63–4 God as beyond being 123–5 as beyond characterization 125–8 attributes of 10–12; see also divine attribute concepts applied to, see concepts, human existence of 120–1, 127–8, 130–6 gender of 11–12, 96, 98–101; see also traditionalism and inclusivism inclusive characterizations of, see inclusivism masculine characterizations of, see traditionalism perfection of 11–12 tri-unity of, see Trinity, doctrine of Greek Trinitarianism 146–7, 152–7 Gregory of Nyssa 130, 162n.47, 163–5 Grelling’s Paradox 195–6 Griffin, Nicholas 188–90, 195, 214 Hampson, Daphne 83, 95–6, 104n.15, 121n.3 Hasker, William 152n.28, 162n.48

 Haslanger, Sally 105n.16 Hector, Kevin 36–50, 94 Heidegger, Martin 37–8, 51–2, 51n.10, 94–5 Heliopolitan cosmogony 173 Hellman, Geoffrey 133n.14 Hermopolitan cosmogony 173 Hick, John 121n.3 Hodge, Charles 143n.4 Homer 54–5 homoousion 142, 156–7, 165 Hornung, Erik 175–9 Howard-Snyder, Dan 125–7, 131 hylomorphism 205–7, 209–11, 215–18 idolatry 36–7, 39, 41–2, 45, 48–52, 94, 115–16 images, see metaphor imago Dei doctrine 98, 101–6, 109 inclusivism 98–103, 116 incomprehensibility 120 indiscernibility of identicals, see Leibniz’s Law ineffability 120, 127n.8, 130 inerrancy of scripture, see biblical inerrancy Jacobs, Jonathan 127n.8 James, William 67–8 Jantzen, Grace 96, 99n.6, 105n.17 Jenson, Robert 83, 102–3 Jesus, see Christ Johnson, Elizabeth 98n.1, 100–1, 104n.15, 111n.24, 121n.3 Jones, Tamsin 39 Julian, of Norwich 105n.17 Kaufman, Gordon 85–9, 91, 131 Kelly, J. N. D. 142 Khonsu Cosmogony 173–4 Latin Trinitarianism 146–52 Layman, C. Stephen 152n.28, 185n.1, 204n.6 Lebens, Samuel R. 127n.8, 133n.14 Leftow, Brian 150–2, 155nn.31,32, 186n.3, 204n.6 Leibniz’s Law 187–9, 208, 215 Lesko, Leonard 173n.8 Levinas, Emmanuel 38, 40–2 Lewis, David 208n.17 Lindbeck, George A. 2n.2 Macdonald, Paul A., Jr. 5n.6, 180n.25 Marion, Jean-Luc 37–9, 49–50 Martinich, A. P. 157n.39, 185n.1, 204n.6 masculine ideals, privileging of 95–6 material constitution 159, 201–2, 205–15 Mavrodes, George 180n.25

225

McCall, Thomas H. 5n.6 McFague, Sallie 83, 121n.3 McKim, Donald 53n.1 Memphite cosmogony 173 Merricks, Trenton 149–50, 155n.32, 157n.39, 186n.3, 204n.6 metaphor 120–1, 126–36 metaphysical accuracy 99–116 metaphysical anti-realism, see realism, in theology and metaphysics metaphysical stance 32–4 metaphysics of sex and gender 103 metaphysics 28–32, 36–7, 94 metatheology 2–10 Miller, David 180n.25 modalism 146, 149, 196, 203 modes of being 124, 131–6 Moltmann, Jürgen 152n.28 monotheism 170–2, 176–80, 185–6, 195; see also polytheism Moo, Douglas 53n.1 Moore, G. E. 3 Moreland, J. P. 148n.21, 152n.28, 154–7, 170–1 Morenz, Siegfried 175, 178n.21 Morris, Thomas V. 149–50 Murphy, Mark 56n.4 New Kingdom Amun-Re theology, see Amun-Re theology Nicene Creed 22–3, 142–4, 155–7 Nicole, Roger 74n.38 non-revealed concepts 12 Noonan, Harold 201 O’Collins 145n.14 Oliver, Simon 5n.6 ontological commitment 19–20 ontology, analytic 28–32 ontotheology 38n.2, 85–6, 93–6 otherness 40–2, 120 Owen, Paul 180n.25 pan-symbolism 120–1, 123–33 Panchuk, Michelle 5n.7 panmetaphoricism 120–1 Pannenberg, Wolfhart 83 part-whole trinitarianism 155–7 Patristic studies 141n.2 Pawl, Tim 5n.6 Peirce, C. S. 67 Pereboom, Derk 213n.22 perfect being theology, see theology, perfect being Peterson, Anne 161n.45 philosophy of language 42–3

226



philosophy of religion 4–5 analytic 83, 85–6, 92, 95 Plantinga, Cornelius, Jr. 152n.28, 154–5, 157, 170–1, 170n.2, 185n.1, 204n.6 Plantinga, Alvin 68, 131n.12 polyadic properties 149 polytheism 146, 150n.25, 152–3, 155–6, 170–2, 176–80, 186, 195, 197–8, 203–4 Porter, Barbara 180n.25 Presbyterian Church 22–3 indiscernibility of identicals, principle of, see Leibniz’s Law properties 45–7 Pseudo-Dionysius the Areopagite 120–3, 130 quantification, absolutely unrestricted 131–2 Raphael, Melissa 99n.5 Raz, Joseph 56n.6 Rea, Michael 2n.1, 5–6, 5nn.6,7, 12, 63n.14, 94, 150n.25, 157n.39, 160n.43, 181n.26, 202n.3, 204n.6, 213n.22, 214–15, 214n.23 realism defined 19 in theology and metaphysics 20–32 metaphysical 66–7 realistically interpreted 21–7 reasons for action and for believing, see decisive reasons for action and for believing Régnon, Theodore 146n.15 relative identity impure relative identity strategy 190–5, 197–8, 214n.23 pure relative identity strategy 190–1, 195–8, 214n.23, 215, 217–18 strategy 147, 157–9, 185–7, 190–2, 203–5 theories of 187–90 trinitarians, see relative identity strategy reliability 54, 73–7 Richard’s Paradox 195–6 Robinson, Denis 208n.17 Rogers, Jack 53n.1 Russell, Bertrand 3 Sarrisky, Darren 5n.6 Schaff, Philip 145n.9, 148n.19, 163–5 Schleiermacher, Friedrich Daniel Ernst 43, 89–90 Schwartz 142 science 28–30 scripture as regulative 92

authority of 8–10, 53–4; see also biblical authority truthfulness of 53–4, 72–3; see also biblical inerrancy see also Bible Sider, Theodore 213n.22 Smith, Mark 180n.25 social trinitarianism 146–7, 152–7, 165, 170–2, 179–81, 185–6, 203–5 Sola scriptura 8–10 sortal-relative identity, see sortals sortals 188–90, 194–6, 209, 214 Soskice, Janet Martin 99n.6 Speaks, Jeff 11 Stead, Christopher 141n.2 Stevenson, Leslie 187–8 subordinationism 145, 203 Swinburne, Richard 95, 145n.8, 152n.28, 154–155, 170–1, 170n.2, 185n.1, 204n.6 syncretism 175–81 Synodical Letter 143–4 Tarski, Alfred 195n.25 Taylor, A. E. 95 texts authority of 53–64 definition of 54 truthfulness of 53–4, 69–72 Thebes 172–4 theological realism 7–8, 67; see also realism, in theology and metaphysics theology biblical 19n.1 dogmatic 19n.1 feminist 85–6, 93, 95–6, 99–100 historical 19n.1 perfect being 30–1, 92, 101, 108, 117 philosophical 19n.1, 30–1 systematic 19n.1, 30–1 tasks of 86 types of 85–92 Thiselton, Anthony C. 74n.38 Tillich, Paul 83 Tobin, Vincent 175n.13, 178nn.18,21 Tracy, David 88–9, 91 traditionalism 98–101, 111–16 transcendence, see divine transcendence Trinity, doctrine of and monotheism 170–2, 179–81 psychological models of, see Latin trinitarianism questions about 141 problem of 12–15, 142–6, 202–5

 types of solutions to the problem of 146–57 the constitution view of 159–61 the relative identity view of, see relative identity, strategy social trinitarian views of, see social trinitarianism sameness without identity and 214–19 truth and truthfulness of texts, see texts, truthfulness of coherence theories of 68–9 correspondence theories of 68–9 epistemic and non-epistemic conceptions of 67–8 nature of 64–9 realist and anti-realist conceptions of 65–7 Tuggy, Dale 156n.33, 156n.36, 157n.39 Turcescu, Lucian 141n.2 unknowability 120 unrestricted quantification, see quantification, absolutely unrestricted

227

Van Fraassen, Bas 19–21, 28–32 van Inwagen, Peter 157n.39, 158–9, 185n.1, 192–7, 204n.6, 215n.26 Vanhoozer, Kevin 5n.6, 54, 93–5 violence 36–42, 44–5, 48–52 Wace, Henry 145n.9, 148n.19, 163–5 Watterson, Barbara 179n.24 Wessling, Jordan 5n.6 Westerholm, Martin 5–7, 5n.6 Western Church 146–7 Western Trinitarianism, see Latin Trinitarianism Wierenga, Edward 152n.28, 156n.35 Wilkinson, Richard 178n.21 Williams, C. J. F. 152n.28, 185n.1, 204n.6 Williamson, Timothy 133n.14 Wittgenstein, Ludwig 36 Wolfson, H. A. 141n.2, 185n.1, 204n.6 Wood, William 5n.6, 93 Young, Edward J. 63–4 Zemach, Eddy 187–8 Zimmerman, Dean 205n.10, 213n.22